cytochrome p450 bm3 as versatile biocatalytic tool in drug development

CYTOCHROME P450 BM3 AS VERSATILE

BIOCATALYTIC TOOL IN DRUG DEVELOPMENT

Cytochrome P450 BM3 as versatile biocatalytic tool in drug development

Vanina Rea

Printed by Wöhrmann Print Service, Zutphen, the Netherlands

Cover design: Tree of life, Gustav Klimt, 1905

VRIJE UNIVERSITEIT

Cytochrome P450 BM3 as versatile biocatalytic tool in drug

development

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad Doctor aan

de Vrije Universiteit Amsterdam,

op gezag van de rector magnificus

prof.dr. L.M. Bouter,

in het openbaar te verdedigen

ten overstaan van de promotiecommissie

van de faculteit der Exacte Wetenschappen

op donderdag 20 december 2012 om 9.45 uur

in de aula van de universiteit,

De Boelelaan 1105

door

Vanina Rea

geboren te Cercola, Italië

promotor : prof.dr. N.P.E. Vermeulen

copromotor : dr. J.N.M. Commandeur

“In the midst of winter, I finally learned that there was in me an invincible summer”.

Albert Camus

Leescommissie: prof.dr. P.D.J. Grootenhuis

prof.dr. M. Honing

prof.dr. R. Leurs

prof.dr. W.M.A. Niessen

dr.B.M.A. van Vugt-Lussenburg

The research described in this thesis was performed within the framework of project D2-

102 “Metabolic stability assessment as a new tool in the Hit-to-Lead selection process and

the generation of new lead compound libraries” of the Dutch Top Institute Pharma and

carried out in the Leiden Amsterdam Center for Drug Research (LACDR), Division of

Molecular Toxicology, Department of Chemistry and Pharmaceutical Sciences, Faculty of

Sciences and Amsterdam Institute of Molecules, Medicines and Systems (AIMMS), Vrije

Universiteit, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands.

CONTENTS

CHAPTER 1: General introduction

CHAPTER 2: Role of residue 87 in substrate selectivity and

regioselectivity of drug-metabolizing cytochrome

P450 BM3 M11

CHAPTER 3: Role of residue 87 in the activity and regioselectivity

of clozapine metabolism by drug metabolizing BM3:

application for structural characterization of

clozapine GSH-conjugates

CHAPTER 4: Active site substitution A82W improves

regioselectivity of steroid hydroxylation by

cytochrome P450 BM3 mutants as rationalized by

spin relaxation NMR studies

CHAPTER 5: Combination of biotransformation by P450 BM3

mutants with on-line post-column bioaffinity and

mass spectroscopical profiling as a novel strategy to

diversify and characterize p38 kinase inhibitors

CHAPTER 6: Summary, conclusions and perspectives

Nederlandse samenvatting

APPENDICES: List of abbreviations

List of publications

Curriculum Vitae

Dankwoord

46

2

74

98

124

152

168

CHAPTER 1

GENERAL INTRODUCTION

Chapter 1 General Introduction

3

Introduction

1. Role of drug metabolism in drug development

1.1. Metabolites in safety testing (MIST)

1.2. Enzyme involved in drug metabolism

1.3. Biosynthesis of reactive metabolites

1.4. Biological screening for active metabolites

2. Microbial conversion of steroid compounds

3. Cytochrome P450s

3.1. Human P450s

3.2. Structural organization of P450s

3.3. The catalytic cycle

4. Cytochrome P450 BM3

4.1. Engineering Cytochrome P450 BM3

4.2. P450 BM3 in biocatalysis

4.3. Key residues in the active site of P450 BM3

5. Spin lattice relaxation NMR

6. Scope and objective of this thesis

7. Outline of this thesis

4

12

15

20

27

29

30


4

1. Role of drug metabolism in drug discovery and development

Most xenobiotic substances such as drugs undergo an array of biotransformation

reactions after internal exposure to humans. Drug metabolism by the host system is one

of the most important determinants of the pharmacokinetic profile of a drug.

Biotransformation of drugs can have different effects, such as the formation of chemically

stable metabolites, which are devoid of pharmacological or toxicological activities, or the

generation of short lived chemically reactive metabolites, which can lead to toxicological

side effects (1). Furthermore, drug metabolism can also lead to formation of

pharmacologically active metabolites which might contribute significantly to drug action in

vivo, or may be responsible for pharmacologically-based adverse drug reactions.

A relatively novel approach in modern drug development is the use of pharmacologically

active metabolites as potential resources for drug discovery and development. There are

several advantages for screening drug candidates for active metabolites during drug

discovery. For example drug metabolites can show superior properties compared to the

lead itself, such as improved pharmacodynamics, improved pharmacokinetics, lower

probability of drug-drug interactions, less variable pharmacokinetics and/or

pharmacodynamics, improved overall safety profile and improved physicochemical

properties (e.g. solubility) (1).

Examples of active metabolites of previously marketed drugs that have been developed

into drugs are listed in Table 1.

Table 1. Pharmacologically active metabolites developed as new drugs (2)

Parent drugs Metabolite drugs Biotransformation Brand name

Allopurinol Oxypurinol Oxidation of xanthine Oxyprim

Amitriptyline Nortryptyline N-Demethylation Aventyl

Bromhexine Ambroxol N-Demethylation &

Hydroxylation

Mucosovan

Diazepam Oxazepam N-demethylation &

Hydroxylation

Serax

Etretinate Acitretin Deesterification Soriatane

Hydroxyzine Cetirizine Carboxylation Zyrtec

Imipramine Desimipramine N-Demethylation Norpramin

Loratadine Desloratadine Descarboethoxylation Clarinex

Loxapine Amoxapine N-Demethylation Asendin

Phenacetin Acetaminophen O-Deethylation Tylenol

Terfenadine Fexofenadine Carboxylation Allegra

Thioridazine Mesoridazine S-Oxidation Serentil


5

For example, fexofenadine, the primary metabolite of terfenadine was developed as anti-

histamine agent with selective peripheral histamine H1-receptor antagonist activity.

Fexofenadine is a safer drug than terfenadine because it does not has the ability to cause

cardiotoxicity by hERG inhibition.

In some instances, metabolic transformation can also produce reactive or toxic

intermediates or metabolites, with potential toxicological implications (3). Hence, a good

understanding of the metabolism of a new chemical entity is needed early in the drug

discovery process.

1.1. Metabolites in safety testing (MIST)

Drug metabolism and pharmacokinetics are crucial factors for successful drug

development. Inappropriate pharmacokinetics previously was one of the main reasons of

attrition in drug development. Although inappropriate pharmacokinetics has become a

less important reason for attrition, less progress has been made in decreasing the attrition

due to drug toxicity. The preparative synthesis of drug metabolites is currently of primary

importance in industry in order to assess potential toxicity, drug-drug interactions and to

examine metabolic pathways (4). In fact, testing the toxicities and biological activities of

human metabolites of drugs is important in drug development to assure effectiveness and

safety (5).

The importance of metabolite identification and quantification is stressed by the

guidelines published in 2008 by the US food and drug administration (FDA). The so-called

Metabolites in Safety Testing (MIST) guidelines describe the studies recommended to

perform on metabolites having more than 10% systemic exposure of the parent drug to

support human safety. A decision diagram of this is depicted in Figure 1.

These guidelines challenged the development of biocatalytic systems for the generation of

large amounts of drug metabolites as well as analytical technologies dealing with their

quantification and identification of drug metabolites in an efficient and intelligent way (7).


6

Figure 1. Decision tree flow diagram of the MIST guidelines as determined by the FDA, based on (6).

1.2. Enzymes involved in drug metabolism

Drug metabolism is typically divided in two phases. Phase I metabolism includes many

types of functionalization reactions, e.g. oxidation, reduction, hydrolysis, hydration and

dehalogenation reactions (9). Cytochrome P450s represent the most important class of

enzymes involved in phase I metabolism, being involved in 75-80% of metabolism of

marketed drugs, Figure 2. Other phase I enzymes include monoamine oxidases, flavin-

containing oxygenases, amidases and esterases (9). Phase II metabolism involves

conjugation of polar groups (e.g. glucuronic acid, sulfate, and amino acids) to drugs and/or

phase I drug metabolites to further increase their hydrophilicity and facilitate excretion.

These biotransformation reactions involve glucuronidation, sulfation, GSH conjugation,

acetylation, amino acid conjugation and methylation reaction (10). Phase II enzymes

include UDP-glucuronosyltransferase (UGTs), sulfotransferases (STs), N-acetyl transferases

(NATs), methyl transferases and glutathione S-transferases (GSTs) (10), Figure 2.

Almost all phase I and phase II enzymes can catalyze the formation of stable metabolites,

as well as chemically reactive intermediates that may lead to the generation of adverse

drug reactions (11).


7

Figure 2. Phase I and phase II metabolism of currently marketed drugs (adapted from (8)). The percentage of phase I and phase II metabolism of drugs that each enzyme contributes is estimated by the relative size of each section of the corresponding chart. ADH, alcohol dehydrogenase; ALDH, aldehyde dehydrogenase; DPD, dihydropyrimidine dehydrogenase; NQO, NADPH quinine oxidoreductase. COMT, catechol O-methyltransferase; GST, glutathione S-transferase; HMT, histamine methyltransferase; NAT, N-acetyltransferase; TPMT, thiopurine methyltransferase; UGTs, uridine 5’-triphosphate glucuronosyltransferases.

1.3. Role of reactive metabolites in adverse drug reactions

Adverse drug reactions (ADRs), as defined by the World Health Organization, are noxious,

unintended and undesirable effects of a drug, which occurs at therapeutical doses used in

humans and which require hospitalization of patients to recover (11). Although much

effort has been spent to the development of safer drugs, ADRs still constitute a major

reason for attrition in drug development and withdrawal or restriction of marketed

products. Therefore, prevention of ADRs constitutes a major challenge for the

pharmaceutical industry (3).

Approximate 75% of ADR in patients are classified as Type A ADR which involve life-

threatening exaggerated pharmacological activity (on-target and/or off-target) due to

unanticipated increased plasma concentrations of the parent drug. This class of ADR often

results from drug-drug interactions or genetic deficiency of an enzyme which is involved in

the major pathway of metabolism. For the other classes of ADR formation of reactive drug

metabolites are considered to play a crucial role (12).

The main difficulty for pharmaceutical industry is the fact that ADRs cannot always be

predicted in preclinical animal studies. In particular idiosyncratic drug reactions (IDRs;

Type B ADR) are difficult to predict as they normally have a very low incidence, have a


8

delayed time of onset, do not necessarily show classic dose-response relationships and are

not predictable from the known pharmacology of the drug. No predictive animal model is

currently available and IDRs therefore remain poorly understood.

Figure 3 shows a general scheme depicting the role of reactive metabolites in the

development of toxicity. Drugs bioactivation can be catalyzed by both phase I and phase II

enzymes.

Different types of reactive metabolites can be distinguished such as electrophiles, radicals

and reactive oxygen species. Electrophilic metabolites can react with nucleophilic sites in

proteins and/or DNA. Free radicals possess unpaired electrons and can abstract hydrogen

atoms from polyunsaturated fatty acids in membranes which will initiate the destructive

process of lipid peroxidation, and can disrupt the cellular redox-state leading to oxidative

stress and subsequent toxicity (12).

Figure 3. Scheme depicting the role of drug metabolism in determining the toxic outcome in ADRs.

Table 2 gives examples of drugs whose toxicities are attributed to covalent binding of

electrophilic metabolites to proteins. Bioactivation of xenobiotics which contain certain

functional groups such as tertiary amine, furan ring or acetylene group, can also cause

mechanism-based inactivation of P450, leading to adverse clinical drug-drug interactions

(13-14).


9

Table 2. Examples of drugs and reactive metabolites possibly involved in ADRs (12-15).

Drugs Reactive intermediate Toxicity

Acetaminophen Quinone-imine Hepatotoxicity

Carbamazepine 2-Hydroxy, Quinone-imine Agranulocytosis, Aplastic

anemia

Clarithromycin Nitroalkane Hypersensitivity

Clozapine Nitrenium ion Agranulocytosis

Dapsone Hydroxulamine, Nitroso Hemolysis, hypersensitivity

Diclofenac Acylglucuronide, Benzoquinone imine Hepatotoxicity

Halothane Trifluoroacetyl Hepatitis

Indomethacin Iminoquinone Hypersensitivity,

Agranulocytosis,

Hepatotoxicity

Isonizaid Isonicotinic acid, acetylating species Hypersensitivity

Phenacetin p-Nitrosophenetole Hepatotoxicity

Procainamide Hydroxylamine, Nitroso Agranulocytosis

Tacrine 7-OH-tacrine Hepatotoxicity

Tamoxifen N-oxide, N-oxide-epoxide Carcinogenesis

Ticlodipine Keto, S-oxide Agranulocytosis, aplastic

anemia

Tielinic acid Thiopene S-oxide Hypersensitivity, hepatitis

Troglitazone Conjugates, Benzoquinone, Quinone

epoxide

Hepatotoxicity

Valproic acid Acyl glucuronides Hypersensitivity,

Hepatotoxicity

Detection of reactive metabolites is difficult due to the chemical reactivity of the short-

lived intermediates and to the usually low amounts present in incubations. The most

common way to screen for formation of reactive metabolites is to trap them with model

nucleophiles such as GSH and to analyze the formed adducts with spectroscopic

techniques. The endogenous tripeptide GSH can trap different types of RIs including

quinones, quinoneimines, iminoquinone, methides, epoxides, areneoxides and nitrenium

ions (15). The corresponding GSH adducts are typically analyzed by liquid-chromatography

mass spectrometry (LC-MS). However, a major disadvantage is that structural information

of the adducts obtained by LC-MS/MS experiments is limited, as it may be insufficient in

identifying the exact position of oxidation, to differentiate stereoisomers, or to provide

the exact structure of the metabolite (16). Complementing MS data with Nuclear

Magnetic Resonance (NMR) experiments is required for complete structural elucidation of


10

drug metabolites and/or resulting adducts. A limitation of NMR, however, is its relative

insensitivity, as it requires several milligrams of pure metabolite.

The generation of pure metabolites can be achieved by organic synthesis,

electrochemical oxidation of the parent drug or by biocatalysis using appropriate enzymes.

The synthesis by organic chemistry is often complicated by the lack of appropriate

synthetic routes for specific metabolites and may yield only low amounts of the desired

product that has to be purified subsequently.

Electrochemical oxidation has been used to generate drug metabolites and has the

advantage that it can be easily scaled up. However, with this technique not all the relevant

enzymatic metabolites can be formed. Electrochemistry has been previously used to

generate GSH adducts of clozapine (17) and troglitazone (18).

Biocatalysis by human or bacterial P450s can also be used for the scaling up of metabolite

production (19). For example, human relevant diclofenac metabolites were synthesized by

microbial fermentation in large amount (up to 170 mg) and their structures were

elucidated by MS and NMR (20).

More recently the use of highly active bacterial P450 BM3 mutants for the generation of

reactive metabolites of drugs has been explored. Damsten et al. showed that BM3

mutants could be used for the generation of human relevant reactive metabolites of

clozapine, acetominophen, diclofenac (21), and trimethoprim (22). Boerma et al. showed

that P450 BM3 mutants with high capacity to activate drugs (clozapine, acetaminophen

and troglitazone) into relevant reactive metabolites can be employed to produce protein

adducts to study the nucleophilic selectivity of highly reactive electrophiles (23). Dragovic

et al. applied highly active BM3 mutant as bioactivation system to study the role of human

GSTs in the inactivation of reactive intermediates of clozapine (24).

1.4. Biological screening for active metabolites

To assess the biological activity of drug metabolites there are several approaches

available. The traditional approach is to isolate and purify all the metabolites from in vitro

incubations with subcellular fractions such as liver microsomes or intact cellular systems

(e.g. epatocytes) containing a full complement of drug metabolizing enzymes.

An alternative approach is to use bioassay-guided methods where biological samples

containing biotransformation products are first evaluated for their pharmacological

activity prior to isolation of the metabolites. The bioassay methods may be based on the

assessment of the pharmacological activity using in vitro ligand binding (25, 26), cell based

assays (1), or in vivo pharmacological assays (1).

One of the major bottlenecks in in vitro assays is the lack of possibility to screen affinity or

activity of individual compounds in complex mixtures.

In general, two approaches can be used to measure the effect of individual

compounds in a mixture. One strategy is to combine HPLC with fractionation techniques


11

to enable the compounds in the mixture to be separated and allow screening of the

individual components using plate reader assays.

The other available strategy is the on-line high resolution screening technique (HRS)

depicted in Figure 4 (27-29).

HRS enables the screening of individual compounds in complex mixtures by coupling a

separation technology, usually gradient HPLC, to post-column biochemical detection

assays on-line. This technique has already been successfully used to develop a number of

antibody- (30) and biotin- (31), receptor- (32), enzyme- (33) (34) (35), and MS-based (36)

screening assays. The HRS platforms are especially attractive when complex mixtures such

as metabolic incubations (37), combinatorial mixtures (38), or natural extracts (39) are

screened. In contrast to conventional high-throughput screening (HTS), isolation of

individual metabolites not showing affinity can be avoided. Thus, the laborious

development of high-yield synthesis and isolation methods can be limited to compounds

with the desired affinity.

Figure 4. Scheme of the online setup. The system enables the separation of mixtures and subsequent parallel detection of enzyme binding and accurate mass. The samples are injected into and separated by an HPLC system. The eluent is split between HR-MS and enzyme binding detection. The fraction which enters the enzyme binding detection is

mixed with the target (e.g. p38) and the tracer and incubated after each mixing step, 24 s for the target–ligand interaction and 11 s for the target–tracer interaction. Finally, the fluorescence is measured as readout of affinity towards the target. In parallel, the second part of the eluent is analyzed by HR-MS delivering structural information to identify the small molecules tested. Adapted from (28).


12

2. Microbial conversion of steroid compounds

Totally, about 300 approved steroid drugs are known to date, and this number tends to

grow (40). Steroid pharmaceuticals are ranked among the most marketed pharmaceuticals

and represent the second largest category next to antibiotics.

The research on steroid drugs started in 1950, with the discovery of the pharmacological

effects of cortisol and progesterone, two endogenous steroids, and with the identification

of the 11-hydroxylation activity of a Rhizopus species (41-42). Microbial steroid

transformation is a powerful tool for generation of novel steroidal drugs, as well as for

efficient production of steroid-containing active pharmaceutical ingredients (APIs) and key

intermediates.

Steroid hydroxylation is a key tool to develop steroid drugs with improved potency, longer

half-lives in the blood stream, simpler delivery methods and reduced side effects (43). The

therapeutic area of hydroxy-steroids is very wide: they are used as anti inflammatory,

immunosuppressive, progestational, diuretic, anabolic and contraceptive agents (41, 44-

47). Applications are found in the treatment of adrenal insufficiencies, in the prevention of

coronary heart disease, as anti-fungal agents, as anti-obesity agents, and in the inhibition

of HIV integrase, the prevention and treatment of infection by HIV and in the treatment of

declared AIDS (43).

The complex structure of the steroid molecule requires complicated, multi-step schemes

for the chemical synthesis of steroid compounds. The selective oxidation of an

unactivated, aliphatic C-H bond to an alcohol is a challenging problem in synthesis (48).

Chemical synthesis requires the use of reagents as pyridine, sulfur trioxide or selenium di-

oxide, whose disposal constitutes an environment issue (43). Moreover, the chemical

catalysts used often show very low regioselectivity.

Steroid hydroxylation by microorganisms represents a powerful alternative to chemical

synthesis. Current trends in microbial hydroxylation studies cover the search for novel

biocatalysts capable of performing reactions targeting specific positions in the steroid

molecule that are particularly important for the pharmaceutical industry (7α, 9α, 11α,

11β, 16α, 17α), production of novel hydroxysteroids with potent therapeutic properties,

and the construction of novel biocatalysts by genetic engineering (40).

Examples are given in Table 3.

Cytochrome P450s play a very important role in endogenous steroid metabolism.

Testosterone hydroxylation in particular has been studied extensively with human liver

microsomal P450s. For example, CYP3A4, the most abundant P450 in human liver and

small intestine, catalyzes the monohydroxylation at 10 different positions of testosterone

(8). Hence testosterone hydroxylation patterns have been utilized as probes for the

presence and characterization of P450s (49) (50).

Hydroxylation of testosterone at 1-, 2-, 2-, 6-, 6-, 7-, 7-, 15-, 15-, 16-, 16-

and 17-positions has been reported in human and rat liver microsomes (Figure 5).


13

Table 3. Examples of steroid hydroxylation in microrganisms

Substrate Product Microrganism Ref.

7-Hydroxylation

Testosterone 7-OH-testosterone Botrytis cinerea AM235 (49)

Pregnenolone 7-OH-pregnenolone Fusarium oxysporum var.cubense

Epiandrosterone 7-OH-epiandrostenone Mortierella isabellina AM212 (50)

7-Hydroxulation

DHEA 7- and 7-derivatives M.racemosus (51)

7-OH DHEA Botryodiplodia malorum CBS 134.50 (52)

7-15-Dihydroxylation

DHEA 7-15-diOH-DHEA (3,7,15-triOH-androst-5-ene-17-one)

C.lini CBS 112.21 (52)

7-15-Dihydroxylation

17-OH-progesterone 8-OH-derivative, along with

15-hydroxy derivative

Corynespora cassiicola CBS 161.60 (53)

9-hydroxylation

Testosterone 9-OH-AD, 9-OH-testosterone R.equi ATCC 14887 (54)

Progesterone 9-OH-progesterone Expression of 9-hydroxylase from M.smegmatis in E.coli BL21

(55)

Androsterone 9-OH-adrenosterone, 9-OH-11-keto-testosterone

Cunninghamella elegans TSY-0865 (56)

11-Hydroxylation

Testosterone 11-OH-testosterone Rhizopus stolonifer ATCC 10404, Fusarium lini NRRL 68751

(57)

11-hydroxylation

Progesterone 11-OH-progesterone T.harzianum, T.hamatum (58)

14-Hydroxylation

Testosterone 14-OH-testosterone A.wentii MRC 200316 (59)

15-Hydroxylation

Testosterone 15-OH testosterone F.oxysporum var. cubense (60)

Progesterone 15-OH progesterone

15-Hydroxylation

Testosterone 15-OH-testosterone Solvent tolerant Pseudomonas putida S12 expressing B.megaterium steroid hydroxyase CYP106A2

(61)

16-Hydroxylation

Testosterone 2,16,17-triOH-androst-4-en-3-one

Whetzelinia sclerotiorum ATCC 18687

(62)

17-Hydroxylation

Progesterone 17-OH-progesterone B.sphaericus ATCC 245, B.sphaericus ATCC 7063, B.sphaericus ATCC 13805

(63)


14

Figure 5. Hydroxylation of Testosterone by human P450s (adapted from (51))

Figure 6. Testosterone hydroxylation by bacterial P450s (adapted from (40)).


15

Hydroxylation of Testosterone by bacterial P450s has also been extensively studied (Figure

6) (51). Profiling this hydroxylation provides useful information for characterizing bacterial

P450s as monogenic traits and for comparing bacterial P450s with human P450s.

The features and functionality of steroid hydoxylases have been recently reviewed by, e.g.

Bernhardt (2006) (52), Hannemann et al (2007) (53), Novikova et al. (2009) (54) and

Kristan and Lanisnik-Rizner (2012) (55).

3. Cytochrome P450s

Cytochrome P450s (P450s or CYPs) constitute a family of monoxygenases involved in the

biotransformation of drugs, xenobiotics, alkanes, terpenes, and aromatic compounds, in

the metabolism of chemical carcinogens, the biosynthesis of physiologically relevant

compounds as steroids, fatty acids, eicosanoids, fat-soluble vitamins, bile acids and in the

degradation of pesticides (52).

They are able to catalyze a number of difficult oxidative reactions, such as C-H bond

hydroxylation, N-dealkylation, N-hydroxylation, O-dealkylation, S-oxidation and

epoxidation of numerous endogenous and exogenous compounds (3). An overview of the

different reactions catalyzed by P450s is listed in Table 3.

The broad substrate acceptance and this huge diversity in catalyzed reactions make the

class of P450 very promising biocatalysts in industrial processes. In fact, in recent years

there has been an increasing interest in the application of CYP biocatalysts for the

industrial synthesis of bulk chemicals, pharmaceuticals, agrochemicals, and food

ingredients, especially when a high grade of stereo and regioselectivity is required (57).

Moreover P450s are of great interest in drug metabolism. They are ubiquitous enzymes

often involved in the oxidation of drugs resulting in the generation of metabolites that are

more easily excreted, thus modulating the toxicity of these compounds (57).

3.1. Human P450s

The human genome encodes for 57 functional P450s (58) of which approximately a dozen

are involved in drug metabolism. The P450s mostly involved in metabolism of drugs in

humans are listed in Table 5, where the main reaction and substrate specificities are

indicated. Genetic polymorphisms that may alter the enzyme activity have been reported

for a number of P450s (59). While for the P450s 1A2, 2A6, 2B6 and 2C8 the clinical effects

due to these polymorphisms are minor, major effects are observed for 2C9, 2C19, and 2D6

(60). Polymorphisms in these isoenzymes affect metabolism of 20% to 30% of the clinically

used drugs and this has become a major issue in the discovery and development of drugs

and other NCEs (61).


16

Table 4. List of different reactions catalyzed by P450s (56)

Reaction type Example

Aliphatic hydroxylation

Aromatic hydroxylation

Epoxidation

Heteroatom dealkylation

Alkyne oxygenation

Heteroatom oxygenation

Aromatic epoxidation and

NIH-shift

Dehalogenation

Dehydrogenation

Reduction

Cleavage of esters


17

Table 5.Substrate specificity of human CYPs

CYP Substrate properties Drug/Substrates Reaction

CYP1A2 Planar, polyaromatic

compounds

Clozapine

Phenacetine

Olanzepine

Imipramine Propranolol

Theophylline

Oxidation

O-deethylation

Oxidation

Demethylation

4-hydroxylation

Hydroxylation

CYP2B6

Cyclophosphamide

Efavirenz

Nevirapine

Bupropion

Artemisisin

Methadone

Profofol

Hydroxylation

Hydroxylation

3-hydroxylation

Hydroxylation

Unclear

N-demethylation

Oxidation

CYP2C9 Polar, weakly acidic

compounds

Diclofenac

Flurbiprofen

Ibuprofen

Naproxen

Phenytoin

Piroxicam

Tolbutamide

Warfarin

4-Hydroxylation

4-Hydroxylation

Oxidation

O-demethylation

4-Hydroxylation

5-Hydroxylation

Hydroxylation

7-Hydroxylation

CYP2C19 Polar acidic compounds

Amitriptyline

Cyclophosphamide

Diazepam

Imipramine

Omeprazole

Phenytoin

Demethylation

Oxidation

Demethylation

n-demethylation

Demethylation

4-Hydroxylation

CYP2D6

Lipophilic, basic,

medium sized,

positively charged at

neutral pH

Amitriptyline

Imipramine

Propranolol

Codeine

Dextromethorphan

Desipramine

Bufaralol

Hydroxylation

2-Hydroxylation

4-Hydroxylation

O-Demethylation

O-Demethylation

Hydroxylation

Hydroxylation

CYP2E1 Small molecular weight

compounds

Acetaminophen

Chlorzoxazone

Dehydrogenation

6-Hydroxylation

CYP3A4

Lipophilic, large

molecular weight

compounds

Alprazolam

Carbamazepine

Testerone

Cyclosporine

Midazolam

Simvastatin

Triazolam

Diazepam

Hydroxylation

Epoxydation

6-Hydroxylation

Oxidation

Oxidation

1-Hydroxylation

Hydroxylation

N-demethylation


18

3.2. Structural organization of P450s

All P450s are heme-containing proteins containing an iron atom at the center of the heme

which is also complexed by a cysteine residue as axial ligand. Based on their redox partner,

P450s can be classified in different classes (Figure 7):

Figure 7. Structural organization of P450s: (a) Class I (mostly bacterial P450s), (b) Class II (mostly microsomal P450s) and (c) P450 BM3. (Adapted from (62))

Class I P450s (a) require two redox proteins: a small redox 2Fe-2S iron-sulfur ferredoxin

and a FAD or FMN containing reductase. Mitochondrial P450s and most soluble bacterial

P450s belong to this class (63).

In class II (b), electrons are transferred via only one reductase having a FAD and a FMN

domain. Microsomal P450s that are attached to the endoplasmatic reticulum belong to

this class (63).

Most bacterial P450s belong to class I. However, there are exceptions: in 1981

Cytochrome P450 BM3 (CYP102A1) from Bacillus megaterium was identified. P450 BM3

(c) is soluble and uses a Class II redox system. This soluble bacterial enzyme contains the

P450 monoxygenase domain fused to the electron transfer flavin mononucleotide

(FMN)/flavin adenine dinucleotide (FAD) reductase domain in a single polypeptide (63). In

contrast, mammalian CYPs require additional redox partners as cytochrome P450

reductase (CPR) and often cytocrome b5 and lipids (64, 65).


19

3.3. The catalytic cycle

The general catalytic cycle for substrate hydroxylation by P450s is depicted in Figure 8.

Figure 8. The catalytic cycle of P450s, adapted from (66)

In the resting state, in absence of substrate, the ferric iron atom is six-coordinated,

complexed with a water molecule (I). Binding of a substrate displaces the water molecule

(II). This causes the ferric spin state to convert from low spin to high spin, a transition that

can be monitored spectrophotometrically since the Soret band undergoes a “type I” shift

from ca. 418 nm to ca. 390 nm. A single electron supplied by a reduced pyridine

nucleotide (NADPH or NADH) reduces the heme iron to the ferrous state (III).

Subsequently, dioxygen binds to the ferrous iron generating the oxy-complex (IV). A

second electron is then transferred, creating a ferric peroxy complex (V). Protonation of

the terminal oxygen atom gives “Compound 0”, a hydroxyperoxy adduct (VI). Addition of a

second hydrogen atom followed by loss of water gives “Compound I” (VII), a ferryl species

thought to be the active entity in most P450 oxidations, which has been characterized

recently (67). This complex abstracts hydrogen from the substrate (VIII) after which the

hydroxyl group formed rebounds to the substrate radical, resulting in formation of the

hydroxylated product which releases from the active site. NAD(P)H consumption is not

necessarily fully coupled to product formation (68). If dioxygen binding is hindered, the

second electron transfer will be too slow, resulting in the loss of superoxide from (IV)


20

(superoxide uncoupling). If substrate fitting in the active site cannot prevent water

encroachment, (VI) will be protonated at the iron-bound oxygen, resulting in the loss of

hydrogen peroxide (peroxide uncoupling). If a substrate binds with no hydrogen atom

conveniently positioned for abstraction, the oxygen atom in (VII) could be reduced to

water (oxidase uncoupling). Those alternative pathways can hamper the use of P450s in

biocatalysis because the expensive electron donating cofactor is used in futile redox

cycling instead of product formation. Moreover the peroxide and superoxide formed by

uncoupling can damage the biocatalyst.

4. Bacterial Cytochrome P450 BM3

Bacterial P450 systems often exhibit much higher catalytic activities than membrane-

bound eukariotic P450s and are easy to handle in the laboratory due to their solubility and

high expression level in heterologous hosts (69). In particular Cytochrome P450 BM3

(BM3) is considered as one of the most promising monoxygenases for biotechnological

applications as it possesses the highest activity ever recorded for a P450 (63, 70, 71) and

can be easily expressed at high yield in Escherichia coli.

The natural substrates of BM3 are C12-C18 fatty acids which are hydroxylated at very high

activity at subterminal positions (72). However, over the recent years, several research

groups have succeeded in expanding the substrate selectivity of P450 BM3 and improving

its catalytic properties by site-directed and/or random mutagenesis. Through rational

redesign and directed evolution, BM3 mutants have been obtained that are able to oxidize

aromatics (73), alkanes (74), hydrocarbons (75), carboxylic acids (76) and pharmaceuticals

(21, 22, 77-84).

Recently a large body of research has been conducted to broaden and alter the substrate

specificity of bacterial CYPs in an effort to generate “human like” P450 activities (57). The

demand of “humanized” bacterial P450 with high activity is constantly growing to meet

the needs of the fine-chemical synthesis, pharmaceutical production and bioremediation

industries (85).

4.1. Engineering P450 BM3

The application of engineered bacterial P450s also have application in pharmacology and

toxicology as it enable the generation of substantial amounts of drug metabolites as well

as in optimization of lead compounds (86). Figure 9 shows examples for substrates of

engineered cytochrome P450 BM3 variants.

In order to improve the selectivity of oxidation to a specific product, the substrate from

which it is formed must be constrained to bind in a particular orientation (87). This can be

obtained by rational redesign of the active site, for example introducing polar residues or

bulky side-chains (88-90).


21

Figure 9. Examples for substrates of engineered cytochrome P450 BM3 variants. (Adapted from (62)).

Directed evolution, in which mutations are introduced at random positions using error-

prone PCR, is a valuable complementary tool to site-directed mutagenesis. This approach

has the advantage of being able to generate unpredictable mutations with unanticipated

beneficial properties. The benefits of the two strategies can be combined using rational

evolution, in which selected residues are subjected to random or semi-random

mutagenesis. Alternatively, directed evolution can be used to identify sites suitable for

targeting, followed by site-saturation mutagenesis to establish which mutations are most

effective in those positions.

For example, Dietrich et al. (91) developed a novel semi biosynthetic route for the

production of artemisinin, a highly important antimalarian agent. By rational redesign of


22

the active site they introduced mutations F87A/A328L in the active site of substrate

promiscuous P450 BM3: mutation F87A appeared to relieve the steric hindrance imposed

upon the substrate amorphadiene and allowed an increased access to the heme group,

while mutation A328L decreased the mobility of amorphadiene in the active site,

promoting epoxidation.

A similar approach has been used by Seifert et al. (75), who focused on the same hotspots

F87 and A328, creating a minimal and highly enriched P450 BM3 mutant library, by

mutating those two positions to all possible combinations of five hydrophobic aminoacids.

In this way, by systematically altering the size of the side chains in those two positions, a

broad range of binding site shapes was generated to convert a range of differently sized

and shaped terpenes (75). Recently, Seifert et al. (92) set up an iterative approach

comprising successive rounds of modeling, designing of focused mutant libraries and

screening, to identify a mutant for the production of the highly valuable product perillyl

alcohol.

It is widely accepted that increased regio-, stereoselectivity is the result of a reduced

number of substrate orientations close to the heme, as discussed in chapter 4 of the

present thesis (82, 90, 93).

Considerable effort has been invested for improving the coupling efficiency (ratio of

product formed to NAD(P)H cofactor consumed) of P450 BM3 by rational mutagenesis.

However the efficiencies obtained are still far lower than with native enzyme-substrate.

Detailed information on the structural determinants of coupling with novel substrates are

missing, therefore it is difficult to rationalize mutations to improve coupling efficiencies.

However, cost-effective, high-throughput screening methods will need to be developed to

assess relative coupling efficiency among P450s in mutant libraries (86).

Properties that need to be improved to enable the application of P450s in industrial

processes are: TTN (total turnover number), substrate acceptance, regio-and stereo-

selectivity, activity (kcat, KM), inhibition and coupling efficiency (70). Also improvement of

physical properties as thermostability, solvent tolerance, oxidative stability and substrate

and product tolerance should be addressed (94).

4.2. P450 BM3 in biocatalysis

One emerging application of P450s for the pharmaceutical industry is the generation of

large amount of drug metabolites, as discussed above (86). Several groups showed drug

metabolite preparation using human P450s expressed in Escherichia coli and in insect cells

(95). The Arnold group showed that human metabolites of propranolol could be produced

by mutants of bacterial P450 BM3 heme domain in a reaction driven by hydrogen

peroxide (5). Yields obtained were comparable to those produced by recombinant human

P450s in bacterial or baculovirus bioreactors (96). Recently, Dubey et al. set up a

biotechnological production of anticancer drug (colchicine derivatives) using P450 BM3 as

biocatalyst obtaining a yield of 7.3 g/L in 70 L fermentor (97). Kim at al. applied BM3 wt


23

and a set of mutants for the generation of the human metabolite piceatannol from the

anticancer-preventive agent resveratrol (77), for the generation of human chiral

metabolites of simvastatin and lovastatin (79) and for the generation of human

metabolites of 7-ethoxycoumarin (78). The Arnold group compiled a panel of 120 mutants

to prepare nearly all human metabolites and a number of novel hydroxylated derivatives

for the drugs verapamil and astemizole (81). This panel of enzymes could also be applied

for the generation of new chemical entities (NCEs) to diversify lead compounds. Within

any catalyst panel both extremes of regioselectivity can be useful: enzymes selectively

producing individual metabolites and less selective enzymes to survey metabolite

possibilities (81).

A recent area of interest is the design and development of novel P450s as biocatalysts of

reactions of commercial interest, such as in pharmaceutical or fine chemical synthesis (86)

as well as for the optimization of lead compounds.

An example of the application of BM3 mutants in chemical synthesis is the

chemoenzymatic elaboration of monosaccharides applied by the Arnold group (98).

Regioselective deprotection of monosaccharide substrates using engineered P450 BM3

demethylases provides a highly efficient method to obtain valuable intermediates that can

be converted to a wide range of substituted monosaccharides and polysaccharides (98).

Engineered Cytochrome P450 BM3s were also applied for the enantioselective -

hydroxylation of 2-arylacetic acid derivatives and could produce an authentic human

metabolite of buspirone with high activity and selectivity (80).

Rentmeister et al. showed that products hydroxylated by BM3 can be subsequently

fluorinated in a chemo-enzymatic process that greatly simplifies the insertion of fluorine

into these structures at specific positions (99). Fluorination has become an important tool

in drug discovery and development as it can modulate the pharmacokinetic and the

pharmacological properties of drugs and lead compounds. In particular, fluorination can

improve membrane permeability, metabolic stability and/or receptor-binding properties

of bioactive molecules (99).

4.3. Key residues in the P450 BM3 active site

Control of regio- and stereoselectivity of biocatalysts is one of the major challenges in

biotechnology, as a limited number of residues in the substrate pocket appear to possess

significant selectivity-influencing powers. Ala82 and Phe87 are among those.

In Figure 10 key residues of the substrate channel and the active site of substrate-bound

P450 BM3 are shown. P450 BM3 crystallizes in a “precatalytic conformation” in which the

substrate is located too far to the heme iron for the oxidation to take place (87). Therefore

it is unclear whether the residues located in the substrate access channel are also

catalytically relevant. Also the spin relaxation NMR experiments of reduced P450 BM3

suggested that the substrates are 6 Å closer to the porphyrin ring than in the oxidized

form, indicating that structural rearrangements occur in order to allow the substrate to


24

move closer to the heme iron (100). In the present thesis mutagenesis studies at position

87 (101) (84) and position 82 (89) will be presented.

Figure 10. The substrate access channel (A) and active site (B) of NPG-bound P450BM3 (Adapted from (66))


25

Phe 87. Phe 87 is one of the most studied residues of the active site of BM3 and it is object

of chapter 2 and 3 of the present thesis. In the substrate-free crystal structure, this

residue lies perpendicular to the porphyrin system, but in the substrate-bound crystal, the

orientation of the aromatic ring changes to parallel to the plane of the heme (102) (103)

(63). It has been proposed that this motion influences the orientation of the substrate

relative to the catalytic center and it is responsible for the displacement of the axial water

ligand from its coordination site (104).

Phe87 forms with Phe81 a small hydrophobic pocket within the active site that sequesters

the -end of fatty acids. It seems likely that small hydrophobic substrates could

preferentially bind in the pocket generating non-productive complexes (105).

Phe87 is the most commonly mutated residue of P450 BM3 (Table 6).

Site-directed mutagenesis studies proved that Phe87 plays a key role in the control of

reaction selectivity (106) (107).

Typically, Phe87 was mutated to aminoacids with less bulky side chains, in order to

destroy the hydrophobic pocket, creating more space close to the heme, to accomodate

larger substrates (73-75, 77, 80, 103, 105, 108-116). Such substitutions, however, lead to a

decreased coupling efficiency of NADPH consumption to product formation due to a less

efficient exclusion of water from the active site (105).

The F87V mutation is a common component of several drug metabolizing variants (69)

(117). Moreover, it has been shown that mutation F87V significantly affected the

stereoselectivity of arachidonic acid epoxidation (106) and the activity towards indoles

(118). Li et al. also showed that mutation F87V is beneficial for oxidation of polycyclic

aromatic hydrocarbons, such as naphthalene, fluorene, acenaphthene, acenaphthylene,

and 9-methylanthracene (119).

In the present thesis saturation mutagenesis was applied to evaluate the effect of all

possible amino acids in this position on the metabolism of alkoxyresorufins, testosterone

and clozapine (84, 101).

Ala 82. The existence of a hydrophobic pocket between Phe81 and Phe87 clearly affects

the attempts to engineer BM3 towards novel target substrates. Huang et al. (90) were the

first to try to “fill” the hydrophobic pocket rather than destroying it, by mutating Ala82 to

large hydrophobic residues like Phe and Trp.

The mutants A82F and A82W showed a remarkable increase in affinity (800-fold) for fatty

acids. Moreover, the efficiency of indole hydroxylation increased, due both to increased

Kcat/KM values and to increased coupling efficiency, achieved by a more efficient exclusion

of water from the active site. Mutation at this position will be object of chapter 4 of the

present thesis.


26

Table 6. Reported mutations at position Phe87:

ALA (29 )(73-75) ( 81) (98, 99) (107 )(120-134)

ARG (101, 135)

ASN (101)

ASP (101)

CYS (101, 133)

GLN (101, 135)

GLU (101, 135)

GLY (5, 74, 81, 98, 101, 122, 130, 135-143)

HIS (101, 135)

ILE (75, 81, 92, 98, 99, 135, 144, 145)

LEU (75, 81, 92, 121, 124, 135, 144, 145)

LYS (101, 135)

MET (101, 135)

PRO (101, 135)

SER (101, 135)

THR (101, 135)

TRP (101, 135)

TYR (101) (106) (135) (136) (140) (146-147)

VAL (69,74, 75, 81, 82, 98, 99, 106, 118, 119, 142, 144, 145,148-153)

SS (5, 91, 101, 118, 144, 145, 154, 155)

Table 7. Reported mutations at position Ala82:

CYS (81, 135, 144)

GLY (81, 99, 144, 155)

ILE (81, 90, 144, 156)

LEU (5, 80, 81, 99, 121, 134, 144, 156, 157)

PHE (81, 90, 144, 156, 158)

PRO (5, 81)

SER (81, 144, 155)

THR (81, 144)

TRP (90, 99, 135, 159, 160)

VAL (156)

SS (5, 144, 155)


27

5. Spin lattice relaxation NMR (or T1 paramagnetic relaxation NMR)

In order to understand the structural basis that governs the regioselectivity of the

metabolism of a certain substrate or the specificity of a given P450 for a particular drug, it

is important to determine the structure of the substrate binding sites. For that purpose 1H

NMR appears as a powerful and attractive technique, since measurements of

paramagnetic relaxation can be used to determine the distances between the heme iron

and protons of the bound substrate in P450-substrate complexes (161).

The heme iron atom of a CYP is paramagnetic when in oxidized state. Hydrogen atoms in a

magnetic field will re-align in the direction of this field after their orientation has been

flipped with a radio-frequency pulse. The velocity of this re-alignment (relaxation rate) is

dependent on the local strength of the magnetic field in the direct proximity of the

hydrogen atom. So a hydrogen atom close to the heme in the active site of a CYP, will

relax faster due to the magnetic moment of the heme iron atom.

The relationship between the iron atom to hydrogen atom distance and the rate of

relaxation has been established by Solomon and Bloembergen (162) and can be used to

measure substrate orientations in CYP active sites. A disadvantage of the technique is that

high concentrations (~mM) of ligands need to be present to measure NMR spectra.

Furthermore, average substrate active site orientations do not always match the

metabolic profile (163).

An overview of reported spin lattice relaxation studies using CYPs is given in Table 6.

In chapter 4 of the present thesis this technique has been successfully applied to

determine the orientation of testosterone in two BM3 mutants (M11 and M11 A82W) to

rationalize the different regioselectivity obtained in the metabolic profile.


28

Table 8. Spin lattice relaxation NMR studies used to determine active site ligand orientations in CYPs.

*the minimal and maximal hydrogen atoms to heme iron atom measured are given. **Measured in the F483I mutant of CYP2D6, the wild type is inactive towards testosterone.

CYP Ligand Distances (A)* Major reaction Reference

102 12-Br-lauric acid 7.8 (C1H)-16.3 (C12H) C2-OH (163)

1A Paracetamol 5.9 (phenylH)-

6.7(methylH)

Benzoquinone

formation (164)

1A2 Caffeine 6.5 (N1methylH)-

6.7 (N3 and N7 methylH) N1-demethylation (163)

2B Paracetamol 5.8 (methylH)- 6.3

(phenylH) Inactive (164)

2C9 Tienilic acid 5.4 (C5H of thiophene)-

7.1 (phenylH) Thiophene-OH (165)

Lauric acid 6.0 (C4-11H) –

6.6 (C2H and C12H) -1-OH (165)

Diclofenac 6.0 (C4'H) –

6.9 (phenylacetateH) 4'-OH (165)

2D6 Codeine 3.1 (OCH3)-12.1 (NCH3) O-demethylation (166)

MPTP 3.4 (NCH3)-11.9 (C9H) N-demethylation (167)

Testosterone** 3.6 (15)- 9.2 (2 and 2) 15-OH (168)

MDMA 6.3 (methylene)- 7.2

(NCH3) O-demethylation (169)


29

6. Scope and objectives

The research presented in this thesis was part of a wider interdisciplinary project

“Metabolic stability assessment as a new tool in the Hit-to-Lead selection process and the

generation of new lead compound libraries” financed by the Dutch Top-Institute Pharma

consortium (grant D2-102). Several industrial (Merck/MSD (Oss, NL) and QPS

(Groeningen, NL)) and academic (Vrije Universiteit Amsterdam and Radboud Universiteit

Nijmegen) partners participated in this project that led to several publications and four

PhD theses.

The aim of the project was the application of new, more efficient and selective P450 BM3

mutants for the generation of biotransformation products, being new chemical starting

points for lead optimization (lead libraries), followed by the determination of their binding

affinity (e.g. competitive fluorescence detection assay) and chemical structural

characterization (LC-MS/MS-NMR) in a hyphenated approach.

The general objectives of the present thesis are:

Engineering drug metabolizing P450 BM3 mutants mimicking human P450s and

mutants with unique catalytic properties based on regio- and stereoselective

metabolism of diagnostic substrates.

Utilization of “humanized” drug metabolizing P450 BM3 mutants for large scale

production of physiologically relevant human metabolites of lead compounds.

Figure 11 shows the general strategy applied in this research: by site-directed, random or

saturation mutagenesis BM3 mutants are engineered to obtain highly active mutants able

to metabolize drugs and drug-like molecules. The metabolic mixtures obtained by

incubation of diagnostic substrates with engineered BM3 mutants are then screened to

verify the acquisition of certain properties: HRS enables the selection of mutants able to

produce metabolites that show pharmacological activity, GSH trapping enables the

identification of mutants able to produce reactive metabolites (subsequently trapped by

GSH), screening by UPLC allows the identification of changes in regio- and stereoselectivity

in a quick and informative way.

Selected mutants that acquired a certain property (e.g. production of high amounts of a

pharmacologically active metabolite, high selectivity for the production of reactive

metabolites, high selectivity of the production of a certain hydroxy-steroid) are then

applied as biocatalysts for the large scale metabolite production in order to enable

pharmacological and toxicological evaluation of drug metabolites and for their structural

elucidation by NMR.


30

peak area 1-20 dextromethorphan demethyl

0

2000000

4000000

6000000

8000000

10000000

12000000

14000000

16000000

1 2 3 4 5 6 7 8 9 10 11 C 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 C 41 42 43 44 45

kolonie nr.

are

a

N-demeth

1. Site-directed

or random mutagenesis of P450 BM3-gene

2. Transformation of E.Coli

3. Colony picking and incubation

4. Activity and/or diversity screening

(HRS, GSH Trapping, UPLC)

5. Selection of most active/diverse BM3-mutant for

next round of random mutagenesis

6. Large scale metabolite production

7. Metabolite profiling and identification

(NMR, MS, affinity screening)

Figure 11. General strategy applied for the research presented in this thesis, leading to a “metabolite production and profiling platform”

7. Outline of this thesis

The work presented in this thesis shows application of BM3 mutants as biocatalyst for

three purposes:

For the regio- and stereoselective hydroxylation of lead compounds (e.g.

steroids)

For the generation of reactive metabolites

For the generation of pharmacologically active metabolites / lead optimization

Regio- and stereo-selective hydroxylation of unactivated C-H bonds of natural or synthetic

compounds is one of the biggest challenges in synthetic organic chemistry. In particular

steroid hydroxylation is extremely challenging due to the complexity of the steroid

molecule that requires complicated, multi-step schemes involving protecting groups and

subsequent regeneration.

In chapter 2 and 4 of the present thesis cytochrome P450 BM3 mutants are

applied as biocatalysts for the regioselective hydroxylation of testosterone and

norethisterone. In chapter 2 a site saturation mutagenesis library of BM3 mutants was

applied to evaluate the effect of a single mutation in the active site of BM3 on the

regioselective hydroxylation of testosterone and on the coupling efficiency in

alkoxyresorufin oxidation. In chapter 4, a single active site substitution was applied in

order to reduce the substrate mobility of testosterone, thus improving the regioselectivity


31

of hydroxylation. Spin lattice relaxation NMR was applied to determine how this single

mutation affected the orientation of the substrate within the active site of two BM3

mutants, M11 and M11 A82W.

The generation of reactive metabolites is important from a toxicological point of

view to be able to identify potentially toxic compounds that may be involved in ADRs or

IDRs.

Chapter 3 of the present thesis focuses on the application of P450 BM3 mutants

for the generation of reactive metabolites of clozapine, which is a drug known to be

responsible of severe ADRs. By screening of a minimal library of BM3 mutants, a mutant

was selected for the generation of large amounts of all major human relevant GSH-

conjugates of clozapine to enable their structural elucidation by NMR.

The generation of pharmacologically active metabolites is important from a “lead

optimization” point of view. Sometimes drug metabolites possess improved

pharmacological activity and improved physico-chemical properties, compared to the lead

itself.

In chapter 5 of the present thesis a focused library of BM3 mutants has been

applied for the generation of human relevant bioactive metabolites of p38 kinase

inhibitor Tak-715. The HRS screening applied enabled the identification and

pharmacological characterization of the drug metabolites in a quick and informative online

mode. Large scale incubation with BM3 mutants and isolation of active metabolites by

Prep-LC allowed their structural elucidation by NMR.

The last part of this thesis, chapter 6 concerns an overall summary of the work

described in the thesis including general conclusions and perspectives for future work.


32

References

1. Fura, A., Shu, Y. Z., Zhu, M., Hanson, R. L., Roongta, V., and Humphreys, W. G.

(2004) Discovering drugs through biological transformation: role of

pharmacologically active metabolites in drug discovery, J Med Chem 47, 4339-

4351.

2. Kang, M. J., Song, W. H., Shim, B. H., Oh, S. Y., Lee, H. Y., Chung, E. Y., Sohn, Y., and Lee, J. (2010) Pharmacologically active metabolites of currently marketed drugs: potential resources for new drug discovery and development, Yakugaku zasshi : Journal of the Pharmaceutical Society of Japan 130, 1325-1337.

3. Kumar, S. Engineering cytochrome P450 biocatalysts for biotechnology, medicine and bioremediation, Expert Opin Drug Metab Toxicol 6, 115-131.

4. Schroer, K., Kittelmann, M., and Lutz, S. Recombinant human cytochrome P450 monooxygenases for drug metabolite synthesis, Biotechnol Bioeng 106, 699-706.

5. Otey, C. R., Bandara, G., Lalonde, J., Takahashi, K., and Arnold, F. H. (2006) Preparation of human metabolites of propranolol using laboratory-evolved bacterial cytochromes P450, Biotechnol Bioeng 93, 494-499.

6. www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/ Guidances/ucm079266.pdf. (2008). 7. Nedderman, A. N. (2009) Metabolites in safety testing: metabolite identification

strategies in discovery and development, Biopharmaceutics & drug disposition 30, 153-162.

8. Evans, W. E., and Relling, M. V. (1999) Pharmacogenomics: translating functional genomics into rational therapeutics, Science 286, 487-491.

9. Woolf, T. F., and Jordan, R. A. (1987) Basic concepts in drug metabolism: Part I, Journal of clinical pharmacology 27, 15-17.

10. Jordan, R. A., and Woolf, T. F. (1987) Basic concepts in drug metabolism: Part II, Journal of clinical pharmacology 27, 87-90.

11. Gomes, E. R., and Demoly, P. (2005) Epidemiology of hypersensitivity drug reactions, Current opinion in allergy and clinical immunology 5, 309-316.

12. Williams, D. P., Kitteringham, N. R., Naisbitt, D. J., Pirmohamed, M., Smith, D. A., and Park, B. K. (2002) Are chemically reactive metabolites responsible for adverse reactions to drugs?, Current drug metabolism 3, 351-366.

13. Fontana, E., Dansette, P. M., and Poli, S. M. (2005) Cytochrome p450 enzymes mechanism based inhibitors: common sub-structures and reactivity, Current drug metabolism 6, 413-454.

14. Zhou, S., Chan, E., Lim, L. Y., Boelsterli, U. A., Li, S. C., Wang, J., Zhang, Q., Huang, M., and Xu, A. (2004) Therapeutic drugs that behave as mechanism-based inhibitors of cytochrome P450 3A4, Current drug metabolism 5, 415-442.

15. Caldwell, G. W., and Yan, Z. (2006) Screening for reactive intermediates and toxicity assessment in drug discovery, Current opinion in drug discovery & development 9, 47-60.

16. Prakash, C., Shaffer, C. L., and Nedderman, A. (2007) Analytical strategies for identifying drug metabolites, Mass spectrometry reviews 26, 340-369.


33

17. DiMagno, E. P. (2007) The pancreatic cyst incidentaloma: management consensus?, Clinical gastroenterology and hepatology : the official clinical practice journal of the American Gastroenterological Association 5, 797-798.

18. Madsen, K. G., Gronberg, G., Skonberg, C., Jurva, U., Hansen, S. H., and Olsen, J. (2008) Electrochemical oxidation of troglitazone: identification and characterization of the major reactive metabolite in liver microsomes, Chem Res Toxicol 21, 2035-2041.

19. Chen, Y., Monshouwer, M., and Fitch, W. L. (2007) Analytical tools and approaches for metabolite identification in early drug discovery, Pharm Res 24, 248-257.

20. Osorio-Lozada, A., Surapaneni, S., Skiles, G. L., and Subramanian, R. (2008) Biosynthesis of drug metabolites using microbes in hollow fiber cartridge reactors: case study of diclofenac metabolism by Actinoplanes species, Drug Metab Dispos 36, 234-240.

21. 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.

22. Damsten, M. C., de Vlieger, J. S., Niessen, W. M., Irth, H., Vermeulen, N. P., and Commandeur, J. N. (2008) Trimethoprim: novel reactive intermediates and bioactivation pathways by cytochrome p450s, Chem Res Toxicol 21, 2181-2187.

23. Boerma, J. S., Vermeulen, N. P., and Commandeur, J. N. (2011) Application of CYP102A1M11H as a tool for the generation of protein adducts of reactive drug metabolites, Chem Res Toxicol 24, 1263-1274.

24. Dragovic, S., Boerma, J. S., van Bergen, L., Vermeulen, N. P., and Commandeur, J. N. (2010) Role of human glutathione S-transferases in the inactivation of reactive metabolites of clozapine, Chem Res Toxicol 23, 1467-1476.

25. Lim, H. K., Stellingweif, S., Sisenwine, S., and Chan, K. W. (1999) Rapid drug metabolite profiling using fast liquid chromatography, automated multiple-stage mass spectrometry and receptor-binding, J Chromatogr A 831, 227-241.

26. Soldner, A., Spahn-Langguth, H., Palm, D., and Mutschler, E. (1998) A radioreceptor assay for the analysis of AT1-receptor antagonists. Correlation with complementary LC data reveals a potential contribution of active metabolites, Journal of pharmaceutical and biomedical analysis 17, 111-124.

27. de Vlieger, J. S., Kolkman, A. J., Ampt, K. A., Commandeur, J. N., Vermeulen, N. P., Kool, J., Wijmenga, S. S., Niessen, W. M., Irth, H., and Honing, M. (2010) Determination and identification of estrogenic compounds generated with biosynthetic enzymes using hyphenated screening assays, high resolution mass spectrometry and off-line NMR, J Chromatogr B Analyt Technol Biomed Life Sci 878, 667-674.

28. Falck, D., de Vlieger, J. S., Niessen, W. M., Kool, J., Honing, M., Giera, M., and Irth, H. (2010) Development of an online p38alpha mitogen-activated protein kinase binding assay and integration of LC-HR-MS, Analytical and bioanalytical chemistry 398, 1771-1780.


34

29. Reinen, J., Kalma, L. L., Begheijn, S., Heus, F., Commandeur, J. N., and Vermeulen, N. P. (2011) Application of cytochrome P450 BM3 mutants as biocatalysts for the profiling of estrogen receptor binding metabolites of the mycotoxin zearalenone, Xenobiotica; the fate of foreign compounds in biological systems 41, 59-70.

30. Miller, K. J., and Herman, A. C. (1996) Affinity chromatography with immunochemical detection applied to the analysis of human methionyl granulocyte colony stimulating factor in serum, Anal Chem 68, 3077-3082.

31. Przyjazny, A., Hentz, N. G., and Bachas, L. G. (1993) Sensitive and selective liquid chromatographic postcolumn reaction detection system for biotin and biocytin using a homogeneous fluorophore-linked assay, J Chromatogr A 654, 79-86.

32. Oosterkamp, A. J., Irth, H., Beth, M., Unger, K. K., Tjaden, U. R., and van de Greef, J. (1994) Bioanalysis of digoxin and its metabolites using direct serum injection combined with liquid chromatography and on-line immunochemical detection, Journal of chromatography. B, Biomedical applications 653, 55-61.

33. Kool, J., van Liempd, S. M., Ramautar, R., Schenk, T., Meerman, J. H., Irth, H., Commandeur, J. N., and Vermeulen, N. P. (2005) Development of a novel cytochrome p450 bioaffinity detection system coupled online to gradient reversed-phase high-performance liquid chromatography, Journal of biomolecular screening 10, 427-436.

34. Ingkaninan, K., Hazekamp, A., de Best, C. M., Irth, H., Tjaden, U. R., van der Heijden, R., van der Greef, J., and Verpoorte, R. (2000) The application of HPLC with on-line coupled UV/MS-biochemical detection for isolation of an acetylcholinesterase inhibitor from narcissus 'Sir Winston Churchill', J Nat Prod 63, 803-806.

35. Schenk, T., Appels, N. M., van Elswijk, D. A., Irth, H., Tjaden, U. R., and van der Greef, J. (2003) A generic assay for phosphate-consuming or -releasing enzymes coupled on-line to liquid chromatography for lead finding in natural products, Anal Biochem 316, 118-126.

36. Krabbe, J. G., Lingeman, H., Niessen, W. M., and Irth, H. (2003) Ligand-exchange detection of phosphorylated peptides using liquid chromatography electrospray mass spectrometry, Anal Chem 75, 6853-6860.

37. Van Liempd, S. M., Kool, J., Meerman, J. H., Irth, H., and Vermeulen, N. P. (2007) Metabolic profiling of endocrine-disrupting compounds by on-line cytochrome p450 bioreaction coupled to on-line receptor affinity screening, Chem Res Toxicol 20, 1825-1832.

38. Giera, M., de Vlieger, J. S., Lingeman, H., Irth, H., and Niessen, W. M. (2010) Structural elucidation of biologically active neomycin N-octyl derivatives in a regioisomeric mixture by means of liquid chromatography/ion trap time-of-flight mass spectrometry, Rapid communications in mass spectrometry : RCM 24, 1439-1446.

39. Giera, M., Heus, F., Janssen, L., Kool, J., Lingeman, H., and Irth, H. (2009) Microfractionation revisited: a 1536 well high resolution screening assay, Anal Chem 81, 5460-5466.


35

40. Donova, M. V., and Egorova, O. V. (2012) Microbial steroid transformations: current state and prospects, Appl Microbiol Biotechnol 94, 1423-1447.

41. Hogg, J. A. (1992) Steroids, the steroid community, and Upjohn in perspective: a profile of innovation, Steroids 57, 593-616.

42. Sedlaczek, L. (1988) Biotransformations of steroids, Crit Rev Biotechnol 7, 187-236.

43. Fernandes, A., Cruz, A., Angelova, B., Pinheiro, H. M., and Cabral, A. (2003) Microbial conversion of steroid compounds: recent developments, Enzyme Microb Technol 32, 688-705.

44. Mahato, S. B., Banerjee, S., and Sahu, N. P. (1984) Metabolism of progesterone and testosterone by a Bacillus sp, Steroids 43, 545-558.

45. Mahato, S. B., and Mukherjee, A. (1984) Microbial transformation of testosterone by Aspergillus fumigatus, J Steroid Biochem 21, 341-342.

46. Ko, D., Heiman, A. S., Chen, M., and Lee, H. J. (2000) New steroidal anti-inflammatory antedrugs: methyl 21-desoxy-21-chloro-11beta,17alpha-dihydroxy-3,20-dioxo-1, 4-pregnadiene-16alpha-carboxylate, methyl 21-desoxy-21-chloro-11beta-hydroxy-3,20-dioxo-1, 4-pregnadiene-16alpha-carboxylate, and their 9alpha-fluoro derivatives*, Steroids 65, 210-218.

47. Tuba, Z., Bardin, C. W., Dancsi, A., Francsics-Czinege, E., Molnar, C., Csorgei, J., Falkay, G., Koide, S. S., Kumar, N., Sundaram, K., Dukat-Abrok, V., and Balogh, G. (2000) Synthesis and biological activity of a new progestogen, 16-methylene-17alpha-hydroxy-18-methyl-19-norpregn-4-ene-3, 20-dione acetate, Steroids 65, 266-274.

48. Newhouse, T., and Baran, P. S. (2011) If C-H bonds could talk: selective C-H bond oxidation, Angew Chem Int Ed Engl 50, 3362-3374.

49. Dutton, D. R., McMillen, S. K., Sonderfan, A. J., Thomas, P. E., and Parkinson, A. (1987) Studies on the rate-determining factor in testosterone hydroxylation by rat liver microsomal cytochrome P-450: evidence against cytochrome P-450 isozyme:isozyme interactions, Archives of biochemistry and biophysics 255, 316-328.

50. Sonderfan, A. J., Arlotto, M. P., and Parkinson, A. (1989) Identification of the cytochrome P-450 isozymes responsible for testosterone oxidation in rat lung, kidney, and testis: evidence that cytochrome P-450a (P450IIA1) is the physiologically important testosterone 7 alpha-hydroxylase in rat testis, Endocrinology 125, 857-866.

51. Agematu, H., Matsumoto, N., Fujii, Y., Kabumoto, H., Doi, S., Machida, K., Ishikawa, J., and Arisawa, A. (2006) Hydroxylation of testosterone by bacterial cytochromes P450 using the Escherichia coli expression system, Bioscience, biotechnology, and biochemistry 70, 307-311.

52. Bernhardt, R. (2006) Cytochromes P450 as versatile biocatalysts, J Biotechnol 124, 128-145.

53. Hannemann, F., Bichet, A., Ewen, K. M., and Bernhardt, R. (2007) Cytochrome P450 systems--biological variations of electron transport chains, Biochim Biophys Acta 1770, 330-344.


36

54. Novikova, L. A., Faletrov, Y. V., Kovaleva, I. E., Mauersberger, S., Luzikov, V. N., and Shkumatov, V. M. (2009) From structure and functions of steroidogenic enzymes to new technologies of gene engineering, Biochemistry. Biokhimiia 74, 1482-1504.

55. Kristan, K., and Rizner, T. L. (2012) Steroid-transforming enzymes in fungi, The Journal of steroid biochemistry and molecular biology 129, 79-91.

56. Guengerich, F. P. (2001) Common and uncommon cytochrome P450 reactions related to metabolism and chemical toxicity, Chem Res Toxicol 14, 611-650.

57. O'Reilly, E., Kohler, V., Flitsch, S. L., and Turner, N. J. (2010) Cytochromes P450 as useful biocatalysts: addressing the limitations, Chem Commun (Camb) 47, 2490-2501.

58. Drnelson.utmem.edu. 59. Pirmohamed, M., and Park, B. K. (2003) Cytochrome P450 enzyme

polymorphisms and adverse drug reactions, Toxicology 192, 23-32. 60. Ingelman-Sundberg, M. (2004) Pharmacogenetics of cytochrome P450 and its

applications in drug therapy: the past, present and future, Trends in pharmacological sciences 25, 193-200.

61. Kirchheiner, J., and Brockmoller, J. (2005) Clinical consequences of cytochrome P450 2C9 polymorphisms, Clin Pharmacol Ther 77, 1-16.

62. Urlacher, V. B., and Girhard, M. (2011) Cytochrome P450 monooxygenases: an update on perspectives for synthetic application, Trends Biotechnol 30, 26-36.

63. Munro, A. W., Leys, D. G., 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.

64. Isin, E. M., and Guengerich, F. P. (2007) Complex reactions catalyzed by cytochrome P450 enzymes, Biochim Biophys Acta 1770, 314-329.

65. Anzenbacher, P., and Anzenbacherova, E. (2001) Cytochromes P450 and metabolism of xenobiotics, Cell Mol Life Sci 58, 737-747.

66. Whitehouse, C. J., Bell, S. G., and Wong, L. L. (2012) P450(BM3) (CYP102A1): connecting the dots, Chemical Society reviews 41, 1218-1260.

67. Rittle, J., and Green, M. T. (2010) Cytochrome P450 compound I: capture, characterization, and C-H bond activation kinetics, Science 330, 933-937.

68. 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.

69. 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.

70. Jung, S. T., Lauchli, R., and Arnold, F. H. (2011) Cytochrome P450: taming a wild type enzyme, Curr Opin Biotechnol 22, 809-817.

71. 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


37

mechanism of a biotechnologically important enzyme, Biochem Soc Trans 33, 747-753.

72. 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.

73. 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.

74. 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.

75. Seifert, A., Vomund, S., Grohmann, K., Kriening, S., Urlacher, V. B., Laschat, S., and Pleiss, J. (2009) Rational design of a minimal and highly enriched CYP102A1 mutant library with improved regio-, stereo- and chemoselectivity, Chembiochem 10, 853-861.

76. Munzer, D. F., Meinhold, P., Peters, M. W., Feichtenhofer, S., Griengl, H., Arnold, F. H., Glieder, A., and de Raadt, A. (2005) Stereoselective hydroxylation of an achiral cyclopentanecarboxylic acid derivative using engineered P450s BM-3, Chem Commun (Camb), 2597-2599.

77. Kim, D. H., Ahn, T., Jung, H. C., Pan, J. G., and Yun, C. H. (2009) Generation of the human metabolite piceatannol from the anticancer-preventive agent resveratrol by bacterial cytochrome P450 BM3, Drug Metab Dispos 37, 932-936.

78. Kim, D. H., Kim, K. H., Liu, K. H., Jung, H. C., Pan, J. G., and Yun, C. H. (2008) Generation of human metabolites of 7-ethoxycoumarin by bacterial cytochrome P450 BM3, Drug Metab Dispos 36, 2166-2170.

79. Kim, K. H., Kang, J. Y., Kim, D. H., Park, S. H., Kim, D., Park, K. D., Lee, Y. J., Jung, H. C., Pan, J. G., Ahn, T., and Yun, C. H. (2011) Generation of human chiral metabolites of simvastatin and lovastatin by bacterial CYP102A1 mutants, Drug Metab Dispos 39, 140-150.

80. Landwehr, M., Hochrein, L., Otey, C. R., Kasrayan, A., Backvall, J. E., and Arnold, F. H. (2006) Enantioselective alpha-hydroxylation of 2-arylacetic acid derivatives and buspirone catalyzed by engineered cytochrome P450 BM-3, J Am Chem Soc 128, 6058-6059.

81. Sawayama, A. M., Chen, M. M., 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 15, 11723-11729.

82. van Vugt-Lussenburg, B. M., Damsten, M. C., Maasdijk, D. M., Vermeulen, N. P., and Commandeur, J. N. (2006) Heterotropic and homotropic cooperativity by a drug-metabolising mutant of cytochrome P450 BM3, Biochem Biophys Res Commun 346, 810-818.

83. van Vugt-Lussenburg, B. M., Stjernschantz, E., Lastdrager, J., Oostenbrink, C., Vermeulen, N. P., and Commandeur, J. N. (2007) Identification of critical residues in novel drug metabolizing mutants of cytochrome P450 BM3 using random mutagenesis, J Med Chem 50, 455-461.


38

84. Rea, V., Dragovic, S., Boerma, J. S., de Kanter, F. J., Vermeulen, N. P., and Commandeur, J. N. (2011) Role of residue 87 in the activity and regioselectivity of clozapine metabolism by drug-metabolizing CYP102A1 M11H: application for structural characterization of clozapine GSH conjugates, Drug Metab Dispos 39, 2411-2420.

85. Yun, C. H., Kim, K. H., Kim, D. H., Jung, H. C., and Pan, J. G. (2007) The bacterial P450 BM3: a prototype for a biocatalyst with human P450 activities, Trends Biotechnol 25, 289-298.

86. Gillam, E. M. (2008) Engineering cytochrome p450 enzymes, Chem Res Toxicol 21, 220-231.

87. Papouchado, B. G., Myles, J., Lloyd, R. V., Stoler, M., Oliveira, A. M., Downs-Kelly, E., Morey, A., Bilous, M., Nagle, R., Prescott, N., Wang, L., Dragovich, L., McElhinny, A., Garcia, C. F., Ranger-Moore, J., Free, H., Powell, W., Loftus, M., Pettay, J., Gaire, F., Roberts, C., Dietel, M., Roche, P., Grogan, T., and Tubbs, R. (2010) Silver in situ hybridization (SISH) for determination of HER2 gene status in breast carcinoma: comparison with FISH and assessment of interobserver reproducibility, The American journal of surgical pathology 34, 767-776.

88. Feenstra, K. A., Hofstetter, K., Bosch, R., Schmid, A., Commandeur, J. N., and Vermeulen, N. P. (2006) Enantioselective substrate binding in a monooxygenase protein model by molecular dynamics and docking, Biophysical journal 91, 3206-3216.

89. Rea, V., Kolkman, A. J., Vottero, E. R., Stronks, E. J., Ampt, K. A., Honing, M., Vermeulen, N. P., Wijmenga, S. S., and Commandeur, J. N. (2011) Active site substitution A82W improves regioselectivity of steroid hydroxylation by cytochrome P450 BM3 mutants as rationalised by spin relaxation NMR studies, Biochemistry.

90. Huang, W. C., Westlake, A. C., Marechal, 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.

91. Dietrich, J. A., Yoshikuni, Y., Fisher, K. J., Woolard, F. X., Ockey, D., McPhee, D. J., Renninger, N. S., Chang, M. C., Baker, D., and Keasling, J. D. (2009) A novel semi-biosynthetic route for artemisinin production using engineered substrate-promiscuous P450(BM3), ACS chemical biology 4, 261-267.

92. Seifert, A., Antonovici, M., Hauer, B., and Pleiss, J. (2011) An efficient route to selective bio-oxidation catalysts: an iterative approach comprising modeling, diversification, and screening, based on CYP102A1, Chembiochem 12, 1346-1351.

93. Koenig, R., Lesemann, D. E., Loss, S., Engelmann, J., Commandeur, U., Deml, G., Schiemann, J., Aust, H., and Burgermeister, W. (2006) Zygocactus virus X-based expression vectors and formation of rod-shaped virus-like particles in plants by the expressed coat proteins of Beet necrotic yellow vein virus and Soil-borne cereal mosaic virus, The Journal of general virology 87, 439-443.

94. Julsing, M. K., Cornelissen, S., Buhler, B., and Schmid, A. (2008) Heme-iron oxygenases: powerful industrial biocatalysts?, Curr Opin Chem Biol 12, 177-186.


39

95. Parikh, A., Gillam, E. M., and Guengerich, F. P. (1997) Drug metabolism by Escherichia coli expressing human cytochromes P450, Nat Biotechnol 15, 784-788.

96. Rushmore, T. H., Reider, P. J., Slaughter, D., Assang, C., and Shou, M. (2000) Bioreactor systems in drug metabolism: synthesis of cytochrome P450-generated metabolites, Metab Eng 2, 115-125.

97. 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, New biotechnology 28, 79-85.

98. Lewis, J. C., Bastian, S., Bennett, C. S., Fu, Y., Mitsuda, Y., Chen, M. M., Greenberg, W. A., Wong, C. H., and Arnold, F. H. (2009) Chemoenzymatic elaboration of monosaccharides using engineered cytochrome P450BM3 demethylases, Proc Natl Acad Sci U S A 106, 16550-16555.

99. Rentmeister, A., Arnold, F. H., and Fasan, R. (2009) Chemo-enzymatic fluorination of unactivated organic compounds, Nat Chem Biol 5, 26-28.

100. 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.

101. Vottero, E., Rea, V., Lastdrager, J., Honing, M., Vermeulen, N. P., and Commandeur, J. N. (2011) Role of residue 87 in substrate selectivity and regioselectivity of drug-metabolizing cytochrome P450 CYP102A1 M11, J Biol Inorg Chem 16, 899-912.

102. Okubo, Y., Bessho, K., Fujimura, K., Kusumoto, K., Ogawa, Y., and Iizuka, T. (2002) Expression of bone morphogenetic protein in the course of osteoinduction by recombinant human bone morphogenetic protein-2, Clinical oral implants research 13, 80-85.

103. Ogawa, M., Kumagai, K., Gondo, N., Matsumoto, N., Suyama, K., and Saku, K. (2002) Novel electrophysiologic parameter of dispersion of atrial repolarization: comparison of different atrial pacing methods, J Cardiovasc Electrophysiol 13, 110-117.

104. Fedorova, A., and Ogawa, M. Y. (2002) Site-specific modification of de novo designed coiled-coil polypeptides with inorganic redox complexes, Bioconjug Chem 13, 150-154.

105. Hoehn, R., Lamparter, J., Pfeiffer, N., and Vossmerbaeumer, U. (2011) Seizures following subconjunctival 5-FU therapy, Graefe's archive for clinical and experimental ophthalmology = Albrecht von Graefes Archiv fur klinische und experimentelle Ophthalmologie 249, 145-146.

106. 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.

107. Oliver, C. F., Modi, S., Sutcliffe, M. J., Primrose, W. U., Lian, L. Y., and Roberts, G. C. (1997) A single mutation in cytochrome P450 BM3 changes substrate orientation in a catalytic intermediate and the regiospecificity of hydroxylation, Biochemistry 36, 1567-1572.


40

108. Nulens, E., Gould, I., MacKenzie, F., Deplano, A., Cookson, B., Alp, E., Bouza, E., and Voss, A. (2005) Staphylococcus aureus carriage among participants at the 13th European Congress of Clinical Microbiology and Infectious Diseases, European journal of clinical microbiology & infectious diseases : official publication of the European Society of Clinical Microbiology 24, 145-148.

109. Schotten, U., Schumacher, C., Conrads, V., Braun, V., Schondube, F., Voss, M., and Hanrath, P. (1999) Calcium-sensitivity of the SR calcium release channel in failing and nonfailing human myocardium, Basic research in cardiology 94, 145-151.

110. de Smet, M. D., Buffam, F. V., Fairey, R. N., and Voss, N. J. (1990) Prevention of radiation-induced stenosis of the nasolacrimal duct, Canadian journal of ophthalmology. Journal canadien d'ophtalmologie 25, 145-147.

111. Voss, K., Barz, H., and Wagner, H. (1977) [Image processing in pathology. III. Structure and use of an automatic morphometry system for recording morphometric features of nuclei (author's transl)], Experimentelle Pathologie 13, 145-161.

112. Voss, R., Rickart, A., Ruile, K., Schoen, H. R., and Kunz, W. (1970) [Hyperbaric organ preservation], Verhandlungen der Deutschen Gesellschaft fur Pathologie 54, 145-150.

113. Vossschulte, K. (1951) [The intervertebral disks as affected by the movements of the spine], Langenbecks Archiv fur klinische Chirurgie ... vereinigt mit Deutsche Zeitschrift fur Chirurgie 267, 145.

114. Ruperti, B., Bonghi, C., Rasori, A., Ramina, A., and Tonutti, P. (2001) Characterization and expression of two members of the peach 1-aminocyclopropane-1-carboxylate oxidase gene family, Physiol Plant 111, 336-344.

115. Bidoli, V., Casolino, M., De Pascale, M. P., Furano, G., Minori, M., Morselli, A., Narici, L., Picozza, P., Reali, E., Sparvoli, R., Fuglesang, C., Sannita, W., Carlson, P., Castellini, G., Galper, A., Korotkov, M., Popov, A., Navilov, N., Avdeev, S., Benghin, V., Salnitskii, V., Shevchenko, O., Boezio, M., Bonvicini, W., Vacchi, A., Zampa, G., Zampa, N., Mazzenga, G., Ricci, M., Spillantini, P., and Vittori, R. (2002) The Sileye-3/Alteino experiment for the study of light flashes, radiation environment and astronaut brain activity on board the International Space Station, Journal of radiation research 43 Suppl, S47-52.

116. Furalev, V. A., Kravchenko, I. V., Vasil'eva, L. A., and Pylev, L. N. (2002) Cytotoxic effects of macrophages and asbestos on transformed rat mesothelial cells, Bulletin of experimental biology and medicine 133, 71-73.

117. Walldal, E., Anund, I., and Furaker, C. (2002) Quality of care and development of a critical pathway, Journal of nursing management 10, 115-122.

118. Li, Q. S., Schwaneberg, U., Fischer, P., and Schmid, R. D. (2000) Directed evolution of the fatty-acid hydroxylase P450 BM-3 into an indole-hydroxylating catalyst, Chemistry 6, 1531-1536.

119. Li, Q. S., Ogawa, J., Schmid, R. D., and Shimizu, S. (2001) Engineering cytochrome P450 BM-3 for oxidation of polycyclic aromatic hydrocarbons, Appl Environ Microbiol 67, 5735-5739.


41

120. 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.

121. Chen, C. K., Shokhireva, T., Berry, R. E., Zhang, H., and Walker, F. A. (2008) The effect of mutation of F87 on the properties of CYP102A1-CYP4C7 chimeras: altered regiospecificity and substrate selectivity, J Biol Inorg Chem 13, 813-824.

122. Eiben, S., Kaysser, L., Maurer, S., Kuhnel, K., Urlacher, V. B., and Schmid, R. D. (2006) Preparative use of isolated CYP102 monooxygenases -- a critical appraisal, J Biotechnol 124, 662-669.

123. Rock, D. A., Perkins, B. N., Wahlstrom, J., and Jones, J. P. (2003) A method for determining two substrates binding in the same active site of cytochrome P450BM3: an explanation of high energy omega product formation, Archives of biochemistry and biophysics 416, 9-16.

124. 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.

125. 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.

126. Rock, D. A., Boitano, A. E., Wahlstrom, J. L., and Jones, J. P. (2002) Use of kinetic isotope effects to delineate the role of phenylalanine 87 in P450(BM-3), Bioorganic chemistry 30, 107-118.

127. Feenstra, K. A., Starikov, E. B., Urlacher, V. B., Commandeur, J. N., and Vermeulen, N. P. (2007) Combining substrate dynamics, binding statistics, and energy barriers to rationalize regioselective hydroxylation of octane and lauric acid by CYP102A1 and mutants, Protein science : a publication of the Protein Society 16, 420-431.

128. 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.

129. Roccatano, D., Wong, T. S., Schwaneberg, U., and Zacharias, M. (2006) Toward understanding the inactivation mechanism of monooxygenase P450 BM-3 by organic cosolvents: a molecular dynamics simulation study, Biopolymers 83, 467-476.

130. Urlacher, V. B., Makhsumkhanov, A., and Schmid, R. D. (2006) Biotransformation of beta-ionone by engineered cytochrome P450 BM-3, Appl Microbiol Biotechnol 70, 53-59.

131. 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.

132. 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.


42

133. 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, Journal of biomolecular screening 10, 246-252.

134. Chen, C. K., Berry, R. E., Shokhireva, T., Murataliev, M. B., Zhang, H., and Walker, F. A. (2010) Scanning chimeragenesis: the approach used to change the substrate selectivity of fatty acid monooxygenase CYP102A1 to that of terpene omega-hydroxylase CYP4C7, J Biol Inorg Chem 15, 159-174.

135. 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, Journal of biomolecular screening 16, 239-250.

136. 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 ( Pt 2), 371-379.

137. Raner, G. M., Hatchell, J. A., Dixon, M. U., Joy, T. L., Haddy, A. E., and Johnston, E. R. (2002) Regioselective peroxo-dependent heme alkylation in P450(BM3)-F87G by aromatic aldehydes: effects of alkylation on cataysis, Biochemistry 41, 9601-9610.

138. 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.

139. Raner, G. M., Thompson, J. I., Haddy, A., Tangham, V., Bynum, N., Ramachandra Reddy, G., Ballou, D. P., and Dawson, J. H. (2006) Spectroscopic investigations of intermediates in the reaction of cytochrome P450(BM3)-F87G with surrogate oxygen atom donors, J Inorg Biochem 100, 2045-2053.

140. Noble, M. A., Quaroni, L., Chumanov, G. D., Turner, K. L., Chapman, S. K., Hanzlik, R. P., and Munro, A. W. (1998) Imidazolyl carboxylic acids as mechanistic probes of flavocytochrome P-450 BM3, Biochemistry 37, 15799-15807.

141. Branco, R. J., 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 73, 597-607.

142. 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.

143. 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.

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

145. Kubo, T., Peters, M. W., Meinhold, P., and Arnold, F. H. (2006) Enantioselective epoxidation of terminal alkenes to (R)- and (S)-epoxides by engineered cytochromes P450 BM-3, Chemistry 12, 1216-1220.


43

146. 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.

147. Daiber, A., Herold, S., Schoneich, C., Namgaladze, D., Peterson, J. A., and Ullrich, V. (2000) Nitration and inactivation of cytochrome P450BM-3 by peroxynitrite. Stopped-flow measurements prove ferryl intermediates, Eur J Biochem 267, 6729-6739.

148. Cowart, L. A., Falck, J. R., and Capdevila, J. H. (2001) Structural determinants of active site binding affinity and metabolism by cytochrome P450 BM-3, Archives of biochemistry and biophysics 387, 117-124.

149. Li, Q. S., Ogawa, J., Schmid, R. D., and Shimizu, S. (2005) Indole hydroxylation by bacterial cytochrome P450 BM-3 and modulation of activity by cumene hydroperoxide, Bioscience, biotechnology, and biochemistry 69, 293-300.

150. Lewis, J. C., Mantovani, S. M., Fu, Y., Snow, C. D., Komor, R. S., Wong, C. H., and Arnold, F. H. (2010) Combinatorial alanine substitution enables rapid optimization of cytochrome P450BM3 for selective hydroxylation of large substrates, Chembiochem 11, 2502-2505.

151. Weber, E., Seifert, A., Antonovici, M., Geinitz, C., Pleiss, J., and Urlacher, V. B. Screening of a minimal enriched P450 BM3 mutant library for hydroxylation of cyclic and acyclic alkanes, Chem Commun (Camb) 47, 944-946.

152. Sulistyaningdyah, W. T., Ogawa, J., Li, Q. S., Maeda, C., Yano, Y., Schmid, R. D., and Shimizu, S. (2005) Hydroxylation activity of P450 BM-3 mutant F87V towards aromatic compounds and its application to the synthesis of hydroquinone derivatives from phenolic compounds, Appl Microbiol Biotechnol 67, 556-562.

153. Misawa, N., Nodate, M., Otomatsu, T., Shimizu, K., Kaido, C., Kikuta, M., Ideno, A., Ikenaga, H., Ogawa, J., Shimizu, S., and Shindo, K. (2011) Bioconversion of substituted naphthalenes and beta-eudesmol with the cytochrome P450 BM3 variant F87V, Appl Microbiol Biotechnol 90, 147-157.

154. Nazor, J., and Schwaneberg, U. (2006) Laboratory evolution of P450 BM-3 for mediated electron transfer, Chembiochem 7, 638-644.

155. Fasan, R., Chen, M. M., Crook, N. C., and Arnold, F. H. (2007) Engineered alkane-hydroxylating cytochrome P450(BM3) exhibiting nativelike catalytic properties, Angew Chem Int Ed Engl 46, 8414-8418.

156. 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, 13442-13450.

157. Murataliev, M. B., Trinh, L. N., Moser, L. V., Bates, R. B., Feyereisen, R., and Walker, F. A. (2004) Chimeragenesis of the fatty acid binding site of cytochrome P450BM3. Replacement of residues 73-84 with the homologous residues from the insect cytochrome P450 CYP4C7, Biochemistry 43, 1771-1780.

158. Huang, W. C., Cullis, P. M., Raven, E. L., and Roberts, G. C. (2011) Control of the stereo-selectivity of styrene epoxidation by cytochrome P450 BM3 using structure-based mutagenesis, Metallomics 3, 410-416.

159. 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


44

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.

160. 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.

161. Novak, R. F., and Vatsis, K. P. (1982) 1H Fourier transform nuclear magnetic resonance relaxation rate studies on the interaction of acetanilide with purified isozymes of rabbit liver microsomal cytochrome P-450 and with cytochrome b5, Mol Pharmacol 21, 701-709.

162. Solomon, I., and Bloembergen, N. (1956) Nuclear magnetic interactions in the HF molecule, J Chem Phys 25, 261-266.

163. Regal, K. A., and Nelson, S. D. (2000) Orientation of caffeine within the active site of human cytochrome P450 1A2 based on NMR longitudinal (T1) relaxation measurements, Archives of biochemistry and biophysics 384, 47-58.

164. van de Straat, R., de Vries, J., de Boer, H. J., Vromans, R. M., and Vermeulen, N. P. (1987) Relationship between paracetamol binding to and its oxidation by two cytochromes P-450 isozymes--a proton nuclear magnetic resonance and spectrophotometric study, Xenobiotica; the fate of foreign compounds in biological systems 17, 1-9.

165. Poli-Scaife, S., Attias, R., Dansette, P. M., and Mansuy, D. (1997) The substrate binding site of human liver cytochrome P450 2C9: an NMR study, Biochemistry 36, 12672-12682.

166. Modi, S., Paine, M. J., Sutcliffe, M. J., Lian, L. Y., Primrose, W. U., Wolf, C. R., and Roberts, G. C. (1996) A model for human cytochrome P450 2D6 based on homology modeling and NMR studies of substrate binding, Biochemistry 35, 4540-4550.

167. Modi, S., Gilham, D. E., Sutcliffe, M. J., Lian, L. Y., Primrose, W. U., Wolf, C. R., and Roberts, G. C. (1997) 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine as a substrate of cytochrome P450 2D6: allosteric effects of NADPH-cytochrome P450 reductase, Biochemistry 36, 4461-4470.

168. Smith, G., Modi, S., Pillai, I., Lian, L. Y., Sutcliffe, M. J., Pritchard, M. P., Friedberg, T., Roberts, G. C., and Wolf, C. R. (1998) Determinants of the substrate specificity of human cytochrome P-450 CYP2D6: design and construction of a mutant with testosterone hydroxylase activity, Biochem J 331 ( Pt 3), 783-792.

169. Keizers, P. H., de Graaf, C., de Kanter, F. J., Oostenbrink, C., Feenstra, K. A., Commandeur, J. N., and Vermeulen, N. P. (2005) Metabolic regio- and stereoselectivity of cytochrome P450 2D6 towards 3,4-methylenedioxy-N-alkylamphetamines: in silico predictions and experimental validation, J Med Chem 48, 6117-6127

46

CHAPTER 2 ROLE OF RESIDUE 87 IN SUBSTRATE- AND REGIOSELECTIVITY OF

DRUG METABOLIZING CYTOCHROME P450 BM3 M11

Eduardo Vottero, Vanina Rea, Jeroen Lastdrager, Maarten Honing., Nico

P.E. Vermeulen and Jan N.M. Commandeur

adapted from Journal of Biological Inorganic Chemistry, 2011, 16(6), 899-

912

Chapter 2 Role of residue 87

47

Abstract:

BM3, originating from Bacillus megaterium, is a highly active enzyme which has attracted

much attention because of its potential applicability as a biocatalyst for oxidative

reactions. Previously we developed a drug metabolizing mutant BM3 M11 by a

combination of site-directed and random mutagenesis. BM3 M11 contains ten mutations,

when compared to wild-type BM3 and is able to produce human-relevant metabolites of

several pharmaceuticals. In this study, active-site residue 87 of drug metabolizing mutant

BM3 M11 was mutated to all possible natural amino acids in order to investigate its role in

substrate selectivity and regioselectivity. With alkoxyresorufins as substrates, large

differences in substrate-selectivities and coupling efficiencies were found, dependent on

the nature of residue 87. For all combinations of alkoxyresorufins and mutants extremely

fast rates of NADPH-oxidation were observed (up to 6000 min-1

). However, the coupling

efficiencies were extremely low: even for the substrates showing highest rates of O-

dealkylation, coupling efficiencies were lower than 1%. With testosterone as substrate, all

mutants were able to produce three hydroxytestosterone metabolites although at

different activities and with remarkably different product ratios. The results show that the

nature of amino acid at position 87 has a strong effect on activity and regioselectivity in

the drug-metabolizing mutant BM3 M11. Because of the wide substrate selectivity of

BM3 M11 when compared to wild-type BM3, this panel of mutants will be useful both as

biocatalysts for metabolite production and as model proteins for mechanistic studies on

the function of P450s in general.


48

2.1. Introduction

Cytochrome P450 BM3 (BM3; EC 1.14.14.1) from Bacillus megaterium is a soluble protein

that catalyzes the hydroxylation and epoxidation of several long-chain fatty acids. BM3

contains a heme and reductase domain fused in a single polypeptide, which might explain

why this enzyme has the highest activity ever reported for a P450 (1, 2). By using site-

directed and/or random mutagenesis, several research groups have succeeded in

broadening the substrate selectivity of this enzyme with the objective to create highly

efficient biocatalysts to be used for several biotechnological applications (3-5).

In parallel, many mechanistic studies have been devoted to rationalize the roles of various

active site residues in substrate selectivity and in the catalytic cycle (1). One of the most

studied active-site residues is Phe87, which is in close proximity to the heme according to

the crystal structures of BM3 (1).

In the substrate-free crystal, this residue lies perpendicular to the heme at the end of the

substrate access channel. Upon binding of substrate, the orientation of Phe87 changes to

parallel to the plane of the heme, thereby influencing the orientation of the substrate

relative to the catalytic centre (6). Many studies have been performed in which Phe87 of

wild-type BM3 or mutants of BM3 has been mutated to other amino acids, as summarized

in Table 1. Typically, in these studies one to maximally four different mutations at position

87 were compared. In most cases Phe87, was mutated to amino acids with small non-

polar side chains (Gly, Ala, Leu, Ile, Val), whereas only two uncharged polar amino acids

(Ser, Tyr) were evaluated. Dependent on the substrate tested and the enzyme template,

the type of mutation of Phe87 has differential effects on catalytic properties of P450

(Table 1), which are rationalized by the changes in active site volume, thereby restricting

or improving access to the reactive oxygen species at the heme-center (7-21).

In this study, all 20 natural amino acids were evaluated at position 87 of BM3 M11. Twelve

of the possible amino acids substitutions have not been described previously in wild-type

or mutants of BM3. The mutants were characterized using a homologous series of

alkoxyresorufins and testosterone as substrates. Alkoxyresorufines are sensitive and

useful probes to determine substrate selectivity of both bacterial and mammalian P450

isoforms by the continuous fluorimetric assay of resorufin formed by O-dealkylation (23-

25). Testosterone and other steroids have been shown to be hydroxylated by BM3

mutants (26, 27). Testosterone is an excellent probe substrate to study the effect of

position 87 on regioselectivity because it can be hydroxylated at multiple positions

dependent on the nature of P450-isoenzymes (28). Furthermore, there is a great interest

in biocatalysts capable to stereo- and regioselectively hydroxylate steroids, because

steroid compounds rank among the most widely marketed products from pharmaceutical

industry (29). We therefore studied whether regioselectivity of steroid hydroxylation by

BM3 can be manipulated by mutations at position 87.


49

Table 1. Effect of mutation of residue 87 on regio-/stereoselectivity of wild-type and mutant BM3

Mutation Template Substrate Effect Ref.

F87G wild type lauric acid less active; biphasic kinetics; changed regioselectivity 9, 19

wild type propylbenzene,

chlorostyrene

more active; changed regio-/stereoselectivity 11

F87A wild type lauric acid, palmitic acid

less active; changed regioselectivity 7, 18, 19

wild type farnesol more active; changed regioselectivity 18

wild type fluoranthene,

phenanthrene, pyrene

more active; changed regioselectivity 10

wild type propylbenzene changed regioselectivity 11

wild type chlorostyrene less active; higher enantioselectivity 11

wild type (+)-valencene changed regioselectivity 21

wild type resveratrol less active 20

A328I (+)-valencene more active 21

A328F limonene more active 21

R47L/Y51F (+)-valencene changed regioselectivity 14

R47L/Y51F resveratrol less active 20

9-10A phenyl acetic acid esters,

buspirone

more active; higher enantioselectivity

16

C(73-78) lauric acid, palmitic acid,

farnesol

less active 18


farnesol

less active; unchanged regioselectivity 18


farnesol

less active 18


50

Table 1 (continued)

F87V wild type lauric acid less active; changed regioselectivity 19

wild type arachidonic acid,

eicopentenoic acid

less active; changed regioselectivity 8

wild type aromatic and phenolic

compound

more active; changed regioselectivity

15

wild type, A328L geranyl acetone more active; changed regioselectivity 21

A328I, A328V (+)-valencene more active; changed regioselectivity 21

R47L; R47L/L188Q benzyl- and

pentoxyresorufin

more active 13

F87L wild type lauric acid less active; changed regioselectivity 18

wild type geranyl acetone less active; changed regioselectivity 21

A328I (+)-valancene more active 21

A328V limonene, geranyl acetone less active 21

C(73-78); C(75-80) lauric acid, palmitic acid,

farnesol

less active 18

C(78-82) lauric acid, palmitic acid less active 18

C(78-82) farnesol more active; changed regioselectivity 18

F87I wild type geranyl acetone more active; changed regioselectivity 21

A328I (+)-valancene more active 21

A328V limonene, geranyl acetone less active 21

F87S wild type lauric acid changed regioselectivity 19

F87Y wild type lauric acid lower NADPH-consumption; biphasic kinetics 9

wild type N-palmitoyl glycine inactive; 100% uncoupled 17

wild type arachidonic acid,

eicopentenoic acid

inactive; 100% uncoupled 8


51

2.2. Materials and Methods

2.2.1. Materials

All chemicals were of analytical grade and obtained from standard suppliers.

Alkoxyresorufins (methoxy- to n-octyloxyresorufin) and benzyloxyresorufin were

synthesized as described previously (24). The pET28a+ vector containing wild-type BM3

was kindly provided by dr. V. Urlacher (Institut für Technische Biochemie, Universität

Stuttgart, Germany).

2.2.2 Library construction

Site-directed mutants of BM3 M11 at position 87 were constructed by mutagenic PCR

using the Stratagene QuikChange XL site–directed mutagenesis kit (Stratagene, La Jolla,

CA, USA) using 20 complementary pairs of mutagenesis primers. The mutagenic PCR was

applied to a pBluescript vector (pBS P450 BM3 M11) containing the gene of the drug

metabolizing BM3 M11, flanked by EcoRI and BamHI restriction sites. BM3 M11 contains

mutations R47L, E64G, F81I, F87V, E143G, L188Q, Y198C, E267V, H285Y and G415S when

compared to wild-type BM3 (10). The sequence of the forward primers was as follows: 5’-

GCA GGA GAC GGG TTA XXX ACT AGT TGG ACG CAT-3’. XXX represents the codon that

was used to introduce the specific mutation at position 87. The reverse primer for the

mutagenic PCR was a 34-mer 5’-CAT GCG TCC AAC TAG TYY YTA ACC CGT CTC CTG C-3’ in

which YYY is the reverse complement of codon XXX. The underlined bases indicate a new

SpeI digestion site. The following codons (XXX) were used: Ala, GCC; Arg, CGG; Asn, AAC;

Asp, GAC: Cys, UGC; Gln, CAG; Glu, GAG; Gly, GGG; His, CAC; Ile, AUC; Leu, CUG, Lys, AAG;

Met, AUG; Phe, UUC; Pro, CCC, Ser, UCC; Thr, ACC; Trp, UGG; Tyr, UAC, and Val, GUG.

After mutagenic PCR, the plasmids were digested with EcoRI and BamHI restriction

enzymes and the genes of mutated BM3 M11 were cloned into a pET28a+ vector, which

encodes for a N-terminal His-tag. The desired mutations in the P450 domain were

confirmed by DNA sequencing (Baseclear, Leiden, The Netherlands).

2.2.3 Expression and purification of the mutants

Expression of the BM3 mutants and wild-type BM3 was performed by transforming

competent Escherichia coli BL21 cells with the corresponding pET28+-vectors, as described

previously (21). Proteins were purified using nickel nitroacetic acid agarose, after which

P450 concentrations were determined according to Omura and Sato (30). The purity of the

enzymes was checked by SDS-PAGE electrophoresis on 12% gel and subsequent

Coomassie-staining.

2.2.4 Metabolism of alkoxyresorufins by BM3-mutants

The enzyme activities of the mutants and wild-type BM3 toward a homologues series of

alkoxyresorufins were measured according to Burke et al. (23) with modifications. To

determine O-dealkylation activities, fluorescence cuvettes (1.5 mL volume) were filled


52

with 860 µL of 100 mM potassium phosphate buffer (pH 7.4), 20 µL of 1 µM BM3 (20 nM

final concentration) and 20 µL of 1 mM alkoxyresorufine in DMSO (final concentrations: 20

µM alkoxyresorufin, 2% DMSO). After 30 seconds of preincubation, reactions were

started by addition of 100 µL of a mixture of 2 mM NADPH, 3 mM glucose-6-phosphate

and 0.4 units/mL glucose-6-phosphate dehydrogenase. The increase in fluorescence was

monitored at an excitation wavelength of 532 nm and an emission wavelength of 586 nm

for 2 minutes at 25°C. Specific activities were determined in triplicate by measuring the

initial slopes of resorufin formation. The Shimadzu RF-1501 spectrofluorimeter used was

calibrated by addition of 10 µL of 1 µM resorufine to the cuvet, and by recording the

increase in fluorescence (23). The results are means of duplicate determinations, with

duplicates typically differing by less than 5% from the mean.

To investigate the involvement of hydroxylation of n-alkoxy-substituents at other

carbons, 10 µL samples of incubations were analyzed by an Agilent 2000 UPLC-system

using a Zorbax Eclipse XDB-C18 column (1.8 µm, 50 x 4.6 mm; Agilent, USA). Samples were

eluted at a flow rate of 2 mL/min using a gradient composed of solvent A (99.8% water,

0.2% formic acid) and solvent B (99.8% acetonitrile, 0.2% formic acid). The gradient was

programmed as follows: from 0 to 1 minutes isocratic 40% B; 1 to 8 minutes linear

increase from 40% to 100% B; from 8 to 8.5 minutes isocratic 100% B; from 8.5 to 9

minutes, linear decrease from 100% to 40% B; from 9 to 10 minutes isocratic 40% B.

Alkoxyresorufines were detected at 460 nm. Under these conditions the following

retention times were obtained: methoxyresorufin, 2.84 min; ethoxyresorufin, 3.57 min; n-

propyloxyresorufin, 4.20 min; n-butoxyresorufin, 5.27 min; n-pentoxyresorufin, 6.06 min;

n-hexoxyresorufin, 6.82 min; n-heptoxyresorufin, 7.55 min; n-octoxyresorufin, 8.21 min.

2.2.5. Metabolism of testosterone by BM3 mutants

Incubations of BM3 mutants (200 nM) with testosterone (0.5 mM) as substrate were

performed at 25 oC in a total volume of 250 µL in 100 mM potassium phosphate buffer (pH

7.4). Reactions were started by adding 25 µL of a mixture of 2 mM NADPH, 3 mM glucose-

6-phosphate and 4 units/mL glucose-6-phosphate dehydrogenase (final concentrations:

0.2 mM NADPH, 0.3 mM glucose-6-phosphate and 0.4 units/mL glucose-6-phosphate

dehydrogenase), and were terminated after 60 minutes by addition of 250 µL of cold

methanol. Samples were centrifuged for 15 minutes at 4000 rpm, after which

supernatants were transferred to HPLC vials.

Testosterone and metabolites were analyzed by high-resolution UPLC (31) using an Agilent

2000-system using a Zorbax Eclipse XDB-C18 column (1.8 µm, 50 x 4.6 mm; Agilent, USA).

Metabolites and substrate were eluted at a flow-rate of 1.3 mL/min using a gradient

composed of solvent A (99.8% water, 0.2% formic acid) and solvent B (99.8% methanol,

0.2% formic acid) The gradient was programmed as follows: from 0 to 2 minutes linear

increase from 60% to 100% B; from 2 to 3 minutes isocratic 100% B; from 3 to 3.2 minutes


53

linear decrease from 100% to 60% B; from 3.2 to 5 minutes isocratic 60% B. Metabolites

and substrate were detected at 254 nm.

2.2.6. NADPH-Consumption

NADPH-consumption was quantified spectroscopically by monitoring the decrease in

absorbance at 340 nm using an extinction coefficient of 6210 M-1

cm-1

. Cuvettes (1.5 mL

volume; 1 cm path length) were filled with 880 µL of 100 mM potassium phosphate buffer

(pH 7.4) and 20 µL of 1 µM BM3 mutant (final concentration 20 nM). Reactions were

performed at 25 oC and started by addition of 100 µL of 2 mM NADPH (final concentration

200 µM), after which the decrease of NADPH was monitored for 2 minutes. To determine

substrate-induced NADPH-consumption, cuvettes were filled with 860 µL of 100 mM

potassium phosphate buffer (pH 7.4), 20 µL of 1 mM of alkoxyresorufin (dissolved in

DMSO) and 20 µL of 1 µM BM3 mutant. Reactions were started by addition of 100 µL of 2

mM NADPH, after which the decrease of NADPH was monitored for 2 minutes. To correct

for the solvent DMSO, incubations were also performed by adding 20 µL DMSO.

The coupling efficiencies of the BM3 mutants which were able to metabolize

alkoxyresorufins were calculated from the ratio of the initial rate of product formation and

the initial rate of NADPH consumption.

To investigate whether NADPH consumption is induced by binding to the

substrate binding site, incubations were also performed in presence of 1 µM

ketoconazole, which was shown to be a very potent inhibitor of BM3 mutants, including

BM3 M11 (13, 32). To test whether wild-type BM3 can also be inhibited by ketoconazole,

inhibition of NADPH consumption induced by 100 µM lauric acid was studied.

2.2.7. UV-vis Spectroscopy

For a selection of the BM3 mutants, the type of substrate binding and spectral dissociation

constants (KD) were determined for four substituted resorufins (methoxyresorufin, n-

butyloxyresorufin, n-heptoxyresorufin and benzyloxyresorufin) using UV-vis difference

spectroscopy. UV-vis spectra were recorded at room temperature using a Shimadzu UV-

2501PC spectrophotometer (Shimadzu Duisburg, Germany).

All substrate-induced binding spectra were recorded using tandem cuvettes (10 mm path

length) to eliminate absorbance by alkoxyresorufins.

Briefly, 1 mL of 2 μM enzyme in 100 mM potassium phosphate buffer (pH 7.4) was added

to one of the chambers of the sample and reference tandem cuvette. The other chambers

were filled with an equal volume of 100 mM potassium phosphate buffer. The enzyme-

containing chamber of the sample cuvette was titrated with microliter volumes of ethanol

solutions of alkoxyresorufins, resulting in final concentrations ranging from 1 to 30 μM. In

the reference cuvette, the same volume of alkoxyresorufin-solutions was added to the

chamber containing buffer only. To correct for effects of ethanol, equal volumes of


54

ethanol were added to the chambers without alkoxyresorufin. At each concentration of

substrate a UV-vis difference spectrum was recorded from 500 to 350 nm.

The observed differences in absorption between 390 nm (peak) and 419 nm (trough),

A390-419, were plotted versus alkoxyresorufin concentration after correction for dilution

and were analyzed by nonlinear regression, using GraphPad Prism 4 (GraphPad Sotware

Inc., San Diego, CA, USA). The spectral binding constants (KD) were determined by fitting

the titration binding curves using to the following equation:

A390-419= A∞*[S]n

KDn+[S]

n

Where A∞ represents the maximal difference at saturating alkoxyresorufin

concentration, KD is the dissociation constant of the enzyme-substrate complex and n is

the Hill coefficient.

2.3. Results

2.3.1. Expression and characterization of the BM3 M11-mutants.

Transformation of E.coli BL21 with the pET28+-vectors encoding the different BM3 M11-

mutants and subsequent purification resulted in significantly different yields of BM3

mutants, ranging from 27 to 666 nmol P450 per liter of original growth medium (Table 2).

After purification by nickel nitriloacetic acid agarose, protein purity was always higher

than 98% according to SDS-PAGE. For four of the purified mutants containing Pro87,

Asp87, Glu87 and Ser87, the reduced CO difference spectra only showed a peak at 420

nm. Formation of P420 is considered to result from coordination of a neutral thiol (in case

of BM3 Cys400) to the heme-iron, instead of a thiolate which is required to produce the

active P450 (33).

Mutation F87S, when applied to wild-type BM3 was previously shown to be an active

enzyme with changed regioselectivity in lauric acid hydroxylation (19). However, carbon-

monoxide difference spectroscopy only showed P420, suggesting that Ser87 is denaturing

owing to sodium dithionite treatment [V. Urlacher, personal communication]. Because

the concentration of these mutants could not be quantified they were not further

evaluated in our metabolism experiments.

The mutant containing Asn87 showed a significant peak at 420 nm with intensity of almost

equal to that at 450 nm. Mutants containing Met87, His87 and Gly87 showed a small

shoulder at 420 nm next to the peak at 450 nm. All other mutants only produced peaks

with maxima ranging from 448 and 450 nm.


55

Table 2: Spectroscopic properties and yield of the BM3 M11-mutants Reduced-CO difference spectra P450 Yield # Residue Max (nm) A450 /A420 ratio (nmoles/L of growth media) Non-polar side chain: 1. Phe87 450 > 100 165 2. Gly87 423, 449 78 148 3. Ala87 449 > 100 51 4. Leu87 449 > 100 237 5. Ile87 449 > 100 581 6. Val87

a) 449 > 100 666

7. Met87 423, 448 12.1 344 8. Pro87 420 n.d. n.q. 9. Trp87 450 > 100 27 Uncharged polar side chain: 10. Ser87 422 n.d. n.q. 11. Thr87 449 > 100 203 12. Asn87 421, 450 1.2 132 13. Gln87 448 > 100 49 14. Tyr87 448 > 100 585 15. Cys87 450 > 100 41 Charged polar side chain: 16. Lys87 448 > 100 91 17. Arg87 449 > 100 173 18. His87 423, 450 5.8 80 19. Asp87 420 n.d. n.q. 20. Glu87 420 n.d. n.q. a) Val87, BM3 M11, containing R47L, E64G, F81I, F87V, E143G, L188Q, Y198C, E267V, H285Y and G415S. n.d.: not detectable peak at 450 nm. n.q.: not quantifiable due to absence of peak at 450 nm.


56

2.3.2. Alkoxyresorufin metabolism by the BM3 M11-mutants.

2.3.2.1. O-dealkylation activities

Table 3 shows that the specific activities of O-dealkylation of nine different

alkoxyresorufins measured in incubations with sixteen variants of BM3 M11 were very

strongly affected by the nature of the amino acid at position 87. On average, the highest

O-dealkylation activities were observed with mutants containing non-polar side chains

Ala87, Gly87 and Val87. Of the polar residues, Tyr87 was most active, showing significant

activity with all alkoxyresorufins, except with methoxyresorufine as substrate. The

mutants containing the charged residues only showed low activities to all

alkoxyresorufins. Under the present conditions, Lys87, Met87 and Trp87 did not show

significant O-dealkylation activity with any alkoxyresorufin.

When comparing O-dealkylation activity between the different n-alkoxy-substituted

substrates (methoxyresorufin to n-octoxyresorufin), we found that the compounds with

shortest n-alkoxy-substituent were poorly metabolized. Methoxyresorufin was not O-

demethylated by any of the mutants, whereas for ethoxyresorufin and n-

propyloxyresorufine only two and three mutants, respectively, showed activity. The longer

the n-alkoxy-group, the more enzymes showed activity, although to very different extents.

The three most active mutants, Gly87, Ala87 and Val87, and mutant Tyr87 all showed the

highest activity with n-pentoxyresorufin as substrate; both increasing or decreasing the n-

alkoxy-substituent leads to gradual decrease in activity. For the mutants with polar side

chains, which show low O-dealkylation activity, generally the highest activity was found

with the longer n-alkoxy-substituents (n-hexoxyresorufin to n-octoxyresorufin).

In addition to the n-alkoxy-substituted substrates, benzyloxyresorufin was used to

characterize the different mutants. As shown in Table 3, benzyloxyresorufin is O-

alkylated at relatively high activity by several mutants when compared to their activities

toward the n-alkyoxyresorufins; for mutants Phe87, Leu87, Val87, Asn87, Gln87, Tyr87 and

Arg87 the highest O-dealkylation was observed with this substrate. Similar to the n-alkoxy

compounds, the highest activities were found in the mutants of small non-polar residues,

with the highest activity in the case of mutants containing Val87, Gly87 and Ala87.

The different alkoxyresorufins were also incubated with wild-type BM3 containing a

phenylalanine at position 87. Consistent with our previous study (13), wild-type BM3

showed no O-dealkylation activity with most alkoxyresorufins. Only with n-

heptoxyresorufin and n-octoxyresorufine a low but significant resorufine production was

found. To study whether long-chain alkoxyresorufins are hydroxylated at other positions

of the n-alkoxy-chain, which would not result in O-dealkylation, incubations were analysed

by UPLC with detection at 460 nm, which is absorption maximum of all alkoxyresorufins,

independent of the nature of n-alkoxy-substituent. Chromatograms of these incubations

measured at 460 nm, showed only strong peaks of the unchanged alkoxyresorufin


57

substrates. Only extremely small peaks (less than 3% of parent compound) were found

[data not shown], indicating that side-chain hydroxylations did not occur to a significant

extent.

2.3.2.2. NADPH-consumption and coupling efficiency

Table 4 shows the specific activities of the alkoxyresorufin-dependent NADPH

consumption in incubations with BM3 M11 mutants and wild-type BM3.

The incubation conditions were similar to those of the fluorimetric O-dealkylation assays,

except that the NADPH-regenerating system was excluded to enable NADPH-consumption

to be measured. Surprisingly, all of the mutants tested showed a very high specific activity

in NADPH-consumption. The highest activities were found in incubations with the mutant

containing residue His87. This mutant showed specific activities higher than 6000 nmol

NADPH/min/nmol P450, leading to complete consumption of NADPH within 4 minutes.

The lowest activities were found in mutants containing residues Tyr87, Ile87, Leu87 and

Met87.

Interestingly, for each mutant NADPH consumption was more or less similar for the

substrates with the shorter n-alkoxy substituents, methoxyresorufin to n-pentoxyresorufin

and allyloxyresorufin. For each mutant, the activities with these substrates varied by only

10 to 15%, which is close to the analytical error. These results suggest that for these

alkoxyresorufins, NADPH-oxidation is not influenced by the length of the n-alkoxy-group.

For substrates with longer n-alkoxy substituents, n-hexoxyresorufin to n-octoxyresorufin,

NADPH oxidation gradually decreased with increasing n-alkoxy chain length. Also, in case

of benzyloxyresorufin, having a bulky O-substituent, significantly lower NADPH-

consumption was observed when compared to the other substrates.

Because no significant other metabolic pathways (side-chain hydroxylation) were found,

the coupling efficiencies of the mutants were calculated from the ratio of the specific

activities of resorufin production and NADPH-consumption (Table 5). Because of the very

high rates of NADPH-consumption for all combinations of substrates and enzymes,

extremely low coupling efficiencies were found. The highest coupling efficiencies were

found with mutants containing Ala87 and Val87, with the highest coupling efficiencies of

0.68% (n-pentyl) and 0.75% (benzyl), respectively. For most reactions, the coupling

efficiencies were too low to quantify because O-dealkylation activities were below the

limit of detection. To test whether NADPH-consumption induced by alkoxyresorufin is the

result of binding to the substrate binding site, it was tested whether 1 µM ketoconazole

was able to block NADPH-consumption in incubations of wild-type BM3 and mutant Val87

(M11).


58

Table 3. Specific activitiesa)

of alkoxyresorufin O-dealkylation by BM3 M11-mutants and wild-type BM3.

a) Specific activities (nmol resorufin/min/nmol P450) observed at 20 µM alkoxyresorufin and 20 nM of P450 BM3. Values represent averages of three measurements; standard deviations were less than 10%. b) Val87 = BM3 M11. c) n.d., not detectable (< 0.01 nmol/min/nmol P450). d) WT (Phe87), wild-type BM3.

Methyl Ethyl n-Propyl n-Butyl n-Pentyl n-Hexyl n-Heptyl n-Octyl Benzyl # Residue (MROD) (EROD) (PrROD) (BuROD) (PROD) (HxROD) (HpROD) (OROD) (BROD)

Non-polar side chain: 1. Phe87 n.d.

c) n.d. n.d. n.d. n.d. 0.11 0.16 0.060 0.29

2. Gly87 n.d. n.d. n.d. 1.01 3.80 2.59 0.74 0.14 4.02 3. Ala87 n.d. 0.16 1.53 4.37 11.05 1.83 0.34 0.10 3.79 4. Leu87 n.d. n.d. n.d. n.d. 0.042 0.024 0.051 0.024 0.23 5. Ile87 n.d. n.d. n.d. n.d. 0.35 0.010 0.062 n.d. 0.031 6. Val87

b) n.d. n.d. 0.75 1.08 1.64 0.42 0.11 0.068 5.43

7. Met87 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 8. Pro87 - - - - - - - - - 9. Trp87 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. Uncharged polar side chain:

10. Ser87 - - - - - - - - - 11. Thr87 n.d. n.d. n.d. 0.12 n.d. 0.10 0.055 0.060 0.12 12. Asn87 n.d. n.d. n.d. n.d. n.d. 0.021 0.053 0.057 0.14 13. Gln87 n.d. n.d. n.d. n.d. n.d. 0.077 0.097 0.064 0.51 14. Tyr87 n.d. 0.024 0.13 0.14 0.55 0.076 0.017 0.031 0.92 15. Cys87 n.d. n.d. n.d. n.d. n.d. 0.026 n.d. n.d. n.d. Charged polar side chain: 16. Lys87 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 17. Arg87 n.d. n.d. n.d. n.d. n.d. 0.013 0.010 0.028 0.038 18. His87 n.d. n.d. n.d. n.d. n.d. 0.046 0.022 0.046 n.d. 19. Asp87 - - - - - - - - - 20. Glu87 - - - - - - - - - 21. WT (Phe87)

d) n.d. n.d. n.d. n.d. n.d. 0.17 0.18 n.d. n.d.


59

Table 4. Specific activities a)

of alkoxyresorufin-induced NADPH-consumption by BM3 M11 mutants and wild-typeBM3

a) Specific activities (nmol resorufin/min/nmol P450). Values represent averages of three measurements; standard deviations were less than 10%. b) Val87 = BM3 M11. c) n.d., not detectable (< 0.01 nmol/min/nmol P450). d) WT (Phe87), wild-type BM3.

Methyl Ethyl n-Propyl n-Butyl n-Pentyl n-Hexyl n-Heptyl n-Octyl Benzyl # Residue (MROD) (EROD) (PrROD) (BuROD) (PROD) (HxROD) (HpROD) (OROD) (BROD) DMSO

Non-polar side chain: 1. Phe87 4100 4590 3640 4520 4290 3120 1150 370 2170 < 10 2. Gly87 4380 3910 3950 4440 4170 2200 960 300 2360 105 3. Ala87 1860 1800 1510 1910 1620 550 300 130 950 90 4. Leu87 1180 1120 960 1110 540 260 270 190 190 95 5. Ile87 820 930 770 730 360 180 100 110 200 85 6. Val87

b) 2930 2660 2590 3180 2100 900 290 310 720 190

7. Met87 1030 760 660 920 880 140 110 75 400 25 8. Pro87 - - - - - - - - - - 9. Trp87 2090 1790 1450 1880 1980 870 320 130 1450 45 Uncharged polar side chain: 10. Ser87 - - - - - - - - - - 11. Thr87 1910 1570 1530 1900 1890 800 470 260 480 <10 12. Asn87 3260 2950 2450 2920 2970 1530 730 250 2260 <10 13. Gln87 1860 1740 1510 1770 1920 700 420 200 1130 40 14. Tyr87 700 680 620 710 690 400 130 87 370 <10 15. Cys87 3570 3490 3030 3430 3470 1590 560 150 2370 35 Charged polar side chain: 16. Lys87 3570 3400 3080 4100 3780 1580 600 210 2220 <10 17. Arg87 3740 3780 2950 3400 2970 840 240 97 1330 25 18. His87 6110 6080 5650 5600 5600 2240 1650 890 5100 <10 19. Asp87 - - - - - - - - - - 20. Glu87 - - - - - - - - - - 21. WT (Phe87) 4650 3970 3290 4220 4020 2260 1030 730 4110 45


60

Table 5. Coupling efficiency (%)a)

of alkoxyresorufin O-dealkylation by BM3 M11-mutants and wild-type BM3.

a) Coupling efficiencies were calculated by the formula: specific activity of resorufine production (Table 3) / specific activity of NADPH consumption (Table 4) x 100%. b) n.d., not detectable (< 0.001 %). c) Val87 = P450 BM3 M11. d) WT (Phe87), wild-type P450 BM3.

Methyl Ethyl n-Propyl n-Butyl n-Pentyl n-Hexyl n-Heptyl n-Octyl Benzyl # Residue (MROD) (EROD) (PrROD) (BuROD) (PROD) (HxROD) (HpROD) (OROD) (BROD)

Non-polar side chain: 1. Phe87 n.d.

b) n.d. n.d. n.d. n.d. 0.004 0.014 0.016 0.013

2. Gly87 n.d. n.d. n.d. 0.023 0.091 0.118 0.077 0.047 0.170 3. Ala87 n.d. n.d. 0.101 0.229 0.682 0.333 0.113 0.077 0.399 4. Leu87 n.d. n.d. n.d. n.d. 0.008 0.009 0.019 0.013 0.121 5. Ile87 n.d. n.d. n.d. n.d. 0.097 0.006 0.062 n.d. 0.016 6. Val87

c) n.d. n.d. 0.029 0.034 0.078 0.047 0.038 0.022 0.754

7. Met87 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 8. Pro87 - - - - - - - - - 9. Trp87 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d Uncharged polar side chain: 10. Ser87 - - - - - - - - - 11. Thr87 n.d. n.d. n.d. 0.006 n.d. 0.013 0.012 0.023 0.025 12. Asn87 n.d. n.d. n.d. n.d. n.d. 0.001 0.007 0.023 0.006 13. Gln87 n.d. n.d. n.d. n.d. n.d. 0.011 0.023 0.032 0.045 14. Tyr87 n.d. 0.004 0.021 0.020 0.080 0.019 0.013 0.036 0.249 15. Cys87 n.d. n.d. n.d. n.d. n.d. 0.002 n.d. n.d. n.d Charged polar side chain: 16. Lys87 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. 17. Arg87 n.d. n.d. n.d. n.d. n.d. 0.002 0.042 0.029 0.003 18. His87 n.d. n.d. n.d. n.d. n.d. 0.002 0.001 0.005 n.d 19. Asp87 - - - - - - - - - 20. Glu87 - - - - - - - - - 21. WT (Phe87)

d) n.d. n.d. n.d. n.d. n.d. n.d. 0.016 0.025 n.d.


61

To test whether this concentration of ketoconazole was able to inhibit BM3, we

first investigated whether it was able to block NADPH-consumption induced by 0.1 mM

lauric acid. With both wild-type BM3 and mutant Val87, lauric acid strongly stimulated

NADPH- consumption. Addition of 1 µM ketoconazole completely blocked lauric acid

induced NADPH-consumption of both enzymes. When wild-type BM3 and mutant Val87

(M11) were tested with benzyloxyresorufine as substrate, ketoconazole was also able to

almost completely block substrate-induced NADPH consumption as well as O-dealkylation

reaction [data not shown].

2.3.2.3. Binding spectra.

Titration of alkoxyresorufins to solutions of BM3 and the two selected BM3 M11 mutants

(Ala87 and Val87) in all cases resulted in typical type I difference spectra with a peak at

390 nm and a trough at 419 nm [data not shown], indicative of the conversion of the

heme iron from low spin to high spin. Table 6 shows the dissociation constants (KD)

obtained by nonlinear fitting of the A390–419 versus substrate concentration curves, after

correction for dilution. Surprisingly, methoxyresorufin, which is not O-demethylated by

any of the mutants, binds to the selected mutants with relatively high affinity with binding

constants ranging from 2.4 to 8 M (Table 6).

Table 6. Alkoxyresorufin binding by wild-type BM3 and BM3 M11-mutants.

a) Dissociation constants were determined by difference spectroscopy using 2 µM enzyme in 100 mM potassium phosphate buffer, pH 7.4, titrated with a solution of substrate. All substrates produced a Type I binding spectrum with a peak at 390 nm and trough at 419 nm. The data for the absorbance at 419 nm minus the absorbance at 390 were corrected for dilution and fitted to an equation for a bimolecular association reaction to obtain the dissociation constant.

Because enzyme activity measurements were carried out with 20 M substrate, it can be

concluded that the lack of O-dealkylation is most likely due to nonproductive binding

rather than lack of affinity for the enzyme. Wild-type BM3, which only showed very low O-

dealkylation activity toward n-heptoxyresorufin, showed higher affinity for the other

alkoxyresorufins which are not O-dealkylated. BM3 M11 mutants containing Ala87 and

Val87, which generally showed relatively high O-dealkylation activity, showed slightly

Methyl

(MROD)

n-Butyl

(BuROD)

n-Heptyl

(HpROD)

Benzyl

(BROD)

Substrate KD (µM) a)

KD (µM) a)

KD (µM) a)

KD (µM) a)

wt BM3 8.4 ± 1.4 1.6 ± 0.2 22 ± 6.4 2.7 ± 0.4

BM3 M11 Ala87 4.0 ± 0.5 1.1 ± 0.1 4.9 ± 0.4 1.4 ± 0.3

BM3 M11 Val87 2.4 ± 0.2 0.8 ± 0.1 1.4 ± 0.2 1.8 ± 0.3


62

higher affinity for the alkoxyresorufins than wild-type BM3. The only substrate showing

significantly lower affinity for wild-type BM3 when compared with the BM3 mutants was

n-heptoxyresorufin (Figure 1).

Figure 1. Structures and reactions of substrates used to characterize the BM3-mutants. (a) Alkoxyresorufine O-dealkylation: R = H (MROD), CH2- (EROD), C2H4 (PrROD), C3H6 (BuROD), C4H8 (PROD), C5H10 (HxROD), C6H12 (HpROD), C7H14 (OROD), CH2=CH- (AllROD), phenyl (BROD). (b) Testosterone hydroxylation: arrows indicate identified positions of hydroxylation.

2.3.3. Testosterone hydroxylation by the BM3 M11-mutants.

Analysis of incubations of the BM3 mutants with testosterone by UPLC showed that three

major metabolites were produced eluting at 0.93, 1.48 and 1.69 minutes. These

metabolites corresponded to the three hydroxytestosterone metabolites previously

observed with triple mutant BM3 R47L/F87V/L188Q (26). Figure 2 shows the

chromatograms obtained after incubations of testosterone with mutants containing

residues Tyr87, Ala87, Phe87 and Ile87, since they are representative of the diversity in

the metabolic profiles obtained.

The metabolite eluted at 1.48 min was previously identified as 16-hydroxytestosterone

(16-OH-T) (26). For structure elucidation of the other metabolites, large scale (50 mL)

incubations were performed for 3 hours containing 250 nM of mutant Val87 (P450 M11),


63

500 µM testosterone, 0.2 mM NADPH and a regenerating system (0.3 mM glucose-6-

phosphate and 0.4 U/mL glucose-6-phosphate dehydrogenase).

Metabolites were extracted by dichloromethane and isolated by preparative HPLC.

Structure identification was performed by a combination of 1D-

1H,

1H-

1H-DQF-COSY,

1H-

13C-HSQC,

1H-

13C-HMQC,

1H-

13C-HMBC and

1H-

1H-NOESY NMR-experiments, which will be

detailed in (34). The metabolite eluted at 0.93 min was identified as 15-

hydroxytestosterone (15-OH-T), whereas the metabolite eluted at 1.69 min was

identified as 2-hydroxytestosterone (2-OH-T). These two metabolites were previously

also identified by similar NMR experiments as products formed in incubations of

testosterone with housefly cytochrome P450 (CYP6A1) (35). Very minor metabolites were

eluting at 1.85 and 1.95 minutes; however, the concentration of these metabolites was

too low to allow structural assignment.

Figure 2. Ultraperformance liquid chromatography chromatograms obtained after incubations of testosterone (0.5 mM) with a selection of BM3 mutants (200 nM), representing the diversity in regioselectivity.

As shown in Table 7, all mutants were able to hydroxylate testosterone with widely

different activities and with significant differences in metabolic profiles.


64

Table 7. Total activity and regioselectivity of testosterone hydroxylation by BM3 M11-mutants and wild-type BM3.

Metabolite formationa)

# Residue 15-OH-T 16-OH-T 2-OH-T Total activity

a)

Non-polar side chain 1. Phe87 86 275 499 860 2. Gly87 31 13 12 56 3. Ala87 356 25 244 625 4. Leu87 2 9 35 46 5. Ile87 21 597 92 710 6. Val87

b) 641 351 218 1210

7. Met87 11 33 36 81 8. Pro87 - - - - 9. Trp87 15 6 8 29 Uncharged polar side chain 10. Ser87 - - - - 11. Thr87 126 367 557 1050 12. Asn87 n.d. n.d. 35 35 13. Gln87 21 n.d. 119 140 14. Tyr87 625 288 337 1250 15. Cys87 1 n.d. 4 5 Charged polar side chain 16. Lys87 3 9 13 25 17. Arg87 32 n.d. 33 65 18. His87 40 n.d. 135 175 19. Asp87 - - - - 20. Glu87 - - - - 21. WT (Phe87)

c) n.d. n.d. n.d. n.d.

a) Specific activities (nmol hydroxytestosterone/nmol P450/60 minutes) observed at 0.5 mM testosterone and 200 nM of P450 BM3. Values represent averages of two measurements; variability was always less than 10%. b) BM3 M11. c) WT (Phe87), wild-type BM3. d) n.d., not detectable.

The most active mutants were Tyr87, Val87, Thr87, Phe87, Ile87 and Ala87 (in decreasing

order); the other mutants showed activities less than 10% of that of Tyr87. In general,

mutants having a very low activity with alkoxyresorufins as substrates also showed a low

activity with testosterone as substrate, except for the mutants containing Phe87 and

Thr87 which showed relatively high activity of testosterone hydroxylation. Incubation of

wild-type BM3 did not show formation of any metabolites.

Figure 3 shows the metabolic profiles of the mutants ordered according to

similarities in metabolic profiles, and by preference for hydroxylation of D-ring (e.g. 15ß-


65

and 16ß-hydroxylation) versus the A-ring (2ß-hydroxylation). Numbers below the bars

represent the ratio of D-ring to A-ring hydroxylation, as calculated from the sum of peak

areas of 15ß-OH-T and 16ß-OH-T, divided by the peak are of 2ß-OH-T.

Figure 3. Effect of amino acid at position 87 on regioselective hydroxylation of testosterone by BM3 mutants. Bars represent percentage product relative to the total amount of products. Numbers below each bar represent the ratio of D-ring hydroxylation (sum of 15ß- and 16ß-hydroxylation) and A-ring hydroxylation (2ß-hydroxylation).

Mutants containing residues Ile87, Val87, Gly87, Tyr87, Trp87 and Ala87 all

showed a preference for hydroxylation of the D-ring of testosterone. For mutants

containing Val87, Gly87, Tyr87 and Trp87 very similar profiles were obtained with 15-OH-

T as major metabolite (50-55%), and with 16-OH-T and 2-OH-T formed in approximately

equal amounts (20-25%). Compared to these mutants, the mutant containing Ala87 and

Ile87 showed remarkable differences in regioselectivity of D-ring hydroxylation. The

mutant containing Ile87 was the only mutant having high selectivity for 16ß-hydroxylation,

whereas the mutant containing Ala87 showed predominantly 15ß-hydroxylation, in

addition to significant 2ß-hydroxylation at the A-ring.

Mutants containing Phe87 and Thr87, which had relatively high activity, showed

2-OH-T as major metabolite, with lower amounts of 16-OH-T and the lowest amounts of

15-OH-T. Similar results were obtained with the much less active mutants with residues

Met87, Lys87 and Leu87. Mutants containing Cys87, His87, Gln87 and Asn87 also showed

the highest preference of A-ring hydroxylation with 2-OH-T as major metabolite. For

three of these mutants D-ring hydroxylation was only found at the 15-position; for Asn87

no significant D-ring hydroxylation was observed.


66

2.4.1 Discussion The role of the residue at position 87 has been investigated in many studies using wild-

type BM3 and mutants containing one or two other mutations, Table 1. These studies are

hampered by the fact that only a limited number of substrates are accepted, such as

medium-chain and long-chain fatty acids and small terpenoid and aromatic substrates.

Also, so far only a limited number of amino acid substitutions have been evaluated at

position 87. By far most studies have been performed with F87A and to a less extent with

F87V and F87L. When applied to mutant 9-10A, containing R47C, V78A, K94I, P142S,

T175I, A184V, F205C, S226R, H236Q, E252G, R255S, A290V and L353V, addition of

mutation of F87A resulted in high enantioselectivity in buspirone hydroxylation (16),

showing that residue 87 also has an important role in more promiscuous mutants of BM3.

However, the effects of other amino acid substitutions were not reported.

In the present study, the role of position 87 in controlling substrate and

regioselectivity of drug-metabolizing mutant BM3 M11 was evaluated, by creation of

mutants with all twenty natural amino acids at this position. Twelve of the amino acid

substitution had not yet been reported previously in any BM3 variant. To characterize the

catalytic properties of these novel mutants, nine different alkoxyresorufins and

testosterone were used as probe substrates.

As shown in Table 3 and 7, very significant different rates of product formation

were found between the different mutants. In case of alkoxyresorufins as substrate, the

mutant containing Phe87 only showed a very low O-dealkylation activity with longer-chain

alkoxyresorufins starting from n-hexoxyresorufin, with activities similar to that obtained

with wild-type BM3, which also contains Phe87. The mutants with the small non-polar

amino acids Gly87, Ala87, Val87 and Ile87, all showed much higher O-dealkylation

activities with an optimal activity with n-pentoxyresorufine as substrate (Table 3). As

summarized in Table 1, replacement of the bulky Phe87 in wild-type BM3 by these small

amino acids also showed generally an increased activity toward non-fatty acid substrates.

Furthermore, the fact that mutants containing Val87, Leu87 and Ile87 showed large

differences in activity towards n-alkoxyresorufins, and strongly different regioselectivity in

testosterone hydroxylation, demonstrates that even minor changes in amino acid

properties of the residue at position 87 can have a large effect on the catalyic properties.

It was proposed previously that replacement of the bulky Phe87 by smaller non-

polar residues creates space for bulky substrates, allowing better positioning with respect

to the activated oxygen species, and as a result higher activities and coupling efficiencies

(8, 10). To test whether coupling efficiencies were affected by amino acid 87 substutions,

the ratio of product formation to NADPH-consumption was studied. As shown in Table 4,

extremely high activities of substrate-induced NADPH-consumption were found with all

mutants and, unexpectedly, wild-type BM3. For the most active enzymes, the specific


67

activities of NADPH-consumption were close to the high specific activities found with wild-

type BM3 and lauric acid and arachidonic acid as substrates (8, 9). Because resorufin was

the only product found, the coupling efficiency was always less than 1% for all productive

enzymes (Table 5). The fact that 20 µM of substrate resulted in complete depletion of 200

µM of NADPH within 2-3 minutes indicates that NADPH-consumption resulted from a

catalytic process in which the substrate triggers NADPH-consumption without being

converted, such as uncoupling, or redox-cycling in which the substrate undergoes one-

electron reduction followed by autoxidation by molecular oxygen.

Uncoupling of P450 is still a poorly understood process, and it is considered to

result from premature release of iron-bound molecular oxygen before completing the

catalytic cycle, producing superoxide anion radical (one-electron reduction) or hydrogen

peroxide (two-electron reduction; 'peroxide shunt') or by reduction of the [FeO]3+

intermediate producing water (oxidase pathway) (36, 37). Factors which might determine

the mode of uncoupling are active-site hydration, which might favor the peroxide shunt,

and the position of the substrate in the active site and/or a large distance of the [FeO]3+

intermediate to the substrate (36, 37). Because the mutants contain amino acids at

position 87 with different polarities and size, different modes of uncoupling might underly

the high NADPH-consumptions observed in the present study. An alternative mechanism

which might explain alkoxyresorufin-induced NADPH-consumption is reduction of

alkoxyresorufin at the level of the reductase domain. Previously it was shown that rat liver

microsomal NADPH-cytochrome P450 reductase was able to catalyze redox-cycling of

resorufin by one-electron reduction of the quinoneimine moiety (38). However, the fact

that ketoconazole was able to almost completely inhibit alkoxyresorufin-induced NADPH-

consumption by wild-type BM3 and mutant Val87, does not support this alternative

mechanism. Also, the fact that alkoxyresorufins produce type I binding spectra when

titrated to BM3 and BM3 mutants (Table 6) indicates that these substrates bind with

relatively high affinity to the substrate binding site of these enzymes. These results

strongly suggest that alkoxyresorufins bind to BM3 at the active site mainly in a

nonproductive orientation. However, the mechanism by which alkoxyresorufine

stimulated extremely high NADPH-consumption in these BM3 mutants and wild-type BM3

still remains to be elucidated.

Previously, several mutations of position 87 in wild-type BM3 were shown to change

regioselectivity and stereoselectivity of several reactions, Table 1. In this study,

testosterone was used as substrate to characterize the regioselectivity of the different

mutants. It was found that all mutants tested were able to catalyze testosterone

hydroxylation although at widely different activities and with different regioselectivities,

Table 6. Only three different metabolites were formed at significant amount, as was

shown previously in incubations with the triple-mutant of BM3, containing mutations

R47L, F87V and L188Q (26). Interestingly, the mutant containing Phe87 appeared to be a


68

relatively active enzyme, whereas the wild-type BM3 which also contains Phe87, was

completely inactive.

Structural identification of the metabolites by NMR revealed that two of the metabolites

result from hydroxylation of the D-ring, at positions 15ß and 16ß; the third metabolite

results from hydroxylation of the A-ring at position 2ß. Available crystal structures of BM3

indicate that its substrate-binding site is a long hydrophobic channel, with the heme iron

located at the bottom of this channel (6). This might explain why only the protons of the

A-ring and D-ring of testosterone might approach the reactive iron-oxo species close

enough to become hydroxylated.

As shown in Figure 3, several mutants have a strong preference for D-ring hydroxylation

whereas others prefer hydroxylation at the A-ring. The mutant containing isoleucine at

position 87 catalyzes predominantly 16ß-hydroxylation, whereas in case of the closely

related leucine amino acid testosterone hydroxylation takes place predominantly as 2ß-

hydroxylation. Why these relatively small change in amino acid side-chain have such a

large effect on regioselectivity remains to be established however. Screening of over 200

microbial P450s (27) and genetic engineering of bacterial P450s previously enabled the

identification of several other P450s that can catalyze 2ß- and 15ß-hydroxylation of

testosterone (27, 39). Although several enzymes were shown to catalyze

16hydroxylation, so far no bacterial P450s have been reported that can catalyse 16ß-

hydroxylation of steroids. The present study shows that the BM3 mutant containing Ile87

is the first mutant with high selectivity for 16ß-hydroxylation.

As summarized in Table 1, by far most amino acid substitutions at position 87 which have

been reported involved mutations of Phe87 to amino acids with small non-polar side

chains. Only two amino acids (Tyr and Ser) with uncharged polar side chains have been

studied, whereas charged polar side chains have not been evaluated so far. Mutation

F87Y, when applied to wild-type BM3, resulted in an unproductive enzyme when long-

chain fatty acids were used as substrate (8, 9, 17).

The increased polarity caused by the hydroxyl-group at the phenyl-ring of Tyr87 was

considered to restrict heme accessibility of the fatty acid substrates, and as a result causes

full uncoupling (8). In the present study the mutant containing Tyr87 appeared more

active than that containing Phe87 with both alkoxyresorufins and testosterone as

substrate. Although Ser87 was not evaluated because carbon monoxide difference

spectrum only showed P420 spectrum, the mutant containing Thr87 showed a P450

spectrum, and was one of the most active mutants with testosterone as substrate, Table

7. The other uncharged and charged polar amino acids also showed enzyme activity,

although this was generally low when compared with the enzyme activity of the non-polar

amino acids. However, the substrates used in the present study all concerned uncharged

substrates. Therefore, it remains to be evaluated wheter the mutants containing the polar


69

amino acids at position 87 will have higher affinity and activity with charged and polar

substrates.

In conclusion, the results of the present study show that, consistent with the

results of studies performed with wild-type enzyme, the nature of amino acid at position

87 has a strong effect on activity and regioselectivity of drug metabolizing mutants of

BM3. Several amino acid substitutions not previously evaluated were shown to be,

dependent on the substrate tested, active enzymes with different substrate- and regio-

selectivity. Because of the wide substrate selectivity of BM3 M11 when compared to wild-

type BM3, this panel of mutants will be useful as biocatalysts for metabolite production.

Furthermore, these mutants might be valuable model proteins for mechanistic studies on

the function of P450s in drug metabolism.


70

References

1. Rankin, S. C., Aceto, H., Cassidy, J., Holt, J., Young, S., Love, B., Tewari, D., Munro, D. S., and Benson, C. E. (2002) Molecular characterization of cephalosporin-resistant Salmonella enterica serotype Newport isolates from animals in Pennsylvania, J Clin Microbiol 40, 4679-4684.

2. 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.

3. Eiben, S., Kaysser, L., Maurer, S., Kuhnel, K., Urlacher, V. B., and Schmid, R. D. (2006) Preparative use of isolated CYP102 monooxygenases -- a critical appraisal, J Biotechnol 124, 662-669.



6. 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.


8. 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.

9. 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 ( Pt 2), 371-379.

10. Carmichael, A. B., and Wong, L. L. (2001) Protein engineering of Bacillus megaterium CYP102. The oxidation of polycyclic aromatic hydrocarbons, European journal of biochemistry / FEBS 268, 3117-3125.

11. 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 letters 508, 249-252.

12. Li, Q. S., Ogawa, J., Schmid, R. D., and Shimizu, S. (2001) Engineering cytochrome P450 BM-3 for oxidation of polycyclic aromatic hydrocarbons, Applied and environmental microbiology 67, 5735-5739.


71

13. 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, Analytical biochemistry 341, 148-155.

14. 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, Organic & biomolecular chemistry 3, 57-64.

15. Sulistyaningdyah, W. T., Ogawa, J., Li, Q. S., Maeda, C., Yano, Y., Schmid, R. D., and Shimizu, S. (2005) Hydroxylation activity of P450 BM-3 mutant F87V towards aromatic compounds and its application to the synthesis of hydroquinone derivatives from phenolic compounds, Applied microbiology and biotechnology 67, 556-562.

16. Landwehr, M., Hochrein, L., Otey, C. R., Kasrayan, A., Backvall, J. E., and Arnold, F. H. (2006) Enantioselective alpha-hydroxylation of 2-arylacetic acid derivatives and buspirone catalyzed by engineered cytochrome P450 BM-3, Journal of the American Chemical Society 128, 6058-6059.

17. 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.

18. Chen, C. K., Shokhireva, T., Berry, R. E., Zhang, H., and Walker, F. A. (2008) The effect of mutation of F87 on the properties of CYP102A1-CYP4C7 chimeras: altered regiospecificity and substrate selectivity, Journal of biological inorganic chemistry : JBIC : a publication of the Society of Biological Inorganic Chemistry 13, 813-824.

19. 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, Journal of biotechnology 139, 115-117.

20. Kim, D. H., Ahn, T., Jung, H. C., Pan, J. G., and Yun, C. H. (2009) Generation of the human metabolite piceatannol from the anticancer-preventive agent resveratrol by bacterial cytochrome P450 BM3, Drug metabolism and disposition: the biological fate of chemicals 37, 932-936.

21. Seifert, A., Vomund, S., Grohmann, K., Kriening, S., Urlacher, V. B., Laschat, S., and Pleiss, J. (2009) Rational design of a minimal and highly enriched CYP102A1 mutant library with improved regio-, stereo- and chemoselectivity, Chembiochem : a European journal of chemical biology 10, 853-861.

22. 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, Chemico-biological interactions 171, 96-107.

23. Burke, M. D., Thompson, S., Weaver, R.J., Wolf, C.R. and Mayer, R.T. (1994) Cytochrome P450 specificities of alkoxyresorufin O-dealkylation in human and rat liver, Biochem. Pharmacol. 48, 923-936.

24. Burke, M. D., and Mayer, R. T. (1983) Differential effects of phenobarbitone and 3-methylcholanthrene induction on the hepatic microsomal metabolism and


72

cytochrome P-450-binding of phenoxazone and a homologous series of its n-alkyl ethers (alkoxyresorufins), Chemico-biological interactions 45, 243-258.

25. Burke, M. D., Thompson, S., Elcombe, C.R., Halpert, J., Haaparanta, T., Mayer, R.T. (1983) Ethoxy-, pentoxy- and benzyloxyphenoxazones and homologues: a series of substrates to distinguish between different induced cytochromes P450, Biochem Pharmacol 34, 3337-3345.


27. de Vlieger, J. S., Kolkman, A. J., Ampt, K. A., Commandeur, J. N., Vermeulen, N. P., Kool, J., Wijmenga, S. S., Niessen, W. M., Irth, H., and Honing, M. (2010) Determination and identification of estrogenic compounds generated with biosynthetic enzymes using hyphenated screening assays, high resolution mass spectrometry and off-line NMR, Journal of chromatography. B, Analytical technologies in the biomedical and life sciences 878, 667-674.


29. Fernandes, P., Cruz, A., Angelova, B., Pinheiro, H.M., Cabral, J.M.S. (2003) Microbial conversion of steroid compounds: recent developments, Bioscience, biotechnology, and biochemistry 70, 307-311.

30. Omura, T., and Sato, R. (1964) The Carbon Monoxide-Binding Pigment of Liver Microsomes. Ii. Solubilization, Purification, and Properties, The Journal of biological chemistry 239, 2379-2385.

31. Wang, D., and Zhang, M. (2007) Rapid quantitation of testosterone hydroxyl metabolites by ultra-performance liquid chromatography and mass spectrometry, Journal of chromatography. B, Analytical technologies in the biomedical and life sciences 855, 290-294.

32. 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, Journal of biomolecular screening 16, 239-250.

33. 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, Proceedings of the National Academy of Sciences of the United States of America 100, 3641-3646.

34. 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.

35. Jacobsen, N. E., Kover, K. E., Murataliev, M. B., Feyereisen, R., and Walker, F. A. (2006) Structure and stereochemistry of products of hydroxylation of human


73

steroid hormones by a housefly cytochrome P450 (CYP6A1), Magnetic resonance in chemistry : MRC 44, 467-474.

36. 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.

37. Yeom, H., and Sligar, S. G. (1997) Oxygen activation by cytochrome P450BM-3: effects of mutating an active site acidic residue, Archives of biochemistry and biophysics 337, 209-216.

38. Dutton, D. R., Reed, G. A., and Parkinson, A. (1989) Redox cycling of resorufin catalyzed by rat liver microsomal NADPH-cytochrome P450 reductase, Archives of biochemistry and biophysics 268, 605-616.

39. Virus, C., Lisurek, M., Simgen, B., Hannemann, F., and Bernhardt, R. (2006) Function and engineering of the 15beta-hydroxylase CYP106A2, Biochemical Society transactions 34, 1215-1218.

CHAPTER 3

ROLE OF RESIDUE 87 IN THE ACTIVITY AND REGIOSELECTIVITY OF

CLOZAPINE METABOLISM BY DRUG METABOLIZING BM3 M11:

APPLICATION FOR STRUCTURAL CHARACTERIZATION OF CLOZAPINE

GSH CONJUGATES

Vanina Rea, Sanja Dragovic, Jan Simon Boerma, Frans J.J. de Kanter, Nico

P.E.Vermeulen and Jan N.M. Commandeur

adapted from Drug Metabolism and Disposition, (2011), 39 (12), 2411-2420

Chapter 3 Generation of reactive intermediates

Abstract

In the present study, a site-saturation mutagenesis library of drug metabolizing BM3 M11

with all 20 amino acids at position 87 was applied as biocatalyst for the production of

stable and reactive metabolites of clozapine. Clozapine is an atypical antipsychotic drug

where formation of reactive metabolites is considered to be responsible for several

adverse drug reactions. Reactive intermediates of clozapine can be inactivated by GSH to

multiple GSH conjugates, by non-enzymatic and glutathione S-transferase (GST) mediated

conjugation reactions. The structures of several GST dependent metabolites have not yet

been elucidated unequivocally. The present study shows that the nature of amino acid at

position 87 of BM3 M11 strongly determines both activity and regioselectivity of clozapine

metabolism. Some mutants showed preference for N-demethylation and N-oxidation,

whereas others showed high selectivity for bioactivation to reactive intermediates. The

mutant containing Phe87 showed both high activity and high selectivity for the

bioactivation pathway and was used for the large scale production of GST dependent GSH

conjugates by incubation in presence of recombinant human glutathione S-transferase P1-

1. Five human relevant GSH adducts were produced at high levels enabling structural

characterization by 1H-NMR. This work shows that drug metabolizing BM3-mutants, in

combination with GSTs, are very useful tools for the generation of GSH conjugates of

reactive metabolites of drugs in order to enable their isolation and structural elucidation.


76

3.1. Introduction

Cytochrome P450 (P450s) are involved in the metabolism of approximately 80% of the

drugs currently on the market (1, 2). In some cases drugs can be oxidized by P450s to

electrophilic reactive intermediates, which subsequently can react with nucleophilic

functional groups in biomolecules such as proteins and DNA. Also, stable metabolites

might possess pharmacological activities that might be responsible for undesired adverse

drug reactions. It is for these reasons that also the characterization of the biological

properties of major metabolites is considered to be important for drug safety assessment

(3). Therefore, methods are required to obtain the relevant drug metabolites in sufficient

yield to allow structural elucidation and to study their pharmacological and toxicological

properties.

Metabolite production can be achieved by organic synthesis, electrochemical oxidation of

parent drug and by biosynthesis using specific P450s. In particular, mutants of the

bacterial cytochrome P450 BM3 (P450 BM3) are considered to have good perspective for

the large scale production of human relevant drug metabolites, as this very stable enzyme

possesses the highest activity ever recorded for a P450 (4). By combinations of site-

directed and random mutagenesis many BM3 mutants have been obtained which are able

to convert drugs and drug-like molecules to human relevant metabolites (5-8). In our

previous work, four mutants of BM3 have been evaluated as biocatalysts for the

bioactivation of several drugs to reactive intermediates (9). Drugs tested were

acetaminophen, diclofenac and clozapine (CLZ), and formation of reactive intermediates

was analyzed by measurement of GSH conjugates. For all drugs tested, most stable

metabolites and reactive intermediates were produced at much higher activity by the BM3

mutants than by human and rat liver microsomes, supporting their potential for use in

characterization of toxicologically relevant metabolites (9).

Recently, the highly active drug-metabolizing mutant BM3 M11 was used to investigate

the role of human glutathione S-transferases in the inactivation of CLZ (10). CLZ is an

atypical antipsychotic drug showing a low incidence of extrapyramidal side effects

combined with excellent antipsychotic efficacy in schizophrenic and manic treatment-

resistant patients (11-13). Approximately 1-2% of patients develop agranulocytosis.

Enhanced serum transaminases were monitored with 37% of the patients while 0.06% of

the patients had liver failure (14). It is still unknown which factors predispose part of the

patient population to these forms of CLZ toxicity. Based on the identification of several

GSH conjugates, formation of a reactive nitrenium ion by peroxidases, hypochlorite and

P450s has been proposed as a possible explanation for these ADRs (15-18). In vitro and in

vivo studies of CLZ have shown the formation of four GSH-conjugates with identical mass

with MH+ ion at m/z 632.2 and one deschlorinated GSH conjugate with MH

+ ion at m/z

598.3. All conjugates can be explained by direct conjugation of GSH at different positions

of a reactive nitrenium ion and by chloro-substitution of the nitrenium ion followed by

reduction, see Figure 1 (10, 17). Unequivocal structure determination by 1H-NMR has been


77

published for only two of the glutathione conjugates of CLZ (15, 16, 19). The major GSH

conjugate (CG-1, Figure 1) was found to be conjugated at position 6 of the chlorinated

aromatic ring, whereas a minor GSH conjugate (CG-3, Figure 1) was found to be

conjugated at position 9 (15, 16). A third GSH conjugate with a MH+ ion at m/z 632.2 was

only identified in in vitro incubations with human and rat liver microsomes and was

tentatively assigned to position 7 (CG-4, Figure 1).

Two GSH-conjugates, with MH+ ion at m/z 632.2 and 598.3, were first discovered

in bile of treated mice and rats and were originally proposed to originate from

unidentified reactive intermediates formed in vivo (17). However, we recently

demonstrated that these GSH conjugates, CG-5 and CG-6 in Figure 1, are formed at high

levels when CLZ incubations with purified BM3 M11 and human liver microsomes were

supplemented with human glutathione S-transferases (10). Three of the four tested

human glutathione S-transferases (hGSTs) showed strongly increased total GSH

conjugation and also resulted in formation of different regioisomeric GSH conjugates of

CLZ, Figure 1 (10). For two of the GSH conjugates that have been found previously, the

structure has not been elucidated by NMR. Conjugate CG-4, which was found in

incubations of CLZ with human liver microsomes, was tentatively assigned as conjugate at

position 7 (17). GSH conjugate CG-5, which was identified in bile of rats and mice, was

proposed to result from GSH conjugation to the non-chlorinated ring (17).

The aim of the present study was to identify the structures of these GSH

conjugates by 1H-NMR, by performing large scale incubations of CLZ with selective BM3

M11-mutants in presence of glutathione S-transferase. Because GST P1-1 appeared to be

the most active hGST in the formation of enzyme-dependent GSH conjugates (10), this

enzyme was selected for large scale production of GSH conjugates. Although BM3 M11

was previously shown to produce high levels of CLZ metabolites, the most abundant

metabolites appeared to be N-demethylclozapine and CLZ N-oxide (9, 10). As a

consequence, also GSH conjugates derived from these metabolites were produced, which

strongly complicates isolation of CG-4 and CG-5. Recently, we showed that by changing

the active site residue at position 87 of BM3 M11 the regioselectivity of testosterone

hydroxylation was strongly modified (20).

Therefore, in the present study, we first evaluated the effect of mutation at position 87 on

the regioselectivity of CLZ metabolism in order to identify the most suitable biocatalyst for

the bioactivation of CLZ and subsequent structural characterization of the formed GSH

conjugates. The results show that the nature of the residue at position 87 strongly

influences regioselectivity of CLZ metabolism and that by using a more selective P450 BM3

M11-mutant all five human relevant GSH conjugates of CLZ could be produced in high

levels, enabling structural elucidation by 1H-NMR.

Fig

.ure

1.

Ox

ida

tiv

e p

ath

wa

ys

of

met

ab

oli

sm o

f C

LZ

by

cy

toch

rom

e P

450

an

d n

on

-en

zym

ati

c a

nd

en

zym

ati

c co

nju

ga

tio

n r

ea

ctio

ns

of

rea

ctiv

e C

LZ

nit

ren

ium

io

n b

y g

luta

thio

ne

(GS

H)

an

d g

luta

thio

ne

S-t

ran

sfer

ase

(G

ST

): a

) N

-dem

eth

yla

tio

n;

b)

N-o

xid

ati

on

; c)

ox

ida

tiv

e o

pen

ing

of

pip

era

zin

e-

rin

g;

d)

deh

yd

rog

ena

tio

n t

o n

itre

niu

m i

on

.


Figure 1. Oxidative pathways of metabolism of CLZ by cytochrome P450 and non-enzymatic and enzymatic conjugation reactions of reactive CLZ nitrenium ion by glutathione (GSH) and glutathione S-transferase (GST): a) N-demethylation; b) N-oxidation; c) oxidative opening of piperazine-ring; d) dehydrogenation to nitrenium ion.


3.2. Materials and methods

3.2.1. Materials

All chemicals were of analytical grade and obtained from standard suppliers.

3.2.2. Library construction

Site-directed mutants of BM3 M11 at position 87 were constructed by mutagenic PCR

using the Stratagene QuikChange XL site–directed mutagenesis kit (Stratagene, La Jolla,

CA, USA) using 20 complementary pairs of mutagenesis primers (20). The mutagenic PCR

was applied to a pBluescript vector containing the gene of the drug metabolizing BM3

M11, flanked by EcoR1 and BamH1 restriction sites. BM3 M11 contains mutations R47L,

E64G, F81I, F87V, E143G, L188Q, Y198C, E267V, H285Y and G415S when compared to

wild-type BM3 (9).

The sequence of the forward primer was as follows: 5’-GCA GGA GAC GGG TTA XXX ACT

AGT TGG ACG CAT-3’. The XXX represents the codon that was used to introduce the

specific mutation at position 87. The reverse primer for the mutagenic PCR was a 34-mer

5’-CAT GCG TCC AAC TAG TYY YTA ACC CGT CTC CTG C-3’ in which the YYY is the reverse

complement of codon XXX. The underlined bases indicate a new SpeI digestion site. The

following codons (XXX) were used: Ala, GCC; Arg, CGG; Asn, AAC; Asp, GAC: Cys, UGC; Gln,

CAG; Glu, GAG; Gly, GGG; His, CAC; Ile, AUC; Leu, CUG, Lys, AAG; Met, AUG; Phe, UUC;

Pro, CCC, Ser, UCC; Thr, ACC; Trp, UGG; Tyr, UAC, and Val, GUG.

After mutagenic PCR, the plasmids were digested with EcoR1 and BamH1 restriction

enzymes and the genes of mutated BM3 M11 were cloned into a pET28a+ vector, which

encodes for a N-terminal His-tag. The desired mutations in the P450 domain were

confirmed by DNA sequencing (Baseclear, Leiden, The Netherlands).

3.2.3. Expression, isolation and purification of enzymes

Expression of the P450 BM3 M11 mutants was performed by transforming competent

E.coli BL21 cells with the pET28+-vectors, as described previously (20). Proteins were

purified using Ni-NTA agarose, after which P450 concentrations were determined using

carbon monoxide (CO) difference spectrum assay. Purity of the enzymes was checked by

SDS-PAGE electrophoresis on 12% gel and Coomassie-staining. Protein purity was higher

than 98% in all samples obtained.

Human GST P1-1 was prepared and purified as described previously (10). Protein

concentration was determined according to the method of Bradford (21) with reagent

obtained from Bio-Rad. The specific activity of the purified GST was assayed according to

Habig et al. (22). The specific activity of the purified recombinant human GST P1-1 using

CDNB as a substrate was 27.9 µmol/min/mg protein.


80

3.2.4. Metabolism of CLZ by BM3 M11-mutants in presence of human GST P1-1

Incubations using BM3 M11 mutants as bioactivation system were performed at a final

enzyme concentration of 250 nM, as described previously (9). All incubations were

performed in 100 mM potassium phosphate buffer (pH 7.4) and at a final volume of 250

L. The substrate CLZ was incubated at a concentration of 500 M. GST P1-1 (8 M) and

GSH (100 M) were added to the incubations in order to trap reactive CLZ nitrenium ion.

Reactions were initiated by the addition of a NADPH regenerating system (0.2 mM

NADPH, 1 mM glucose-6-phosphate, 0.4 U/mL glucose-6-phosphate dehydrogenase, final

concentrations) and performed for 30 min at room temperature. In this time period,

product formation was linear, as described previously (9). Reactions were terminated by

addition of 25 L of 10% HClO4, and centrifuged for 15 min at 14000 rpm. The

supernatants were analyzed by reversed-phase liquid chromatography using a Luna 5 m

C18 column (150 mm x 4.6 mm i.d.; Phenomenex, Torrance, CA, USA) as stationary phase,

protected by a 4.0 mm × 3.0 mm i.d. security guard (5 μm) C18 guard column

(Phenomenex). The gradient used was constructed by mixing the following mobile phases:

eluent A (0.8 % acetonitrile, 99 % water, and 0.2 % formic acid) and eluent B (99 %

acetonitrile, 0.8 % water, and 0.2 % formic acid). The first 5 min were isocratic at 0 %

eluent B; from 5 to 30 min the percentage of eluent B linearly increases to 100 %; from 30

to 35 min linear decrease to 0 % B and maintained at 0 % for re-equilibration until 40 min.

The flow rate was 0.5 mL/min.

Samples were analyzed using LC-MS/MS for identification and using UV/vis

detection at 254 nm for quantification. The Shimadzu Class VP 4.3 software package

(Shimadzu, Kyoto, Japan) was used for determination of peak areas in the UV

chromatograms. A standard curve of CLZ was used to estimate the concentrations of the

metabolites, assuming that the extinction coefficients of the metabolites at 254 nm are

equal to that of CLZ. UV/vis spectra of clozapine and its metabolites, as determined online

by diode array detection (180-400 nm), all showed similar spectra with maxima at 240,

260 and 295 nm [data not shown]. The standard curve of CLZ was linear between 1 and

100 µM; the limit of quantitative detection by UV/vis was estimated to be 0.1 µM [data

not shown].

For identification of the metabolites, an Agilent 1200 series rapid resolution LC

system was connected to a hybrid quadrupole-time-of-flight (Q-TOF) Agilent 6520 mass

spectrometer (Agilent Technologies, Waldbronn, Germany), equipped with an

electrospray ionization (ESI) source and operating in the positive mode. The MS ion source

parameters were set with a capillary voltage at 3500 V; nitrogen was used as the

desolvation and nebulizing gas at a constant gas temperature of 350°C, drying gas 8 L/min

and nebulizer 40 psig. Nitrogen was used as a collision gas with collision energy of 25V. MS

spectra were acquired in full scan analysis over an m/z range of 50 to 1000 using a scan

rate of 1.003 spectra/s. The MassHunter Workstation Software (version B.02.00, Agilent


81

Technologies) was used for system operation and data collection. Data analysis was

performed using Agilent MassHunter Qualitative analysis software.

3.2.5. Preparative scale biotransformation

The CLZ GSH conjugates produced on a preparative scale by large scale incubation with

the most selective BM3 M11 mutant as biocatalyst. A 5 mL reaction volume containing

purified enzyme (1 M), CLZ (500 M), GSH (100 M), GST P1-1 (8 M), and an NADPH

regenerating system (0.2 mM NADPH, 1 mM glucose-6-phosphate, 0.4 U/mL glucose-6-

phosphate dehydrogenase) was prepared in potassium phosphate buffer (100 mM, pH

7.4). The reaction was allowed to continue for 6 h at 25°C. To achieve maximal conversion

of CLZ, the incubation was supplemented every hour with 40 µL of 120 µM M11 Phe87, 20

µL of 25 mM GSH, 500 L of the NADPH regenerating system and 100 µL of 200 M GST

P1-1. The final incubation volume was 8.4 mL. The reaction was terminated by adding 0.84

mL of 10 % HClO4 and centrifuged for 15 min at 14000 rpm. The supernatant was applied

to a Strata X C-18 solid-phase extraction column (200mg/3mL, Phenomenex).

The column was washed with 5 ml of H2O to remove salts and proteins. CLZ and

its metabolites were subsequently eluted using 2 mL of methanol. The sample was

evaporated to dryness and reconstituted in 2 mL of eluent A (0.8% acetonitrile, 99%

water, and 0.2% formic acid). The sample was applied by manual injection on a

preparative chromatography column Luna 5 m C18(2) column (250 mm x 100 mm i.d.)

from Phenomenex, which was previously equilibrated with 100% eluent A. A flow rate of 2

mL/min and a gradient using eluent A (0.8% acetonitrile, 99% water, and 0.2% formic acid)

and B (99% acetonitrile, 1% water, and 0.2% formic acid) was applied for separation of

formed CLZ metabolites. The first 10 min were isocratic at 0% eluent B; from 10 to 65 min,

the percentage of eluent B increased linearly to 65%; from 65 to 70 min further increase

of eluent B to 100%; from 70 to 80 min, there was a linear decrease to 0% B, and re-

equilibration was maintained until 120 min. Metabolites were detected using UV

detection (254 nm) and collected manually.

Collected fractions were first analyzed for purity and identity by the analytical

HPLC and LC-MS/MS methods as described above. The samples were evaporated to

dryness under nitrogen stream and dissolved in 1 mL deuterium oxide to exchange acidic

hydrogen atoms by deuterium atoms. Samples were evaporated to dryness in the vacuum

concentrator, the residues were redissolved in 500 L of methyl alcohol-d4 and 1H-NMR

spectra were recorded at room temperature. 1H-NMR-analysis was performed on Bruker

Avance 500 (Milan, Italy), equipped with cryoprobe. 1H-NMR measurements were carried

out at 500.23 MHz.


82

3.3. Results

3.3.1. Expression of BM3 M11 mutants

A saturation mutagenesis library with a different residue at position 87 of BM3 M11 was

recently created in our laboratory (20). All 20 mutants were expressed in E.coli BL21 with

pET28+-vectors; the P450 quantification was done by CO difference spectrum. For the

mutants Pro87, Asp87 and Ser87, the reduced CO difference spectra only showed a peak

at 420 nm, suggesting that these amino acids negatively affect the folding and/or stability

of BM3 M11. The mutant containing Asn87 showed a significant peak at 420 nm with

intensity of almost equal to that at 450 nm. Mutants containing Met87, His87 and Gly87

showed a small shoulder at 420 nm next to the peak at 450 nm. All other mutants only

produced peaks with maxima ranging from 448 nm to 450 nm (20).

3.3.2. Activity and regioselectivity of metabolism of CLZ by BM3 M11-mutants

When CLZ was incubated at analytical scale with the 20 different BM3 M11 mutants in

presence of GSH and hGST P1-1, 13 different metabolites were found in total (Table 1),

which is consistent with our previous studies (9, 10). Five of the metabolites result from

N-oxidation (C-1), N-demethylation (C-2) and piperazine ring cleavage (C-3), and

combinations of these (C-4 and C-5), Figure 1.

In total, eight different GSH conjugates were found resulting from bioactivation of CLZ to

reactive intermediates. Five of these GSH conjugates result from addition reactions of the

CLZ nitrenium ion with GSH (CG-1, CG-3, CG-4, CG-5) and chlorine-substitution (CG-6). The

LC-MS/MS spectra of these five GSH-conjugates are tabulated in Table 2; the assignment

of the fragments have been described elsewhere (9, 10, 17). GSH conjugates designated

CG-2 and CG-8 were found to be secondary GSH conjugates, resulting from bioactivation

of N-demethylclozapine to its corresponding nitrenium ion and subsequent addition (CG-

2) and chlorine-substitution (CG-8). CG-7 corresponds to a di-GSH conjugate that most

likely results from GSH conjugation to the GSH containing nitrenium ion formed after

chlorine substitution of the CLZ nitrenium ion, Figure 1.

As shown in Table 1, the nature of amino acid residue at position 87 has strong influence

on both the activity and regioselectivity of formation of CLZ metabolites. As indicated in

the last column, the highest activity was generally observed with mutants containing

apolar amino acids at position 87. The mutants containing Ala87, Val87 and Ile87 showed

the highest activity, followed by Phe87 and Trp87. Mutants containing Leu87, Met87,

Gly87 and Pro87 showed only very low activity (<6% conversion). Among the mutants

containing polar uncharged residues, Tyr87 and Gln87 were the most active, showing

25.7% and 9.4% conversion, respectively. The mutants containing negative charged

residue Asp87 and Glu87 had low activity with 7.4% and 2.4% conversion, respectively.


Table 1. Effect of aminoacid at position 87 in BM3 M11H on the GSH-conjugation to reactive metabolites formed by oxidative bioactivation of clozapine. C-1 C-2 C-3 C-4 C-5 CG-1 CG-2 CG-3 CG-4 CG-5 CG-6 CG-7 CG-8 Specific

activity* % conversion

m/z (MH

+)

343 313 301 287 329 632 618 632 632 632 598 903 584

Apolar side chain Gly87 0.4 0.7 0.1 0.0 0.0 1.1 0.0 0.1 0.0 1.2 2.7 0.5 0.4 1.2 1.4 Ala87 35.1 82.5 1.1 2.0 7.8 13.7 0.7 1.0 0.3 14.4 26.1 4.0 2.5 31.9 38.2 Val87 15.1 90.1 0.9 1.0 7.1 9.3 0.7 0.8 0.3 11.7 15.8 5.5 2.9 26.9 32.3 Leu87 1.0 9.0 0.3 0.1 0.0 3.2 0.1 0.5 0.3 4.7 6.3 1.2 0.6 4.6 5.5 Ile87 6.8 71.4 0.5 0.3 0.6 9.6 0.5 0.8 0.4 0.0 20.7 2.9 1.6 21.6 25.9 Phe87 8.7 10.5 0.2 0.2 0.0 18.2 0.2 0.9 0.3 16.7 37.2 5.3 2.2 16.8 20.1 Trp87 11.6 46.9 0.0 0.0 0.0 5.0 0.0 0.0 0.0 5.4 7.6 2.7 0.8 13.3 16.0 Met87 0.7 5.5 0.1 0.1 0.0 2.8 0.0 0.3 0.0 3.5 5.1 0.0 0.7 3.2 3.8 Pro87 1.4 4.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 n.d. ** 1.2 Polar uncharged side chain Tyr87 16.5 73.4 0.9 0.9 5.2 6.7 0.3 0.5 0.2 8.1 11.5 3.0 1.1 21.4 25.7 Gln87 3.2 6.4 0.1 0.1 0.0 8.5 0.1 0.7 0.3 9.7 13.9 1.9 2.1 7.9 9.4 Asn87 0.7 3.2 0.2 0.2 0.0 2.3 0.0 0.2 0.0 2.8 4.7 1.0 0.7 2.6 3.2 Cys87 0.2 1.0 0.3 0.0 0.0 1.0 0.0 0.0 0.0 1.0 2.0 0.0 0.0 0.9 1.1 Ser87 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. Thr87 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. Polar charged side chain Asp87 8.5 13.6 0.0 0.0 0.0 3.7 0.0 0.0 0.0 3.7 7.1 0.3 0.3 n.d. ** 7.4 Glu87 0.6 4.3 0.1 0.1 0.0 1.7 0.0 0.0 0.0 1.7 3.1 0.5 0.0 2.0 2.4 His87 0.9 1.3 0.3 0.1 0.0 0.9 0.0 0.0 0.0 1.1 1.6 0.0 0.0 1.0 1.2 Arg87 1.7 4.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.2

Concentration of metabolites (μM) formed after 30 minutes incubations of 200 nM BM3 with 500 μM clozapine, 8 μM GST P1-1 in presence of 100 μM GSH. Values represent averages of three individual experiments; standard errors were always lower than 10%. *(nmol product/nmol BM3/30min);** P450 concentration could not be quantified (CO difference spectrum only showed a peak at 420


Figure 2. Relative amounts of CLZ metabolites formed by BM3 M11 mutants with different amino acid residues at position 87.

Incubations were carried out for 30 minutes in presence of 100 µM GSH and 8 µM recombinant human GST P1-1. (A) relative amounts of stable CLZ-metabolites (C-1 to C-5) and total of GSH-conjugates (CG-total); (B) relative amounts of individual GSH-conjugates.


For the mutant Asp87, no total activity could be calculated, as the P450 concentration

could not be measured. The mutants containing positive charged residue His87, Arg87

were poorly active in the metabolism of CLZ (<2% conversion). The mutants having Lys87,

Ser87, and Thr87 at position 87 did not show any activity.

Figure 2 shows the effect of amino acid residue 87 on the relative amounts of the

stable metabolites formed via pathways a, b and c (Figure 1) and the relative amounts of

GSH conjugates resulting from bioactivation of CLZ (pathway d, Figure 1).

For all the mutants the major stable metabolite was N-demethylclozapine (C-2) followed

by CLZ N-oxide (C-1), Figure 2A. Significant differences were observed in the ratios of C-2

to C-1. In case of mutants containing Ile87, Leu87, Met87 and Glu87, N-demethylation

was up to 10-fold more abundant than N-oxidation. In contrast, with mutants containing

Phe87, His87, Asp87, Gly87 and Gln87 the ratio of N-demethylation to N-oxidation ranged

from 1.2 to 2. The other stable metabolites (C-3, C-4 and C-5) represented only minor

metabolites for all the mutants.

As illustrated in Figure 2B, the metabolic profile of the formed GSH adducts appears to be

relatively constant for all the active mutants, supporting the hypothesis that all GSH

conjugates originate from the CLZ nitrenium ion (10). In all cases, GSH conjugate CG-6 is

the major metabolite and accounts for on average 41 ± 5 % of the total GSH conjugates.

Considering the fact that conjugate CG-7 most likely also originates from the same

intermediate nitrenium ion (Figure 1), the chlorine substitution pathway represents 47 ±

4% of the total GSH conjugation in presence of hGST P1-1. GSH conjugate CG-5, which was

tentatively assigned to the conjugate in which the GSH moiety is attached to the non-

chlorinated aromatic ring, represents on average 25 ± 3% of the GSH conjugates.

Conjugate CG-1, which is the major conjugate formed in non-enzymatic GSH-conjugation,

represents on average to 22 ± 2% of the GSH conjugates. GSH conjugates CG-3 and CG-4,

as shown in Figure 1, represented less than 2% of the total of GSH conjugates. GSH

conjugate CG-8 that most likely results from the chlorine substitution of the nitrenium ion

of N-demethylclozapine represents 3.5 ± 2.3% of the total GSH conjugates.

The secondary GSH conjugates, CG-2 (MH+ ion at m/z 618.23) and CG-8 (MH

+ ion at m/z

584.25) derived from N-demethylclozapine, and CG-7 (MH+ ion at m/z 903.35) derived

from CG-6 (MH+ ion at m/z 598.27), have not been found in human studies and therefore

were not further characterized.

To select the most appropriate BM3 M11 mutant for large scale production of GSH

conjugates, it was investigated which mutant showed a combination of high overall

activity and high selectivity towards the bioactivation pathway (route d in Figure 1).

Figure 3 shows the ratio of total of GSH conjugates to stable metabolites for each active

mutant, ranked from low to high ratio. As shown in Figure 3, for the four most active

mutants, having active site amino acids Ala87, Val87, Tyr87 and Ile87, less than 35% of the

total of metabolites represented GSH conjugates.


86

Figure 3. Percentage of total of GSH-conjugates formed in incubations of CLZ with BM3 M11 mutants with different amino acid residues at position 87, ranked in order of activity. Numbers below each bar represent the activity of the BM3 M11 mutants relative to the most active mutant BM3 M11 Ala87.

For mutants containing Gln87, Phe87 and Gly87, however, approximately 80% of the

metabolites found were GSH conjugates, indicating that for these mutants bioactivation to

CLZ nitrenium ion (route d, Figure 1) is the major pathway of metabolism. Because the

mutant containing Phe87 has the highest activity, this mutant was selected for large scale

production of GSH conjugates for structural elucidation by 1H-NMR.

3.3.3. Large scale incubation and NMR identification of isolated GSH conjugates

Figure 4 shows the preparative HPLC chromatogram, with UV-detection at 254 nm,

obtained after large scale incubation of CLZ with mutant Phe87. After isolation of the

individual metabolites by preparative HPLC, their purity and identity was first analyzed by

analytical HPLC and LC/MS/MS method, resulting in the assignment of metabolites and

parent compound as presented in Figure 5.

By hourly additions of enzymes and cofactors, over 90% of CLZ was converted,

according to the strong decrease in parent compound. Based on the peak areas

approximately 98% of the metabolites found were GSH conjugates. This higher

percentage of GSH conjugation, compared to the analytical scale incubations, can be

explained by further bioactivation of the stable metabolite C-2, producing CG-8 and CG-2.

The low yield of CLZ N-oxide (C-1) might be explained by non-enzymatic reduction of the

N-oxide by NADPH and GSH that was added hourly to the incubation (18).

For the five primary GSH conjugates of CLZ, having MH+ ion at m/z 632.23 (CG-1, CG-3, CG-

4 and CG-5) and MH+ ion at m/z 598.25 (CG-6),

1H NMR spectra were recorded to identify

the position of GSH conjugation. Figure 5 shows the signals of the aromatic hydrogen

atoms of the CLZ-moiety of these GSH conjugates.


87

Figure 4. Preparative HPLC-UV chromatogram of metabolites obtained by large scale incubation of CLZ with BM3 M11 Phe87 in presence of hGST P1-1 and GSH.

Figure 5A shows the

1H-NMR spectrum of conjugate CG-1, which is the major GSH

conjugate formed in absence of glutathione S-transferases (9, 10). This conjugate, which

eluted after 32.5 min in the preparative HPLC (Figure 4), was previously identified as C-6

glutathionyl clozapine. The spectrum shown in Figure 5A is in full agreement with the 1H-

NMR spectra of C-6 glutathionyl clozapine which was identified previously as the major

GSH conjugate formed by peroxidases and electrochemical oxidation of CLZ (15, 16, 19).

Two doublets at 6.96 and 7.23 ppm correspond to the protons at positions 9 and

7, with a small coupling constant of 2.5 Hz due to proton in the meta-position. Fischer et

al., previously assigned doublet at 6.96 ppm to H7 and the signal at 7.23 ppm to H9 (15).

Madsen et al., however, assigned doublet at 6.96 ppm to H9 and the signal at 7.23 ppm to

H7 (19). Which signal corresponds to which proton could not be determined

unequivocally, only based on chemical shift and coupling pattern (16). However, this does

not affect the identification of the position of GSH conjugation because each theoretically

possible GSH conjugate is expected to have its own unique combination of multiplicity and

coupling pattern. Therefore, for signals that could not be assigned unequivocally to

specific aromatic protons, two possibilities are shown in Table 2. The assignments before

the slashes correspond to the first possibility, whereas the assignments after the slashes

correspond to the second possibility.


88

Figure 5. Aromatic regions of the

1H-NMR spectra of clozapine GSH-conjugates. The

conjugates were obtained by purification by preparative HPLC of metabolites formed by BM3 M11 Phe87, GSH and hGSTP1-1. (A) CG-1, C-6 glutathionyl CLZ; (B) CG-3, C-9 glutathionyl CLZ; (C) CG-4, C-7 glutathionyl CLZ; (D) CG-6, C-8 glutathionyl deschloroclozapine; (E) CG-5, C-2 glutathionyl CLZ or C-3 glutathionyl CLZ.

Figure 5B shows the

1H-NMR spectrum of purified CG-3, which eluted at 32.2 min

with preparative HPLC. This conjugate was previously found to be a minor GSH conjugate,

with MH+ ion at m/z of 632.2, in incubations of CLZ with liver microsomes and BM3 M11

when incubated in absence of glutathione S-transferases (9, 10). The spectrum shown in

the Figure 5B can only be explained by conjugation of GSH at the C-9 position of clozapine.

Two signals at 6.83 and 7.02 ppm showed only a coupling of 8 Hz and are therefore


89

assigned to the neighboring H6 and H7 protons. The corresponding chemical shifts and

coupling constants for the observed signals are given in Table 2.

Two more conjugates having a MH+ ion at m/z 632.2 eluted at retention times

30.9 min and 33.1 min (Figure 4) and appeared to correspond to conjugates CG-4 and CG-

5 (10). On the basis of the order of elution and small differences in fragmentation

patterns in LC-MS/MS (17), the structure of these GSH conjugates were previously

tentatively assigned to C-7 glutathionyl clozapine (CG-4) and a conjugate with GSH bound

to the non-chlorinated ring (CG-5). However, so far no 1H NMR spectra have been

reported confirming the exact positions. Spectra for this two conjugates are shown in the

Figure 5C and 5E, respectively.

Figure 5E shows the 1H-NMR-spectrum of the aromatic region of CG-5, which is

found at high levels when GSH conjugation is catalyzed by GST P1-1 (Figure 4). This

spectrum can only be explained by conjugation of GSH at position 2 or 3 of the non-

chlorinated aromatic ring. In each conjugate in which the non-chlorinated aromatic ring is

not substituted, two triplets are found corresponding the protons H2 and H3 that are both

strongly coupled by two ortho-protons, Figure 5A-D. These typical triplets could not be

found in the spectrum of CG-5 (Figure 5E), indicating that one of these protons is

substituted. Furthermore, the signals at 6.96, 7.37 and 7.45 ppm could be attributed to

protons H6, H9 and H7 of the chlorinated ring, according to the COSY-spectrum [data not

shown]. This confirms that addition of the GSH is at the non-chlorinated aromatic ring.

Although the loss of characteristic triplets show that GSH is conjugated to position 2 or 3,

the 1H-NMR and COSY-spectra could not differentiate between positions 2 or 3 for GSH

binding. When GSH is bound at the 2 position, the signal at 6.85 ppm will correspond to

proton H3, because this signal has a ortho-coupling of 8.5 Hz due to H4 and a weak

coupling of 1.5 Hz due to the meta proton in position 1. The doublet with ortho-coupling

of 8.5 Hz at 6.78 ppm corresponds to H4, coupled by H3, whereas the doublet with weak

coupling of 1.5 Hz at 6.95 ppm would correspond to H1 by meta-coupling by proton H3.

When GSH is bound at the 3-position, the signal at 6.85 ppm will correspond to proton H2,

because with ortho-coupling of 8.5 Hz due to H1 and a weak coupling of 1.5 Hz due to the

meta proton in position 4.

The doublet with an ortho-coupling of 8.5 Hz at 6.78 ppm corresponds to H1,

coupled by H2, whereas the doublet with a weak coupling of 1.5 Hz at 6.95 ppm

corresponds to H4 by meta-coupling by proton H2.


Table 2. LC-MS/MS characteristics and 1H NMR spectra of the aromatic hydrogen atoms of the GSH conjugates of clozapine

1H-NMR-spectra LC-MS/MS- mass spectra Conjugate Assignment

a) δ ppm (intensity) multiplicity: coupling constants m/z (intensity)

b)

CG-1 H1/H4 7.30 (1H) d of d:

2JHH, 7 Hz;

3JHH, 1.5 Hz 623.23 (MH

+; 100%);

H2/H3 7.08 (1H) d of t: 2JHH 7.5 Hz,

3JHH 1.5 Hz 614.22 (8%); 575.17 (6%);

H3/H2 7.41 (1H) d of t: 2JHH 7.5 Hz, 3JHH 1.5 Hz 503.18 (37%); 446.12 (1%); H4/H1 7.10 (1H) d:

2JHH 7.5 Hz 359.12 (29 %); 302.06 (12%)

H7/H9 7.23 (1H) d: 2JHH 2.5 Hz

H9/H7 6.96 (1H) d: 2JHH 2.5 Hz

CG-3 H1/H4 7.36 (1H) d: 2JHH 8 Hz 623.23 (MH

+; 100%);

H2/H3 7.07 (1H) d of t: 2JHH 7 Hz, 3JHH 1.5 Hz 614.22 (1%); 503.18 (23%); H3/H2 7.37 (1H) t:

2JHH 7 Hz 446.12 (1%); 359.12 (41%);

H4/H1 6.99 (1H) d: 2JHH 8 Hz 302.06 (6%) H6 7.02 (1H) d:

2JHH 8 Hz

H7 6.83 (1H) d: 2JHH 8 Hz

CG-4 H1/H4 7.28 (1H) d of d: 2JHH 8 Hz,

3JHH 1.5 Hz 623.23 (MH

+; 100%);

H2/H3 7.04 (1H) t: 2JHH 7 Hz 614.22 (8%); 575.17 (3%) H3/H2 7.37 (1H) d of t: 2JHH 7 Hz, 3JHH 1.5 Hz 503.18 (44%); 446.12 (4%); H4/H1 7.07 (1H) d: 2JHH 8 Hz 359.12 (17 %); 302.06 (4%); H6/H9 7.01 (1H) s 243.06 (5%) H9/H6 7.03 (1H) s CG-5 H1/H4 6.95 (1H) d: 3JHH 1 Hz 623.23 (MH+; 100%); H3 or H2 6.85 (1H) d of d: 2JHH 8.5 Hz, 3JHH 2 Hz 614.22 (6%); 575.17 (10%); H4/H1 6.78 (1H) d: 2JHH 9 Hz 503.18 (27%); 446.12 (25%); H6 6.96 (1H) d: 2JHH 9 Hz 359.12 (15 %); 302.06 (13%); H7 7.45 (1H) d of d: 2JHH 9 Hz, 3JHH 2 Hz 243.06 (34%) H9 7.37 (1H) d: 2JHH 2 Hz CG-6 H1/H4 7.00 (1H) d of d:

2JHH 8 Hz,

3JHH 2 Hz 598.25 (MH

+; 100%);

H2/H3 7.35 (1H) d of t: 2JHH 7.5 Hz, 3JHH 1.5 Hz 580.24 (8%); 469.20 (77%); H3/H2 7.04 (1H) d of t: 2JHH 7.5 Hz, 2JHH 1 Hz 412.20 (4%); 325.15 (24%); H4/H1 7.30 (1H) d of d: 2JHH 7.5 Hz, 3JHH 1 Hz 268.09 (6%) H6 6.82 (1H) d: 2JHH 8 Hz H7 7.01 (1H) d: 2JHH 8 Hz

H9 7.12 (1H) d: 3JHH 2 Hz

a) Absolute assignment of protons was not possible; therefore each signal is indicated by two assignments corresponding to two possible solutions. Assignments before the slash correspond to solution 1; assignments after the slash correspond to solution 2. b) Electrospray spectra (LC-MS/MS) were acquired using nitrogen as collision gas with collision energy of 25V.


Conjugate CG-4 is the fourth GSH conjugate with MH+ ion at m/z 632.2, and was

previously tentatively assigned to C-7 glutathionyl clozapine (10). However, theoretically

this can also represent a conjugate with GSH bound to one of the other positions of the

non-chlorinated aromatic ring. This GSH conjugate eluted at 30.9 min with preparative

HPLC, Figure 4.

Because this conjugate is produced at very low yield, several large scale

incubations were performed to obtain enough material to record a 1H-NMR-spectrum with

sufficient signal-to-noise ratio. Although this small amount of the conjugate appeared to

be contaminated by an unknown compound, the COSY-spectrum allowed us to solve the

spectrum, despite the strong contaminant signal at 7.10 ppm, Figure 5C. Based on these

spectra this GSH conjugate is identified as C-7 glutathionyl clozapine, consistent with the

previous proposals (10, 17). Firstly, two triplets at 7.04 and 7.37 could be assigned to the

protons H2 and H3. Two sharp singlets at 7.01 and 7.03 ppm are attributed to isolated

protons that do not have ortho- or meta-coupling. If GSH was bound to position 7, protons

at position 6 and 9 would lose their ortho and meta couplings by H7. Two signals centered

at 7.28 and 7.07 ppm represent proton H1 and H4, which showed both ortho- and meta-

coupling by H2 and H3, Table 2. Although the signal at 7.07 partially overlaps with the

signal of the impurity, the COSY-spectrum confirmed that signal centered at 7.07 ppm is a

doublet with ortho-coupling [data not shown].

The fifth GSH-conjugate for which a 1H-NMR spectrum is recorded is CG-6, which

was the major GSH conjugate found in the incubations in presence of hGST P1-1, and

which showed a MH+ ion at m/z 598.25 by LC-MS/MS-analysis. In the preparative HPLC

system used, this GSH conjugate eluted at 27.6 min, Figure 4. Figure 5D show the 1H-NMR

obtained. Two triplets centered at 7.04 and 7.35 ppm with small meta-couplings

correspond to the protons H3 and H2 at the non-chlorinated aromatic ring. The signals

centered at 7.00 and 7.30 ppm correspond to protons H1 and H4, as demonstrated by the

combination of both ortho and meta coupling by protons H2 and H3. The signals at 6.82,

7.01 and 7.12 ppm correspond to protons H6, H7 and H9 respectively, based on the

coupling patterns and COSY-spectrum [data not shown]. On the basis of this spectrum and

mass spectrum it was confirmed that this conjugate correspond to C-8 glutathionyl

deschloroclozapine.

3.4. Discussion

Currently, there is an increasing interest in developing novel methodologies to produce

human relevant drug metabolites on a large scale in order to enable structural

characterization of metabolites and test their pharmacological and toxicological

properties. One of the approaches is to make use of genetically engineered cytochromes

P450s that are developed for the catalysis of regio- and stereoselective hydroxylation of

chemicals at high activity. In particular the bacterial cytochrome P450 BM3 from Bacillus

megaterium has high potential as a biocatalyst for these purposes because this enzyme is


92

the most active P450 discovered so far and because the substrate selectivity and

metabolic profile can be manipulated by site-directed and/or random mutagenesis (6, 7).

One of the BM3 mutants that show high activity in drug metabolism is BM3 M11, which

contains 10 different amino acid substitutions compared to wild-type BM3. This BM3

mutant was shown to be highly active in metabolizing a variety of drugs to human relevant

metabolites, including reactive intermediates (5, 9, 10). Recently, we have performed a

saturation mutagenesis study in which the active-site residue at position 87 was mutated

to all 20 possible amino acids (20). In BM3 M11 the residue at this position is Val87, which

was introduced at an early stage of the mutagenesis process, to expand the substrate

selectivity to drug metabolism (23). In the saturation mutagenesis study in which all amino

acids were evaluated at position 87, we recently demonstrated that the type of amino acid

at position 87 has strong effect on substrate selectivity when comparing a series of

alkoxyresorufins (20). In this study it was also demonstrated that the nature of the amino

acid at position 87, strongly influences the regioselectivity of testosterone hydroxylation

of BM3 M11.

In the present study, the library of BM3 M11 mutants with different amino acids

at position 87 was evaluated with CLZ as substrate. CLZ is a drug that can be metabolized

by peroxidases and P450s to multiple metabolites, including reactive nitrenium ions that

might be involved in adverse drug reactions associated with CLZ therapy. BM3 M11 has

been shown to produce high levels of most human relevant metabolites of CLZ, which are

represented in Figure 1. All metabolites can be explained by four different initial oxidative

pathways: N-demethylation (a), N-oxidation (b), piperazine-ring opening (c) and

dehydrogenation to a reactive nitrenium ion (d). Although BM3 M11 with residue Val87

produced significant amounts of reactive nitrenium ion (as identified as GSH conjugates),

the major pathways of metabolism are N-demethylation and N-oxidation, which explain

approximately 70% of the total of metabolites. The aim of this study was to investigate

whether residue 87 also controls the regioselectivity of CLZ metabolism, and to investigate

whether a mutant could be identified with higher selectivity toward the bioactivation to

the toxicologically relevant CLZ nitrenium ion. A more selective P450 BM3 mutant would

be more useful for the generation of high levels of CLZ GSH-conjugates that still require

structural confirmation by 1H-NMR. So far, the structure of only two of the GSH-

conjugates shown in Figure 1 has been elucidated by 1H-NMR and mass spectrometry.

However, the structures of the GSH-conjugates found in bile of rats and mice, and which

also have MH+ ions at m/z 632.2 (17) have not yet been characterized by

1H-NMR.

As shown in Table 1, changing the amino acid residue at position 87 of BM3 M11 has

strong effects on the total activity and regioselectivity of CLZ oxidation. The mutants

Ala87, Val87, and Ile87 were found to be the most active, as was found previously with

alkoxyresorufins and testosterone (20). These mutants have a small and apolar residue in

position 87. This seems consistent with the previous hypothesis that replacement of the


93

bulky Phe87 in wild type BM3 by smaller amino acids creates space for the bulky

substrates that allows better positioning with respect to the activated oxygen species,

resulting in higher activities and coupling efficiencies (24-26). However, in the present

study relatively high activities were also found in the BM3 M11 mutants containing the

relatively bulky amino acids Phe87, Tyr87 and Trp87. Previously, replacing Phe87 by Tyr87

in wild-type BM3 was found to be detrimental for activity towards long-chain fatty acids,

probably by disruption of the hydrophobic interaction by the phenol-group. In case of

BM3 M11 the presence of these bulky amino acids apparently is less restrictive for bulky

substrates because in presence of Phe87 and Tyr87 both testosterone (20) and CLZ are

metabolized at high activity. Apparently, by the combination of ten mutations present in

BM3 M11 the topology of the active site and/or substrate access channel has changed

significantly, explaining the much wider substrate selectivity compared to wild-type BM3.

As shown in Figure 2A, the nature of amino acid 87 has strong effect on regioselectivity of

CLZ metabolism. When considering the stable metabolites that are formed via pathways

a, b and c (Figure 1), the major metabolite with all mutants was N-demethylclozapine (C-

2), although the N-oxide (C-1) was also produced at significant levels. However, the ratio

of N-demethylation to N-oxidation appeared to be quite dependent on the nature of

amino acid residue at position 87. For example, for the mutants containing Leu87 and

Ile87, N-demethylation was almost 10-fold higher than N-oxidation. In the mutant

containing Phe87, N-demethylclozapine and CLZ N-oxide were formed in almost the same

amount. However, the relative contribution of N-oxidation in all incubations might be

somewhat underestimated, because all incubations were performed in presence of GSH

which is known to reduce CLZ N-oxide back to CLZ (27). Therefore, in case of Phe87 N-

oxidation of CLZ might even be higher in absence of reductive agents. In case of the

human P450s, it has been shown that CYP1A2 preferentially metabolizes CLZ by N-

demethylation, whereas CYP3A4 is mainly responsible for production of CLZ N-oxide (27).

However, the factors that determine the ratio of N-demethylation and N-oxidation are still

unclear. Different presentation of the piperazine N-methyl group to the oxidative species

at the active site might explain why some human P450s preferentially catalyze N-

demethylation, whereas others predominantly catalyze N-oxidation.

One of the aims of the present study was to identify mutants with high activity

and selectivity for bioactivation of CLZ to the reactive nitrenium ion. As shown in Figure

2A several mutants produced high levels of GSH conjugates (CG-total), indicative for

relative high selectivity in the formation of the reactive nitrenium ion. Other mutants

showed strong preference in catalyzing formation of N-demethylclozapine and CLZ N-

oxide. However, from the results it is unclear what features of the amino acid side-chain

determines selectivity for bioactivation. For example, the BM3 M11-mutant containing

the bulky Phe87 showed high selectivity and activity in formation of GSH conjugates,

whereas the mutants containing the bulky Tyr87 and Trp87 preferentially catalyzed


94

formation of stable metabolites. Future detailed protein modeling studies including those

evaluating protein dynamics and substrate mobility might help to rationalize the different

regioselectivities observed. Figure 2B shows the relative amounts of the different GSH

conjugates that are formed in incubations of CLZ with BM3 mutants in presence of

recombinant hGST P1-1.

Consistent with our previous study, the major pathway of GST P1-1 catalyzed GSH

conjugation is substitution of the chlorine-atom of the CLZ nitrenium ion (10). The

resulting GSH bound nitrenium ion is subsequently reduced by NADPH or GSH, to form CG-

6, or further conjugated to GSH, to form CG-7 (Figure 1). The fact that with all mutants the

same ratio of GSH conjugates is formed strongly suggests that all form from the same

reactive intermediate.

Mutant Phe87 was selected for large scale biosynthesis of GSH conjugates because this

mutant combined high activity with high preference for the bioactivation pathway, Figure

3. Previous studies, aiming at the characterization of GSH conjugates of CLZ, showed that

non-enzymatic GSH conjugation to the CLZ nitrenium ion, formed by peroxidases or

electrochemically, mainly produced a GSH conjugate bound at the C-6 position of CLZ and

minor amounts of conjugate bound at the C-9 position. The structures of these two GSH

conjugates have been elucidated by 1H-NMR. However, in vivo studies with rats and mice

have shown that in bile two major GSH conjugates are excreted that do not correspond to

these two conjugates (17). Also, incubations with rat liver microsomes showed small

amounts of a fifth GSH conjugate with MH+ ion at m/z 632.2 (17).

It was initially concluded that these GSH conjugates might originate from as yet

unidentified reactive intermediate produced in vivo. However, we recently demonstrated

that these alternative GSH conjugates probably are resulting from GST catalyzed

inactivation of the CLZ nitrenium ion (10). By using mutant Phe87, we were able to

produce significant amounts of all GSH conjugates, for which the structures were not yet

elucidated unequivocally by 1H-NMR. Because four GSH conjugates were found with MH

+

ion at m/z 632.2, it was previously concluded that for at least one of the conjugates, GSH

is bound to the non-chlorinated aromatic ring of CLZ. The present study shows that

conjugate designed CG-5, which is a major product in presence of hGST P1-1 has the GSH

moiety bound to the non-chlorinated ring at the position 2 or 3 (Figure 5). For the minor

conjugates CG-4, we were able to confirm binding at the 7 position, as it was tentatively

assigned based on fragmentation pattern in LC-MS/MS (17).

In conclusion, the present study shows that mutation of residue 87 in drug

metabolizing mutant BM3 M11 has strong influence on activity and regioselectivity of CLZ

metabolism. Using a mutant that combined high activity and high selectivity for CLZ

bioactivation, we were able to produce sufficient amounts of as yet tentatively assigned

GSH conjugates to characterize their structures by 1H-NMR. This study confirms the high

potential of BM3 mutants as tool to characterize human-relevant metabolites.


95

References

1. Shimada, T., Yamazaki, H., Mimura, M., Inui, Y., and Guengerich, F. P. (1994) Interindividual variations in human liver cytochrome P-450 enzymes involved in the oxidation of drugs, carcinogens and toxic chemicals: studies with liver microsomes of 30 Japanese and 30 Caucasians, The Journal of pharmacology and experimental therapeutics 270, 414-423.

2. Evans, W. E., and Relling, M. V. (1999) Pharmacogenomics: translating functional genomics into rational therapeutics, Science 286, 487-491.

3. Smith, D. A., and Obach, R. S. (2009) Metabolites in safety testing (MIST): considerations of mechanisms of toxicity with dose, abundance, and duration of treatment, Chemical research in toxicology 22, 267-279.

4. 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 letters 486, 173-177.


6. Yun, C. H., Kim, K. H., Kim, D. H., Jung, H. C., and Pan, J. G. (2007) The bacterial P450 BM3: a prototype for a biocatalyst with human P450 activities, Trends Biotechnol 25, 289-298.


8. Kim, K. H., Kang, J. Y., Kim, D. H., Park, S. H., Kim, D., Park, K. D., Lee, Y. J., Jung, H. C., Pan, J. G., Ahn, T., and Yun, C. H. (2011) Generation of human chiral metabolites of simvastatin and lovastatin by bacterial CYP102A1 mutants, Drug metabolism and disposition: the biological fate of chemicals 39, 140-150.



11. Buchanan, R. W. (1995) Clozapine: efficacy and safety, Schizophrenia bulletin 21, 579-591.

12. Safferman, A., Lieberman, J. A., Kane, J. M., Szymanski, S., and Kinon, B. (1991) Update on the clinical efficacy and side effects of clozapine, Schizophrenia bulletin 17, 247-261.

13. Wagstaff, A., and Perry, C. (2003) Clozapine: in prevention of suicide in patients with schizophrenia or schizoaffective disorder, CNS drugs 17, 273-280; discussion 281-273.


96

14. Hummer, M., Kurz, M., Kurzthaler, I., Oberbauer, H., Miller, C., and Fleischhacker, W. W. (1997) Hepatotoxicity of clozapine, Journal of clinical psychopharmacology 17, 314-317.

15. Fischer, V., Haar, J.A., Greiner, L., Lloyd, R.V., and Mason, R.P. (1991) Possible role of free radical formation in clozapine (Clorazil) induced agranulocytosis., Mol. Pharmacol. 40, 846-853.

16. Liu, Z. C., and Uetrecht, J. P. (1995) Clozapine is oxidized by activated human neutrophils to a reactive nitrenium ion that irreversibly binds to the cells, The Journal of pharmacology and experimental therapeutics 275, 1476-1483.

17. Maggs, J. L., Williams, D., Pirmohamed, M., and Park, B. K. (1995) The metabolic formation of reactive intermediates from clozapine, a drug associated with agranulocytosis in man, The Journal of pharmacology and experimental therapeutics 275, 1463-1475.

18. Pirmohamed, M., Williams, D., Madden, S., Templeton, E., and Park, B. K. (1995) Metabolism and bioactivation of clozapine by human liver in vitro, The Journal of pharmacology and experimental therapeutics 272, 984-990.

19. Madsen, K. G., Olsen, J., Skonberg, C., Hansen, S. H., and Jurva, U. (2007) Development and evaluation of an electrochemical method for studying reactive phase-I metabolites: correlation to in vitro drug metabolism, Chemical research in toxicology 20, 821-831.

20. Vottero, E., Rea, V., Lastdrager, J., Honing, M., Vermeulen, N. P., and Commandeur, J. N. (2011) Role of residue 87 in substrate selectivity and regioselectivity of drug-metabolizing cytochrome P450 CYP102A1 M11, Journal of biological inorganic chemistry : JBIC : a publication of the Society of Biological Inorganic Chemistry 16, 899-912.

21. Bradford, M. M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding, Analytical biochemistry 72, 248-254.

22. Habig, W. H., Pabst, M. J., and Jakoby, W. B. (1974) Glutathione S-transferases. The first enzymatic step in mercapturic acid formation, The Journal of biological chemistry 249, 7130-7139.

23. 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, Analytical biochemistry 341, 148-155.

24. Carmichael, A. B., and Wong, L. L. (2001) Protein engineering of Bacillus megaterium CYP102. The oxidation of polycyclic aromatic hydrocarbons, European journal of biochemistry / FEBS 268, 3117-3125.


26. Mann, R. E., Suurvali, H. M., and Smart, R. G. (2001) The relationship between alcohol use and mortality rates from injuries: a comparison of measures, Am J Drug Alcohol Abuse 27, 737-747.


97

27. Tugnait, M., Hawes, E. M., McKay, G., Eichelbaum, M., and Midha, K. K. (1999) Characterization of the human hepatic cytochromes P450 involved in the in vitro oxidation of clozapine, Chemico-biological interactions 118, 171-189.

98

CHAPTER 4

ACTIVE SITE SUBSTITUTION A82W IMPROVES REGIOSELECTIVITY

OF STEROID HYDROXYLATION BY CYTOCHROME P450 BM3

MUTANTS AS RATIONALIZED BY SPIN RELAXATION NMR STUDIES

Vanina Rea, Ard J. Kolkman, Eduardo Vottero, Erik J. Stronks, Kirsten A.M. Ampt, Maarten Honing, Nico P.E. Vermeulen, Sybren S. Wijmenga and Jan

N.M. Commandeur

adapted from Biochemistry, (2012), 51, 750-760

99

Abstract

Cytochrome P450 BM3 from Bacillus megaterium is a monoxygenase with great potential

for biotechnological applications. In this work, we present engineered drug metabolizing

P450 BM3 mutants as a novel tool for regioselective hydroxylation of steroids at position

16ß. In particular, we show that by replacing alanine at position 82 with a tryptophan in

P450 BM3 mutants M01 and M11, the selectivity toward 16ß-hydroxylation for both

testosterone and norethisterone was strongly increased. The A82W mutation led to a <

42-fold increase in Vmax for 16ß-hydroxylation of these steroids. Moreover, this mutation

improves the coupling efficiency of the enzyme, which might be explained by a more

efficient exclusion of water from the active site. The substrate affinity for testosterone

increased at least 9-fold in M11 with tryptophan at position 82. A change in the

orientation of testosterone in the M11 A82W mutant as compared to the orientation in

M11 was observed by T1 paramagnetic relaxation nuclear magnetic resonance.

Testosterone is oriented in M11 with both the A- and D-ring protons closest to the heme

iron. Substituting alanine at position 82 with tryptophan results in increased A-ring

proton-iron distances, consistent with the relative decrease in the level of A-ring

hydroxylation at position 2.

Chapter 4 Regioselective steroid hydroxylation

100

4.1. Introduction

Cytochrome P450 (CYPs) constitute a large superfamily of monoxygenases that are

involved in the oxidation and reduction of a wide variety of primary and secondary

metabolites and in the biotransformation of xenobiotics (1). Because of their broad

substrate range and catalytic diversity, there is an increasing interest to use P450 enzymes

in biotechnology (2-4). Among others, their application in the production of

pharmaceuticals and the optimization of lead compounds by P450 enzymes has been

explored recently (5). An important class of compounds for which P450s plays an

important role in biosynthesis and catabolism consists of steroids, which have a wide

range of therapeutic activities, including anti-inflammatory, immunosuppressive,

progestational, diuretic, anabolic and contraceptive activities (6). Both chemical and

biological approaches are used for the production of steroids.

Chemical synthesis of steroids requires highly complicated, multistep schemes that are

expensive and time-consuming. Many microorganisms that are able to hydroxylate

steroids at specific positions have been identified. Therefore, microbial biosynthesis can

be an efficient alternative for large-scale synthesis of steroid drugs (6). Very often, an early

step in these biotransformations is a regio- and stereospecific hydroxylation of a steroid

core by microbial P450s. However, microbial hydroxylation of steroids is often hampered

by the formation of byproducts because of non-specific and secondary metabolism.

Another approach is to clone a genetically engineered steroid hydroxylating P450 into a

suitable expression system that is devoid of secondary steroid catabolism. Several

mammalian P450s are known to regioselectively hydroxylate steroids (7). However,

because of their low stability and catalytic activity they are not very suitable as

biocatalysts for steroid hydroxylations (8, 9). Genetically engineered bacterial P450s are

often much more stable and exhibit higher catalytic activities, which make them more

promising candidates for biocatalysis (10). Of the bacterial P450s, cytochrome P450 BM3

from Bacillus megaterium is considered as one of the most promising monoxygenases for

biotechnological applications because it is the most active P450 so far identified (8). In

recent years, several mutants of P450 BM3 have been shown to catalyze the hydroxylation

of steroids.

A triple mutant of P450 BM3 (R47L/F87V/L188Q) and several mutants obtained by

random mutagenesis were found to hydroxylate testosterone, norethisterone,

nandrolone, progesterone and androstendione (13-15). However, the regioselectivity of

steroid hydroxylation by most P450 BM3 mutants was still poor, resulting in multiple

mono- and dihydroxy metabolites. Testosterone hydroxylation by P450 BM3 mutants has

been shown to occur at the 2, 15 and 16 positions (16) and recently other mutants

were obtained that are highly selective for 2- and 15-hyroxylation (17) .

Previously, it was shown that increasing the size of the active site residue at position 82 in

wild-type P450 BM3, by mutations A82F and A82W, strongly improves binding of fatty


101

acids and indoles, resulting in substantially increased catalytic efficiency and binding

affinity (18). However, no significant change in regioselectivity in metabolism was

observed. The aim of this study was to investigate whether the introduction of mutation

A82W in the P450 BM3 mutants M01 and M11 affects activity and/or regioselectivity of

steroid hydroxylation. Two steroid substrates, testosterone and norethisterone, were

selected because they have been shown to be metabolized to three and two

monohydroxy-metabolites, respectively, by M01 and M11 (15, 19).

The orientation of a substrate in the active site of P450 enzymes can be determined by T1

paramagnetic relaxation NMR (or Spin Relaxation NMR) (20). T1 paramagnetic relaxation

experiments have been used previously to determine proton-heme iron distances in P450

enzymes (21-23), which give information about the orientation(s) and mobility of ligands

in the active site of P450 enzymes. For P450 BM3, these experiments have been used to

determine substrate orientation of the highly flexible substrates lauric acid and 12-

bromolauric acid (21, 24, 25).

Steroid compounds might be more informative compounds for determining the

substrate orientation because they are large compounds with a dense network of protons;

hence, many protons can be used to determine the proton-heme iron distances. Because

they are rigid structures, the obtained distances give more information about the binding

and mobility in the active site, rather than the conformational flexibility of the steroid. No

T1 paramagnetic relaxation NMR experiments have been used to study the binding of

steroids to P450 BM3s so far. The only T1 paramagnetic relaxation NMR study addressing

the binding orientation of steroids in P450s, involved binding of testosterone to a

genetically engineered mutant of CYP2D6 (F438I), which had acquired the ability to

catalyze testosterone hydroxylation (26). In this study, Smith et al. determined the

distances of several protons of testosterone to the heme iron of the CYP2D6-mutant,

including protons H2, H15 and H16. Recently, we and others have demonstrated that

BM3 mutants catalyze testosterone hydroxylation only at these three positions but with

different regioselectivities (16, 17). In the study presented here, T1 paramagnetic

relaxation NMR studies were performed to investigate whether the change in

regioselectivity of testosterone hydroxylation that is observed in the A82W mutants can

be explained by different binding orientations of testosterone.

4.2. Materials and Methods

4.2.1. Materials

D-Glucose 6-phosphate dipotassium salt hydrate (100% pure), glucose-6-phosphate

dehydrogenase, dimethylsulfoxide-d6 (100% pure; 99.9 atom %D), carbon monoxide

(>99% pure) and sodium dithionite (>86% pure) were purchased from Sigma-Aldrich

(Schnelldorf, Germany). All other chemicals were of analytical grade and obtained from

standard suppliers.


102

4.2.2. Site-directed mutagenesis

The A82W mutation was introduced by site-directed mutagenesis in two templates, M11

(R47L, E64G, F81I, F87V, E143G, L188Q, Y198C, E267V, H285Y and G415S) and M01 (R47L,

F87V, L188Q, E267V, G415S) (10). The mutation was introduced into pBS-p450 BM3

mutant using the QuikChange XL Site – Directed mutagenesis kit (Stratagene). The

sequences of the forward primer for the mutation in M11 was:

5’-GT.CAA.GCG.CTT.AAA.TTT.GTT.CGC.GAT.ATT.TGG.GGA.GAC.GGG-3’ with the altered

residue in bold italic. The reverse primer for this positions was 5’-

CCC.GTC.TCC.CCA.AAT.ATC.GCG.AAC.AAA.TTT.AAG.CGC.TTG.AC-3’.

The sequences of the primers, for the mutation in M01 were as follows: forward primer:

5’-GT.CAA.GCG.CTT.AAA.TTT.GTT.CGC.GAT.TTT.TGG.GGA.GAC.GGG-3’; reverse primer:

5’-CCC.GTC.TCC.CCA.AAA.ATC.GCG.AAC.AAA.TTT.AAG.CGC.TTG.AC-3’.

The PCR product was digested with EcoR1 and BamH1 restriction enzymes and cloned into

a pET28a+ vector, which encodes an N-terminal His-tag to allow facile purification. The

desired mutations were confirmed by DNA sequencing of the heme domain (Baseclear,

Leiden, The Netherlands).

4.2.3. Expression, isolation and purification of P450 BM3 mutants

Expression of the BM3 mutants was performed by transforming competent E.coli BL21

cells with the corresponding pET28+-vectors, as described previously (27). Proteins were

purified using Ni-NTA agarose (27), after which P450 concentrations were determined

according to methods described by Omura and Sato (28). The purity of the enzymes was

checked by SDS PAGE electrophoresis on 12% gel and Coomassie-staining.

4.2.4. Determination of kinetic parameters of the biotransformation of testosterone

and norethisterone by P450 BM3 mutants

Testosterone and norethisterone incubations were performed in 100 mM potassium

phosphate buffer (pH 7.4), with 200 nM purified P450 BM3 mutants (10). The final volume

of the incubation was 200 μL, with a substrate concentration of 200 µM. The reactions

were initiated by addition of an NADPH regenerating system (final concentrations of 0.2

mM NADPH, 0.3 mM glucose-6-phosphate and 0.4 U/mL glucose-6-phosphate

dehydrogenase). The reaction was allowed to proceed for 60 min at 25°C and terminated

by the addition of 200 L of cold methanol. Precipitated protein was removed by

centrifugation (15 min at 14000 g), and the supernatant was analyzed by UPLC.

Metabolites were separated using a C18 column (Zorbax Eclipse XDB-C18, 4.6x50 mm, 1.8

m, Agilent Technologies) with a flow rate of 1 mL/min, isocratic 60% B (99% MeOH, 0.8%

H2O, 0.2% formic acid), 40% A (99% H2O, 0.8% MeOH, 0.2% formic acid), and UV

detection at 254 nm.


103

To determine the kinetic parameters, the ranges where the enzyme activity is

linear with enzyme concentration and incubation time were first determined. Based on

these experiments [data not shown], the most suitable conditions were chosen for the

determination of the kinetic parameters. The enzyme concentration was 300 nM and the

incubation time was 10 min; 10 substrate concentrations were used in the range 20-200

M for testosterone and 10-250 M for norethisterone. Specific activities were calculated

and plotted against substrate concentrations. The data were fitted with the Michaelis

Menten equation, using Origin Pro8.

4.2.5. Determination of coupling efficiency

To measure the NADPH consumption rate, P450 BM3 mutant (final concentration 200 nM)

was mixed with 870 L 100 mM potassium phosphate buffer (pH 7.4) containing 200 M

testosterone or norethisterone in DMSO (final concentration of 2%). The reaction was

started by adding 200 M NADPH. NADPH consumption was monitored at 340 nm with a

Libra S12 Biochrom spectrophotometer at 25°C for 10 min. The NADPH concentration was

calculated using = 6,200 M-1

cm-1

(29).

To measure the coupling efficiency, product formation and substrate consumption were

quantified by HPLC. The percentage of coupling was determined by the ratio of the

amount of product formed and the amount of NADPH consumed.

4.2.6. Preparative-scale biotransformation and isolation of testosterone metabolites

The testosterone metabolites were produced on a preparative scale by large-scale

incubation with M11 as biocatalyst. A 50 mL reaction volume containing 250 nM M11

mutant, 500 µM testosterone and NADPH regenerating system was prepared in 100 mM

potassium phosphate buffer (pH 7.4). The reaction was allowed to continue for 3 hours at

25°C. The reaction products were extracted by using 3 x 100 mL of dichloromethane. The

organic layers were combined and evaporated using a rotary evaporator. The dried

product was redissolved in 0.5 mL DMSO and injected into a Waters preparative LC

system, equipped with a HP Liquid Handler for injection and fractionation, a Waters

Xbridge Prep C18-MS (10 x 55 mm 5 µm particles) column and a Waters photodiode array

detector. At a flow-rate of 5 mL/min the following gradient was used: 0-1 minute 100% A

(99.95% H2O, 0.05% TFA); 1 - 15 minutes linear increase to 60 % B (30% MeOH, 60%

acetonitrile, 10 % H2O, 0.05% TFA), 15 – 15.5 minutes linear increase to 100% B, and then

was kept constant for 2 minutes, and finally a column equilibration time of 2 minutes with

100% A. Fractions were collected, triggered by UV absorbance at 210 nm. All fractions

were evaporated to dryness and redissolved in 550 µL DMSO-d6 with tetramethylsilane for

internal referencing in the subsequent NMR analysis.


104

4.2.7. Structural characterization of testosterone metabolites by NMR

The chemical structures of the testosterone metabolites formed by M11 were determined

by a combination of one dimensional (1D) 1H,

1H-

1H DQF-COSY,

1H-

13C HSQC,

1H-

13C HMQC,

1H-

13C HMBC and

1H-

1H NOESY NMR experiments. The experiments were conducted on a

Varian INOVA 500MHz spectrometer operating at a 1H frequency of 499.85 and a Varian

INOVA 600MHz spectrometer operating at a 1H frequency of 599.76 both equipped with a

5-mm probe and operating at 298 K. The proton and carbon chemical shifts were

referenced to the internal reference TMS (proton, δ 0.00; carbon, δ 0.00). Data were

processed using ACDLabs/SpecManager software, version 12.01.

4.2.8. UV-vis spectroscopy

UV-visible spectra were obtained using a Varian Cary 300 UV-visible spectrophotometer

with a temperature controller. All spectra were recorded at 25˚C in quartz cuvets (10 mm

path length) for samples containing 100 mM phosphate buffer (pH 7.4). UV-vis

spectroscopy was used to determine the concentration of enzyme, by measuring the

concentration of the CO-reduced complex (28), the spectral binding constants (Ks) and

spin states of both M11 and M11 A82W. Experimental details of the former two will be

discussed below.

4.2.9. Determination of spectral binding constant by UV-vis difference spectroscopy

Spectral binding constants (Ks) for both M11 and M11 A82W with testosterone were

determined by UV-vis difference spectroscopy in a fashion similar to that described

previously (21, 30). Briefly, different enzyme concentrations were used (ranging between

0.5 – 10.0 μM). The enzyme was placed in the sample and reference cuvettes. The sample

cuvette was titrated with 1 μL aliquots of 10 mM testosterone dissolved in DMSO.

Subsequently, the reference cuvette was titrated with an equal amount of DMSO and this

was used for background correction. The DMSO concentration was always kept below 2%.

After the samples had been mixed, the contents of the sample cuvette were allowed to

equilibrate for 2 minutes prior to analysis.

UV-vis difference spectra were recorded between 350 and 600 nm. The difference in

absorption between 390 nm (peak) and 419 nm (trough) was plotted versus the

testosterone concentration.

The spectral binding constants were determined by fitting the difference in

absorption according to the formula:

nn

s

n

SK

SAA

][

][*419390

(1)

where A390-419 and A∞ represent the difference in absorption between the peak and

trough at a specific testosterone concentration [S] and a saturating testosterone

concentration, respectively; Ks is the spectral binding constant of the enzyme-substrate


105

complex and n is the Hill coefficient. Ks and A∞ were estimated by non-linear curve

fitting, using OriginPro8.

4.2.10. Determination of spin states from UV spectra

The heme iron spin state of both M11 and M11 A82W in the substrate-bound form was

determined by deconvoluting absolute UV-vis spectra. Absolute UV-vis spectra were

deconvoluted between 350 and 600 nm with an enzyme concentration of 500 nM and a

testosterone concentration of 200 µM (concentrations were identical to those used for

NMR experiments). UV-vis spectra were deconvoluted by using the multiple-Gaussian

curve fitting program available in OriginPro8. Both enzymes were deconvoluted by using

four components: a low spin Soret band (419 nm), a high spin Soret band (390 nm), δ-

bands (~360 nm) and an additional ‘broad shoulder’ between 440 and 490 nm, caused by

the flavin prosthetic groups (31).

4.2.11. Preparation of NMR samples

The samples used for NMR experiments contained 50, 100, 150 or 200 µM testosterone,

added from a DMSO stock solution, and 500 nM purified enzyme in 100 mM phosphate

buffer (pH 7.4) with 10% D2O for locking. As a diamagnetic control, 200 µM testosterone

was dissolved in 100 mM phosphate buffer (pH 7.4). Prior to analysis, all samples were

flushed with nitrogen to remove possible dissolved oxygen, which might influence the

relaxation rates. The final volume of the NMR tube was 550 µL.

4.2.12. T1 relaxation experiments

The relaxation rate (R1 = 1/T1) experiments were conducted on a Varian Unity INOVA600

and a Bruker Avance 600 spectrometer operating at a 1H frequency of 599.76 MHz using a

saturation recovery pulse sequence with water flip back and Watergate to suppress the

water signal (32). Eight spectra were recorded for each sample, with an interpulse delay τ

ranging from 0.1 to 10s. For each spectrum, 64 scans were acquired. The data was

processed by using ACDLabs/SpecManager Version 12.01; i.e. zero-filling, line-broadening,

phasing, baseline correction and peak picking. OriginPro8 was used to determine R1

relaxation rates from the peak intensities.

4.2.13. Deriving distances from R1 relaxation data

Under fast-exchange conditions, the longitudinal relaxation rate is the weighted average

of the relaxation rates of the free and bound substrate (R1,f and R1,b , respectively) (21):

R1,obs = pf.R1,f + pb.R1,b (2)

where pf and pb are the fraction of the substrate in the free and bound state, respectively.

Under the conditions of the present experiment, the substrate concentration (50 - 200


106

µM) is much higher than the protein concentration (0.5 µM), so pf ≈ 1.

R1,b has a

contribution of both paramagnetic and diamagnetic relaxation rates (R1,b = R1,D + R1,P). As

discussed by Luz and Meiboom (33), R1,P depends on the relaxation rate caused by the

paramagnetic ion (R1,M) and the residence time of a ligand in the active site (M):

R1,obs - R1,f = pb.R1,D + pb /(T1,M + M) (3)

Under fast-exchange conditions, T1,M >> M, so that R1,M = 1/T1,M is:

R1,M = (R1,obs - R1,f - pb.R1,D)/pb (4)

R1,M is related to the proton-iron distance (rIS) by using the Solomon-Bloembergen

equation (34):

22226

222

2

0

,11

7

1

3)1(

4

15

2

cS

c

cI

c

IS

SI

Mr

SS

R

(5)

Where, 0 the permeability of free space, I and S are the gyromagnetic ratios of the

proton and electron, respectively, S is the spin state of the heme iron (obtained from

deconvoluting UV-vis spectra), and are the nuclear and electronic Larmor

frequencies, respectively, rIS is the distance from proton nuclei to the heme iron and c is

the correlation time which describes the dipolar interaction between ligand and

paramagnetic iron in solution. The correlation timec mainly originates from the electron-

spin relaxation timeS and has been estimated previously for P450 enzymes (21, 23, 26).

The average value of c reported for P450 enzymes is 3.0 e-10

s, and this value was used

here. Note that even if the used c value would be different by a factor of 2, this error

would result in a systematic over- or underestimation of the calculated distances by only

10% (23). The validity of fast exchange was confirmed from the observation of a positive

temperature dependence of R1,obs at five temperatures (288, 293, 298, 303 and 308 K) (22,

23).

Often, R1,M is calculated via equation 4 from R1,obs, by subtracting the relaxation

rate caused by the CO-reduced diamagnetic form of the enzyme (R1,f + pb.R1,D) (21, 30).

However, as noted by Smith et al. (26) and Jacobs et al. (35), the diamagnetic relaxation

rates were similar in the CO-reduced complex and enzyme-free buffer, so the diamagnetic

relaxation effect caused by the enzyme (pb.R1,D) is extremely small and can be neglected.

Therefore equation 6 was used to calculate R1,M:

R1,M = (R1,obs - R1,f)/pb (6)


107

where pb = [P450 BM3]/(Kd + [testosterone])

with Kd being the dissociation constant of the P450 BM3-testosterone complex.

In a case of multiple orientations, R1,M in Equation 5 can be rewritten as:

R1,M (fa/ra6 + fb/rb

6 + fc/rc

6 + ...)

with fa, fb, fc etc. are the fractions, and ra6

, rb6, rc

6 the distances of orientations a, b, c, etc.,

respectively, in the active site. R1,M is therefore dominated by the term with the shortest

distances. If there is one orientation present, i.e. fa =1, the distances determined for all

protons are representative for that specific orientation. The calculated distance

corresponds to the orientation with the proton closest to the heme iron. This also implies

that the distances should be internally consistent with a single orientation. Therefore,

when distances are internally inconsistent, this indicates the presence of multiple

orientations and suggests dynamic binding.

The rIS values for each proton were calculated using equations 2 to 6 with R1,obs

and R1,f as input variables (for the other required parameters, see above). To obtain robust

and reliable error estimates, we used Monte Carlo simulations (36, 37) given the

complexity of equations 2 to 6. For each proton, the distance was calculated in 10.000

iterations with R1,obs, R1,f and spectral binding constants Ks values varying in a Gaussian

distribution around their means with a width determined by the experimental errors

which were determined from at least three experiments. From the resulting histograms,

the average distances and distance error (standard deviation) were determined.

Results

4.3.1. Identification of testosterone and norethisterone metabolites.

The structures of the testosterone and norethisterone metabolites formed by the four

enzymes are displayed in Figure 1. The structures were determined on the basis of a

combination of 1D-1H,

1H-

1H DQF-COSY,

1H-

13C HSQC,

1H-

13C HMQC,

1H-

13C HMBC and

1H-

1H NOESY NMR and MS experiments. The assignments of the NMR spectra and

1H and

13C

chemical shifts of 15β-hydroxytestosterone, 16β-hydroxytestosterone and 2β-

hydroxytestosterone are consistent with previously published data (38, 39). Interestingly,

the NOESY spectrum of 2ß-hydroxytestosterone showed that the A-ring is folded into an

inverted half-chair conformation, as identified by a strong NOE interaction between the

proton at position 2α and the proton at position 9, which is consistent with the

observations by Jacobsen et al. (40).

The norethisterone metabolites were assigned to 15β-hydroxynorethisterone and 16β-

hydroxynorethisterone, as recently discussed by de Vlieger et al. (15).


108

10

5

1

4

2

3

8

7

9

6

13

14

12

1117

16

15

19

18

M11

M11 A82W

M01

M01 A82W

OH

O

H

H H

OH

Testosterone

M11

M11 A82W

M01

M01 A82W

CH3OH

O

H

H H

CH

OH

+ +

+

OH

O

H

H H

CH

OH

OH

O

H

H H

CH

OH

O

H

H HOH

CH3

CH3

OH

O

H

H H

CH3

CH3OH

O

H

H H

CH3

CH3

CH3

CH3

CH3CH3

Norethisterone

OH

Figure 1. Chemical structures of testosterone, norethisterone and the monohydroxy-metabolites formed by P450 BM3 mutants M11, M11 A82W, M01 and M01 A82W.

4.3.2. Effect of mutation A82W on regioselectivity of steroid hydroxylation

To evaluate the ability of this set of mutants to oxidize steroids, incubations were first

measured at a fixed substrate concentration (200 M). Figure 2 shows the relative

amounts of monohydroxy metabolites after incubation of 200 µM testosterone and

norethisterone in presence of the four studied mutants of P450 BM3.

In the case of testosterone, three metabolites were produced, corresponding to 15-

hydroxytestosterone, 16-hydroxytestosterone and 2-hydroxytestosterone. For both

M01 and M11, the major metabolite was 15-hydroxytestosterone. Addition of mutation

A82W to both M01 and M11 resulted in a strong increase in the level of 16-hydroxylation

of testosterone. In the case of M01, mutation A82W resulted in an increase in the level of

16-hydroxylation from 22% to 85% of total metabolism. In the case of M11, mutation

A82W increased the selectivity for 16-hydroxylation from 25% to 75%. In the case of

norethisterone, two metabolites were produced, corresponding to 15-

hydroxynorethisterone and 16-hydroxynorethisterone. As was found with testosterone,

mutation A82W strongly increased selectivity for 16-hydroxylation. In M01, mutation

A82W increased selectivity for 16-hydroxylation from 42% to 88%; in M11 selectivity for

16-hydroxylation increased from 58% to 77%.

4.3.3. Enzyme kinetic characterization of the P450 BM3 mutants

To evaluate whether the strong increase in selectivity for 16-hydroxylation was related to

an increase in affinity or in catalytic efficiency, we determined enzyme kinetic parameters

KM and Vmax for all mutants with testosterone and norethisterone (Table 1). For

testosterone, mutation A82W in M01 led to a 4.2 ± 0.8-fold increase in the KM values. The

Vmax for 16-hydroxylation increased 42.5 times, and the catalytic efficiency (Vmax/KM) for

16-hydroxylation increased 9.3 times. In the case of M11, mutation A82W led to a 1.7 ±

0.5 fold decrease in the KM value. The Vmax for 16-hydroxylation increased 1.5 times


109

whereas for 15-hydroxylation and 2-hydroxylation it decreased 7.3 times. The Hill

coefficient is never significantly different than 1, therefore cooperativity appears not to

take place. In contrast to testosterone, no A-ring hydroxylation was found in the case of

norethisterone. In the case of norethisterone, in M01, mutation A82W did not affect KM

values significantly. Therefore the different selectivity was due to a change in Vmax values.

The Vmax for 16-hydroxylation increased 15.9-fold. In the case of M11, mutation A82W

led to a 2.4 ± 0.2-fold increase in the KM value. The Vmax for 16-hydroxylation increased 3-

fold, while Vmax for 15-hydroxylation did not change significantly. The KM values for

norethisterone are always smaller than for testosterone. Therefore, it seems that

norethisterone has higher affinity for the enzymes, although the product formation is

higher for testosterone.

Figure 2. Relative amounts of monohydroxy metabolites of testosterone and norethisterone metabolism formed by the P450 BM3 mutants M11, M11 A82W, M01 and M01 A82W. Metabolites were analyzed by UPLC with UV detection at 254 nm and results from triplicate experiments are shown. Quantification is based on LC-UV hromatograms assuming that the extinction coefficients of the substrate and the metabolites are similar.


110

Table 1. Enzyme kinetic parameters of testosterone and norethisterone metabolism by P450 BM3 mutants M01, M01 A82W, M11 and M11 A82W

Testosterone Norethisterone

KMa

Vmaxb

Vmax/ KM KMa Vmax

b Vmax/ KM

M01 15-OH 52.6 ± 5.5 0.4 ± 0.02 7.8 ± 1.2 63.4 ± 5.6 0.7 ± 0.02 10.7 ± 1.3

16-OH 59.5 ± 3.2 0.2 ± 0.005 3.7 ± 0.3 102.8 ± 11.2 0.8 ± 0.04 7.8 ± 0.9

2-OH 52.3 ± 2.5 0.3 ± 0.004 5.5 ± 0.3

M01 A82W 15-OH 178.2 ± 30.7 0.7 ± 0.05 3.7 ± 0.9 64.5 ± 9.8 1.2 ± 0.02 18.6 ± 3.1

16-OH 248.1 ± 3.9 8.5 ± 0.12 34.5 ± 1 97.4 ± 4.6 12.7 ± 0.6 130.4 ± 12.3

2-OH 256.7 ± 52.1 0.6 ± 0.05 2.2 ± 0.5

M11 15-OH 149.6 ± 20.4 18.2 ± 1.2 121.7 ± 24.6 37.5 ± 4.2 1.1 ± 0.05 29.3 ± 4.6

16-OH 145.1 ± 11.8 7.0 ± 0.27 48.5 ± 5.8 39.5 ± 4.6 1.6 ± 0.07 40.5 ± 6.5

2-OH 224.6 ± 34.5 5.8 ± 0.4 24.9 ± 5.5

M11 A82W 15-OH 106.7 ± 5.5 2.5 ± 0.06 23.4 ± 1.8 85 ± 12.1 1.2 ± 0.08 14.1 ± 3

16-OH 109.9 ± 13.6 10.2 ± 0.6 92.8 ± 16.9 100.8 ± 6.7 4.8 ± 0.3 47.6 ± 6.1

2-OH 99.6 ± 7.5 0.8 ± 0.03 8.1 ± 0.9

a) KM values are in M; b) Vmax values are nmol product/minute/nmol enzyme.


111

4.3.4. Determination of coupling efficiency

To evaluate the effect of the restriction of the active site size on the coupling efficiency of

the enzymes, the rates of NADPH consumption and product formation were measured in

presence of 200 M of testosterone and norethisterone (Table 2).

Table 2. NADPH consumption rate, product formation rates and coupling efficiency of P450 BM3 mutants with testosterone (TST) and norethisterone (NET)

NADPH consumption

ratea

Total product

formation rateb

Coupling efficiency (%)

TST NET TST NET TST NET

M11 22.5+2.2 5.4+0.1 17.1+0.4 2.2+0.04 76+9.2 41+1.6

M11 A82W 10.6+2.7 6.2+1.3 9.2+0.2 4.1+0.1 83.5+8.2 66.9+7.2

M01 5.4+0.6 4.3+0.2 0.7+0.03 0.6+0.03 13.5+2 14.2+1.3

M01 A82W 11.2+0.06 14.8+0.7 4.3+0.2 9.6+0.04 38.5+1.8 64.7+3.5 anmol NADPH/min nmol BM3

bnmol/min nmol BM3

For testosterone, mutation A82W in M01 led to a 2.1-fold increase in the level of NADPH

consumption and to a 6.1-fold increase in the level of product formation. Thus, this

mutation resulted in a 2.8-fold increase in the coupling efficiency of this enzyme. In the

case of M11, after mutation A82W had been introduced, levels of both NADPH

consumption and product formation decreased 2-fold; therefore, the coupling efficiency

did not change significantly. For norethisterone, mutation A82W in M01 led to a 3.4-fold

increase in the level of NADPH consumption and a 16-fold increase in the level of product

formation, resulting in a 4.6-fold increase in coupling efficiency. In the case of M11,

mutation A82W did not affect NADPH consumption significantly and led to a 1.9-fold

increase in product formation, which corresponds to a 1.9-fold increased coupling

efficiency.

4.3.5. Binding affinity and spin state from optical absorption

As mentioned above, the A82W mutants of both M11 and M01 produced significantly

larger amounts of 16β-hydroxytestosterone as well as of 16β-hydroxynorethisterone. For

a more in-depth analysis of the metabolism in terms of binding affinity, spin-state

determination, and orientation of ligand in the heme active site, optical binding studies

and T1 paramagnetic relaxation NMR studies were performed. Testosterone was chosen

over norethisterone because it also showed changes in 15-hydroxylation to 16-

hydroxylation ratio as well as in D-ring to A-ring hydroxylation ratio. Its higher solubility

also allows for more sensitivity in the optical and NMR experiments. In addition, the NMR

spectrum of testosterone displays a smaller degree of proton resonance overlap than that

of norethisterone, so that T1 relaxation times can be measured more accurately. Although


112

the increase in D-ring selectivity caused by A82W mutation was greater for M01 than for

M11, the NMR studies were performed with M11 and M11 A82W because of their higher

coupling efficiency, which suggests a more productive mode of binding.

Binding of testosterone to both M11 and M11 A82W resulted in a type I binding spectrum

(Figure 3) corresponding to a substrate-induced low-spin to high-spin conversion.

Figure 3. UV-vis difference spectra of testosterone binding to M11 (A) and M11 A82W

(B).

Table 3. Characteristics of binding of testosterone to P450 BM3 Mutants M11 and M11 A82W

Ks (µM)a Hill coefficient High Spin (%)

b

M11 > 450 0.8 ± 0.3 27

M11 A82W 51 ± 4 1.01 ± 0.04 16 a

Spectral binding constant obtained by optical titration experiments; see Materials and Methods

b In presence of 500 nM enzyme and 200 μM testosterone

The spectral binding constants of testosterone binding to M11 and M11 A82W and the

percentage of high-spin state in the bound form are summarized in Table 3.

For M11 A82W, a dissociation constant of 51 ± 4 µM (Table 3) was determined for

testosterone with good accuracy thanks to the saturation of substrate binding being

almost reached. In contrast, when testosterone was titrated to M11, no complete enzyme

saturation was reached, which is indicative of a much higher dissociation constant

compared to that of M11 A82W. Nevertheless, via a fit of the nonlinear curve, the spectral

dissociation constant of testosterone binding to M11 was estimated to be >450 μM. This

value lies outside the range of testosterone concentrations used in the titration. The


113

highest concentration employed was 200 µM and was limited by the solubility of

testosterone under our conditions. Smith et al. (26) were able to record binding spectra to

a CYP2D6 mutant with testosterone concentrations of < 1.5 mM by including 13% DMSO

in the medium. However, DMSO concentrations > 2% induced strong spectral changes

with our P450 BM3 mutants [data not shown], consistent with previous observations (41).

The percentage of low- and high-spin of the heme iron was determined by deconvolution

of the absolute spectra at a protein concentration of 500 nM and in the presence of 200

µM testosterone. For M11 A82W the percentage of the high-spin form was estimated to

be 16% (Table 3) while for M11 the high-spin content of substrate-bound M11 was

estimated to be 27%. In the absence of testosterone, the high-spin content was estimated

to be ~10% in both enzymes.

4.3.6. Determination of fast exchange

Equation 4 assumes that T1,M >> M, which means that testosterone should be in fast-

exchange with M11 and M11 A82W. The fact that the fast-exchange conditions hold

follows from the following observations. R1,obs of the testosterone protons increases

linearly with an increasing reciprocal temperature in both M11 and M11 A82W [data not

shown] (22, 23). Titration of testosterone into a constant concentration of M11 and M11

A82W shows chemical shift changes of the methyl resonances [data not shown] which

indicates fast-exchange (42). We further note that the ratio of unbound to bound

substrate is around 400:1 in our case. Under these conditions, binding affects T1 only if M

is not significantly smaller than 1,M. As mentioned by Myers et al.(43), a M value of > 10-4

s seems unlikely for a substrate under these circumstances.

4.3.7. T1 relaxation and proton-iron distances

The spin-lattice (R1) relaxation rates of testosterone protons were measured in the

presence and absence of M11 and M11 A82W to obtain information about the individual

proton-iron distances in each mutant. Only for the protons for which the chemical shift

does not overlap with those of other protons, relaxation rates can be measured. Figure 4

shows the assignment of the signals in the 1D- 1H NMR spectrum of testosterone that do

not overlap with other signals. For eleven protons, spread over the rigid testosterone

structure, the relaxation rate and thus proton-iron distances could be determined.

Unfortunately, the signals of protons H2 and H16the sites of 2- and 16-

hydroxylation, respectively, overlapped with those of protons H6 and H11, respectively,

according to the 1H-

13C HSQC-spectrum of testosterone [data not shown].


114

Figure 4. 1D-

1H NMR spectrum of testosterone in aqueous buffer. The resonances used

for the T1 NMR relaxation experiments are indicated.

Table 4 includes the relaxation rates for each proton of testosterone in the absence and

presence of mutants M11 and M11 A82W and the calculated proton-iron distances. The

values presented are averages and standard deviations of at least three independent

experiments with 200 µM testosterone and 500 nM P450 BM3 mutant. Proton-iron

distances and their corresponding standard deviations were determined by Monte Carlo

simulations, in which the experimental errors originating from the relaxation rates and

binding constants were simulated in a Gaussian distribution around their means.

As one can see in Table 4, the shortest proton-iron distances in M11 are for the

protons in the A-ring and for protons in the D-ring, while longer distances are found for

the B- and C-rings. Because it is not possible to derive one specific orientation for

testosterone that is consistent with all measured distances, the binding must be dynamic;

i.e. testosterone must bind in multiple orientations in M11. In the case of multiple

orientations, there is a possibility that each binding orientation has a different binding

constant (44).

The spectral binding constant that was obtained by difference UV-vis does not

provide information about the affinity of different orientations. By measuring R1,obs as a

function of ligand concentration for each individual proton, and by fitting the obtained

data with equation 6, we determined previously independent estimates of the binding

constant by NMR (44). However, because of the limited solubility of testosterone and the

low signal-to-noise ratios at low testosterone concentration, we could measure the

concentration dependence of R1,obs only over a small concentration range (100- 200 µM).

Although this small range and small number of data points did not permit the assessment

of accurate Kd values for each proton, for each mutant no significant differences in the

slope of the plot R1,obs versus testosterone concentration were found for each proton


115

[data not shown]. This might point to comparable binding constants for each orientation.

From the significantly different slopes of the R1,obs versus testosterone concentration

curves between M11 and M11 A82W, we can conclude that the affinity of testosterone for

M11 is much lower than that for M11 A82W, consistent with the difference in the spectral

binding constant.

Upon comparison of the proton-iron distances of testosterone bound to M11

A82W with those of M11, it is observed that the A-ring protons H1, H1 and H4 show

small but significantly increased proton-iron distances in the M11 A82W mutant. In

contrast, the distances from D-ring protons H18 and H15 to the heme iron were

decreased by the A82W mutation. The results of the relaxation experiments at the lower

testosterone concentrations were consistent to those obtained at 200 µM testosterone.

Again, the introduction of the A82W mutation was shown to cause an increase in the

proton-iron distances of A-ring protons, and a decrease of proton-iron distances of some

D-ring protons [data not shown]. Because of the better signal-to-noise ratios, Table 4

shows only the results obtained at 200 µM testosterone.


116

Table 4. R1 relaxation rates (s

-1) of testosterone

a protons in presence and absence of M11 and M11 A82W and calculated

proton-iron distances (Å)

Ring buffer M11 M11 A82W M11 M11 A82W

R1,f (s-1

) b R1,obs (s

-1)

b R1,obs (s

-1)

b r (Å)

c r (Å)

c

H1α A 1.88 ± 0.04 2.20 ± 0.07 2.19 ± 0.06 6.0 ± 0.2 6.7 ± 0.2 #

H1ß A 2.11 ± 0.05 2.21± 0.08 2.19 ± 0.05 7.1 ± 0.5 8.2 ± 0.5 #

H4 A 0.53 ± 0.01 1.26 ± 0.03 1.54 ± 0.02 5.26 ± 0.04 5.53 ± 0.02 #

H19 A/B 1.27 ± 0.01 1.49 ± 0.01 1.76 ± 0.01 6.4 ± 0.1 6.3 ± 0.1

H7ß B 1.93 ± 0.05 2.04 ± 0.09 2.26 ± 0.07 6.9 ± 0.5 6.6 ± 0.2

H12α C 1.90 ± 0.04 2.07 ± 0.09 2.24 ± 0.08 6.5 ± 0.2 6.5 ± 0.2

H12ß C 1.74 ± 0.03 1.86 ± 0.08 2.22 ± 0.04 6.9 ± 0.4 6.2 ± 0.1 #

H18 C/D 1.14 ± 0.01 1.37 ± 0.02 1.72 ± 0.01 6.4 ± 0.1 6.1 ± 0.1 #

H15ß D 1.73 ± 0.04 1.84 ± 0.09 2.11 ± 0.16 7.1 ± 0.5 6.4 ± 0.3 #

H16α D 1.44 ± 0.05 1.89 ± 0.12 2.23 ± 0.09 5.6 ± 0.3 5.7 ± 0.1

H17α D 1.11 ± 0.02 1.38 ± 0.05 1.72 ± 0.03 6.2 ± 0.2 6.0 ± 0.1 a R1 relaxation rates were determined at a 200μM testosterone concentration;

b averages and standard deviations on R1

relaxation rates were determined from three to five independent R1 measurements; c distances and standard deviations

were estimated based on Monte Carlo simulations considering the standard deviations of R1,obs, R1,f and Kd; #

Statistically significant difference in proton-iron distances between M11 and M11 A82W.


117

4.3. Discussion Because hydroxysteroids have important pharmaceutical applications, there is a great

need for biocatalysts with high activity and regioselectivity in steroid hydroxylation.

Recently, by laboratory evolution starting from P450 BM3 (F87A) mutants were obtained

with high selectivity for 2- and 15-hydroxylation (45). In this work, we present two

novel P450 BM3 enzymes, M01 A82W and M11 A82W, which can catalyze regioselective

hydroxylation of steroids at position 16. The importance of modification of steroids at

this position has been discussed previously by Laplante et al. (46) and Vicker et al. (47).

Huang et al. (18) previously showed that substitution of alanine 82 in wild-type

P450 BM3 by the larger and more hydrophobic amino acids, tryptophan and

phenylalanine, improves binding affinity and KM for long-chain fatty acids by orders of

magnitude. In this study, we demonstrate that substitution of alanine 82 with tryptophan

in P450 BM3 mutants M01 and M11 results in improved regioselectivity and catalytic

efficiency of testosterone hydroxylation. The binding affinity of testosterone for M11

A82W is significantly higher than its binding affinity for M11 (Table 3), which might be

explained by the addition of a hydrophobic interaction between testosterone and the

tryptophan at position 82. A more efficient exclusion of water from the active site by the

stronger substrate-enzyme interaction might also explain the increased coupling efficiency

of the A82W mutants (Table 2).

Uncoupling of the catalytic cycle of P450s can occur when mutations are

introduced into the active site of P450 or in the presence of non-native substrates (48),

leading to the formation of reactive oxygen species and rapid enzyme inactivation (8).

Interestingly, the recently published P450 BM3 mutants with high selectivity for 2ß- and

15ß-hydroxylation of testosterone also contained substitutions at position 82, including

A82F substitution (17). However, only small amounts of 16ß-hydroxytestosterone were

found. In this study, mutations at position 82 were always combined with mutations at

position 78, because this combination was expected to act synergistically. Therefore, the

simultaneous mutation at position 78 might explain the preference for 15ß-hydroxylation.

T1 NMR relaxation experiments were performed to investigate whether mutation

A82W leads to changes in proton-iron distances that might reflect differences in the

orientation and dynamics of testosterone in the active sites of M11 and M11 A82W. Using

the same approach, Smith et al. (26) previously studied the orientation of binding of

testosterone to a CYP2D6 mutant that had acquired the ability to catalyze testosterone

hydroxylation by a F483I mutation. In the present study, first the proton resonances of

testosterone were assigned on the basis of DQF-COSY, 1H-

13C HSQC,

1H-

1H NOESY and 1D-

1H NMR spectra. All assignments were consistent with those of Kirk et al. (39).

As shown in Table 4, for M11 the shortest proton-iron distances are obtained for

protons located in the A- and D-rings (H1α, H4 and H16α). The fact that it is not possible to

assign a specific orientation of testosterone in M11 that is consistent with all measured


118

distances, implies that the binding must be dynamic; i.e. testosterone is able to bind in

multiple orientations.

The distance data, which are well spread over the rigid testosterone, appear to

be consistent with two classes of orientations of testosterone in the M11 active site,

namely, a group of orientations with the A-ring closest to the heme iron and a group of

orientations with the D-ring closest to the heme iron. Also, in case of M11 A82W, the

distances imply the presence of multiple orientations and dynamics involved in binding.

Statistically significant differences in testosterone orientation were found upon

comparison of M11 A82W with M11, as visualized in Figure 5.

Figure 5. Differences in distances of testosterone binding to M11 and M11 A82W (rM11A82W (Å) – rM11(Å)). The protons for which proton-iron distances were derived are colored according to the difference in distance, see legend. The directly attached carbon and C-H bond are colored similar to the color of the proton. Protons for which no proton-iron distance could be derived are white. A negative difference in distance (reddish) means a shorter proton-iron distance in M11 A82W compared to M11. A positive difference in distance (blueish) means a longer proton-iron distance in M11 A82W compared to M11.

In M11 A82W protons H1α, H1ß and H4 moved further away from the heme iron, whereas

H12ß, H15ß and H18ß moved closer to the heme iron. Although the changes in substrate

orientation is mainly deduced by changes in distances of protons that are not

hydroxylated, the changes are consistent with the increased selectivity for D-ring

hydroxylation and reduced selectivity for A-ring hydroxylation for M11 A82W. The

decrease in 15ß-hydroxylation in the M11 A82W mutant relative to that in M11 (Figure 2)

might be explained by an orientation of the D-ring in which H16ß is able to approach the

heme iron significantly closer than H15ß. Alternatively, the C-H16ß bond might orient in a

more favorable position for abstraction of hydrogen by the reactive FeO species: a linear


119

C-H-O arrangement is considered a typical feature in C-H bond activation by FeO species,

which precedes aliphatic hydroxylation (49). In both cases, H16 will be the preferred site

of hydroxylation at the expense of 15ß hydroxylation. Unfortunately, it was not possible to

determine the difference in distances between the heme iron and H15 and H16

protons, because the signal of the latter proton overlaps with that of proton H11.

From the distance data, it can be seen that the average proton-iron distances of

testosterone in M11 and M11 A82W is 6.3Å, while the shortest distances are ~5.3Å. A

distance of less than 4 Å is considered to be necessary to allow hydrogen abstraction (50).

However, it should be realized that the proton-iron distances in Table 4 represent binding

of testosterone to the initial ferric enzyme-substrate complexes that do not necessarily

represent the orientation of testosterone in the oxygenated ferrous form of P450 BM3. It

has been shown previously that fatty acids bind P450 BM3 enzymes via a two-step binding

process, with initial binding to a site distant from the heme iron and movement closer to

the heme iron upon iron reduction (50).

Nonflexible compounds, like codeine and testosterone, were considered to bind

via a one-step binding mechanism to CYP2D6 (class I P450) (51). The question of whether

nonflexible ligands, like steroids, bind in P450 BM3 enzymes via a two-step binding

mechanism or directly to the proximal position then arises. The average distances and

shortest distances for binding of testosterone to M11 and M11 A82W indicate that

testosterone binds distant in the non-reduced form in both enzymes, similar to fatty acids

binding to nonreduced P450 BM3 (21). Thus, it is most likely that steroids bind to P450

BM3 enzymes via a two-step binding mechanism.

In summary, we showed that applying substitution A82W to P450 BM3 mutants

M01 and M11 improves regioselective hydroxylation of norethisterone and testosterone

at position 16ß. This mutation improved both binding affinity and coupling efficiency,

thereby strongly improving the catalytic activity. As shown by T1 paramagnetic relaxation

NMR, this single mutation appears to change the orientation of testosterone in the M11

A82W mutant as compared to the orientation in M11. Testosterone is oriented in M11

with both the A- and D-ring closest to the heme iron, whereas testosterone is mainly

oriented with its D-ring closest to the heme iron in the M11 A82W mutant, which might

explain its increased selectivity of D-ring hydroxylation.


120

References

1. Guengerich, F. P. (2001) Common and uncommon cytochrome P450 reactions related to metabolism and chemical toxicity, Chem Res Toxicol 14, 611-650.

2. Guengerich, F. P. (2002) Cytochrome P450 enzymes in the generation of commercial products, Nat Rev Drug Discov 1, 359-366.

3. Bernhardt, R. (2006) Cytochromes P450 as versatile biocatalysts, J Biotechnol 124, 128-145.

4. Urlacher, V., and Schmid, R. D. (2002) Biotransformations using prokaryotic P450 monooxygenases, Curr Opin Biotechnol 13, 557-564.


6. Fernandes, P., Cruz, A., Angelova, B., Pinheiro, H. M., and Cabral, J. M. S. (2003) Microbial conversion of steroid compounds: recent developments, Enzyme Microb Tech 32, 688-705.


8. Liu, H., Munro, E., Rouse, J., and Zhu, X. (2002) Simulation methods for multipole imaging systems and aberration correctors, Ultramicroscopy 93, 271-291.

9. 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, 13442-13450.

10. Lussenburg, B. M. A., Babel, L., Vermeulen, N. P. E., and Commandeur, J. N. M. (2005) Evaluation of alkoxyresorufins as fluorescent substrates for high-throughput screening of cytochrome P450BM3 and site-directed mutants, Drug Metab Rev 37, 20-21.

11. Lewis, J. C., Mantovani, S. M., Fu, Y., Snow, C. D., Komor, R. S., Wong, C. H., and Arnold, F. H. (2010) Combinatorial Alanine Substitution Enables Rapid Optimization of Cytochrome P450(BM3) for Selective Hydroxylation of Large Substrates, Chembiochem 11, 2502-2505.

12. Lewis, J. C., Bastian, S., Bennett, C. S., Fu, Y., Mitsuda, Y., Chen, M. M., Greenberg, W. A., Wong, C. H., and Arnold, F. H. (2009) Chemoenzymatic elaboration of monosaccharides using engineered cytochrome P450BM3 demethylases, Proc Natl Acad Sci U S A 106, 16550-16555.




121

15. de Vlieger, J. S., Kolkman, A. J., Ampt, K. A., Commandeur, J. N., Vermeulen, N. P., Kool, J., Wijmenga, S. S., Niessen, W. M., Irth, H., and Honing, M. (2010) Determination and identification of estrogenic compounds generated with biosynthetic enzymes using hyphenated screening assays, high resolution mass spectrometry and off-line NMR, J Chromatogr B Analyt Technol Biomed Life Sci 878, 667-674.

16. Vottero, E., Rea, V., Lastdrager, J., Honing, M., Vermeulen, N. P., and Commandeur, J. N. (2011) Role of residue 87 in substrate selectivity and regioselectivity of drug-metabolizing cytochrome P450 CYP102A1 M11, J Biol Inorg Chem 16, 899-912.

17. Kille, S., Zilly, F. E., Acevedo, J. P., and Reetz, M. T. (2012) Regio- and stereoselectivity of P450-catalysed hydroxylation of steroids controlled by laboratory evolution, Nat Chem 3, 738-743.

18. Huang, W. C., Westlake, A. C., Marechal, 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.

19. Vottero, E., Rea, V., Lastdrager, J., Vermeulen, N.P.E., Commandeur J.N.M. (2011) Role Of Residue 87 In Substrate- And Regioselectivity Of Drug Metabolizing Cytochrome P450 CYP102A1 M11. Unpublished experiments.

20. Yao, H., McCullough, C. R., Costache, A. D., Pullela, P. K., and Sem, D. S. (2007) Structural evidence for a functionally relevant second camphor binding site in P450cam: Model for substrate entry into a P450 active site, Proteins 69, 125-138.

21. Modi, S., Primrose, W. U., Boyle, J. M. B., Gibson, C. F., Lian, L. Y., and Roberts, G. C. K. (1995) Nmr-Studies of Substrate-Binding to Cytochrome-P-450 (Bm3) - Comparisons to Cytochrome-P-450 (Cam), Biochemistry-Us 34, 8982-8988.

22. Regal, K. A., and Nelson, S. D. (2000) Orientation of caffeine within the active site of human cytochrome P450 1A2 based on NMR longitudinal (T-1) relaxation measurements, Arch Biochem Biophys 384, 47-58.

23. Cameron, M. D., Wen, B., Allen, K. E., Roberts, A. G., Schuman, J. T., Campbell, A. P., Kunze, K. L., and Nelson, S. D. (2005) Cooperative binding of midazolam with testosterone and alpha-naphthoflavone within the CYP3A4 active site: A NMR T-1 paramagnetic relaxation study, Biochemistry-Us 44, 14143-14151.

24. Oliver, C. F., Modi, S., Sutcliffe, M. J., Primrose, W. U., Lian, L. Y., and Roberts, G. C. K. (1997) A single mutation in cytochrome P450 BM3 changes substrate orientation in a catalytic intermediate and the regiospecificity of hydroxylation, Biochemistry-Us 36, 1567-1572.

25. Oliver, C. F., Modi, S., Primrose, W. U., Lian, L. Y., and Roberts, G. C. K. (1997) Engineering the substrate specificity of Bacillus megaterium cytochrome P-450 BM3: hydroxylation of alkyl trimethylammonium compounds, Biochem J 327, 537-544.

26. Smith, G., Modi, S., Pillai, I., Lian, L. Y., Sutcliffe, M. J., Pritchard, M. P., Friedberg, T., Roberts, G. C., and Wolf, C. R. (1998) Determinants of the substrate specificity of human cytochrome P-450 CYP2D6: design and construction of a mutant with testosterone hydroxylase activity, Biochem J 331 ( Pt 3), 783-792.


122

27. Damsten, M. C., van Vugt-Lussenburg, B. M. A., Zeldenthuis, T., de Vlieger, J. S. B., Commandeur, J. N. M., and Vermeulen, N. P. E. (2008) Application of drug metabolising mutants of cytochrome P450BM3 (CYP102A1) as biocatalysts for the generation of reactive metabolites, Chem-Biol Interact 171, 96-107.

28. Omura, T., and Sato, R. (1964) Carbon Monoxide-Binding Pigment of Liver Microsomes .I. Evidence for Its Hemoprotein Nature, J Biol Chem 239, 2370-2378.

29. Truan, G., and Peterson, J. A. (1998) Thr268 in substrate binding and catalysis in P450BM-3, Arch Biochem Biophys 349, 53-64.

30. Hummel, M. A., Gannett, P. M., Aguilar, J., and Tracy, T. S. (2008) Substrate proton to heme distances in CYP2C9 allelic variants and alterations by the heterotropic activator, dapsone, Arch Biochem Biophys 475, 175-183.

31. Li, H. Y., Darwish, K., and Poulos, T. L. (1991) Characterization of Recombinant Bacillus-Megaterium Cytochrome-P-450bm-3 and Its 2 Functional Domains, J Biol Chem 266, 11909-11914.

32. Liu, M. L., Mao, X. A., Ye, C. H., Huang, H., Nicholson, J. K., and Lindon, J. C. (1998) Improved WATERGATE pulse sequences for solvent suppression in NMR spectroscopy, J Magn Reson 132, 125-129.

33. Luz, Z., and Meiboom, S. (1964) Proton Relaxation in Dilute Solutions of Cobalt(2) + Nickel)2) Ions in Methanol + Rate of Methanol Exchange of Solvation Sphere, J Chem Phys 40, 2686-&2692.

34. Solomon, I., and Bloembergen, N. (1956) Nuclear Magnetic Interactions in the Hf Molecule, J Chem Phys 25, 261-266.

35. Jacobs, R. E., Singh, J., and Vickery, L. E. (1987) Nmr-Studies of Cytochrome-P-450scc - Effects of Steroid Binding on Water Proton Access to the Active-Site of the Ferric Enzyme, Biochemistry-Us 26, 4541-4545.

36. Bevington (1992) Data reduction and error analysis for the pysical sciences, McGraw-Hill, London.

37. Press, W. H. (2007) Numerical Recipes: The Art of Scientific Computing, Cambridge University Press Chapter 7 Random numbers.

38. Mahato, S. B., and Mukherjee, A. (1984) Microbial Transformation of Testosterone by Aspergillus-Fumigatus, J Steroid Biochem 21, 341-342.

39. Kirk, D. N., Toms, H. C., Douglas, C., White, K. A., Smith, K. E., Latif, S., and Hubbard, R. W. P. (1990) A Survey of the High-Field H-1-Nmr Spectra of the Steroid-Hormones, Their Hydroxylated Derivatives, and Related-Compounds, J Chem Soc Perk T 2, 1567-1594.

40. Jacobsen, N. E., Kover, K. E., Murataliev, M. B., Feyereisen, R., and Walker, F. A. (2006) Structure and stereochemistry of products of hydroxylation of human steroid hormones by a housefly cytochrome P450 (CYP6A1), Magn Reson Chem 44, 467-474.

41. Roccatano, D., Wong, T. S., Schwaneberg, U., and Zacharias, M. (2005) Structural and dynamic properties of cytochrome P450BM-3 in pure water and in a dimethylsulfoxide/water mixture, Biopolymers 78, 259-267.

42. Jardetzky, O., and Roberts, G. C. K. (1981) NMR in Molecular Biology, Academic Press New York.


123

43. Myers, T. G., Thummel, K. E., Kalhorn, T. F., and Nelson, S. D. (1994) Preferred Orientations in the Binding of 4'-Hydroxyacetanilide (Acetaminophen) to Cytochrome-P450 1a1 and 2b1 Isoforms as Determined by C-13-Nmr and N-15-Nmr Relaxation Studies, J Med Chem 37, 860-867.

44. Modi, S., Gilham, D. E., Sutcliffe, M. J., Lian, L. Y., Primrose, W. U., Wolf, C. R., and Roberts, G. C. K. (1997) 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine as a substrate of cytochrome P450 2D6: Allosteric effects of NADPH-cytochrome P450 reductase, Biochemistry-Us 36, 4461-4470.

45. 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.

46. Laplante, Y., Cadot, C., Fournier, M. A., and Poirier, D. (2008) Estradiol and estrone C-16 derivatives as inhibitors of type 1 17beta-hydroxysteroid dehydrogenase: blocking of ER+ breast cancer cell proliferation induced by estrone, Bioorg Med Chem 16, 1849-1860.

47. Vicker, N., Lawrence, H. R., Allan, G. M., Bubert, C., Smith, A., Tutill, H. J., Purohit, A., Day, J. M., Mahon, M. F., Reed, M. J., and Potter, B. V. (2006) Focused libraries of 16-substituted estrone derivatives and modified e-ring steroids: inhibitors of 17beta-hydroxysteroid dehydrogenase type 1, ChemMedChem 1, 464-481.

48. Loida, P. J., and Sligar, S. G. (1993) Engineering Cytochrome-P-450 (Cam) to Increase the Stereospecificity and Coupling of Aliphatic Hydroxylation, Protein Eng 6, 207-212.

49. Kamachi, T., and Yoshizawa, K. (2003) A theoretical study on the mechanism of camphor hydroxylation by compound I of cytochrome p450, J Am Chem Soc 125, 4652-4661.

50. Modi, S., Sutcliffe, M. J., Primrose, W. U., Lian, L. Y., and Roberts, G. C. K. (1996) The catalytic mechanism of cytochrome P450 BM3 involves a 6 angstrom movement of the bound substrate on reduction, Nat Struct Biol 3, 414-417.


124

CHAPTER 5

COMBINATION OF BIOTRANSFORMATION BY P450 BM3

MUTANTS WITH ON-LINE POST-COLUMN BIOAFFINITY AND

MASS SPECTROSCOPIC PROFILING AS A NOVEL STRATEGY TO

DIVERSIFY AND CHARACTERIZE P38 KINASE INHIBITORS

Vanina Rea, David Falck, Jeroen Kool, Frans J. de Kanter, Jan N.M.

Commandeur, Nico P.E.Vermeulen, Wilfried M.A.Niessen and Maarten

Honing

Submitted

Chapter 5 Generation and characterization of modified p38 inhibitors

125

Abstract

Mutants of bacterial Cytochrome P450 BM3 from Bacillus megaterium have gained

increasing interest to support drug metabolism studies by producing human relevant

biotransformation products and/or to support drug discovery by producing libraries of

new chemical entities based on initial lead compounds. In the present study, a library

of 33 P450 BM3 mutants was used to diversify TAK-715, a representative (highly

lipophilic) inhibitor of p38α mitogen-activated protein kinase (p38α MAPK or p38α

kinase). To simultaneously determine the individual bioaffinity and identity of the

different metabolites produced, an analytical high-resolution screening (HRS) approach

was used, based on post-column on-line bioaffinity profiling with parallel MS/MS

identification. The screening of metabolic mixtures produced by the P450 BM3

mutants demonstrated that introducing mutations in the active site of the P450 BM3

resulted in different profiles of biotransformation products. Several P450 BM3 mutants

were mimicking the metabolic profile obtained by human liver microsomes. The major

biotransformation products of TAK-715 could be produced in semi-preparative

amounts with the most active P450 BM3 mutant, enabling full structure elucidation of

metabolites by 1H-NMR and quantification of their p38α kinase affinity.

The present results demonstrate that the combination of catalytically diverse

P450 BM3 mutants and the HRS technology for on-line bioaffinity profiling and

structure identification of metabolic mixtures might be a valuable platform for the

generation of relevant quantities of human relevant drug biotransformation products

for pharmacological and toxicological evaluation. Furthermore, linking structural

modifications by metabolism to changes in drug target affinity might be efficient tool

to construct the pharmacophore model next to the more conventional medicinal

chemistry approach of synthesizing structural analogues of lead compounds.


126

5.1. Introduction

For more than 70% of the top 200 most prescribed drugs hepatic metabolism

represents the major mechanism of clearance from the organism (1). The majority of

drug metabolism reactions involved are catalyzed by the cytochrome P450 (CYP)

enzymes of which CYP1A2, CYP2D6, CYP2C9, CYP2C19 and CYP3A4 in particular are

responsible for the bulk of metabolism of known drugs in humans (1). Because it has

been established that sometimes drug metabolites can be pharmacologically active

towards the same drug target as the parent compound, or towards other targets,

leading to undesired off-target pharmacology, it is nowadays required for drug

registration to perform safety testing on circulating metabolites if they represent a

high proportion of the dose in humans (2). Full pharmacological assessment of

biotransformation products usually requires organic synthesis to obtain sufficient

amounts of pure compound. As an alternative to organic synthesis, biosynthetical

approaches are upcoming methodologies. Biocatalysts (i.e. enzymes) can perform

reactions with a high degree of regio- and stereoselectivity that often cannot be

achieved by organic synthesis (3, 4). Relatively new in drug discovery is the application

of biocatalysts for the generation of new chemical entities or even unique compound

libraries (5). When using mono-oxygenase systems, such biotransformation products

are particularly interesting in lead optimization as they can show improved

physicochemical properties such as solubility and target selectivity without significant

loss of their pharmacological activity compared to the parent itself.

Cytochromes P450 (CYPs) are emerging as biocatalysts in biotechnology due to their

catalytic versatility and extremely wide substrate diversity (6, 7). In particular, bacterial

CYPs are preferred to mammalian CYPs because they are soluble, show much higher

turnover rates and can be expressed at high levels in E.coli. A very promising candidate

as biocatalyst for the generation of large quantities of drug metabolites is P450 BM3

(CYP102A1) from Bacillus megaterium. This enzyme shows the highest catalytic activity

ever recorded for a P450 (8). Over the recent years, many research groups have

succeeded in expanding the substrate selectivity of P450 BM3 and improving its

catalytic properties by site-directed and/or random mutagenesis, as thoroughly

reviewed recently (9). Through rational redesign and directed evolution, P450 BM3

mutants have been obtained that are able to oxidize fatty acids, aromatic

hydrocarbons, long-chain and gaseous alkanes, terpenoids, steroids, carboxylic acids

and pharmaceuticals.

Although several P450 BM3 mutants have been shown to perform oxygenation

reaction with a high degree of regio- and stereoselectivity (10-12), biosynthesis of drug

metabolites using engineered P450 BM3s also can lead to the formation of multiple

metabolites (5, 13-18). Structural identification and characterization of

pharmacological and toxicological properties of individual compounds in mixtures

therefore usually depends on elaborate fractionation processes using semi-preparative


127

separation strategies. Alternatively, characterization of compounds in complex

mixtures can be performed using the so-called high-resolution screening (HRS)

approach, that combines state-of-the-art analytical separations with post-column on-

line bioaffinity profiling, based on enzyme inhibition or displacement of tracer

compounds, and parallel identity assessment with MS. As reviewed recently (19), this

approach has been successfully used for the identification of ligands/inhibitors of

several drug targets, such as human estrogen receptors and , acetylcholinesterase,

acetylcholine binding protein, phosphodiesterases, proteases and peptidases,

cathepsine B and drug metabolizing enzymes such as cytochromes P450 and

glutathione S-transferases.

Recently, a HRS methodology has been developed for the on-line screening of

inhibitors of p38α kinase (p38α MAPK or p38α kinase) and simultaneous structure

elucidation by high resolution mass spectrometry (HR-MS) (20). p38α kinase is an

important drug target in contemporary drug discovery because it plays a central role

in inflammatory cellular signaling processes (21). Several companies have reported

preliminary human clinical results for p38α kinase inhibitors, but until now no drug has

reached the market (22). The HRS methodology for p38α kinase affinity has recently

been applied to identify inhibitors in mixtures which were generated by on-line

electrochemical conversion of commercially available p38-inhibitors (23). However,

although electrochemistry in some cases has been shown to produce human-relevant

oxidation products, it cannot mimic all P450-dependent oxygenation reactions, such as

hydroxylation of unactivated sp3-CH-groups, which are synthetically challenging (24).

In the present study, a library of P450 BM3 mutants was used to prepare

metabolic mixtures of the p38 kinase inhibitor TAK-715 (N-[4-[2-ethyl-4-(3-

methylphenyl)-1,3-thiazol-5-yl]-2-pyridyl]benzamide) which were subsequently

characterized using the recently developed HRS platform for p38 kinase inhibitors.

The intrinsic high turnover rates of P450 BM3s and the convenient large scale

production and purification protocols for these enzymes allow semi-preparative

production of human relevant metabolites or new chemical entities for further

pharmacological profiling (5). TAK-715 was one of the most potent p38 kinase

inhibitor from a series of synthetic 4-phenyl-5-pyridyl-1,3-thiazoles (25) , which also

induced LPS-stimulated release of TNF- from human monocytic THP-1 cells (26). TAK-

715 was advanced to clinical Phase II trials for rheumatoid arthritis, but was

discontinued as it did not satisfy criteria for further development (27). However,

recently Verkaar et al. showed that TAK-715 is able to inhibit Wnt-3-stimulated -

catenin signaling pathway (28) whose aberrations are associated with several

malignancies, most notably cancer (29). Therefore, lead diversification of TAK-715

aiming at enhancing its effectiveness, diminish its toxicity, and/or increase its oral

absorption is still required for its anti-inflammatory and the chemotherapeutic


128

indication. In addition, previously reported lead optimization of typical inhibitors was

performed using a library encompassing rather lipophilic compounds. Application of

these mono-oxygenases is anticipated to improve the physical chemical properties of

the lead compounds, while maintaining their bioaffinity.

5.2. Materials and methods

5.2.1. Chemicals

Human recombinant p38α and TAK-715 were a kind gift of MSD Research Laboratories

(Oss, the Netherlands). SKF86002 was delivered by Merck KGaA (Darmstadt,

Germany). Fused silica tubing (250-μm inner and 375-μm outer diameter) covalently

coated with polyethylene glycol (PEG) was purchased from Sigma-Aldrich (Schnelldorf,

Germany). Methanol (LC–MS grade) and formic acid (LC–MS grade) were from Biosolve

(Valkenswaard, the Netherlands). All other chemicals were of analytical grade and

were obtained from Sigma-Aldrich (Schnelldorf, Germany). Human liver microsomes

(HLM) pooled from different individual donors were obtained from BD Gentest TM

(San Jose, USA) and contained 20 mg/mL protein (Cat. No. 452161).

5.2.2. Expression and isolation of P450 BM3 mutants

Expression of the BM3 mutants was performed by transforming competent E.Coli BL21

cells with the corresponding pET28+-vectors, as described previously (30). For

expression, 600 mL Terrific Broth (TB) medium (24 g/L yeast extract, 12 g/L tryptone, 4

mL/L glycerol) with 30 mg/mL kanamicin was inoculated with 15 mL of an overnight

culture. The cells were grown at 175 rpm and 37°C until the OD600 reached 0.6. Then,

protein expression was induced by the addition of 0.6 mM isopropyl--D-

thiogalactopyranoside (IPTG). The temperature was lowered to 20°C and 0.5 mM of

the heme precursor -aminolevulinic acid was added. Expression was allowed to

proceed overnight. Then, cells were harvested by centrifugation (4600 g, 4°C, 25 min),

and the pellet was resuspended in 20 mL KPi-glycerol buffer (100 mM potassium

phosphate (KPi) pH=7.4, 10% glycerol, 0.5 mM EDTA, and 0.25 mM dithiothreitol). Cells

were disrupted by French press (1000 psi, 3 repeats) and the cytosolic fraction was

separated from the membrane fraction by ultracentrifugation of the lysate (120.000 x

g, 4°C , 60 min). CYP concentrations were determined using carbon monoxide (CO)

difference spectrum assay.

5.2.3. Selection of the P450 BM3 mutant library

In the present study, 33 mutants of P450 BM3 were selected which could be expressed

as good levels and which showed catalytic diversity towards a variety of drugs and

steroids. Mutants M01 (R47L, F87V, L188Q, E267V, G415S), M02 (R47L, L86I, F87V,


129

L188Q, N319T), M05 (R47L, F81I, F87V, L188Q, E267V, G415S) and M11 (R47L, E64G,

F81I, F87V, E143G, L188Q, Y198C, E267V, H285Y and G415S) were previously

constructed by a combination of three site-directed mutations and subsequent

random-mutagenesis by error-prone PCR (14). Mutants M01 and M11 where used as

templates for additional site-directed mutagenesis of active site residues, guided by

available crystal structures of P450 BM3 and by computational modeling of the active

site of P450 BM3(31). Positions which were selected for mutagenesis were residues 72,

74, 82, 87 and 437 which appear to be key active site residues which have profound

influence on regio- and stereoselectivity of P450 BM3 mutants.

Positions S72 and A74 are both located in the substrate binding channel (32, 33) and

have been shown to influence regioselectivity and activity (11, 15). The effect of a

negative charged residue (Asp or Glu) in both M01 and M11 in position 72 and 74 is

evaluated in this work.

Mutation A82W was selected because it strongly influenced regioselectivity of steroid

hydroxylation when applied to M01 and M11 (10).

Mutations at position 87 have been extensively studied as the amino acid lies very

close to the heme, according to available crystal structures of P450 BM3. In our

previous studies, we mutated this position in M11 to all possible amino acids, showing

that mutation at position 87 has strong influence in the metabolism of

alkoxyresorufins, and the regioselectivity of testosterone and clozapine metabolism

(30, 34).

Mutation of L437 was shown to have strong influence on regio- and stereoselectivity of

-ionone hydroxylation (12) . This position was mutated in M11 to the negative

charged residue Glu (L437E) and to three polar residues Asn (L437N), Ser (L437S), and

Thr (L437T), which differ in size and hydrogen bonding capabilities.

5.2.4. Metabolism of TAK-715 by P450 BM3 mutants and human liver microsomes.

Incubations of TAK-715 were performed in 100 mM KPi buffer pH 7.4 with 250 nM

P450 BM3 mutants. The final volume of the incubation was 200 μL, with 100 M

substrate concentration. The reactions were initiated by addition of a NADPH

regenerating system with final concentrations of 0.2 mM NADPH, 20 mM glucose-6-

phosphate and 0.4 U/mL glucose-6-phosphate dehydrogenase. The reaction was

allowed to proceed for 60 min at 25°C and was terminated by the addition of 200 L of

cold methanol. For human liver microsomes (HLM), a final microsomal protein

concentration of 5 mg/mL was used and these incubations were performed as

described above, at 37°C instead of 25°C. Precipitated protein was removed by

centrifugation (15 min, 14000 g), and the supernatant was analyzed by high-resolution

UPLC using an Agilent 2000 system using a Zorbax Eclipse XDB-C18 column (1.8 m, 50

mm x 4.6 mm; Agilent, USA). The gradient used was constructed by mixing the


130

following mobile phases: eluent A (0.8 % methanol, 99 % water, and 0.2 % formic acid);

eluent B (99 % methanol, 0.8 % water, and 0.2 % formic acid) with a flow rate of 1

mL/min. The gradient was programmed as follows: from 0 to 8 minutes linear increase

of eluent B from 40% to 100%; from 8 to 9 minutes, isocratic 100% B; from 9 to 9.5

minutes linear decrease from 100% to 40% of eluent B; from 9 to 12 minutes isocratic

40% eluent B. Biotransformation products and substrate were detected at 254 nm.

5.2.5. HRS analysis of metabolites of TAK-715 using the LC-p38 kinase affinity/MS

platform

The HRS analysis of the biotransformation products is conducted with an LC-enzyme

binding detection/MS platform which has been described previously (20, 23). The

platform comprised a LC–MS system from Shimadzu (‘s Hertogenbosch, the

Netherlands), including two LC-20AD and two LC-10AD isocratic pumps, an SIL-20AC, a

CTO-20AC and a CTO-10AC column oven, an RF-10AXL fluorescence detector, an SPD-

AD UV\VIS detector, and a CBM-20A controller coupled to an ion-trap time-of-flight

(IT-TOF) hybrid mass spectrometer equipped with an ESI source. In short, the

biotransformation product mixtures were separated on a Xbridge C18 column

100×2.1 mm with 3.5 m particles (Waters, Milford, MA, USA) at 40°C and a flow rate

of 113 L/min. The gradient used was constructed by mixing the following mobile

phases: eluent A (1 % methanol, 99 % water, and 0.01 % formic acid); eluent B (99 %

methanol, 1 % water, and 0.01 % formic acid). The gradient was programmed as

follows: from 0 to 2 minutes isocratic at 20% B; from 2 to 18 minutes, linear increase of

eluent B from 20% to 90%; from 18 to 22 minutes, isocratic 90% B; from 22 to 23

minutes linear decrease from 90% to 20% of eluent B; from 23 to 30 minutes re-

equilibration isocratic at 20% eluent B. Post-column, the flow was split in a ratio of 1:9

directing 13 L/min to the bioaffinity detection and 100 L/min to UV/vis and ESI-HR-

MS analysis. The bioaffinity detection is based on competition of analytes with a tracer

showing fluorescence enhancement. Affinity towards p38α is assessed by mixing the

analytes with 45 nM of enzyme and 630 nM of tracer, SKF86002. The readout occurred

at excitation 355 nm and emission 425 nm, 15 nm bandwidth each, in the flow-through

fluorescence detector. The UV/vis detector was operated in dual wavelength mode at

210 nm and 254 nm.

For ESI the needle voltage was set to 4.5 kV and the source heating block and the

curved desolvation line were kept at 200°C. A drying gas pressure of 62 kPa and a

nebulizing gas flow-rate of 1.5 L/min assisted the ionization. Full spectra were obtained

in the positive-ion mode between m/z 200 and 650. MS2 and MS

3 spectra were

obtained in data-dependent mode between m/z 100 and 650 with an ion accumulation

time of 10 ms, a precursor isolation width of 3 Da and a collision energy of 75%. For

the analytes in the large scale incubation mixture MS2 and MS

3 spectra were generated


131

with manual precursor selection and data-dependent analysis, respectively. Structure

identification was based on calculating the elemental composition of parent and

fragment ions from the accurate mass measurements.

The IC50 determinations were done in flow-injection analysis (FIA) mode,

which means the separation column was replaced by a low dead volume union (VICI,

Schenkon, Switzerland). The lower dead volume reduces analysis time and minimizes

peak broadening compared to a column elution without retention. In these

experiments an injection volume of 50 µL was used to achieve adequate final

concentrations for full inhibition. Negative peak heights were plotted against the

corresponding final concentrations. The latter were calculated as described earlier

(20).

5.2.6. Large scale metabolite production and isolation of metabolites by preparative

HPLC

The metabolites of TAK-715 were produced on a semi-preparative scale by incubation

with the most active P450 BM3 mutant as biocatalyst. A 100 mL reaction volume

containing 1 M P450 BM3, 100 M TAK-715 and NADPH regenerating system (as

described above) was prepared in 100 mM KPi buffer at pH 7.4. The reaction was

allowed to continue for 5h at 25°C.

To achieve maximal conversion of TAK-715, the incubation was supplemented every

hour with 3 mL of 30 µM P450 BM3 M11 and 5 mL NADPH regenerating system (20 ×

concentrated). The reaction mixture was extracted three times by 100 mL ethyl

acetate. The combined organic layers were collected in a round-bottom flask and

evaporated to dryness using a rotary evaporator. The residue was redissolved in 10 mL

of 50% MeOH/H2O and applied by manual injection on a preparative chromatography

column Luna 5 m C18 (250 × 100 mm i.d.) from Phenomenex (Torrance, CA, USA)

which was previously equilibrated with 40% eluent B. Flow rate of 2 mL/min

and a gradient using the eluent A and B was applied for separation of formed TAK-715

biotransformation products. The gradient was programmed as follows: from 0 to 40

min linear increase of eluent B from 40 to 100%; from 40 to 50 min isocratic 100% B,

from 50 to 55 min linear decrease to 40% B, and then re-equilibration was maintained

until 65 min.

Biotransformation products were detected using UV detection at 254 nm and collected

manually. Collected fractions were first analyzed for purity and identity by the

analytical UPLC. Fractions containing individual metabolites were evaporated to

dryness under nitrogen stream and dissolved in 1 mL deuterium oxide to exchange

acidic hydrogen atoms by deuterium atoms. Finally, after drying by a SpeedVac

evaporator, the residues were redissolved in 500 L of methyl alcohol-d4 and 1H-NMR


132

spectra were recorded at room temperature. 1H-NMR-analysis was performed on

Bruker Avance 500 (Fallanden, Switzerland) at 500.23 MHz.

Afterwards, the samples were dried and redissolved in 30% MeOH and subjected to

LC-enzyme binding detection/MS analysis. LogP values were calculated with

ChemBioDraw Ultra version 12 from the structures obtained.

5.3. Results

5.3.1. Metabolism of TAK-715 by human liver microsomes and P450 BM3-mutants

As shown in Figure 1 in total 8 metabolites were found after incubation of TAK-715

with HLM, encoded M1 to M8 based on their order of elution.

After 60 minutes of incubation 33 + 2 % of the substrate was converted by HLM,

corresponding to 25.2 nmol product/min/CYP. The main biotransformation product

was M7, which represented 29 + 3% of the total metabolism, respectively. Metabolites

M1, M2, M5 and M6 were formed at comparable amounts by HLM, ranging from 14 to

20% of the metabolites, whereas the minor metabolites M3, M4 and M8 were formed

at 3 to 5% of total metabolism.

Figure 1. Ultraperformance liquid chromatography (UPLC) chromatograms obtained after incubations of TAK-715 (100 µM) incubated for 60 minutes with 200 nM P450 BM3 M11 (upper line) and 5 mg/mL human liver microsomes (lower line). UV-detection of metabolites was at 254 nm

TAK-715 was incubated for 60 minutes with pooled human liver microsomes (HLM)

and a library of 33 mutants of bacterial P450 BM3, and subsequently analyzed by

UPLC, equipped with diode array detector to determine UV absorption maxima of all


133

products formed. The results of these incubations, including the percentage of

conversion of substrate and profile of metabolites, are tabulated in Table 1.

As shown in Table 1, the P450 BM3-mutants were able to form almost all

human-relevant metabolites although with different activities and with strongly

different product ratios. Only metabolites M3 and M4 were not observed in the small

scale incubations of P450 BM3-mutants. The most active mutants were M11 and its

mutants at position 74 and 437. Similar to HLM, M11 produced metabolite M7 as

major biotransformation product, Figure 1, representing 42 + 3% of the total

metabolism. Metabolite M5 was produced as 22 + 2% of the total metabolism,

whereas metabolites M1, M2, M6 and M8 were formed at 13% or less of the total

metabolism. Mutation to a negatively charged amino acid (Asp or Glu) or hydrogen

bond donor (Thr) in positions 74 or 437 of M11 has little to no effect on activity but

changed the metabolic profile significantly by producing three-fold higher amounts of

metabolite M1, making this the major metabolite, Table 1. In contrast, mutant M11

V87I did not produce any metabolite M1, making it more selective to produce M7

(60% of total metabolism).

As shown in Table 1, for 22 out of the 33 P450 BM3-mutants studied,

metabolite M5 turned out to be the major metabolite. In particular the mutants

showing relatively low activity appeared to be highly selective in the production of this

metabolite. The only mutant which showed a significantly different profile was mutant

M11 L437N which was the only mutant which produced metabolite M6 as the major

metabolite, being 38% of the total of metabolites.

Based on its high activity and its ability to produce most of the metabolites at

significant amounts, mutant M11 was selected to further evaluate the affinity of the

TAK-715 metabolites towards p38-kinase in the on-line HRS-affinity assay and to

perform a large scale incubation of TAK-715 to obtain sufficient amounts of

metabolites for 1H-NMR and affinity assessments with individual metabolites.


134

Table 1. Percentage of conversion and ratio between the metabolites of TAK-715 incubations with human liver microsomes (HLM) and 33 mutants of P450 BM3

a

Enzyme %

Conversion M1 M2 M3 M4 M5 M6 M7 M8

m/z 432.138b m/z 446.121 b m/z 430.123 b m/z 432.138 b m/z 416.146 b m/z 416.146 b m/z 430.123 b m/z 416.146 b

HLM 33 19 14 5 3 17 10 28 4

M11 38 11 6 0 0 22 3 42 13

M11 L437E 36 35 9 0 0 23 15 16 2

M11 A74E 36 35 8 0 0 18 13 24 2

M11 A74D 33 34 7 0 0 20 14 24 3

M11 L437S 31 20 7 0 0 25 15 27 6

M11 L437T 30 33 5 0 0 23 15 20 3

M11 V87I 25 0 0 0 0 32 0 60 8

M01 L437E 20 18 0 0 0 42 26 9 5

M11 V87F L437S 20 21 0 0 0 44 20 9 6

M11 A82W 19 3 2 0 0 23 4 57 11

M05 19 6 0 0 0 34 6 40 14

M01 S72D 18 11 2 0 0 38 8 28 13

M11 V87F L437N 18 23 0 0 0 31 23 13 10

M01 A82W 18 5 0 0 0 38 5 48 4

M01 S72E 17 14 0 0 0 48 26 7 5

M01 A74D 17 13 0 0 0 48 26 7 6

M11 V87I L437N 16 4 0 0 0 47 6 28 15

M11 A82W V87F 15 0 0 0 0 78 0 8 14


135

(table 1 continued)

M11 V87F 15 0 0 0 0 82 5 6 7

M11 L437N 15 9 0 0 0 29 38 18 6

M11 S72D 12 12 6 0 0 52 17 6 7

M11 V87A 10 10 0 0 0 53 8 25 4

M11 V87I L437S 9 0 0 0 0 70 1 6 23

M01 7 0 0 0 0 60 14 6 20

M11 V87Y 7 0 0 0 0 76 6 6 12

M02 7 0 0 0 0 75 10 6 9

M11 V87L 6 0 0 0 0 79 5 3 13

M11 A82Y L437S 4 0 0 0 0 100 0 0 0

M11 A82Y V87I 4 0 0 0 0 85 15 0 0

M11 A82Y V87F 4 0 0 0 0 100 0 0 0

M11 V87Q 3 0 0 0 0 85 15 0 0

M01 A82C 3 0 0 0 0 100 0 0 0

M11 A82W L437S 0 0 0 0 100 0 0 0 0

a. All values represent averages of three individual experiments; RSDs were always less than 10%; b. [M+H]+-values as determined by

exact mass measurements by IT-TOF.


136

5.3.2. Parallel on-line bioaffinity assay and mass spectrometry of TAK-715

metabolites produced by HLM and P450 BM3 M11

Figure 2 shows the bioaffinity of metabolites of TAK-715 produced by HLM and the

extracted ion chromatograms representing the metabolites observed. Retention times

differ from those of the UPLC-chromatograms shown in Figure 1 because the HRS/MS

system used a conventional reversed phase HPLC system. The order of elution of the

metabolites was unchanged.

As shown in Figure 2, next to TAK-715 seven of its metabolites showed affinity

toward p38 kinase by showing varying degree of decrease of the fluorescence signal

of the tracer SKF86002. TAK-715 eluted at 39.5 minutes showed a strong bioaffinity, as

expected and ([M+H]+ of m/z 400.151. The second strongest decrease in fluorescence

was found at 33 minutes, which appeared to closely correspond to the retention times

of metabolites M6 (tret 32.7 min) and M7 (tret 33.0 min) in this chromatographic

systems. The lower resolution of the HRS-system did not allow attributing this

bioaffinity signal to any of these metabolites, however. Mass spectrometrical analysis

showed that m/z-values of the [M+H]+ of these metabolites corresponded to 416.146

and 430.123, respectively. M6 therefore corresponds to a monooxygenated

metabolite of TAK-715, whereas M7 appears to result from double oxygenation and

dehydrogenation. Metabolite M5 (tret 31.4 min) was a second metabolite with [M+H]+

of 416.146 and also showed a bioaffinity peak at the same retention time.

Metabolites M1 and M4 which eluted at 23.0 and 27.8 minutes, respectively,

both showed bioaffinity peaks and a [M+H]+ of 432.138, Figure 2, and appear to result

from double oxygenation. Metabolite M3 (tret 24.3 min) showed only a very weak

bioaffinity signal and a [M+H]+ of 432.138. Two of the metabolites formed, M2 (tret

24.3 min) and M8 (tret 34.5 min) did not show significant bioaffinity peaks and had

[M+H]+-values of 446.121 and 416.146, respectively.

Next to the eight metabolites previously found by UPLC, two additional

metabolites were identified based on small bioaffinity peaks at 8.4 and 17.4 minutes.

Metabolite designated M9 (tret 8.4 min) showed a [M+H]+

with m/z 312.119 which can

be rationalized by hydrolysis of the amide-bond of one of the monooxygenated TAK-

715 metabolites. The second biotransformation product M10 showed a [M+H]+ with

m/z 448.138 and apparently results from triple oxygenation of TAK-715. The bioaffinity

signal at 4.5 minutes appeared to be unrelated to TAK-715.

Figure 3 shows the p38 bioaffinity traces and extracted ion chromatograms

of the metabolic mixture obtained after incubation of TAK-715 with P450 BM3 mutant

M11. Three of bioactive peaks were corresponded to M1 (tret 23.0 min), M5 (tret 31.4

min) and TAK-715 (tret 39.5 min). Consistent with the results obtained with

metabolites of HLM, a strong bioaffinity peak was observed at the retention times of

metabolites M6 and M7.


137

Figure 2. Correlation between biochemical detection (p38 kinase affinity trace, upper trace) and extracted ion traces from the mass-

spectrometric detection (lower trace) of oxidative metabolites of TAK-715 produced in incubations with human liver microsomes (HLM).


138

Figure 3. Correlation between biochemical detection (p38 kinase affinity trace, upper trace) and extracted ion traces from the mass-spectrometric detection (lower trace) of oxidative metabolites of TAK-715 produced in incubations with P450 BM3 mutant M11.


139

5.3.3. Preparative scale biosynthesis of TAK-715 metabolites

To obtain sufficient amounts of TAK-715 metabolites for structural identification and

affinity testing of isolated metabolites to p38 kinase, a large volume (100 mL)

incubation was performed of TAK-715 with mutant P450 M11 since this appeared to

be the most active mutant and was able to produce multiple metabolites of TAK-715.

Figure 4 shows the UV-chromatogram (recorded at 254 nm) obtained after separation

of the metabolic mixture by preparative HPLC with UV detection at 254 nm. After 5

hours of incubation, with hourly additions of enzyme and cofactors, almost 90%

conversion of 100 M of TAK-715 was obtained.

Figure 4. Preparative HPLC-UV chromatogram of metabolic mixture obtained by large scale incubation of TAK-715 with P450 BM3 mutant M11. UV-Detection of metabolites was at 254 nm.

LC-MS/MS measurements were performed to confirm the identity of isolated

metabolites. The main biotransformation product M1 (35.7 min) formed represents

39% of the total metabolism, corresponding to 1.5 mg of product formed. Two

monohydroxylated biotransformation products (M5 and M6, 42.6 min and 43.5 min)

represent 27% and 12% of the total metabolism, respectively, corresponding to 1 mg

and 0.4 mg of product formed. Biotransformation product M7 (43.8 min) represents

15% of the total metabolism, corresponding to 0.6 mg of product formed.

Biotransformation product M2 with [M+H]+ with m/z 446.146 (37 min) was formed as


140

less than 5% of the total metabolism. Interestingly, also a small amount of metabolite

M4 was found in the large scale incubation. The relative amounts of metabolites

obtained by large scale incubation are slightly different than the ratio obtained in the

screening experiment because of the longer incubation time (5h vs 1h) and higher

enzyme/cofactor concentration. Also the less efficient aeration of large volume

incubations might contribute to the qualitative different metabolic profile.

5.3.4. Structure elucidation of TAK-715 metabolites

For five of the metabolites of TAK-715 enough material was obtained by preparative

HPLC to record a 1H-NMR spectrum, which together with the LC-MS/MS data allowed

full structure elucidation. By comparison with the spectra and fragmentation patterns

of TAK-715, it was possible to assign the structural modifications introduced by drug

metabolism for most of the metabolites.

Metabolites M5 and M6 both showed a [M+H]+ with m/z 416.146, indicative for

monooxygenation of TAK-715 at different positions. The integrals and coupling

pattern of the 1H-NMR spectrum of M5 showed that both all 12 aromatic hydrogen

atoms and the hydrogen atoms of the ethyl moiety were still present. The fact that the

singlet of the benzylic methyl-group was not observed anymore, whereas the chemical

shifts of the aromatic protons were shifted upfield indicates that M5 results from

hydroxylation of the benzylic methyl-group. In case of metabolite M6, the 1H-NMR-

spectrum indicates that a hydroxylation occurred at the methylene carbon of the

ethyl-side chain: the triplet of the terminal methyl-group in TAK-715 was changed to a

doublet, whereas the quartet of original methylene-group was shifted upfield and

showed half the integral. In case of M6 also loss of acetaldehyde was observed,

consistent with hydroxylation on the methylene-group of the ethyl moiety.

Metabolite M7, which was the major metabolite formed by HLM showed a

[M+H]+ with m/z 430.123 in LC-MS. The

1H-NMR spectrum of M7 still showed all

aromatic hydrogen atoms and an unchanged ethyl side-chain. Similar to metabolite M5

the signal of the benzylic methyl-group was not observed, whereas the signals of the

aromatic ring of the benzylic ring were strongly shifted upfield. This, in combination

with the LC-MS/MS data, indicates that the benzylic methyl-group is metabolized to a

carboxylate group which can be explained by further oxidation of M5 by two

successive oxidation reactions, Figure 5.

Metabolite M1, which according to its [M+H]+

with m/z 432.138 results from

two sequental oxygenation reactions, showed a 1H-NMR in which both the signals of

the benzylic methyl-group and the ethyl-side chain of TAK-715 were strongly affected,

suggestion that this metabolite results from hydroxylation of both alkyl side chains.

The coupling pattern and integrals of the signals of the ethyl-side chain show that

hydroxylation occurred at the methylene-group, consistent with the formation of M5.


141

The 1H-NMR spectrum of metabolite M2, which has [M+H]

+ with m/z 446.138,

also showed strong changes in the signals of the benzylic methyl-group and the ethyl-

side chain. The signals of the ethyl-side chain point to hydroxylation at the methylene-

position, similar to M1 and M6, whereas the signals of the benzylic group showed the

same changes as observed in M7, indicative for formation of a carboxylic acid by

sequential oxidations, Figure 5.

For metabolites M3, M4, M8, M9 and M10 no 1H-NMR data could be obtained since

no P450 BM3-mutant was available that can produce these metabolites in sufficient

amounts. Structure elucidation for these metabolites was therefore based on the

fragmentation patterns of the LC-MS/MS measurements. Metabolite M3 has a [M+H]+

with m/z 430.125 and can be rationalized by two sequential hydroxylation reactions,

followed by an oxidative dehydrogenation reaction. Inspection of the MSn data

revealed that M3 showed the same fragmentation as M2 and M7: in all three cases the

[M+H]+ ion showed sequental neutral losses of masses 122.121 (-C7H6O2) and 27.94 (-

CO). The first neutral loss might be explained by a combination of a loss of

benzaldehyde (C7H4O) and water, and might point to the presence of an aromatic

aldehyde moiety in M3. The loss of acrylonitrile after fragmentation of 280.083 might

indicate that the loss of water results from the hydroxylated ethyl-group.

Metabolite M4 has a [M+H]+ with m/z 432.141 which can be explained by two

oxygenation reactions. Upon secondary fragmentation of this ion both a direct loss of

C2H4O2 as well as two sequential losses of formaldehyde were observed which might

point to hydroxylation of both carbon atoms of the ethyl-side chain.

For metabolites M8 and M10 which result from single and triple oxygenation reactions

no fragmentations were found which could assist in structure elucidation.


142

Figure 5. Proposed structures of TAK-715 biotransformation products by human liver microsomes (HLM) and P450 BM3 mutants.

N

O

H

N

N

SCH2

CH3

H3C

N

O

H

N

N

SCH2

CH3

CH2HO

N

O

H

N

N

SCH2

CH3

CHO

N

O

H

N

N

SCH2

CH3

HOOC

N

O

H

N

N

SHC

CH3

H3C

OH

N

O

H

N

N

SHC

CH3

HOOC

OH

N

O

H

N

N

SHC

CH3

CH2

OH

HO

M5

M6 M1

M7

M2

N

O

H

N

N

SHC

CH3

CH

OH

O

TAK-715

M3

N

O

H

N

N

SHC

H2C

H3C

OH

M4

OH


143

5.3.5. Quantitative affinity measurements (IC50)

Metabolites M1, M2, M5, M6 and M7 were isolated by preparative HPLC. After

checking their purity by 1H-NMR and analytical HPLC, the concentration of each stock

solution was determined by using a standard curve of TAK-715, assuming that the

extinction coefficients of each metabolite was equivalent to that of TAK-715.

Subsequently, different concentrations of TAK-715 and TAK-715 metabolites were

analyzed using the on-line HRS system in flow-injection mode in order to assess the

affinity of each compound towards p38-kinase.

Figure 6 shows the concentration-dependence of binding of the individual metabolites

and TAK-715 to p38-kinase, after correction for the dilutions intrinsic to the HRS-

system used. Calculated IC50-values of TAK-715 and its metabolites are presented in

Table 2.

Figure 6. Concentration dependent binding of TAK-715 and purified metabolites M1,

M5, M6 and M7 to p38 kinase as determined by flow-injection analysis. Percentage of binding was calculated by the concentration dependent displacement of the tracer

SKF86002 from p38 kinase. Error bars represent standard deviations of at least three independent experiments.


144

Table 2. Product identification TAK-715 metabolites

treta

(min) code m/z [M+H]+ Structural

modification p38 kinase affinity (nM)

IC50 95% Conf.interval

23 M1 432.138 +2O 460 ± 165 240 to 890 24.3 M2 446.121 +3O-2H No affinity 25.2 M3 430.125 +2O -2H + (n.q.) 27.8 M4 432.141 +2O + (n.q.) 31.4 M5 416.146 +O 310 ± 33 250 to 380 32.7 M6 416.146 +O 38 ± 4 31 to 47

33 M7 430.123 +2O -2H 1500 ± 203 1100 to 1900 34.5 M8 416.146 +O No affinity 8.4 M9 312.119 - C7H4O + (n.q.) 17.4 M10 448.138 +3O + (n.q.) 39.4 TAK-715 400.151 71 ± 26 36 to 140 a

tret = retention time in the HRS analysis (Figures 2 and 3).

n.q., not quantified.

For both TAK-715 and metabolite M1 not a full binding curve could be

obtained. In case of TAK-715, peak tailing due to aspecific binding of this highly

lipophilic compound to materials of the HRS-system, deformed the Gaussian shape so

severely, at concentrations higher than 60 nM, that it was not possible to calculate the

local concentration in the assay. Also, the baseline for the following injection could

not be recovered in a reasonable amount of time.

In the case of M1, (auto)fluorescence of the metabolite at the assay

wavelengths appeared to interfere with the p38 kinase affinity measurement at high

concentrations. Due to the different concentration-dependent behavior of both

contributions, e.g. increase of M1-(auto)fluorescence and decrease of fluorescence

enhancement of the tracer by displacement at increasing M1-concentration, a W-

shaped peak was obtained whose area no longer reflects the fraction of binding.

Although incomplete binding curves were obtained for TAK-715 and M1, their IC50-

values and 95% confidence intervals were estimated by fitting the datapoints by the

same fitting procedure as was used for the other compounds, which showed complete

binding curves. The IC50 obtained for TAK-715, 71 nM is consistent with the value

observed previously with this platform (20) .

As shown in Figure 6, for three of the metabolites, M1, M5 and M7 the

binding curves obtained were shifted to higher concentrations, when compared to that

of TAK-715. The largest shift was observed with metabolite M7. This metabolite

showed an IC50-value of 1500 ± 203 nM which is a 20-fold lower affinity to p38-kinase

compared to TAK-715. Metabolites M1 and M5 showed 5 to 7-fold decreased affinity


145

according to their IC50-values of 460 ± 165 nM and 310 ± 33 nM, respectively.

Interestingly, metabolite M6 showed high affinity to p38kinase according to its IC50-

value of 38 ± 4 nM. However, due to the large standard error of the IC50-value of TAK-

715, it could not be determined whether the increased affinity of M6 is statistically

significant. For metabolite M2 no binding curve was obtained, indicating that this

metabolite completely lacked affinity to p38-kinase.

5.4. Discussion

The aim of this study was to demonstrate the applicability of P450 BM3 mutants to

produce significant amounts of metabolites of drugs and to characterize their affinity

to p38α kinase. TAK-715 was selected as a model compound for the present study. To

determine the affinity of metabolites of TAK-715 to p38α kinase, a recently developed

HRS platform was used which was capable of assessing this property for individual

compounds in a mixture (23). The approach to use P450 BM3-mutants as biocatalyst

for metabolite production in combination with a HRS affinity screening platform to

identify active metabolites has been used previously by de Vlieger et al. (35) and

Reinen et al. (36) with estrogen receptors as drug targets. De Vlieger et al. applied

BM3 mutants for the generation of biotransformation products of the estrogenic

compound norethisterone. The metabolic mixture obtained was analyzed by

hyphenated screening assay towards the human estrogen receptor ligand binding

domain (hER) and (hER) integrating target ligand interaction and LC-MS detection.

(35). Reinen et al. applied P450 BM3 mutants to generate in vitro human relevant

biotransformation products of the mycotoxin zearalenone (36). The hydroxylated

metabolites of zearalenone formed did not display an increased ER bioaffinity.

The results of the present study show that the selected mutants of bacterial P450 BM3

were able to produce significant amounts of metabolites of the p38-kinase inhibitor

TAK-715 allowing both structural characterization by LC-MS/MS and 1H-NMR and

characterization of affinity of metabolites towards p38-kinase. By comparison with

metabolites formed by HLM it was found that the P450 BM3 only produced human

relevant metabolites of TAK-715. Figure 5 shows the metabolic scheme which

rationalizes the formation of the metabolites which were identified in the present

study. Metabolism of TAK-715 appeared to occur almost exclusively at its alkyl side

chains. Metabolites M5 and M7 result from sequential oxygenation reactions on the

benzylic methyl-group of TAK-715. The intermediate aldehyde which was expected

could not be detected, suggesting that this aldehyde undergoes extensive further

oxygenation to M7 and to lesser extent to M3. Although conversion of the alcohol-

groups to carboxylates is often catalyzed by sequential alcohol dehydrogenases,

aldehyde dehydrogenase and oxidase reactions, conversion of M5 to M7 appears to be

fully catalyzed by P450 BM3, since this reaction was also observed when catalyzed by


146

purified P450 BM3 M11 [data not shown]. So far only few examples have been

described on the multistep oxidation of methyl- to carboxylic acid groups by P450

enzymes. Oxidation of a methyl of the t-butyl moiety of terfenatine and ebastine to

their corresponding carboxylic acids fexofenadine and carebastine in HLM was

previously shown to be catalyzed by CYP3A4 and CYP2J2 (37). More recently as yet

unidentified microbial P450s from Streptomyces platensis were also shown to be

capable to produce fenofexadine from terfenatine (38). Similar to our P450 BM3

incubations no intermediate aldehyde was found, indicating that the conversion of

aldehydes to acids by P450s is highly efficient. The availability of biocatalysts for

conversion of methyl groups to carboxylic acids is considered of high value because

chemical synthesis is laborious, and because incorporation of carboxylic acids in drug

molecules might be benificial for water solubility, might decrease the risk of hERG

inhibition (38) and might restrict passage of blood-brain-barrier of drugs.

When comparing the metabolic profiles of the different P450 BM3 mutants,

Table 1, significant differences were observed between the mutants. Considering the

fact that M5 and M7 both only involve modification of the benzylic methyl-group,

whereas M6 and M4 only consider modification of the ethyl-side chain, it can be

concluded that oxygenation of the benzylic methyl-group is by far the major pathway

for all mutants studied. The sum of M5 and M7 represents from 50 to 100% of the

metabolites, dependent on the mutant. Because metabolites M1, M2 and M3 might

be secondary oxidation products of M5 and M7, the total contribution of this pathway

might even be underestimated. Because the metabolic profiles appears to result from

extensive secondary oxidation of the primary metabolites, and the fact that several

metabolites can be formed by different pathways, it was not attempted to rationalize

the effects of the mutations on the metabolic profile.

Changes in profile might result from both changes in substrate orientation and/or

changes in enzyme activity for each metabolic step.

For example, the most selective mutants which almost specifically produce primary

metabolite M5, showed only very low enzyme activity, explaining why no secondary

metabolism was observed. Only for mutants M11 V87I and M11 V87F the change in

metabolic profile might be attributed to changes in substrate orientation, because this

is the only mutant which did not show any oxygenation of the ethyl-side chain: the

benzylic methyl-oxygenation appeared to represent 90% of metabolism. Both mutants

produced a small amount of metabolite M8 which, based on its mass, might result

from N-oxygenation or aromatic hydroxylation. Because this minor metabolite did not

show affinity to p38-kinase no attemps were undertaken to identify the structure.

Mutant M11 L437N showed the highest amount of metabolite M6, and relative low

levels of M5 an M7, which suggests that mutation L437N causes a shift from benzylic

methyl-hydroxylation to hydroxylation of the ethyl-side chain.


147

By using the recently developed hyphenated HRS-system for parallel bioaffinity testing

and LC-MS analysis of complex mixtures (20, 23), we were able to quickly determine to

presence of ligands for p38-kinase in metabolic mixtures of TAK-715, Figures 2 and 3.

By isolating the corresponding compounds by preparative HPLC and subsequent

analysis by the HRS in flow-injection mode (e.g. in absence of analytical column) the

IC50-values of several metabolites could be determined, Table 3. The results of these

studies indicate that oxygenation of the benzylic methyl-group appears to be

detrimental for affinity to p38-kinase. The fact that M7 shows a very low affinity,

whereas M2 shows no affinity at all, indicate that the carboxylic acid most strongly

disturbs binding to p38-kinase. These results are consistent with these results of

Miwatashi et al. which showed that modifying the benzylic methyl-group of TAK-715

by other substituents, or by positioning the methyl-group to the ortho- or para-

position, in all cases led to a strong loss in affinity towards p38-kinase (25).

Apparently, the benzylic methyl-group at the meta-position of TAK-715 is essential for

high affinity. These is further confirmed by the high affinity of M6 in which only the

ethyl-side chain of TAK-715 is hydroxylated. Again this seems consistent with the

study of Miwatashi et al. (26) which showed that modification of the thiazole-2-ethyl

moeity of TAK-715 to other alkyl-substituents (methyl, propyl) or a large substituent

such as 4-methylsulfonylphenyl only showed little effect on p38-kinase.

In conclusion, the results of the present study shows that the combination of

a catalytically diverse set of P450 BM3 mutants as a toolbox to diversify drugs and a

HRS-system capable to rapidly screen for affinity to p38-kinase might be a promising

platform to generate potential novel lead compounds which might have improved

physico-chemical properties (by improved water solubility) and desired

pharmacological properties.

As exemplified by TAK-715 as model compound, the P450 BM3 mutants were

able to produce almost all metabolites produced by human liver microsomes at

significant levels. Furthermore, profiling the effect of structural modifications

introduced by P450 on p38-kinase affinity gives valuable information on which

structural elements are critical and which are non-critical for affinity. This might be a

very potent addition to contruction of pharmacophore models, next to the more

conventional medicinal chemistry approach of synthesizing large series of structural

analogues of drug candidates.

Finally, our results should confirm that biocatalysis with CYPs can provide

compounds with improved pharmacological and/or physicochemical properties

compared to the lead itself which are not easily accessible by classic organic synthesis.

This approach can be considered as an extra tool, next to the chemical synthesis in the

typical medicinal chemistry approach.


148

References

1. Wienkers, L. C., and Heath, T. G. (2005) Predicting in vivo drug interactions from in vitro drug discovery data, Nature reviews. Drug discovery 4, 825-833.


3. Clouthier, C. M., and Pelletier, J. N. (2012) Expanding the organic toolbox: a guide to integrating biocatalysis in synthesis, Chemical Society reviews 41, 1585-1605.

4. Hollmann, F., Arends, I. W. C. E., Buehler, K., Schallmey, A., and Buhler, B. (2011) Enzyme-mediated oxidations for the chemist, Green Chemistry 13, 226-265.


6. Guengerich, F. P. (2002) Cytochrome P450 enzymes in the generation of commercial products, Nature reviews. Drug discovery 1, 359-366.

7. Kumar, S. (2010) Engineering cytochrome P450 biocatalysts for biotechnology, medicine and bioremediation, Expert opinion on drug metabolism & toxicology 6, 115-131.

8. Munro, A. W., Leys, D. G., 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 in biochemical sciences 27, 250-257.


10. 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.

11. 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 : a European journal of chemical biology 13, 520-523.

12. Venkataraman, H., Beer, S. B. A. d., Geerke, D. P., Vermeulen, N. P. E., and Commandeur, J. N. M. (2012) Regio- and Stereoselective Hydroxylation of Optically Active α-Ionone Enantiomers by Engineered Cytochrome P450 BM3 Mutants, Advanced Synthesis & Catalysis 354, 2172-2184.



149

14. van Vugt-Lussenburg, B. M., Stjernschantz, E., Lastdrager, J., Oostenbrink, C., Vermeulen, N. P., and Commandeur, J. N. (2007) Identification of critical residues in novel drug metabolizing mutants of cytochrome P450 BM3 using random mutagenesis, Journal of medicinal chemistry 50, 455-461.

15. Reinen, J., van Leeuwen, J. S., Li, Y., Sun, L., Grootenhuis, P. D., Decker, C. J., Saunders, J., Vermeulen, N. P., and Commandeur, J. N. (2011) Efficient screening of cytochrome P450 BM3 mutants for their metabolic activity and diversity toward a wide set of drug-like molecules in chemical space, Drug metabolism and disposition: the biological fate of chemicals 39, 1568-1576.


17. Damsten, M. C., de Vlieger, J. S., Niessen, W. M., Irth, H., Vermeulen, N. P., and Commandeur, J. N. (2008) Trimethoprim: novel reactive intermediates and bioactivation pathways by cytochrome p450s, Chemical research in toxicology 21, 2181-2187.

18. Yun, C. H., Kim, K. H., Kim, D. H., Jung, H. C., and Pan, J. G. (2007) The bacterial P450 BM3: a prototype for a biocatalyst with human P450 activities, Trends in biotechnology 25, 289-298.

19. Kool, J., Giera, M., Irth, H., and Niessen, W. M. (2011) Advances in mass spectrometry-based post-column bioaffinity profiling of mixtures, Analytical and bioanalytical chemistry 399, 2655-2668.


21. Cohen, P. (2002) Protein kinases--the major drug targets of the twenty-first century?, Nature reviews. Drug discovery 1, 309-315.

22. Dorn, A., Schattel, V., and Laufer, S. (2010) Design, synthesis and SAR of phenylamino-substituted 5,11-dihydro-dibenzo[a,d]cyclohepten-10-ones and 11H-dibenzo[b,f]oxepin-10-ones as p38 MAP kinase inhibitors, Bioorganic & medicinal chemistry letters 20, 3074-3077.

23. Falck, D., de Vlieger, J. S., Giera, M., Honing, M., Irth, H., Niessen, W. M., and Kool, J. (2012) On-line electrochemistry-bioaffinity screening with parallel HR-LC-MS for the generation and characterization of modified p38alpha kinase inhibitors, Analytical and bioanalytical chemistry 403, 367-375.

24. Fasan, R. (2012) Tuning P450 Enzymes as Oxidation Catalysts, ACS Catalysis 2, 647-666.

25. Miwatashi, S., Arikawa, Y., Naruo, K., Igaki, K., Watanabe, Y., Kimura, H., Kawamoto, T., and Ohkawa, S. (2005) Synthesis and biological activities of 4-phenyl-5-pyridyl-1,3-thiazole derivatives as p38 MAP kinase inhibitors, Chemical & pharmaceutical bulletin 53, 410-418.


150

26. Miwatashi, S., Arikawa, Y., Kotani, E., Miyamoto, M., Naruo, K., Kimura, H., Tanaka, T., Asahi, S., and Ohkawa, S. (2005) Novel inhibitor of p38 MAP kinase as an anti-TNF-alpha drug: discovery of N-[4-[2-ethyl-4-(3-methylphenyl)-1,3-thiazol-5-yl]-2-pyridyl]benzamide (TAK-715) as a potent and orally active anti-rheumatoid arthritis agent, Journal of medicinal chemistry 48, 5966-5979.

27. Ozdemir, C., and Akdis, C. A. (2007) Discontinued drugs in 2006: pulmonary-allergy, dermatological, gastrointestinal and arthritis drugs, Expert opinion on investigational drugs 16, 1327-1344.

28. Verkaar, F., van der Doelen, A. A., Smits, J. F., Blankesteijn, W. M., and Zaman, G. J. (2011) Inhibition of Wnt/beta-catenin signaling by p38 MAP kinase inhibitors is explained by cross-reactivity with casein kinase Idelta/varepsilon, Chemistry & biology 18, 485-494.

29. Barker, N., and Clevers, H. (2006) Mining the Wnt pathway for cancer therapeutics, Nature reviews. Drug discovery 5, 997-1014.

30. Vottero, E., Rea, V., Lastdrager, J., Honing, M., Vermeulen, N. P., and Commandeur, J. N. (2011) Role of residue 87 in substrate selectivity and regioselectivity of drug-metabolizing cytochrome P450 CYP102A1 M11, Journal of biological inorganic chemistry : JBIC : a publication of the Society of Biological Inorganic Chemistry 16, 899-912.

31. Stjernschantz, E., van Vugt-Lussenburg, B. M. A., Bonifacio, A., de Beer, S. B. A., van der Zwan, G., Gooijer, C., Commandeur, J. N. M., Vermeulen, N. P. E., and Oostenbrink, C. (2008) Structural rationalization of novel drug metabolizing mutants of cytochrome P450 BM3, Proteins: Structure, Function, and Bioinformatics 71, 336-352.

32. 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, Journal of biotechnology 139, 115-117.

33. Otey, C. R., Bandara, G., Lalonde, J., Takahashi, K., and Arnold, F. H. (2006) Preparation of human metabolites of propranolol using laboratory-evolved bacterial cytochromes P450, Biotechnology and bioengineering 93, 494-499.

34. Rea, V., Dragovic, S., Boerma, J. S., de Kanter, F. J., Vermeulen, N. P., and Commandeur, J. N. (2011) Role of residue 87 in the activity and regioselectivity of clozapine metabolism by drug-metabolizing CYP102A1 M11H: application for structural characterization of clozapine GSH conjugates, Drug metabolism and disposition: the biological fate of chemicals 39, 2411-2420.

35. de Vlieger, J. S., Kolkman, A. J., Ampt, K. A., Commandeur, J. N., Vermeulen, N. P., Kool, J., Wijmenga, S. S., Niessen, W. M., Irth, H., and Honing, M. (2010) Determination and identification of estrogenic compounds generated with biosynthetic enzymes using hyphenated screening assays, high resolution mass spectrometry and off-line NMR, Journal of chromatography. B, Analytical technologies in the biomedical and life sciences 878, 667-674.


151


37. Ling, K. H., Leeson, G. A., Burmaster, S. D., Hook, R. H., Reith, M. K., and Cheng, L. K. (1995) Metabolism of terfenadine associated with CYP3A(4) activity in human hepatic microsomes, Drug metabolism and disposition: the biological fate of chemicals 23, 631-636.

38. Zhu, B. Y., Jia, Z. J., Zhang, P., Su, T., Huang, W., Goldman, E., Tumas, D., Kadambi, V., Eddy, P., Sinha, U., Scarborough, R. M., and Song, Y. (2006) Inhibitory effect of carboxylic acid group on hERG binding, Bioorganic & medicinal chemistry letters 16, 5507-5512.

152

CHAPTER 6

SUMMARY, CONCLUSIONS AND PERSPECTIVES NEDERLANDSE SAMENVATTING

Chapter 6 Summary, conclusions and perspectives

153

6. Summary

Several drugs are biotransformed to active metabolites that can significantly contribute to

their overall pharmacological or adverse effects. Tracking active metabolites is not only

important to correctly interpret the pharmacological and/or adverse effects in preclinical

studies but may also be used as a promising tool to identify potentially new drug

candidates for drug discovery and development (1).

In drug discovery and development it is important to have information on drug

metabolism as early as possible. Knowledge of the metabolic pathways, metabolic

stability, toxicity and the specific enzymes involved in the metabolism are all important

information in the drug discovery and development process but also in planning human

clinical studies.

In chapter 1 the evolving role of drug metabolism in drug discovery and

development is discussed. Because of the impact of biotransformation reactions on the

fate and the effects of drugs, the preparative synthesis of drug metabolites is currently of

primary importance in industry in order to assess potential pharmacological activities,

toxicity, drug-drug interactions and to examine metabolic pathways (2).

Full pharmacological assessment of biotransformation products usually requires organic

synthesis to obtain sufficient amounts of pure compound. As an alternative to organic

synthesis, biosynthetical approaches are upcoming methodologies. Biocatalysts (i.e.

enzymes) can perform reactions with a high degree of regio- and stereoselectivity that

often cannot be achieved by organic synthesis (3).

Bacterial Cytochrome P450 BM3 appears as a powerful and versatile tool for the

generation of high amounts of human relevant drug metabolites that can be isolated,

identified and tested for toxicity and pharmacological activity.

By rational re-design and directed evolution this enzyme can be engineered to catalyze

reactions that are mimicking human P450s or to acquire completely novel catalytic

properties (shown in this thesis).

In recent years, P450 BM3 has emerged not only as an promising tool for the generation

of drug metabolites and commercial products but also as an excellent model system to

study general mechanistic aspects of P450 chemistry.

The main aim of this thesis is to contribute to the development of a metabolic

production and profiling platform encompassing the application of P450 BM3 mutants for

the generation of physiologically relevant drug metabolites and the application of

screening methods for their identification and pharmacological and toxicological

characterization. Moreover it comprises mechanistic insights into P450 catalysis and

chemistry and the geometry of the active site.

The general strategy applied (Introduction, Figure 11) comprises genetic engineering of

P450 BM3 mutants by different mutation strategies (site-directed, site-saturation, random

mutagenesis), screening of mutant libraries (UPLC, fluorescent assay, LC-MS, GSH-


154

trapping, HRS), upscaling and isolation of metabolites by Prep-LC, and structural

elucidation of isolated metabolites by 1H-NMR.

One of the BM3 mutants which had been developed previously in our Molecular

Toxicology group and which showed high activity in drug metabolism is BM3 M11,

containing ten different amino acid substitutions compared to wild-type BM3. This BM3

M11 mutant was shown to be highly active in metabolizing a variety of drugs (e.g.

clozapine, testosterone, MDMA, dextromethorphan, diclofenac) to human relevant

metabolites (4-6), including reactive intermediates (7-9).

In the present thesis, we have performed a saturation mutagenesis study in which the

active-site residue at position 87 in BM3 M11 was mutated to all 20 possible amino acids.

It was demonstrated that the type of amino acid at this position has strong effect on

substrate selectivity when comparing a series of alkoxyresorufins (chapter 2), on the

activity and regioselectivity of testosterone hydroxylation (chapter 2) and on the activity

and regioselectivity of clozapine bioactivation (chapter 3). Twelve of the amino acid

substitutions were not yet been reported previously in any BM3 variant.

In chapter 2, a series of nine alkoxy-substituted substrates (methoxyresorufin to

n-octoxyresorufin, and benzyloxyresorufin) were tested as diagnostic substrates. It was

shown that mutation at position 87 dramaticaly affected not only the substrate selectivity

but also the coupling efficiency of the enzyme. Interestingly, the coupling efficiency with

these substrates was always less than 1% for all productive enzymes suggesting that

alkoxyresorufins bind to BM3 at the active site mainly in a nonproductive orientation.

Uncoupling of P450 is still a poorly understood process. Because the BM3 mutants contain

amino acids at position 87 with different polarities and size, different modes of uncoupling

might underly the high NADPH-consumptions observed. The mechanism by which

alkoxyresorufine stimulated extremely high NADPH-consumption in the position 87

mutants and wild-type P450 BM3 therefore still remains to be elucidated.

Testosterone was hydroxylated by the library of twenty mutants at position 87, at three

different positions, as was shown previously in incubations with the triple mutant of P450

BM3, containing mutations R47L, F87V and L188Q (5).

Structural identification of the metabolites by NMR revealed that two of the metabolites

result from hydroxylation of the D-ring, at positions 15ß and 16ß; the third metabolite

results from hydroxylation of the A-ring at position 2ß. With the triple mutant very poor

regioselectivity was observed. With this library of mutants instead, big changes in

metabolic profile were observed. For example, the mutant containing isoleucine at

position 87 catalyzed 16ß-hydroxylation with very high selectivity whereas in case of the

closely related leucine aminoacid testosterone hydroxylation was taking place

predominantly at the position 2ß. Mutant M11 V87I is the first bacterial P450 able to


155

selectively hydroxylate testosterone at position 16ß. Why these relatively small changes in

amino acid side chain have such a large effect on regioselectivity remains to be

established.

In chapter 3 the site-saturation library of BM3 mutants at position 87 presented

in chapter 2 was applied for the generation of reactive metabolites of clozapine. While in

chapter 2 the library was screened only with non-charged substrates, in chapter 3 a

positively charged molecule, i.e. clozapine, was evaluated. Clozapine is known to be

involved in severe ADRs due to the formation of a reactive metabolite (10). Often reactive

metabolites cannot be synthetised by organic chemistry, thus complicating the availability

of reference compounds and standards. Also, the amounts of reactive metabolites formed

by human P450s are generally too low to allow purification and NMR identification. In

chapter 3, it was investigated whether BM3-mutants at position 87 could be applied for

the generation of reactive metabolites in quantities allowing chemical characterization.

Results showed that the nature of aminoacid at position 87 strongly determines both

activity and regioselectivity of clozapine metabolism. The mutant containing Phe87

showed both high activity and high selectivity for the bioactivation pathway and was used

for the large scale production of GST-dependent GSH conjugates by incubation in presence

of glutathions S-transferase P1-1. Five human relevant GSH conjugates were produced in

high amounts enabling structural characterization by 1H-NMR. This results demonstrated

the applicability of P450 BM3 mutants for the generation of human relevant reactive

metabolites in sufficient amount to allow their structural elucidation by NMR. Detection

and structural elucidation of reactive metabolites early in drug development is very critical

for the development of safer drugs.

In chapter 4, site-directed mutagenesis was applied to improve the

regioselectivity of steroid hydroxylation by BM3 mutants. The strategy applied

encompassed the restriction of the actve site size by mutating Ala82 with a Trp in order to

reduce the substrate mobility, therefore improving the regioselectivity of steroid

hydroxylation. The mutation A82W led to a < 42-fold increase in Vmax for 16-

hydroxylation of testosterone and norethisterone, and improved the coupling efficiency of

the enzyme by a more efficient exclusion of water from the active site. Spin relaxation

NMR was applied to rationalize the change in metabolic profile observed, showing that

the mutation caused a change in the orientation of testosterone in M11 A82W as

compared to the orientation in M11.

In chapter 5, mutants of P450 BM3 were used to support drug development by

producing human relevant drug metabolites analyzed for identity and bioaffinity

assessment by the analytical high-resolution screening (HRS). A panel of BM3 mutants was

applied for the generation of metabolic mixtures of TAK-715, a known p38 inhibitor. HRS

screening allowed the identification and bioaffinity determination of all the metabolites


156

produced. The high turnover rates of BM3 M11 and the convenient large scale production

and purification protocols for this enzyme allowed semi-preparative production of the

most abundant active metabolites that have been identified by NMR, while HRS allowed

the determination of their IC50. These results showed that the combination of a

catalytically diverse set of P450 BM3 mutants as a toolbox to diversify drugs and a HRS-

system capable to rapidly screen for affinity to p38-kinase might be a promising platform

to generate potential novel lead compounds which might have improved physico-chemical

properties (by improved water solubility) and desired pharmacological properties.

6.1. Conclusions and perspectives

In this research, different mutagenesis techniques have been applied for the development

of P450 BM3 mutants that are able to metabolize drug molecules to human relevant

metabolites and that could be used as biocatalysts in drug discovery and synthesis. P450

mutants able to produce high amounts of drug metabolites are very useful biocatalysts to

generate metabolites from novel drug candidates for structural, toxicological and

pharmacological characterization. Also they can be used for the biosynthesis of known

pharmacologically active compounds. In drug discovery, a library of diverse mutant P450s

could be used to functionalize lead compounds in order to identify potential novel drugs

and drug candidates.

When the research described in this thesis started, in 2008, the so-called

Metabolites in Safety Testing (MIST) guidelines were published by the FDA, leading several

research groups to focus their efforts on the development of new biocatalysts for the

generation of large amounts of human relevant drug metabolites (11).

At that time the work of Van Vugt-Lussenburg et al. (4, 5, 12, 13) already had shown that

by site-directed mutagenesis of specific amino acids in the active site of wild-type P450

BM3, the substrate spectrum could be expanded to accept drugs and drug-like molecules.

Damsten et al. (8) showed that P450 BM3 mutants could also be applied for the

generation of reactive drug metabolites, however structural elucidation of these

metabolites was still hampered by the small amounts obtained. BM3 mutants that were

able to hydroxylate steroids were already developed (5), however with very poor regio-

and steroselectivity.

The platform presented in the introduction of this thesis (Chapter 1, Figure 11) has been

applied for the regioselective steroid hydroxylation (chapter 2 and 4), for the generation

of potentially toxic reactive metabolites (chapter 3), for the generation of bioactive drug

metabolites (chapter 5).

6.1.1. Regioselective hydroxylation of steroids by P450 BM3 mutants

The present thesis has shown that one of the current challenges in synthetic organic

chemistry, namely the control of regio- and stereoselective oxidation of unactivated C-H


157

bond of complex organic compounds (14-16) can be met by engineering P450 BM3

enzymes.

In chapter 2 and chapter 4, it was shown that by single mutation in the active site

of P450 BM3 (position 87 and position 82) dramatic changes in regio- and stereoselectivity

of steroid hydroxylation could be obtained. Spin relaxation NMR was used as powerful

tool to determine changes in orientation of testosterone in the active site of BM3 mutants

showing different regioselectivity. The regio- and stereoselectivity of P450-mediated

reactions depends upon the orientation of the substrate relative to the reactive iron-oxo

species, which is in turn determined by the active-site configuration of the P450 enzyme

(17). Small variations in the active site of P450s could alter or improve their substrate

scope, regio- or stereoselectivity, activity and coupling efficiency (18). Interestingly,

recently Venkatamaran et al. showed that a single active site mutation S72I in M01 A82W

and in M11 V87I inverted the stereoselectivity of hydroxylation from 16 β to 16 α (19).

Structure guided redesign of the active site can be used to manipulate the regioselectivity

of P450 enzymes to obtain mutants able to target specific position in the steroid molecule,

therefore creating novel hydroxysteroids. Spin-relaxation NMR, docking and molecular

dynamics can be used as powerful tools to shed light on the origin of regio- and

stereoselectivity.

6.1.2. Biosynthesis of reactive metabolites by P450 BM3 mutants

Measuring the potential bioactivation of drugs and drug candidates leading to chemically

reactive metabolites early in the drug discovery phase is often hampered by the

difficulties in detecting and characterizing low levels of RIs (20). When this research

started, Damsten et al. already showed that BM3 mutants were able to produce reactive

metabolites from the drugs clozapine, diclofenac and acetominophen (8). However, the

amounts obtained were still not sufficient for isolation and structural elucidation by NMR.

Recently, Dragovic et al. (9) identified novel human relevant GST-dependent GSH-

conjugates for which unequivocal structural elucidation by NMR was still missing. Boerma

et al. showed that P450 BM3 mutants with high capacity to activate drugs (clozapine,

acetominophen and troglitazone) into relevant reactive metabolites can be employed to

produce protein adducts to study the nucleophilic selectivity of highly reactive

electrophiles (21).

In chapter 3, by using BM3 mutant M11V87F, we were able to produce significant

amounts of all major human relevant GSH conjugates of clozapine, for which the

structures were not yet elucidated unequivocally by 1H-NMR.

This study confirmed the high potential of BM3 mutants as tool to assist the identification

of reactive metabolites of drugs, the elucidation of novel bioactivation pathways and to

generate high amounts of drug metabolites allowing the isolation of mg amounts of pure

reactive metabolites for their full structural elucidation by 1H-NMR.


158

At current, predicting the potential of new chemical entities to generate IDRs is still not

possible because of the lack of reliable pre-clinical models. However the formation of

reactive metabolites and protein covalent binding are perceived as significant risk factors.

In vitro screening tools for the generation and detection of reactive metabolites presented

in chapter 3 can be applied as novel tools in the development of safer drugs.

6.1.3. Biosynthesis of active metabolites by P450 BM3 mutants

Full characterization of metabolite profiles and elucidation of metabolite structures early

in drug development is often hampered by difficulties in producing sufficient amounts of

metabolites and in detecting active metabolites in a rapid and efficient way (22). The first

problem is tackled by the application of highly active P450 BM3 mutants for the

generation of high amounts of drug metabolites. The second problem is solved by the

application of the High Resolution Screening (HRS), which allows structural

characterization and bioaffinity determination of the metabolites formed in hyphenated

mode (23).

In chapter 5 an integrated strategy encompassing the use of BM3 mutants for

generation of metabolic mixtures and the structural identification and bioaffinity

assessment of metabolites with the HRS platform is presented. The combination of

hyphenated screening, MSn analysis and NMR spectroscopy enabled full structure

elucidation (except stereochemistry) and affinity determination of all the

biotransformation products synthesized in semi-preparative amounts. The panel of BM3

mutants presented is highly suitable to be used in the drug development process as

general reagents for lead diversification. The multidimensional screening approach

described here truly adds valuable information to the more conventional chemical analysis

methods. In a fast and efficient way, data was generated on both structure and biological

activity of the metabolites formed. Furthermore, linking structural modifications by

metabolism to changes in drug target affinity might be efficient tool to construct the

pharmacophore model next to the more conventional medicinal chemistry approach of

synthesizing structural analogues of lead compounds.

6.2. Future perspectives

In this thesis we describe the development of a metabolite production and profiling

platform, where P450 BM3 enzymes are engineered to mimic human P450s or to acquire

novel catalytic properties, based on the metabolism of diagnostic substrates; libraries of

site-directed, site-saturation or random mutants are then screened to select the best

mutant for upscaling and production of large amounts of individual metabolites, that are

purified by Prep-LC, allowing their structural elucidation by NMR and their

pharmacolological evaluation. However, there are some unresolved issues that should be

addressed in further research.


159

The presented examples of P450 BM3 engineering amply demonstrate that the activity of

this enzyme can be “tamed” for particular applications. Much has been made for the

enzyme’s suitability in commercial-scale applications, and the realization of this goal is one

of the researchers’ priorities. However the space-time yields and overall productivity of

cultures still need to be extended considerably for industrial applications (24). Enhancing

the longevity of the enzyme is one of the goals to achieve: total turnover numbers are

limited by a range of factors, including the stability of the variants employed, intrinsic

activity levels, and the response of the enzyme to specific substrates over extended

periods of turnover, including the inhibition rates associated with each product (24). With

non-natural substrates, too little is currently known about exit channels and product

release, and whether degradation is due to denaturation or co-factor loss, heme

modification or other factors. The commercial demands for enzymes that are functional in

non-natural environments (elevated temperature, nonnative pH, high substrate and

product concentrations, and organic solvents) are current challenges that have been only

partially met.

Moreover, the incubations performed in this study still required the addition of the

expensive cofactor NADPH, which is not desirable for large scale incubations. This could be

solved by performing whole-cell incubations, under non-lytic conditions, in order to allow

the use of the endogenous NADPH supply of the cells. To perform whole-cell incubations,

several issues need to be addressed: for example, many organic compounds do not readily

cross cell membranes; this issue could be tackled adding permeability enhancing agents

(e.g. EDTA), or by selecting mutant host strains with greater permeability, or by expressing

BM3 on cell surface (25). Moreover, substrate or reaction products can be toxic for the

host and over-oxidation can lead to the formation of many secondary or tertiary

metabolites. A lot of research is currently done in the field of alternative oxidants, such as

peroxides, metal powders and metal electrodes, to acquire simple, cost-effective and

efficient ways of performing CYP-mediated reactions (26).

Rational and directed evolution approaches have been successfully applied to

engineer BM3 enzymes for enhanced activity, stability and expression in E.coli, as well as

for altered substrate specificity and regio- and steroselectivity. However, the ability of

BM3 to adopt new functions has mechanistic underpinnings that have yet to be fully

elucidated. Recent advances in enzyme engineering have used a combination of random

methods of directed evolution with elements of rational enzyme modification to

successfully by-pass certain limitations of both directed evolution and rational design (27).

Semi-rational approaches targeting multiple, specific residues to create “smart libraries”

have been very successful (28, 29). Combinatorial alanine substitution has been

successfully applied to generate P450 BM3 variants active with large substrates (30).

Mutagenesis with un-natural aminoacids or insertion of cysteine residues that can be

subsequently alkylated with different alkylating agents in the active site can be used to


160

further improve the properties of the enzyme, opening up new chemistry not available

with the standard twenty aminoacids (27).

The synergy between computational and experimental BM3 research led to important

mechanistic insights into the origin of regio- and stereoselectivity: recently, de Beer et al.

successfully applied free energy calculations to get insight into the stereoselective

hydroxylation of -ionones by engineered BM3 mutants (31). Moreover the role of

protein plasticity in molecular dynamics simulation aimed at rationalize the

regioselectivity in testosterone hydroxylation by BM3 mutants has been extensively

studied (32).

In conclusion, P450 BM3 has been presented as a valuable, versatile tool to

support drug development by producing drug metabolites in such amounts that

toxicological and pharmacological characterization is possible. Mutagenesis techniques

are efficient tools to tailor the enzyme activity for a wide variety of applications. The

biocatalytic potential of P450 BM3 mutants to generate human relevant and novel unique

drug metabolites was demonstrated and these mutants were successfully used for the

generation and structural characterization of reactive metabolites, for the

functionalization of lead molecules and for the regioselective hydroxylation of steroid

compounds. The combination of BM3 biosynthesis and HRS screening is a highly valuable

platform to identify potential new lead compounds and to assess pharmacological

properties of drug metabolites. The platform presented in this thesis can be applied in

early stage drug discovery to expand the toolbox of the medicinal chemist for the

generation and optimization of lead compounds and to detect potentially dangerous

reactive metabolites for the development of safer drugs.


161

References

1. Guanaratna, C. (2000) Drug Metabolism & Pharmacokinetics in Drug Discovery: A primer for Bioanalytical Chemists, Part I, Current Separations 19, 17-23.

2. Schroer, K., Kittelmann, M., and Lutz, S. (2010) Recombinant human cytochrome P450 monooxygenases for drug metabolite synthesis, Biotechnol Bioeng 106, 699-706.

3. Clouthier, C. M., and Pelletier, J. N. (2012) Expanding the organic toolbox: a guide to integrating biocatalysis in synthesis, Chemical Society reviews 41, 1585-1605.

4. 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.



7. Damsten, M. C., de Vlieger, J. S., Niessen, W. M., Irth, H., Vermeulen, N. P., and Commandeur, J. N. (2008) Trimethoprim: novel reactive intermediates and bioactivation pathways by cytochrome p450s, Chemical research in toxicology 21, 2181-2187.

8. 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.


10. Fischer, V., Haar, J.A., Greiner, L., Lloyd, R.V., and Mason, R.P. (1991) Possible role of free radical formation in clozapine (Clorazil) induced agranulocytosis., Mol. Pharmacol. 40, 846-853.

11. Nedderman, A. N. (2009) Metabolites in safety testing: metabolite identification strategies in discovery and development, Biopharmaceutics & drug disposition 30, 153-162.


13. Stjernschantz, E., van Vugt-Lussenburg, B. M., Bonifacio, A., de Beer, S. B., van der Zwan, G., Gooijer, C., Commandeur, J. N., Vermeulen, N. P., and


162

Oostenbrink, C. (2008) Structural rationalization of novel drug metabolizing mutants of cytochrome P450 BM3, Proteins 71, 336-352.

14. Newhouse, T., and Baran, P. S. (2011) If C-H bonds could talk: selective C-H bond oxidation, Angew Chem Int Ed Engl 50, 3362-3374.

15. Chen, M. S., and White, M. C. (2007) A predictably selective aliphatic C-H oxidation reaction for complex molecule synthesis, Science 318, 783-787.

16. Gaich, T., and Baran, P. S. (2010) Aiming for the ideal synthesis, The Journal of organic chemistry 75, 4657-4673.

17. Zhang, K., El Damaty, S., and Fasan, R. (2011) P450 fingerprinting method for rapid discovery of terpene hydroxylating P450 catalysts with diversified regioselectivity, J Am Chem Soc 133, 3242-3245.

18. Jung, S. T., Lauchli, R., and Arnold, F. H. (2011) Cytochrome P450: taming a wild type enzyme, Curr Opin Biotechnol 22, 809-817.

19. 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.


21. Boerma, J. S., Vermeulen, N. P., and Commandeur, J. N. (2011) Application of CYP102A1M11H as a tool for the generation of protein adducts of reactive drug metabolites, Chemical research in toxicology 24, 1263-1274.

22. Fura, A., Shu, Y. Z., Zhu, M., Hanson, R. L., Roongta, V., and Humphreys, W. G. (2004) Discovering drugs through biological transformation: role of pharmacologically active metabolites in drug discovery, J Med Chem 47, 4339-4351.



25. Yim, S. K., Kim, D. H., Jung, H. C., Pan, J. G., Kang, H. S., Ahn, T., and Yun, C. H. (2010) Surface display of heme- and diflavin-containing cytochrome P450 BM3 in Escherichia coli: a whole cell biocatalyst for oxidation, J Microbiol Biotechnol 20, 712-717.

26. Kumar, S. (2010) Engineering cytochrome P450 biocatalysts for biotechnology, medicine and bioremediation, Expert Opin Drug Metab Toxicol 6, 115-131.

27. Chica, R. A., Doucet, N., and Pelletier, J. N. (2005) Semi-rational approaches to engineering enzyme activity: combining the benefits of directed evolution and rational design, Curr Opin Biotechnol 16, 378-384.

28. Seifert, A., Vomund, S., Grohmann, K., Kriening, S., Urlacher, V. B., Laschat, S., and Pleiss, J. (2009) Rational design of a minimal and highly enriched CYP102A1 mutant library with improved regio-, stereo- and chemoselectivity, Chembiochem 10, 853-861.


163

29. Seifert, A., Antonovici, M., Hauer, B., and Pleiss, J. (2011) An efficient route to selective bio-oxidation catalysts: an iterative approach comprising modeling, diversification, and screening, based on CYP102A1, Chembiochem 12, 1346-1351.

30. Lewis, J. C., Mantovani, S. M., Fu, Y., Snow, C. D., Komor, R. S., Wong, C. H., and Arnold, F. H. (2010) Combinatorial alanine substitution enables rapid optimization of cytochrome P450BM3 for selective hydroxylation of large substrates, Chembiochem 11, 2502-2505.

31. de Beer, S. B., Venkataraman, H., Geerke, D. P., Oostenbrink, C., and Vermeulen, N. P. (2012) Free Energy Calculations Give Insight into the Stereoselective Hydroxylation of alpha-Ionones by Engineered Cytochrome P450 BM3 Mutants, Journal of chemical information and modeling.

32. de Beer, S. B., van Bergen, L. A., Keijzer, K., Rea, V., Venkataraman, H., Guerra, C. F., Bickelhaupt, F. M., Vermeulen, N. P., Commandeur, J. N., and Geerke, D. P. (2012) The role of protein plasticity in computational rationalization studies on regioselectivity in testosterone hydroxylation by cytochrome P450 BM3 mutants, Current drug metabolism 13, 155-166.

Samenvatting

164

SAMENVATTING

Cytochromen P450 (P450 of CYP) vertegenwoordigen een grote superfamilie van heme-

bevattende monoxygenases die in nagenoeg alle organismen voorkomen. Bij de mens zijn

P450's betrokken bij de biotransformatie (metabolisme) van 80% van de geneesmiddelen

op de markt. Metabolieten die door P450 worden geproduceerd kunnen farmacologische

activiteit vertonen of verantwoordelijk zijn voor de toxiciteit of andere ongewenste

bijwerkingen van geneesmiddelen of geneesmiddelkandidaten. Het is om deze redenen

dat het sinds enkele jaren verplicht is voor geneesmiddel registratie om de biologische

eigenschappen van de belangrijkste metabolieten te karakteriseren. Er is daarom een

grote behoefte aan systemen waarmee deze metabolieten in voldoende hoeveelheden

kunnen worden geproduceerd zodat hun farmacologische en toxicologische

eigenschappen in detail kunnen worden gekarakteriseerd.

Tot nu toe werd productie van metabolieten meestal uitgevoerd door organische

synthese of door grootschalige incubaties met menselijke P450's die zijn verkregen door

heterologe expressie in E.coli, gist, insectcellen of zoogdiercellijnen. Echter, vanwege hun

instabiliteit en intrinsiek lage activiteit is het rendement van metaboliet productie door

menselijke P450's vaak erg laag en de kosten zeer hoog. In vergelijking met hun menselijke

tegenhangers vertonen microbiële P450's over het algemeen een veel hogere stabiliteit en

specifieke activiteit. Één van de meest bestudeerde microbiële P450's is cytochrome P450

BM3 (CYP102A1) van Bacillus megaterium.

Bij dit microbiële P450 is het heme domein en het reductase domein gefuseerd in een

enkele polypeptideketen waardoor het electronentransport van de cofactor NADPH naar

het katalytische centrum zeer efficient is. Dit in tegenstelling tot de P450's van zoogdieren

waar het heme domein en reductase domein als aparte membraangebonden eiwitten

voorkomen. P450 BM3 is een zeer stabiel en oplosbaar enzym en is het meest actieve

P450 dat tot nu toe in de natuur is gevonden. Omdat dit enzym op grote schaal kan

worden geëxpresseerd in E.coli, en daaruit gemakkelijk kan worden gezuiverd, heeft P450

BM3 grote perspectieven voor de toepassing als biokatalysator voor metaboliet productie

op grote schaal. Het feit dat er verschillende kristalstructuren van P450 BM3 met

substraten zijn opgehelderd maakt het tevens mogelijk de substraatselectiviteit van dit

enzym met gerichte site-directed mutagenese te manipuleren.

Het in dit proefschrift beschreven onderzoek is uitgevoerd in de context van het

Top Instituut Pharma-project 'MetStab' (D2-102). Het belangrijkste doel van het

onderzoek was het leveren van een bijdrage aan de ontwikkeling van een platform

waarmee, gebruik makend van een verzameling van hoog-actieve en katalytisch diverse

BM3 mutanten en innovatieve on-line hoge resolutie screeningsmethodes, op een zeer

efficiente manier metabolieten van geneesmiddelen of lead compounds kunnen worden

geproduceerd, geidentificeerd en geprofileerd met betrekking tot hun affiniteit voor

specifieke drug targets. Bovendien beoogde het onderzoek het verkrijgen van meer

Samenvatting

165

mechanistisch inzicht in de werking van de P450 en de geometrie en de topologie van de

substraat bindingsplaats. In dit proefschrift zijn verschillende mutagenese technieken

beschreven waarmee de enzymactiviteit van P450 BM3 kan worden gestuurd in de

richting van verschillende toepassingen, zoals bioactivatie van geneesmiddelen tot

reactieve metabolieten, stereoselectieve hydroxylering van steroïdes en productie van

farmacologisch actieve geneesmiddel metabolieten.

In hoofdstuk 1 wordt bediscussieerd waarom in de afgelopen jaren steeds meer

aandacht wordt besteed aan de rol van metabolisme bij de ontdekking en ontwikkeling

van nieuwe geneesmiddelen. Naast het feit dat metabolisme in belangrijke mate de

farmacokinetiek van geneesmiddelen bepaalt, worden in dit hoofdstuk verschillende

voorbeelden beschreven van geneesmiddelen die worden gemetaboliseerd tot

farmacologisch zeer potente metabolieten die aanzienlijk bijdragen tot de algehele

farmacologische werking van de geneesmiddelen. Anderzijds kan de biologische activiteit

van metabolieten verantwoordelijk zijn voor toxiciteit of andere ongewenste

bijwerkingen van geneesmiddelen.

Karakterisering van actieve metabolieten is daarom van groot belang om zowel de

farmacologische als toxicologische effecten beter te kunnen voorspellen in preklinische en

klinische studies. Het feit dat metabolieten soms veel potenter zijn dan de uitgangsstof

maakt dat structurele modificatie van lead compounds door middel van metabolisme ook

een veelbelovend instrument kan zijn in het drug discovery proces, naast de meer

conventionele benadering van organische synthese van structurele analoga.

Controle van de regio-en stereoselectiviteit van de hydroxylering van niet-

geactiveerde CH bindingen in complexe organische verbindingen is een van de uitdagingen

in de synthetische organische chemie. In dit proefschrift is aangetoond dat genetisch

gemodificeerde P450 BM3 enzymen in staat zijn steroides regio- en stereospecifiek te

hydroxyleren. Door mutaties aan te brengen op specifieke posities in de substraat

bindingsplaats van P450 BM3 werden grote veranderingen in regio-en stereoselectiviteit

van hydroxylering van testosteron verkregen.

In hoofdstuk 2 werd de rol van het aminozuur op positie 87 onderzocht, dat zich

in de onmiddelijke nabijheid van het katalytische centrum van P450 BM3 bevindt. Deze

studie werd uitgevoerd op een mutant van P450 BM3, mutant M11, die door een

combinatie van site-directed en random mutagenese het vermogen heeft verworden om

een groot aantal geneesmiddelen en steroides te metaboliseren met veel hogere activiteit

dan de menselijke P450's.

Met behulp van site-directed mutagenese werden alle 20 mogelijke aminozuren op positie

87 onderzocht. Hieruit bleek dat, afhankelijk van de aard van het aminozuur op positie 87,

niet alleen de enzymactiviteit sterk veranderde maar ook de regioselectiviteit, hetgeen

aantoont dat testosteron zich door de aminozuur-verandering op verschillende manieren

orienteert in de active site.

Samenvatting

166

In hoofdstuk 4 werd de rol van het aminozuur op positie 82 op de regio- en

stereoselectiviteit van steroid metabolisme onderzocht. Vervangen van het

oorspronkelijke alanine-residue door het ruimtelijk grotere aminozuur tryptofaan

resulteerde in een mutant dat testosteron en norethisteron met hoge regioselectiviteit op

de 16ß-positie hydroxyleert. De regio-en stereoselectiviteit van P450-gemedieerde

reacties wordt waarschijnlijk bepaald door de oriëntatie van het substraat ten opzichte

van de reactieve ijzer-oxo species in het katalytische centrum van P450. Met behulp van

spin relaxatie NMR, waarmee de kortste afstand van de waterstofatomen van een

gebonden substraat ten opzichte van het heme-ijzer atoom van P450 kan worden bepaald,

kon worden aangetoond dat de mutatie op positie 82 inderdaad tot een veranderde

oriëntatie van testosteron in de actieve plaats van P450 BM3 heeft geleid, in

overeenstemming was met de waargenomen verandering in regioselectiviteit.

Veel schadelijke bijwerkingen van geneesmiddelen zijn het gevolg van P450-

afhankelijke vorming van hoog-reactieve metabolieten, reactieve intermediairen, die

kunnen reageren met kritische macromoleculen in de cel. Al in een vroeg stadium van

geneesmiddel ontwikkeling wordt daarom al onderzocht of bij metabolisme van een

kandidaat geneesmiddel door P450's reactieve intermediairen ontstaan. Dit wordt veelal

gedaan door de reactieve intermediaren in te vangen met het tripeptide glutathione (GSH)

dat ook in elke lichaamscel als functie heeft de kritische macromoleculen te beschermen

tegen reactieve producten. Echter, door de zeer lage niveau's waarmee de menselijke

P450's de reactieve metabolieten produceren wordt de detectie en structuuropheldering

van gevormde GSH-conjugaten sterk bemoeilijkt.

In hoofdstuk 3 is de serie P450 BM3 mutanten met de verschillende aminozuren

op positie 87 onderzocht op hun vermogen tot selectieve bioactivatie van clozapine. Dit

geneesmiddel veroorzaakt bij sommige patienten ernstig bijwerkingen zoals

agranulocytose en levertoxiciteit. In beide gevallen wordt een reactief nitrenium ion

verantwoordelijk geacht voor de toxiciteit. Uit het onderzoek met de P450 BM3 mutanten

bleek dat de mutant met een phenylalanine op positie 87, M11 V87F, clozapine zeer

selectief en met hoge activiteit kan metaboliseren tot het reactieve nitrenium ion. Dit

reactieve nitrenium ion bleek op verschillende manieren met GSH te kunnen reageren,

resulterend in vijf verschillende GSH-conjugaten. Met behulp van mutant M11 V87F kon

elk GSH-conjugaat van clozapine op voldoende schaal kon worden geproduceerd zodat

elke structuur kon worden opgehelderd met 1H-NMR. Deze studie bevestigt het grote

potentieel van BM3 mutanten als hulpmiddel bij de identificatie van producten van

reactieve metabolieten van geneesmiddelen, en, indirect, de aard van het reactieve

intermediair. Dit nieuwe hulpmiddel voor de karakterisering van reactieve metabolieten

kan tevens worden gebruikt om bioactivatie routes op te helderen van geneesmiddelen

waarvan nog niet bekend is waarom ze toxische bijwerkingen vertonen. Uiteindelijk zal

deze kennis een belangrijke rol spelen in de ontwikkeling van veiligere geneesmiddelen en

kandidaat-geneesmiddelen.

Samenvatting

167

Zoals hierboven reeds aangegeven, kunnen bij het P450-afhankelijke

metabolisme van geneesmiddelen soms producten ontstaan met een veel potentere

farmacologische werking dan de uitgangsstof. Om deze reden kunnen P450's ook worden

toegepast in de drug discovery fase, als alternatief voor organische synthese van

structurele analoga van lead verbindingen. Tevens kan structurele modificatie door P450's

leiden tot producten met verbeterde fysisch-chemische eigenschappen, zoals verbeterde

wateroplosbaarheid.

In hoofdstuk 5 wordt een geïntegreerde strategie gepresenteerd waarbij met

behulp van verschillende P450 BM3 mutanten metabole mengsels van een verbinding

worden gegenereerd, waarna de mengsels met een bioaffiniteitsplatform worden

gescreened op aanwezigheid van metabolieten met hoge affiniteit voor een

farmacologisch belangrijk receptoreiwit. Om het principe van deze nieuwe strategie aan

te tonen was in dit onderzoek gekozen voor TAK-715. Dit geneesmiddel is een potente

remmer van p38 kinase maar heeft als nadeel dat het zeer slecht wateroplosbaar is.

Metabolisme van TAK-715 door P450 kan de wateroplosbaarheid verbeteren maar kan

ook de affiniteit voor p38 kinase beinvloeden. Voor de identificatie van metabolieten van

TAK-715 met hoge affiniteit voor p38 kinase werd gebruik gemaakt van HPLC gekoppeld

aan zowel een massa spectrometer als een hoge resolutie screening (HRS) platform

waarmee on-line de p38 kinase affiniteit van componenten van mengsels kan worden

bepaald. Uit dit onderzoek bleek dat mutanten van P450 BM3 inderdaad in staat waren

om TAK-715 zeer efficient te metaboliseren tot verschillende producten. Met behulp van

het on-line bioaffiniteitsplatform kon worden aangetoond dat verschillende producten

hoge bindingsaffiniteit hadden voor p38 kinase. Door grote schaal productie van de

producten met het meest actieve P450 BM3 mutant kon de volledige structuur (behalve

stereochemie) van de actieve en niet-actieve producten worden bepaald met 1H-NMR

spectroscopie en de affiniteit van elk product voor p38 kinase worden gekwantificeerd.

Hieruit bleek dat sommige producten zelfs een hogere affiniteit vertoonden dan TAK-715

zelf.

Concluderend: het in dit proefschrift beschreven onderzoek laat zien dat het

enzymsysteem P450 BM3 op verschillende manieren een zeer waardevolle bijdrage kan

leveren aan de ontwikkeling van nieuwe geneesmiddelen. Zo kunnen humaan-relevante

geneesmiddel metabolieten met behulp van dit bacteriële enzym op grote schaal worden

geproduceerd, zodat hun structuur kan worden opgehelderd en hun farmacologische en

toxicologische eigenschappen in detail kunnen worden bestudeerd. De regio- en

stereoselectiviteit van de enzymen, die kan worden gestuurd door site-directed

mutagenese, maakt het mogelijk producten te maken die met organische synthese

moeilijk toegankelijk zijn. Daarnaast heeft het geintegreerde platform van P450 BM3

mutanten in combinatie met hoge resolutie bioaffiniteitsscreening veel perspectief om te

worden toegepast voor lead optimalisatie in het vroege stadium van

geneesmiddelontwikkeli.

Appendices

168

LIST OF ABBREVIATIONS

15-OH-N 15-hydroxynorethisterone

15-OH-T 15-hydroxytestosterone

ADR Adverse drug reaction

CLZ Clozapine

CO Carbon monoxide

FAD Flavin adenine dinucleotide

FMN Flavin mononucleotide

GSH Glutathione (reduced)

GST Glutathione S-transferase

HLM Human liver microsomes

HPLC High performance liquid chromatography

HRS High resolution screening

IDR Idiosyncratyc drug reaction

KPi Potassium phosphate

NCE New chemical entity

NET Norethisterone

Ni-NTA agarose Nickel nitroacetic acid agarose

NMR Nuclear Magnetic Resonance

p38α p38α mitogen-activated protein kinase

P450 Cytochrome P450 monooxygenase

P450 BM3 Cytochrome P450 BM3

RI Reactive Intermediate

SD Standard Deviation

SDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis

T1 relaxation NMR Spin relaxation NMR

TES Testosterone

Tr Retention time

UPLC Ultra Performance liquid chromatography

Appendices

169

Appendices

170

PUBLICATIONS

Vottero, E., Rea, V., Lastdrager, J., Honing, M., Vermeulen, N.P.E., and

Commandeur, J.N.M., Role of residue 87 in substrate selectivity and

regioselectivity of drug-metabolizing cytochrome P450 BM3 M11,

J.Biol.Inorg.Chem., 2011, 16 (6), 899-912

Rea, V., Dragovic, S., Boerma, J.S., de Kanter, F., Vermeulen, N.P.E., and

Commandeur, J.N.M., Role of residue 87 in the activity and regioselectivity

of clozapine metabolism by drug metabolising BM3: application for

structural characterization of clozapine GSH-conjugates, Drug Metab Dispos.

2011, 39 (12), 2411-20

Rea, V., Kolkman, A., Vottero, E., Stronks, E., Ampt, K., Honing,

M.,Vermeulen, N.P.E., Wijmenga, S., Commandeur, J.N.M., Active site

substitution A82W improves regioselectivity of steroid hydroxylation by

cytochrome P450 BM3 mutants as rationalized by spin relaxation NMR

studies, Biochemistry, 2012, 51 (3), 750–760

Rea, V., Falck, D., Kool, J., de Kanter, F.J., Commandeur, J.N.M.,

Vermeulen, N.P.E., Niessen, W., and Honing, M., Application of

Cytochrome P450 BM3 mutants for the generation of human relevant active

metabolites of P38 MAP Kinase inhibitor Tak-715, 2012, submitted

de Beer, S.B., van Bergen, L.A., Keijzer, K., Rea, V., Venkataraman, H.,

Guerra, C.F., Bickelhaupt, F.M., Vermeulen, N.P.E., Commandeur, J.N.M.,

Geerke, D.P., The role of protein plasticity in computational rationalization

studies on regioselectivity in testosterone hydroxylation by Cytochrome

P450 BM3 mutants. Curr.Drug Metab. 2012, 13(2), 155-166

Appendices

171

Appendices

172

CURRICULUM VITAE

Vanina Rea was born on the 6th October 1984 in Cercola (NA). In 2002, she graduated

from high school at the Liceo Scientifico “F.Silvestri” in Portici (NA). In the same year she

started studying Biotechnology for Products and Processes at the University of Naples

“Federico II”, obtaining her bachelor degree with honors (cum laude). During her bachelor,

she did an internship in the Department of Organic and Inorganic Chemistry, division of

Mass Spectrometry, working on affinity chromatography and mass spectrometry

techniques for the separation and characterization of glycosylated proteins and peptides.

In 2005, she started her Master in Industrial Biotechnology at the University of Naples

“Federico II”. In 2006, she obtained the Erasmus Exchange Scholarship that allowed her to

spend one year at the Swammerdam Institute for Life Sciences, Universiteit van

Amsterdam, for her master thesis research project, under the supervision of Prof.

K.J.Hellingwerf. In 2007, she obtained her master degree with honors (cum laude) with a

thesis on “Enantioselective biocatalysis for the synthesis of chiral drugs”. In 2008, she

started her PhD at the department of Molecular Toxicology at the Vrije Universiteit

Amsterdam, under the supervision of Prof. Nico P.E. Vermeulen and Dr. Jan N.M.

Commandeur. Her research at the VU was supported by the Dutch Top Institute Pharma

grant D2-102: “Metabolic stability assessment as a new tool in the Hit-to-Lead selection

process and the generation of new lead compound libraries” and resulted in this book that

you have at this very moment in your hands.

Appendices

173

Appendices

174

DANKWOORD

Olè! It’s finally done… so it’s time to thank all the people that made it possible. First I wish to express my gratitude to Nico and Jan. Nico: thank you for giving me the opportunity to work in your team. I am grateful for the scientific knowledge and experience acquired and your contribution towards this achievement. Jan: thank you for being always present and willing to help in my research. The valuable practical and theoretical discussions that we had helped me to work in a better way. Furthermore, I wish to extend my gratitude to all the members of the Reading Committee for the review of this thesis. Chris V., thank you for your sharp comments and valuable input on the work. Franz de Kanter, thank you very much for your help with NMR. Laura thank you for your kindness, we miss you at the VU. Special thanks to my special roommates: Galvin, Jantje and Sanja, that made the room P2.40 a safe haven from the cruel world. Galvin, thanks for always being ready to help your undutchable colleagues and for being the great person you are. Jantje, thanks for all the laughs and your sarcasm. Your special habits will be never forgotten. Sanja, thanks for being the great friend you are, thanks for giving me support every time I needed it, for giving me a roof every time I needed it and for being my human collector in Prep LC every time I needed it! Thanks to the “other” Moltoxers: Harini, Jelle, Kevin and Rokus. Thanks for the great company and the nice time spent together. Harini, thanks for your afternoon visits and updates. Jelle, we had our up&downs, but we could always solve them with a good laugh. Rokus, thanks for the interesting talks and funny remarks. My best wishes to the new moltoxers Angelina, Shalenie and Michiel! I’d like also to thank all former colleagues at Moltox: Eduardo, Jeroen L., Jolanda and Rene. Eduardo, we started this adventure together, thanks for guiding me in the very beginning, for your great help and for all the laughs outside the lab. Jeroen L., you were really precious at the VU, thanks for great company at borrels, dinners, coffee breaks and at the unforgettable practical courses! Thank you to the former computational dream team: Chris O, Eva and Stephanie: thanks for the support in the difficult moments. And also thanks to the new computational team: Daan, Ruben, Bas, Jozef, Lovorka, and Daphne, for making my Wednesday afternoons very puzzling sometimes. Thanks to all the member of my Ti-Pharma group. Ard, Cornelis, David, Geert, Jaap, Jeroen K., Jon, Kirsten, Lionel, Lutgarde, Sybren, Wilfried. Thanks for the fruitful collaboration that made our project so successful. Special thank you to Maarten, the PI of our project: thank you for the support, the enthusiasm, the real keen interest in the results and for the constant encouragement during my research.

Appendices

175

Special thanks to Ard and David, for the nice and fruitful collaboration and for the nice time spent at conferences and TiPharma courses. Thanks to the ACAS members: Ansgar, Ben, David, Ferry, Filipe, Jon, Linda, Lygia, and the MedChem members, Chimed, Ewald, Mark and Oscar for making borrels, courses and conferences so much fun. Special thanks to my “dutch” family: Brenda, Dario, Dave, Eugenie, Miguel without you this would not have been possible. And special thanks to my Italian family: Francesco, Gianpi, Margherita, Martina, Mirta that never made me feel far from home. Thanks to my parents that always supported me: I am as proud of you as you are of me. Thanks to my siblings, Alessio, Antonio, Ciro, Giusi, Marco: I would not be the same person without you. Thanks to all my family. And thanks to all my friends who taught me a very simple lesson:

cytochrome p450 bm3 as versatile biocatalytic tool in drug development

Documents