university of groningen electrochemistry on-line with mass spectrometry … · 2016-03-09 ·...

University of Groningen

Electrochemistry on-line with mass spectrometryJurva, Johan Ulrik

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:2004

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):Jurva, J. U. (2004). Electrochemistry on-line with mass spectrometry: instrumental methods for in vitrogeneration and detection of drug metabolites. Groningen: s.n.

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 26-06-2020

https://www.rug.nl/research/portal/en/publications/electrochemistry-online-with-mass-spectrometry(a5630a6a-6a66-4e8b-b6f2-c09524c1a074).html



Electrochemistry on-line with mass spectrometry

Instrumental methods for in vitro generation and

detection of drug metabolites

Ulrik Jurva

RIJKSUNIVERSITEIT GRONINGEN

Electrochemistry on-line with mass spectrometry:

Instrumental methods for in vitro generation and detection of drug metabolites

PROEFSCHRIFT

ter verkrijging van het doctoraat in de

Wiskunde en Natuurwetenschappen

aan de Rijksuniversiteit Groningen

op gezag van de

Rector Magnificus, dr. F. Zwarts,

in het openbaar te verdedigen op

maandag 10 mei 2004

om 16.15 uur

door

Johan Ulrik Jurva

geboren op 15 juni 1971

te Falköping, Zweden

Promotor Prof. dr. H.V. Wikström Copromotor Dr. A.P. Bruins Beoordelingscommissie

Prof. dr. N. Castagnoli Jr Prof. dr. J.P Franke Prof. dr. N.M.M. Nibbering

ISBN 90-367-1999-2

‘Livet blir så mycket lättare att leva när man vet att saker och ting alltid fixar sig’

(Antti Jurva)

To my parents: Antti & Gunnel

Paranimfen Jonas Eriksson Sofia Jurva ISBN 90-367-1999-2 An electronic version of this thesis is available at http://www.ub.rug.nl/eldoc/dis/science/j.u.jurva/ ISBN electronic version: 90-367-2000-1 Printing: Mediagruppen Intercopy, Göteborg, Sweden Cover design: Protoporphyrin IX, the active site of cytochrome P450 The research described in this thesis was carried out within the framework of the research school GUIDE

Contents

Chapter 1 .............................................................................................................. 1 Introduction Chapter 2............................................................................................................... 17 Comparison between electrochemistry/mass spectrometry and cytochrome P450 catalyzed oxidation reactions Chapter 3............................................................................................................... 39 Electrochemically assisted Fenton reaction: The reaction of hydroxyl radicals with xenobiotics followed by on–line analysis with HPLC/MS/MS Chapter 4............................................................................................................... 53 The electrochemically assisted Fenton system applied to the prenylated flavonoid Xanthohumol Chapter 5............................................................................................................... 93 The oxidative metabolism of the dopamine agonist N–0923 mimicked by EC/MS and electrochemically assisted Fenton reaction. Chapter 6............................................................................................................... 113 Practical aspects of electrochemistry on–line with ESI/MS Chapter 7............................................................................................................... 121 Bioactivation of the catecholamine pro–drugs PD217015, (–)-GMC6650 and (+)-GMC6650 in male Wistar rats Overall conclusions and future prospects.......................................................... 145 Summary............................................................................................................... 147 Samenvatting........................................................................................................ 150 List of publications............................................................................................... 153 Acknowledgements............................................................................................... 154

Chapter 1 Introduction

1

Chapter 1

Introduction

1.1 Electrochemistry coupled on–line with mass spectrometry: An

historic overview

Studies of on–line coupling of electrochemistry with mass spectrometry date back to the work by Bruckstein et al. in the early 1970s.1 In this and other early experiments, porous working electrodes or permeable membranes were used for the measurement of volatile products of electrochemical reactions by means of electron ionization (EI) mass spectrometry.

In 1986, Hambitzer et al. studied the redox reactions of N,N–dimethyl-aniline in an electrochemical cell on–line with thermospray mass spectrometry (TSP/MS).2 In the same year Getek used electrochemistry on–line with TSP/MS to study the oxidation of acetaminophen.3

In 1989, Volk, Yost and Brajter–Toth used the ESA electrochemical cell on–line with TSP/MS/MS to study the electrochemical oxidation of uric acid.4 A couple of years later, the same group used a modified version of this setup in the study of 6–thiopurine.5 An HPLC–column was introduced between the electrochemical cell and the TSP/MS to enable separation of the electrochemical oxidation products prior to the TSP/MS. The early developments in electrochemistry on–line with mass spectrometry in the field of drug metabolism have been reviewed by Volk et al.6

In 1992, Van Berkel and co–workers showed that radical cations observed for various organic compounds in electrospray ionization mass spectrometry (ESI/MS) originate from electrochemical oxidation in the electrospray source.7 A couple of years later, the same group presented a number of different setups for on–line coupling of electrochemistry and electrospray mass spectrometry (ESI/MS).8,9 They also described how the electrospray source can be characterized as a controlled current electrolytic cell.10,11

In 1995, Bond et al. showed how a simple flow cell constructed from two lengths of platinum microtubing could be used on–line with ESI/MS in the analysis of copper, nickel and cobalt diethyldithiocarbamates.

12


2

In 1996, Baczynskyj used the ESA Coulochem cell coupled on–line with LC/ESI/MS to study the electrochemical oxidation of Delavirdine and other organic compounds.13 The same year, Cole and co–workers designed an on–line EC/ESI/MS probe to minimize the time between the electrochemical reaction and the mass spectrometer.14,15

In 1997, Zhou, Hefta and Lee presented a modified version of the source used in 1995 by Zhou and Van Berkel. Electrochemical modifications of 21 phenylthiohydatoin amino acid derivatives in this setup enabled their detection in 50–1000 nM concentrations at low flow rates.16 The same year, Regino and Brajter–Toth connected a home–made electrochemical cell on–line with TSP/MS and particle beam mass spectrometry (PB/MS) to study the electrochemical oxidation of uric acid and dopamine.17,18 The same group also studied the effects of the mobile phase composition on the electrochemical cell conversion efficiency.19 A couple of years later, the EC/PB/MS–setup was also used for the characterization of intermediate radical cations in non–aqueous media.20 Also in 1997, Stassen and Hambitzer used a thin layer electrochemical cell on–line with TSP/MS to study the electrochemical oxidation of aniline and a number of N–alkyl-anilines.21,22 A thorough description of EC/TSP/MS instrumentation was presented a year later.23

In 1999, Jurva, Wikström and Bruins used the ESA Coulochem cell on–line with ESI/MS to mimic a number of metabolic oxidation reactions performed by cytochrome P450.24-26 The same year Iwashi used the ESA Coulochem cell on–line with HPLC/UV/ESI/MS to study the electrochemical oxidation of 3–hydroxyanthranilic acid.27 Also in 1999, Deng and Van Berkel studied the electrochemical oxidation of dopamine in a home made thin layer electrochemical flow cell coupled on–line with ESI/MS.28 The same type of electrochemical cell was also used in a different setup for electrochemically modulated pre–concentration and matrix elimination of tamoxifen and other organic analytes.29

Another interesting approach to EC/ESI/MS was demonstrated by the group of Van Berkel in 2000 when they coupled electrochemically modulated liquid chromatography (EMLC) on–line with ESI/MS.30 This setup allowed the electrochemical oxidation of aniline (on the column) and subsequent separation of the oxidation products prior to the ESI/MS.

In 2001, Diehl, Liesener and Karst used HPLC with post–column electrochemical treatment on–line with ESI/MS for signal enhancement of ferrocenecarboxylic acid esters of various alcohols and phenols.31,32 This technique has also been applied in the analysis of gasoline, diesel and mineral oil samples.33 Hayen and Karst also used the same type of setup for the oxidation of phenothiazine and its derivatives into radical cations or sulfoxides that resulted in lower detection limits.34 Brown, Rollag, Lin and Lee used a similar setup where electrochemical oxidation enabled the detection of a number of compounds that are otherwise difficult to detect by ESI/MS.35 A home made


3

electrochemical cell was used on–line with electrospray ionization Fourier transform ion cyclotron resonance mass spectrometry (ESI–FTICR/MS) by Zhang, Palii, Eyler and Brajter–Toth for electrochemically enhanced ionization of a number of compounds including triphenylamine and polycyclic aromatic hydrocarbons.36,37 Also in 2001, the group of Amster used a home–made electrochemical cell on–line with ESI/MS to study the electrochemical reduction of a number of different metalloproteins.38

Over the last couple of years, the number of publications in the field of EC/ESI/MS has increased. The ease of use and the many advantages that can be gained from adding an electrochemical cell in front of the ESI/MS will probably lead to an extended use in the future. In addition to the obvious use for characterization of oxidation/reduction products and short lived intermediates in various electrochemical reactions, there are several other areas of interest for EC/ESI/MS.

The on–line electrochemical generation of drug metabolites, on which large parts of this thesis are focused, is one field that is expected to find an extended use in the future.

Signal enhancement through electrochemical oxidation of neutral molecules or by addition of electrochemically active tags, such as ferrocene, are areas with great potential to lower detection limits in the analysis of a large range of organic compounds.

Because electrochemical equipment is inexpensive, easy to use and already available in many laboratories, we expect a continued growth for EC/ESI/MS in all of the above mentioned areas in the future.

1.2 Cytochrome P450

The cytochrome P450 enzymes comprise a large family of heme containing proteins with a cysteine in the active site that acts as an axial ligand to the heme iron. Over 500 forms of cytochrome P450 have been found in different forms of life, including plants, bacteria and mammals.39-42 All enzymes in the P450 family are monooxygenases i.e. they catalyze the incorporation of one oxygen atom of dioxygen into substrate while the other oxygen atom is reduced by two electrons to give water. NADPH usually provides the two electrons needed for this process. The overall reaction is presented below.

Substrate + O2 + 2e- +2H+ Substrate(O) + H2OP450

There is a great diversity in the reactivity among different P450 enzymes and a large range of transformations is catalyzed. The most important and commonly encountered


4

cytochrome P450 catalyzed oxidations are hydroxylations, epoxidations, heteroatom oxidations/dealkylations, dehalogenations and alcohol oxidations.43,44

The reactive site of all P450 enzymes contains an iron protoporphyrin IX (Figure 1.1) with cysteine as the fifth ligand, leaving the sixth coordination site to bind and activate molecular oxygen.

Cys

Fe

S

NOH

O

N

N

N

OH

O

Figure 1.1 Iron protoporphyrin IX with cysteine as the sixth ligand The catalytic cycle of cytochrome P450 is given in Figure 1.2. The resting form of the enzyme is a six coordinate (low spin) ferric state (1) with water as the sixth ligand. The catalytic cycle is initiated as the substrate (R–H) enters the active site and the water molecule is displaced, generating a five coordinate (high spin) ferric state (2). The next step is a reduction of the iron(III) center to a ferrous state (3). Dioxygen then reacts with 3 to form an oxyferrous complex (4a,b). One electron reduction of (4), the rate limiting step in all cytochrome P450 catalyzed reactions, is supposed to give a negatively charged iron(III)–peroxo complex (5) which upon protonation yields a hydroperoxide complex (6). A second protonation followed by heterolytic cleavage of the O–O bond generates water and an electrophilic, high valent iron–oxo species (7) generally considered to be the active oxidant in most cytochrome P450 catalyzed oxidations.


5

e-

H+ -H2O

R(O)H

4b

FeIII

SCys

N

N

N

NO

O

4aFeII

SCys

N

N

N

NO

O

O2

N

N

FeII

SCys

N

N3

RH

e-

+.O

FeIV

SCys

N

N

N

N7

FeIII

SCys

N

N

N

NO

OH

6

H+

FeIII

SCys

N

N

N

NO

O_

5

1

FeIII

SCys

N

N

N

NO

HH

HR

2

FeIII

SCys

N

N

N

N

Figure 2.2 The catalytic cycle of cytochrome P450

The first four P450 states in the catalytic cycle are stable enough to allow isolation and characterization by different spectroscopic techniques. Crystal structures have been determined for states 1–4 of bacterial P450cam.45 The reactions following the reduction of 4 are very rapid and the structures within brackets in Figure 1.2 have not yet been fully characterized. A crystal structure presumed to correspond to the oxo–ferryl porphyrin radical intermediate 7 has been presented but the crystal structure might also in part be made up from the hydroperoxo–iron species 6.45 The debate about what is the actual “reactive oxygen” form of cytochrome P450 is not yet resolved. Newcomb et al. have presented data that suggest that the hydroperoxo–iron intermediate 6 can act as an electrophile and be the actual oxygenating intermediate for many oxidations.46 In some P450 enzymes protonation of the hydroperoxo–iron species 5 may be inhibited in the


6

active site. In such enzymes the peroxo–iron intermediate has an extended lifetime and can take part in oxidations initiated by nucleophilic attack. Examples where such a mechanism is believed to take place are the oxidation of cyclohexane carboxaldehyde catalyzed by P450 2B4 and the conversion of human androgens to estrogens by P450 aromatase.44 The oxoferryl porphyrin radical 7 is the most electrophilic species in the catalytic cycle and believed by many to be the reactive species in most P450 catalyzed oxidations. Heterolytic cleavage of the hydroperoxo–iron species 6 would generate a formal FeV=O species, but extraction of an electron from the porphyrin moiety would give the more stable oxoferryl porphyrin radical 7. Some possible resonance forms of 7 are presented in Figure 1.3.39 In the discussion about oxygenation mechanisms of different substrates in chapter 2, this species is simplified as [Fe=O]3+ and is presumed to be the active intermediate in all mechanisms presented.

+.

O

FeIV

SCys

N

N

N

NO

FeIV

SCys

N

N

N

NO

FeV

SCys

N

N

N

N

.

Figure 1.3 Resonance forms of the electrophilic, high valent iron–oxo species, believed by many to be the active intermediate in most cytochrome P450 catalyzed oxidations.

1.3 Reactive oxygen species (ROS)

The importance of reactive oxygen species (ROS) in health and disease has been recognized by every branch of medicine and biological science. Overwhelming evidence indicates that ROS play a role in many major health problems and that inhibition of oxidative damage to molecules, cells, and tissues prevent chronic and degenerative diseases.47 Some examples of the many disorders that have been associated with oxidative stress are Parkinson's disease, Alzheimer’s disease, diabetes, atherosclerosis, HIV and aging.48

Many ROS present in biological systems have the potential to induce damage. These include hydrogen peroxide (H2O2), organic hydroperoxides, hypochlorous acid (HOCl), nitric oxide (NO⋅), peroxynitrite (ONOO–), superoxide (O2⋅–) alkoxyl radicals and the hydroxyl radical (⋅OH).49

There are many endogenous sources for ROS including50i) normal aerobic respiration, i.e. reduction of O2 by mitochondria that ultimately results in H2O2 production;


7

ii) stimulated polymorphonuclear leukocytes and macrophages releasing superoxide, supposedly to protect the body from bacteria;51 iii) peroxisomes, organelles responsible for degrading fatty acids and other molecules that produce hydrogen peroxide as a by-product; iv) induction of cytochrome P450 enzymes resulting in oxidant by-products.52 One example of an enzyme that is very active in the formation of ROS is CYP2E1, a cytochrome P450 isoenzyme that can be induced by chronic alcohol consumption.53 There is also a large range of exogenous sources for reactive oxygen species in vivo such as tobacco smoke, ionizing radiation, pollutants, organic solvents, anesthetics, hyperoxic environment and pesticides.

The most reactive species of the ROS is the hydroxyl radical on which this brief introduction will be focused. Hydroxyl radicals are produced in living organisms in various ways, often as a result of degradation of other reactive oxygen species. Some in vivo sources of hydroxyl radicals are given below. 1) The Haber Weiss reaction.54

OHOH -O2 H2O2+ O2 + + .-. -.

2) Homolysis of peroxynitrite50 which is generated in vivo from the reaction between O2⋅– and NO⋅.

O2

+ NO O N O O - O N O OHH +

.-. -.

The protonated product is unstable and is cleaved homolytically.

OHO N O OH NO2NO2+ ..

3) Reaction between O2⋅– and HClO.55

HOCl O2 Cl -+ + +O2

-.O2

-. -.OH. OH.

The most important source of hydroxyl radicals in a cellular environment is probably the Haber–Weiss reaction. The overall reaction presented above is somewhat more complex


8

and a more thorough description of the iron catalyzed Haber–Weiss reaction is presented below.49

Fe 2+Fe3+ +

OH-

H2O2

+ O2

+ +Fe 2+ + Fe3+ (the Fenton reaction)

O2

-.O2

-. -.

OH. OH.

Superoxide is produced in vivo through reduction of molecular oxygen by various processes as discussed earlier. Hydrogen peroxide is essential for the last step of the Haber–Weiss reaction, the Fenton reaction. A number of enzymes including amino acid oxidase, amine oxidase and glucose oxidase generate hydrogen peroxide directly.56 Other enzymes such as xanthine oxidase, aldehyde oxygenase and NADPH oxygenase generate superoxide that can be converted to hydrogen peroxide as follows.57

H ++ HO2

HO2+ Fe 2+ Fe3++HO2

H2O2H +HO2 +

O2

-.O2

-. -. .

. -. -.

-. -.

A transition metal ion in its lower valence state is needed for the last step in the Haber Weiss reaction, the Fenton reaction. A number of ions, such as Fe2+, Cu+, Ti3+, Cr2+ and Co2+, can donate an electron to hydrogen peroxide.55 In vivo, only iron and copper ions are present in high enough concentrations to play a major role in the generation of hydroxyl radicals. Iron is the most abundant transition metal in vivo, and thereby biologically most relevant. In the mammalian cell nucleus, however, the concentration of Cu2+ is high and, therefore, copper is believed to be largely responsible for oxidative damage to DNA and membranes in the nucleus.58 The location of transition metal ions in vivo is often a crucial parameter that decides where oxidative damage takes place.59,60

Biological systems are very complex and in order to gain information about the reactivity of different molecules towards hydroxyl radicals, several in vitro systems have been developed for hydroxyl radical generation. The most common ways of generating hydroxyl radicals in vitro are presented below.61 1) Radiolysis of aqueous solutions in the presence of N2O or H2O2 is the cleanest way of generating hydroxyl radicals and is often the method of choice for kinetic studies.62 2) Photolysis of hydrogen peroxide.63


9

3) Chemical reactions, mainly from the reaction between hydrogen peroxide and transition metals in their lower valence states as for example Fe2+ and Cu+. This reaction is known as the Fenton reaction and is by far the most used chemical method.64 Although it has been over 100 years since Fenton performed his experiments, the mechanism of the reaction is still debated. It is clear that the net Fenton reaction is a simplification of a more complex process. The easiest way to explain the reaction would be a single electron donation from Fe2+ to hydrogen peroxide (outer sphere mechanism).

L Fe 2+ + H2O2L Fe 3+ + +OH - OHH2O2

-. -. .

On thermodynamic grounds this mechanism is unfavorable and an inner sphere mechanism involving a transient ferrous peroxide complex is more likely.65,66

L Fe (H2O2)2+L Fe 2+ H2O2+

The complexing ligand, represented by (L) in the above scheme, might be any chemical or biological chelator and is of great importance to the outcome of the reaction. The iron–hydrogen peroxide complex could then break down in a number of ways: 1) to give a hydroxide ion and a hydroxyl radical

L Fe (H2O

2)2+

L Fe 3+ + +OH - OH.

2) to give two hydroxide ions65

L Fe (H2O2)2+ L Fe 4+ + 2 OH

-

3) direct reaction with a substrate

L Fe (H2O2)2++ R OH

-+L Fe 3+ + R OH.


10

The choice of the ligand, pH and the type of solvent used are of great importance for the outcome of the reaction. In a chemical Fenton system most variables are easy to control but in a biological system, the iron will form complexes with numerous different biological chelators, and the type of ligand will decide the fate of the reaction. Ferrous ions are stable in acidic solutions but in neutral or basic aqueous solutions a ligand is generally added to a chemical Fenton system to prevent precipitation of ferrous and ferric hydroxides. For the experiments described in chapters 3, 4 and 5 ethylenediaminetetraacetic acid (EDTA) was used. Under these conditions, formation of hydroxyl radicals via break–down of the iron–hydrogen peroxide complex to give a hydroxide ion and a hydroxyl radical (pathway 1) is expected to be the main reaction pathway. Regardless if the Fenton reaction takes place via “free” hydroxyl radicals or via direct reaction with a transient Fe (H2O2)2+ species, the Fenton reaction remains an excellent system for the incorporation of oxygen into organic molecules.

The hydroxyl radical has a half–life of about one nanosecond and reacts very quickly with most organic and biological molecules, generally with rate constants above 107M–1s–1.61 The reactions can be divided into four different types: 1) radical–radical reactions

.OH H2O2.OH +

2) hydrogen atom abstraction; the hydroxyl radical can extract a hydrogen atom from various organic compounds and thereby generating a new radical

.OHR-H + R + H2O.

3) addition reactions; the hydroxyl radical is electrophilic and adds readily to aromatic rings and double bonds67

R CH

CH2.OH R CH C

H2

OH+.

4) electron transfer reactions; the hydroxyl radical is a powerful oxidant and commonly reacts with inorganic compounds by abstracting an electron.

.OH OH_

Fe2+ + Fe3+ +

In complex systems various combinations of the above described reactions often take place. In reactions with organic compounds, hydrogen atom abstraction from aliphatic


11

groups and addition to unsaturated π–systems are the predominant reactions. Since both these processes generate a new radical, the end products of reactions between hydroxyl radicals and organic molecules are dependent on the molecular structure of the organic compound and the nature of the surrounding media.

1.4 Antioxidants

Antioxidant activity can be defined as the protection against oxidative damage.47 There are three major ways for an antioxidant to protect biomolecules from damage induced by oxidative stress in vivo: i) suppression of the formation of ROS i.e. compounds that chelate iron in such a way that it becomes less active in Fenton chemistry, ii) scavenging of radicals and iii) repairing oxidative damage.68

When it comes to the hydroxyl radical, the high reactivity excludes radical scavenging as an option unless the radical scavenger is added in extremely large amounts on the basis of competition kinetics.69 If the radical scavenger on the other hand also acts as an iron chelating agent, it will be very close to the place of radical formation and may be effective in scavenging hydroxyl radicals even at low concentrations.

To guard against damage by ROS, organisms have developed a number of antioxidant enzymes including superoxide dismutase (SOD), catalase and glutathione peroxidase.70 The main non–enzyme antioxidants in the human body are the water soluble ascorbic acid (vitamin C) and uric acid and α–tocopherol (vitamin E) in membranes.71 Free methionine and surface exposed methionine residues of all proteins in a cell may also constitute an important mechanism for the intracellular scavenging of ROS.72

1.5 Scope of the thesis

The introduction of combinatorial chemistry has resulted in a dramatic increase in the number of new chemical entities that support drug discovery efforts. To limit the cost of drug development, it is vital to try to eliminate as many compounds as possible at the early stages of the discovery process. One important consideration in high–throughput screening is metabolic stability, and, in case of instability, identification of the metabolites. The main goal of our studies has been the development of electrochemical techniques on–line with electrospray mass spectrometry for in vitro generation and characterization of drug metabolites. These purely instrumental methods may have advantages over, or may be complementary to, the existing methods of screening, i.e., in vitro studies with purified enzymes or organ fractions e.g. microsomes, hepatocytes and liver slices.


12

References

1. S. Bruckenstein; R. R. Gadde, "Use of a porous electrode for in situ mass spectrometric determination of volatile electrode reaction products", J. Am. Chem. Soc., 1971, 93, 793-794.

2. G. Hambitzer; J. Heitbaum, "Electrochemical thermospray mass spectrometry", Anal. Chem., 1986, 58, 1067-1070.

3. T. A. Getek, "The development of thermospray mass spectrometry combined with electrochemical redox processes", Proceedings of the 34th ASMS Annual Conference on Mass Spectrometry and Allied Topics, Cincinnati, 1986.

4. K. J. Volk; R. A. Yost; A. Brajter-Toth, "On-line electrochemistry/thermospray/tandem mass spectrometry as a new approach to the study of redox reactions: The oxidation of uric acid", Anal. Chem., 1989, 61, 1709-1717.

5. K. J. Volk; R. A. Yost; A. Brajter-Toth; J. A. Freeman, "A dynamic approach for studying redox reactions: electrochemistry on-line with mass spectrometry", Analusis, 1992, 20, 421-425.

6. K. J. Volk; R. A. Yost; A. Brajter-Toth, "Electrochemistry On Line with Mass Spectrometry, Insite into Biological Redox Reactions", Anal. Chem., 1992, 64, 21A-33A.

7. G. J. V. Berkel; S. A. McLuckey; G. L. Glish, "Electrochemical origin of radical cations observed in electrospray ionization mass spectra", Anal. Chem., 1992, 64, 1586-1593.

8. F. Zhou; G. J. V. Berkel, "Electrochemically enhanced electrospray ionization mass spectrometry", Proceedings of the 42nd ASMS Conference on Mass Spectrometry and Allied Topics, Chicago, Illinois, 1994.

9. F. Zhou; G. J. V. Berkel, "Electrochemistry Combined On-Line with Electrospray Mass Spectrometry", Anal. Chem., 1995, 67, 3643-3649.

10. G. J. V. Berkel; F. Zhou, "Electrospray as a controlled-current electrolytic cell: Electrochemical ionization of neutral analytes for detection by electrospray mass spectrometry", Anal. Chem., 1995, 67, 3958-3964.

11. G. J. V. Berkel; F. Zhou, "Characterization of an electrospray ion source as a controlled-current electrolytic cell", Anal. Chem., 1995, 67, 2916-2923.

12. A. M. Bond; R. Colton; A. D'Agostino; A. J. Downard; J. C. Treager, "A role for electrospray mass spectrometry in electrochemical studies", Anal. Chem., 1995, 67, 1691-1695.

13. L. Baczynskyj, "EC-LC-ESI MS: Electrochemistry Combined with Liquid Chromatography and Electrospray Mass Spectrometry", Proceedings of the 44th ASMS Conference on Mass Spectrometry and Allied Topics, Portland, Oregon, 1996.

14. W. Lu; X. Xu; R. B. Cole, "On-line linear sweep voltammetry-electrospray mass spectrometry", Anal. Chem., 1997, 69, 2478-2484.


13

15. X. Xu; W. Lu; R. B. Cole, "On-Line Probe for Fast Electrochemistry/Electrospray Mass Spectrometry. Investigation of Polycyclic Aromatic Hydrocarbons", Anal. Chem., 1996, 68, 4244-4253.

16. J. Zhou; S. Hefta; T. D. Lee, "High sensitivity analysis of phenylthiohydantoin amino acid derivatives by electrospray mass spectrometry.", J. Am. Soc. Mass. Spectrom., 1997, 8, 1165-1174.

17. M. C. S. Regino; A. Brajter-Toth, "An Electrochemical Cell for On-Line Electrochemistry/Mass Spectrometry", Anal. Chem., 1997, 69, 5067-5072.

18. M. C. S. Regino; A. Brajter-Toth, "Real time characterization of catalysis by on–line electrochemistry/mass spectrometry. Investigation of quinone electrocatalysis of amine oxidation", Electroanalysis, 1998, 11, 374-379.

19. M. Regino; C. Weston; A. Brajter-Toth, "Effect of mobile phase composition on the electrochemical cell conversion efficiency in electrochemistry/mass spectrometry", Anal. Chim. Acta, 1998, 369, 253-262.

20. T. Zhang; A. Brajter-Toth, "On-line investigation of the generation of nonaqueous intermediate radical cations by electrochemistry/mass spectrometry", Anal. Chem., 2000, 72, 2533-2540.

21. G. Hambitzer; J. Heitbaum; I. Stassen, "Electrochemical termospray mass spectrometry-anodic oxidations of N,N-dialkylanilines", J. Electroanal. Chem., 1998, 447, 117-124.

22. I. Stassen; G. Hambitzer, "Anodic oxidation of aniline and N-alkylanilines in aqueous sulphuric acid studied by electrochemical thermospray mass spectrometry", J. Electroanal. Chem., 1997, 440, 219-228.

23. G. Hambitzer; J. Heitbaum; I. Stassen, "Electrochemical thermospray mass spectrometry instrumentation for coupling electrochemistry to mass spectrometry", Anal. Chem., 1998, 70, 838-842.

24. U. Jurva; A. P. Bruins; H. V. Wikström, "The ESA electrochemical cell coupled with electrospray ionization mass spectrometry", Proceedings of the 47th ASMS Conference on Mass Spectrometry and Allied Topics, Dallas, Texas, 1999.

25. U. Jurva; A. P. Bruins; H. V. Wikström, "In vitro mimicry of metabolic oxidation reactions by electrochemistry/mass spectrometry", Rapid commun. mass spec., 2000, 14, 529-533.

26. U. Jurva; A. P. Bruins; H. V. Wikström, "In vitro mimicry of metabolic oxidation reactions by electrochemistry/mass spectrometry", Presented at the15th International Mass Specttrometry conference, Barcelona, Spain, 2000.

27. H. Iwahashi, "3-Hydroxyanthranilic acid-derived compounds formed through electrochemical oxidation", J. Chrom. B, 1999, 736, 237-245.

28. H. Deng; G. J. V. Berkel, "A Thin-Layer Electrochemical Flow Cell Coupled On-Line with Electrospray-Mass Spectrometry for the Study of Biological Redox Reactions.", Electroanalysis, 1999, 11, 857-865.


14

29. J. R. Pretty; H. Deng; D. E. Goeringer; G. J. v. Berkel, "Electrochemically modulated preconcentration and matrix elimination for organic analytes coupled on-line with electrospray mass spectrometry", Anal. Chem., 2000, 72, 2066-2074.

30. H. Deng; G. J. v. Berkel; H. Takano; D. Gazda; M. D. Porter, "Electrochemically modulated liquid chromatography coupled on-line with electrospray mass spectrometry", Anal. Chem., 2000, 72, 2641-2647.

31. G. Diehl; A. Liesener; U. Karst, "Liquid chromatography with post-column electrochemical treatment and mass spectrometric detection of non-polar compounds", Analyst, 2001, 126, 288-290.

32. G. Diehl; U. Karst, "Combining HPLC/MS with electrochemistry for sensitive detection of Ferrocene-labeled alcohols and phenols", Proceedings of the 49th ASMS Conference on Mass Spectrometry and Allied Topics, Chicago, Illinois, 2001.

33. G. Diehl, PhD Thesis, University of Twente, the Netherlands, 2002. 34. H. Hayen; U. Karst, "Determination of phenothiazine and its derivatives by means of on–

line HPLC/electrochemistry/mass spectrometry", Proceedings of the 49th ASMS Conference on Mass Spectrometry and Allied Topics, Chicago, Illinois, 2001.

35. P. Brown; J. Rollag; P. Lin; J. Lee, "Oxidation and reduction of compounds using electrochemical detector to enhance LC/MS quantification.", Proceedings of the 49th ASMS Conference on Mass Spectrometry and Allied Topics, Chicago, Illinois, 2001.

36. S. P. Palii; T. Zhang; A. Brajter-Toth; J. R. Eyler, "Investigation of electrochemically assisted ionization (ECAI) in ESI FT-ICR mass spectrometry", Proceedings of the 49th ASMS Conference on Mass Spectrometry and Allied Topics, Chicago, Illinois, 2001.

37. T. Zhang; S. P. Palii; J. R. Eyler; A. Brajter-Toth, "Enhancement of Ionization efficiency by electrochemical reaction products in on-line electrochemistry/electrospray ionization Fourier transform cyclotron resonance mass spectrometry", Anal. Chem., 2002, 74, 1097-1103.

38. K. A. Johnson; B. A. Shira; J. L. Anderson; I. J. Amster, "Chemical and on–line electrochemical reduction of metalloproteins with high resolution electrospray ionization mass spectrometry detection", Anal. Chem., 2001, 73, 803-808.

39. B. Meunier; J. Bernadou, "Active Iron-Oxo and Iron-Peroxo Species in Cytochrome P450 and Peroxidases; Oxo-hydroxo Tautomerism with Water-Soluble Metalloporphyrins", Struct. Bond., 2000, 97, 1-35.

40. P. R. O. d. Montellano "Cytochrome P450, Structure, Mechanism, and Biochemistry" second ed.; Plenum Press, 1995.

41. D. F. V. Lewis; J. M. Pratt, "The P450 catalytic cycle and oxygenation mechanism", Drug Metab. Rev., 1998, 30, 739-786.

42. M. Sono; M. P. Roach; E. D. Coulter; J. H. Dawson, "Heme-Containing Oxygenases", Chem. Rev., 1996, 96, 2841-2887.


15

43. J. T. Groves; Y. Z. Han, "Models and Mechanisms of Cytochrome P450 Action", In "Cytochrome P450, Structure, Mechanism, and Biochemistry" second ed.; Montellano, P. R. O. d., Ed.; Plenum Press:, 1995.

44. D. L. Wertz; J. S. Valentine, "Nucleophilicity of Iron-Peroxo Porphyrin Complexes", Struct. Bond., 2000, 97, 37-60.

45. I. Schlichting; J. Berendzen; K. Chu; A. M. Stock; S. A. Maves; D. E. Benson; R. M. Sweet; D. Ringe; G. A. Petsko; S. G. Sligar, "The Catalytic Pathway of Cytochrome P450cam at Atomic Resolution", Science, 2000, 287, 1615-1622.

46. M. Newcomb; R. Shen; S. Y. Choi; P. H. Toy; P. F. Hollenberg; A. D. N. Vaz; M. J. Coon, "Cytochrome P450-Catalyzed Hydroxylations of Mechanistic Probes that Distinguish between Radicals and Cations. Evidence for Cationic but not for Radical Intermediates.", J. Am. Chem. Soc., 2000, 122, 2677-2686.

47. L. Packer (Ed.) "Oxidants and Antioxidants, Part A" ; Academic Press, 1999. 48. D. Armstrong (Ed.)"Free Radical and Antioxidant Protocols" ; Humana Press, 1998. 49. J. P. Kehrer, "The Haber-Weiss reaction and mechanism of toxicity", Toxicol., 2000, 149,

43-50. 50. H. Alho; J. Leinonen, "Total Antioxidant Activity Measured by Chemiluminiscence

Methods", In "Oxidants and Antioxidants, Part A" ; Packer, L., Ed.; Academic Press:, 1999; Vol. 299.

51. N. Gotho; E. Niki, "Measurement of Superoxide Reaction by Chemiluminiscence", In "Oxygen Radicals in Biological Systems, Part C" ; Packer, L., Ed.; Academic Press:, 1994; Vol. 233.

52. M. S. Paller; H. S. Jacob, "Cytochrome P-450 mediates tissue-damaging hydroxyl radical formation during reoxygenation of the kidney", Proc. Natl. Acad. Sci. USA, 1994, 91, 7002-7006.

53. E. Albano; A. Tomasi; M. Ingelman-Sundberg, "Spin Trapping of Alcohol-derived Radicals in Microsomes and Reconstructed Systems", In "Oxygen Radicals in Biological Systems, Part C" ; Packer, L., Ed.; Academic Press:, 1994; Vol. 233.

54. F. Haber; J. Weiss, "The catalytic decomposition of hydrogen peroxide by iron salts", Proc. R. Soc. London [Biol], 1934, A147, 332-351.

55. P. Wardman; L. P. Candeias, "Fenton Chemistry: An Introduction", Radiat. Res., 1996, 145, 523-531.

56. H. Yokoyama, "In vivo measurement of hydrogen peroxide by microelectrode", In "Oxidants and Antioxidants, Part B" ; Packer, L., Ed.; Academic Press:, 1999; Vol. 300.

57. A. L. Feig; S. J. Lippard, "Reactions of non-heme iron(II) centers with dioxygen in biology and chemistry", Chem. Rev., 1994, 94, 759-805.

58. M. J. Burkitt, "Copper-DNA Adducts", In "Oxygen Radicals in Biological Systems, Part D" ; Packer, L., Ed.; Academic Press:, 1994; Vol. 234.


16

59. C. Demougeot; C. Marie; A. Beley, "Importance of iron location in iron-induced hydroxyl radical production by brain slices", Life Sciences, 2000, 67, 399-410.

60. D. R. Lloyd; D. H. Phillips, "Oxidative DNA damage mediated by copper(II), iron(II) and nickel(II) Fenton reactions: evidence for site-specific mechanisms in the formation of double strand breaks, 8-hydroxydeoxyguanosine and putative intrastrand cross-links", Mutat. Res., 1999, 424, 23-36.

61. G. Czapski, "Reaction of OH radical", In "Oxygen Radicals in Biological Systems, Part A" ; Packer, L., Ed.; Academic Press:, 1984; Vol. 105.

62. K. P. Madden; H. Taniguchi, "The role of the DMPO-hydrated electron spin adduct in DMPO-OH spin trapping", Free Rad. Biol. Med., 2001, 30, 1374-1380.

63. R. V. Lloyd; P. M. Hanna; R. P. Mason, "The origin of the hydroxyl radical oxygen in the Fenton reaction", Free rad. biol. med., 1997, 22, 885-888.

64. H. J. H. Fenton, "Oxidation of tartaric acid in presence of iron", J. Chem. Soc., 1894, 65, 899-910.

65. C. C. Winterbourn, "Toxicity of iron and hydrogen peroxide: the Fenton reaction", Toxicol. Lett., 1995, 82/83, 969-974.

66. S. Goldstein; D. Meyerstein; G. Czapski, "The Fenton Reagents", Free Rad. Biol. Med., 1993, 15, 435-445.

67. C. v. Sonntag; H.-P. Schuchmann, "Pulse radiolysis", In "Oxygen Radicals in Biological Systems, Part C" ; Packer, L., Ed.; Academic Press:, 1994; Vol. 233.

68. L. C. Bourne; C. A. Rice-Evans, "Detecting and Measuring Bioavailability of Phenolics and Flavanoids in Humans: Pharmacokinetics of Urinary Excretion of Dietary Ferulic acid.", In "Oxidants and Antioxidants, Part A" ; Packer, L., Ed.; Academic Press:, 1999; Vol. 299.

69. C. v. Sonntag; H.-P. Schuchmann, "Suppression of hydroxyl radicals in biological systems: Considerations based on competition kinetics", In "Oxygen Radicals in Biological Systems, Part C" ; Packer, L., Ed.; Academic Press:, 1994; Vol. 233.

70. H. Z. Chae; S. W. Kang; S. G. Rhee, "Isoforms of Mammalian Peroxiredoxin that reduce Peroxides in Presence of Thioredoxin", In "Oxidants and Antioxidants, Part B" ; Packer, L., Ed.; Academic Press:, 1999; Vol. 300.

71. J. P.A. Glasscott; J. L. Farber, "Assessment of Physiological Interaction between Vitamin E and Vitamin C.", In "Oxidants and Antioxidants, Part B" ; Packer, L., Ed.; Academic Press:, 1999; Vol. 300.

72. J. Moskovitz; B. S. Berlett; J. M. Poston; E. R. Stadtman, "Methionine Sulfixide Reductase in Antioxidant Defense", In "Oxidants and Antioxidants, Part B" ; Packer, L., Ed.; Academic Press:, 1999; Vol. 300.

Chapter 2 Comparison between EC/MS and cytochrome P450 catalyzed oxidation reactions

17

Chapter 2

Comparison between electrochemistry/mass

spectrometry and cytochrome P450 catalyzed

oxidation reactions

In these studies we have investigated to what extent electrochemistry on–line with electrospray mass spectrometry can be used to mimic cytochrome P450 catalyzed oxidations. Comparisons on a mechanistic level have been made for most reactions in an effort to explain why certain reactions can, and some can not, be mimicked by electrochemical oxidations. The EC/MS/MS system provided successful mimics in cases where the P450 catalyzed reactions are supposed to proceed via a mechanism initiated by a one electron oxidation, such as N–dealkylation, S–oxidation, P–oxidation, alcohol oxidation and dehydrogenation. The P450 catalyzed reactions thought to be initiated via direct hydrogen atom abstraction, such as O–dealkylation and hydroxylation of unsubstituted arenes generally had a too high oxidation potential to be electrochemically oxidized below the oxidation potential limit of water, and were not mimicked by the EC/MS/MS system. Even though the EC/MS/MS system is not able to mimic all oxidations performed by cytochrome P450, valuable information can be obtained concerning the sensitivity of the substrate towards oxidation and the regions on the molecule where oxidations are likely to take place. For small–scale electrochemical synthesis of metabolites, starting from the drug, the EC/MS/MS system should be very useful for quick optimization of the electrochemical conditions. The simplicity of the system, and the ease and speed with which it can be applied to a large number of compounds, make it a useful tool in drug metabolism research. 2.1 Introduction

Drug metabolism is normally divided into phase I, II and III metabolism. Phase I metabolism involves primarily functionalization, including oxidation, reduction, hydrolysis and isomerization. Phase II metabolism is often related to conjugation, including glucuronidation, sulfation, glutathione conjugation, methylation, acetylation and condensation. The term phase III metabolism is used to describe transport processes, such as biliary and renal excretion. The most important phase I metabolic pathways are enzyme


18

catalyzed oxidations. Several enzymes are involved and the single most important enzyme system is cytochrome P450.

There is a large body of knowledge on electrochemical oxidations of organic compounds1 and the aim of our studies was to investigate the potential of an electrochemical system to mimic phase I oxidative metabolism. Results presented by Shono et. al suggest electrochemical oxidation as a mild synthetic route to N–dealkylated drug metabolites and it has been demonstrated that there is analogy between cytochrome P450 catalyzed N–dealkylation and electrochemical oxidation of amines and amides.2-4. Watanabe et al. have shown that the rates of sulfoxide formation for a number of para–substituted thioanisoles by a reconstructed system with purified cytochrome P450 were correlated with the electrochemical oxidation potentials.5 They observed the same correlation for the oxidation of the corresponding sulfoxides to sulfones. A correlation between the cytochrome P450 oxidation rates and the electrochemical oxidation potentials in a series of N,N-dimethyl-anilines has been reported by Guengerich and Macdonald.6,7

In 1986, Hambitzer et al. studied the redox reactions of N,N–dimethyl-aniline in an electrochemical cell on–line with thermospray mass spectrometry.8 In the same year Getek used electrochemistry on–line with thermospray mass spectrometry to study the oxidation of acetaminophen.9 With the addition of glutathione or cysteine, these authors also were able to mimic phase II metabolism by the conjugation of glutathione and cysteine with the oxidation products of acetaminophen.10 The early developments in electrochemistry/mass spectrometry (EC/MS) for the simulation of metabolism have been reviewed by Volk et al.11 Despite the potential of the technique, only a few papers have been published on EC/MS in the field of drug metabolism.12-18 There could be several reasons for this.

First, insufficient quality of instrumentation may have prevented routine use. Thermospray mass spectrometry requires tight control of the vaporizer temperature, which depends on solvent composition and condition of the vaporizer. As a result, long–term stability and ease of use are difficult to achieve. High flow rates used in thermospray lead to a short residence time of samples in the electrochemical cell, and may give rise to low conversion efficiency. Electrospray MS provides high sensitivity and reliability at the low flow rates required for high electrochemical reaction efficiency.

Second, developments in the field of molecular biology have provided time efficient alternatives to in vivo studies. Liver microsomes and organ slices are used widely for the study of enzymatic oxidation reactions, and individual cytochrome P450 enzymes are available for detailed in vitro reaction studies.19

Third, the study of metabolism of drugs and new chemical entities was previously done in the later stages of drug development when the number of drug candidates had been reduced such that the time and cost involved in a real in vivo metabolism study was not considered to be prohibitive. Now, drug metabolism is considered much earlier in drug


19

discovery requiring simple tests of a large number of compounds and, with the developments in instrumentation, a reevaluation of EC/MS may be appropriate.

The introduction of combinatorial chemistry has resulted in a dramatic increase in the number of new chemical entities. To limit the cost of drug development, it is vital to try to eliminate as many compounds as possible at the early stages of the discovery route. One important consideration in high–throughput screening is metabolic stability, and, in case of instability, identification of the main metabolite. EC/MS, because it is a purely instrumental method, may have advantages over, or be complementary to, the existing methods of screening, i.e., in vitro studies with recombinantly expressed enzymes or organ fractions e.g. microsomes, hepatocytes and liver slices.

Cytochrome P450 enzymes are present in virtually every mammalian tissue and organ, and they play an important role in the oxidation of endogenous substances as well as a tremendous range of drugs and xenobiotics.20-23 The cytochrome P450 enzymes catalyze diverse types of reactions, such as hydroxylation of aliphatic and aromatic compounds, dealkylation and oxygenation of heteroatom containing compounds, oxidation of alcohols and aldehydes etc.24,25 The mechanisms for most of the enzymatic reactions are still debated and different mechanistic pathways have been suggested for many of the reactions shown in this article. All of the mechanisms shown here are based on the assumption that the active oxygen intermediate of Cytochrome P450 is the electrophilic oxoferryl porphyrin cation radical shown in Figure 2.1. For simplicity of presentation, this species is presented as [Fe=O]3+ in the figures.

Cys

Fe

S

O

NN

HO

O

O

HON N

+.+.

Figure 2.1 The electrophilic oxoferryl porphyrin cation radical, proposed reactive species of

cytochrome P450. In the experiments described below, we have tried to mimic these different enzymatic oxidations in an EC/MS system. Comparisons on the mechanistic level are made for


20

several reactions in an effort to explain why certain reactions can, and some can not, be mimicked by electrochemical oxidations. 2.2 Experimental

Chemicals N–0923 is the S(–)enantiomer of the dopamine agonist N–0437 [2-(N-propyl-N-2-thienylethylamino)-5-hydroxytetralin] and has been synthesized earlier.26 Amphetamine sulfate and mephenytoin were generous gifts from the Department of Pharmaceutical Analysis, University of Groningen, the Netherlands. Xanthohumol was a generous gift from Dr. J. F. Stevens at the Institute of Plant Biochemistry, Halle, Germany and Dr. C.L. Miranda at Oregon State University, Corvallis, Oregon, USA. The following chemicals were obtained commercially: 2–acetamidofluorene, Sigma; benzo[a]pyrene, 97 %, Aldrich; ammonia, 25 % solution in water, p.a., MERCK; ammonium acetate, p.a., MERCK; coumarin, Aldrich; dibutylsulfide, 96 %, Aldrich; dibutylsulfoxide, 96 %, Aldrich; diclofenac, Ciba–Geigy B.V.; diphenylsulfide, 98 %, Aldrich; diphenylsulfoxide, 96 %, Aldrich; N,N–diethyl nicotineamide, 97 %, Janssen; N,N–dimethylaniline, 99 %, Acros; 7–ethoxycoumarine, >99%, Fluka; glacial acetic acid, p.a., MERCK; hexylamine, 99 %, Fluka; 2–(hydroxymethyl)–pyridine (2–pyridylcarbinol), 98 %, Aldrich; p–hydroxyphenethylamine (tyramine), 97 %, Acros; lauric acid, 99.5 %, Aldrich; lidocaine, >99%, Inter Pharm; 2,3–lutidine, 99 %, Aldrich; 7–methoxycoumarine, 98 %, Aldrich; 6–methylmercaptopurine, Sigma; p–nitrophenol, spectrophotometric grade, Sigma; paracetamol (acetaminophen), ICN Biochemicals Inc.; parathion, Riedel de Haën; phenethylamine, 99 %, Janssen; 2–picoline, 98 %, Aldrich; picolinic acid, 99 %, Aldrich; 2–picolylchloride hydrochloride, 98 %, Aldrich; pyridine, 99 %, Acros; Pyridine–2–aldehyde, 99 %, Aldrich (2–pyridinecarboxaldehyde); thioanisole, 99 %, Janssen. The water used in the experiments was purified in a Maxima Ultrapure water system (ELGA, High Wycombe, Bucks, U.K.) and was sonicated for about 15 minutes before use.


21

Instrumentation Mac Lab

cell control and read-out

ESACoulochem5011 cell

Potentiostat

LoopInjector Sciex

API 3000LC/MS/MS

50 µl/min 0.5 ml 0.0 to +1.5 Vin 5 minutes

TurboIonSprayFull scan spectraProduct Ion scans

Pump

Figure 2.2 The EC/MS system

A schematic overview of the EC/MS system is given in Figure 2.2. A Series 200 micro LC pump (Perkin Elmer, Norwalk, CT, USA) delivered a flow of 50 µl/min. The analytes were injected with a 0.5 ml injection loop and passed through an ESA Coulochem 5011 analytical cell (ESA Inc., Bedford, Massachusetts, USA) connected to a home-made potentiostat. A MacLab system with Chart 3.5.7 software (AD Instruments, Castle Hill, NSW, Australia) was used to control the potentiostat and to apply the desired potential over the electrochemical cell. The ESA working electrode is porous graphite and all reported cell potentials were recorded versus a palladium reference electrode. A program was made in Chart to ramp the potential from 0 to +1500 mV in steps of 10 mV during 5 minutes. The cell current was continuously registered by Chart. Full scan spectra or product ion spectra, were taken continuously with an API 3000 triple quadrupole mass spectrometer (MDS–Sciex, Concord, Ontario, Canada) equipped with a TurboIonSpray interface. The MS was operated at such a low orifice voltage that “up front” collision induced dissociation did not take place. For certain compounds, such as benzo[a]pyrene and the sulfides discussed in the results and discussion section, no ions originating from the sample could be detected with electrospray. For these compounds, an APCI (Atmospheric Pressure Chemical Ionization) interface was used. The oxidation products of the compounds described above were all easily detected. The delay between the electrochemical cell and the mass spectrometer was determined as follows. At a continuous flow of analyte, a potential step from 0 to +1000 mV was performed. The time between the potential step and the appearance of the oxidation product at the mass spectrometer was measured by selected ion monitoring of the main oxidation product. With the delay between the electrochemical cell and the mass spectrometer determined to be 45 s, spectra


22

can be shown for any given potential and the signal from the different oxidation products can be extracted from the full scan data file and plotted against the potential. Electrochemical reaction conditions All compounds were tested in the following aqueous supporting electrolyte solutions: acetic acid, ammonium acetate, and ammonia. The supporting electrolytes were used in 1, 10 and 100 mM concentrations. To enhance solubility of the analyzed compounds and the oxidation products, 20 % acetonitrile was added to all electrolyte solutions. To keep benzo[a]pyrene in solution, 90 % acetonitrile was added. Generally, a high electrolyte concentration gave a better electrochemical conversion of the sample but also gave increased background ion signals and decreased the sample ion signal. Most oxidations were found to be pH dependent. Some amines, for example, are not oxidized at all in 100 mM acetic acid, while they are readily oxidized in neutral or basic solutions. 2.3 Results and discussion

Table 2.1 Aliphatic hydroxylation, aromatic hydroxylation and epoxidation

Enzyme catalyzed oxidation Electrochemical oxidation products

a) Aliphatic hydroxylation Ex. CYP4A11 catalyzed 12-hydroxylation of lauric acid

C9H18 OH

O

mw 200 mw 216

C9H18HO

OH

O

No electrochemical reaction

b) Aromatic hydroxylation (1) Ex. N-0923, aromatic hydroxylation in rat liver experiment

OH

N S

OH

N S

HO

mw 315 mw 331

N S

O

O

mw 329 c) Aromatic hydroxylation (2)

Ex. CYP2C19 catalyzed 4-hydroxylation of mephenytoin

N

NH

O

O

N

NH

O

O

HO

mw 218 mw 234


d) Epoxidation Ex. 4,5-epoxidation of benzo[a]pyrene

Omw 252 mw 266

O

O

mw 282


23

Aliphatic hydroxylation The CYP4A11 catalyzed 12–hydroxylation of lauric acid is described in Table 2.1a.27 No oxidation was observed in the electrochemical cell. The electrochemical oxidation potentials for aliphatic hydrocarbons are generally very high28 and no electrochemical reactions on aliphatic sidechains have been observed for any of the compounds tested throughout these experiments. Hydroxylation of aliphatic C–H bonds is chemically very difficult, and the mechanism for cytochrome P450 catalyzed hydroxylations of this kind is still debated.29 Aromatic hydroxylation The metabolism of the dopamine agonist N–0923 has been studied intensively. Hydroxylation ortho to the phenolic group to give the corresponding catechol metabolite, shown in Table 2.2.1b, is a major metabolic pathway reported from experiments with in vitro rat liver perfusion.30 Electrochemical oxidation of the phenol function yields a catechol or a p–hydroquinone at about +200 mV in 1 mM acetic acid.15 The products are immediately oxidized further to the corresponding quinones. Information about the ortho, meta, para distribution can not be obtained without the application of further techniques such as NMR and HPLC–separation.

The CYP2C19 catalyzed hydroxylation of mephenytoin is given in Table 2.1c.31 No electrochemical reaction was observed on the aromatic ring.

A range of other compounds has been tested for aromatic hydroxylation in the EC/MS system, including coumarin,32 p–nitrophenol,33 diclofenac,34 xanthohumol,35 phenethylamine, p–hydroxy–phenethylamine (tyramine) and N,N–dimethylaniline. The conclusion from these experiments is that, in order to be oxidized electrochemically within the potential limits of water, the aromatic ring has to be activated by an electron donating group, such as a hydroxyl or amino group. For an aromatic ring without an electron donating group, the starting material is considerably more difficult to oxidize than the product. Benzene is for example oxidized at a potential about 1.3 V more anodic than phenol.28,36 As a consequence, the oxidation can not be stopped at the phenol stage, but will proceed to the benzoquinone.

2-Phenylethylamine is an example of a compound containing an aromatic ring without activating groups. 2-p-Hydroxy-phenylethylamine is activated by a hydroxyl group in the para-position. The electrochemical oxidation of 2-phenylethylamine gave no hydroxylation products, but 2-p-hydroxy-phenylethylamine was oxidized to the quinone at 250 mV. Since the intermediate catechol is more easily oxidized than the starting phenol, further oxidation to the quinone can not be prevented.


24

H

+

H

π-complex

i

H+

H

OFe+

OH

H

HO

H

OH

H

O

H

OFe

H

OFe

radical cation radical

σ-complex

radical σ-complex

iii

ii

Fe O[ ]3+ Fe O[ ]3+

Fe O[ ]2+

H

Fe OH[ ]3+

.

..

.

Figure 2.3 Mechanism for oxidation of π-bonded systems by cytochrome P450 as proposed by

Guengerich and Macdonald The mechanism for the cytochrome P450 catalyzed oxidation of π–bonded systems is still debated, and different compounds will most likely react via different mechanistic pathways. Guengerich and co–workers have suggested the mechanism described in Figure 2.3 for π–bonded systems.37 The π–electrons of a C–C double bond form an intermediate with the [Fe=O]3+ form of cytochrome P450. Three reaction steps are possible for the intermediate. i) A single electron transfer to yield a radical cation of the π–system. ii) A radicaloid addition to the π–system to produce a σ–bonded radical complex. iii) An electron addition process that leads directly to a σ–complex. The nature of the π–system and the physical features of the P450 active site will dictate the specific pathway taken. In the first case, the suggested radical cation intermediate is deprotonated to yield the corresponding radical. Insertion of the “OH radical” from the formed [Fe–OH] 3+ (oxygen rebound) gives the hydroxylated compound. P450 catalyzed hydroxylation of the aromatic rings of anilines and phenols most likely proceed via this mechanism.38 Another fate of the radical cation is the formation of a σ–complex with the iron–oxygen species. This complex will react further and give hydroxylated products, epoxides etc. Nonenzymatic opening of arene oxides is probably a common reaction pathway for the formation of hydroxylated aromatic compounds.


25

OH

R

O

R

R

OH

OH

O

O

R

+

R

O

OH

R

e-- H+

-

e--2

-2H+

e--

-

H2O

H+

.

.

Figure 2.4 Electrochemical oxidation of phenols

The suggested mechanism for electrochemical oxidation of phenols is given in Figure 2.4.36 Compared to the cytochrome P450 catalyzed oxidation, it is initiated in the same way as pathway i) in Figure 2.3. Compounds that are easily electrochemically oxidized probably follow this pathway in the cytochrome P450 catalyzed oxidation, while compounds with high oxidation potentials are more likely to proceed via one of the other pathways upon oxidation by cytochrome P450. In an electrochemical system, the nature of the π–system, the physical features of the working electrode and the surrounding electrolyte, will determine the fate of the compound. Different substituents on the aromatic ring often result in different products. Phenols are generally oxidized to quinones as described in Figure 2.4 while other substituents on the aromatic ring will give a different outcome of the reaction. In the electrochemical oxidation of N,N–dimethyl-aniline, formation of dimers was favored over further oxidation and reaction with water. This can be explained by the high stability of the intermediate radical of N,N–dimethyl-aniline.39

In conclusion, the hydroxylation of aromatic compounds, as performed by cytochrome P450, shows some mechanistic resemblance to electrochemical oxidation. However, because of the lack of an active iron–oxygen species in the electrochemical system, the radical intermediate obtained after the initial oxidation and deprotonation steps will react further in various ways that will be determined by the substituents. In cases where hydroxylation takes place, such as for phenols, the products generally have a lower oxidation potential than the starting compound, and will be further oxidized. Epoxidation The cytochrome P450 catalyzed oxidation of benzo[a]pyrene yields an epoxide in the 4,5–position as described in Table 2.1d. The main electrochemical oxidation products were presumably a mixture of 1,6–, 3,6–, and 6,12–benzo[a]pyrene quinones and no epoxides were formed.40 Xu, Lu and Cole analyzed benzo[a]pyrene in an on–line electrochemical cell that generates electrochemical intermediates and products in situ at the tip of the


26

electrospray capillary needle.41 In addition to benzo[a]pyrene quinones, they also observed an ion at m/z 267 as a major oxidation product, corresponding to an intermediate 6–oxobenzo[a]pyrene cation. In our experiments, the ion at m/z 267 was present only in trace amounts. The lower yield of this oxidation product in our experiments is probably due to a higher water content (10 % compared to 0.2 % by Xu) and a longer delay time between the electrochemical cell and the mass spectrometer, that favored hydrolysis of the intermediate. The enzymatic 4,5-epoxidation of benzo[a]pyrene was thus not mimicked by the EC/MS system. It should be noted that the metabolism of polyaromatic hydrocarbons is usually very complex and is not limited to epoxidation. The carcenogenic character of many polyaromatic hydrocarbons is associated with the formation of radical cations through one electron oxidation by cytochrome P450.42-45 The end products of such oxidations are often quinones and the electrochemical oxidation of benzo[a]pyrene is probably very similar to this process.

Benzo[a]pyrene is not easily ionized in solution, and is therefore difficult to analyze with electrospray mass spectrometry. The main electrochemical oxidation products, the benzo[a]pyrene quinones, are readily protonated in aqueous solutions, and can thus be detected with electrospray. This experiment is a good example of how electrochemical oxidation can be used to enhance the signals of compounds that are otherwise difficult to detect with electrospray mass spectrometry.41,46-50 The S–oxidation of dibutylsulfide to dibutylsulfoxide described later is another example of a compound that is impossible to detect before oxidation, but has a detectable oxidation product.

Dealkylation and oxygenation of substrates containing a heteroatom:

Introduction

Fe OH[ ]3+

Fe OH[ ]3+

Fe OH[ ]3+

Fe O[ ]3+

Fe O[ ]2+

Fe O[ ]2+

Fe O[ ]3+

Fe O[ ]2+

Fe3+

Fe3+

e--

X

HH

+

-H

-H +

X

HH

O-

+

XH

+

O

H

A

B

C

D

EX

H

X

OHH

X

HH

Figure 2.5 The mechanism of cytochrome P450-catalyzed heteroatom oxidation and dealkylation

proposed by Guengerich and Macdonald. X=NR, S, or O.


27

In Figure 2.5 the proposed mechanism for cytochrome P450–catalyzed heteroatom oxidation and dealkylation is shown.20,38,51 Depending on the nature of the heteroatom, the starting compound (A) can either undergo a one electron oxidation to give a heteroatom centered cation radical (B), or loose an α–hydrogen atom to give a carbon based radical (C). The same neutral radical (C) is also obtained by deprotonation of the heteroatom centered cation radical (B). Oxygen rebound to the radical (C) gives a compound hydroxylated at the α–carbon (E). Nonenzymatically, this hydroxylated compound decomposes to the final products, the dealkylated heteroatom–containing substrate and an aldehyde. If the substrate for example is a tertiary amine, the final products will be the secondary amine and the corresponding aldehyde. The heteroatom centered cation radical (B) can also react directly with the [Fe=O]2+ complex to give an oxygenated product (D). Substrates without α–hydrogens will obviously follow this route since the pathway A–>B–>C is not possible.

In addition to the presence or absence of α–hydrogens, the outcome of the enzymatic reaction is dependent on the stability of the radical cation (B in Figure 2.5). For alkylamines, dealkylation is the main pathway in the electrochemical oxidation as well as in the enzymatic oxidation. The cation radical is better stabilized by sulfur and phosphorus, than by nitrogen. As a consequence, the main products from cytochrome P450 catalyzed oxidation, as well as from electrochemical oxidation, will be sulfoxides and phosphine oxides. If the heteroatom is oxygen, cytochrome P450 is not able to perform a one electron oxidation, due to the poor ability of oxygen to carry the positive charge. The reaction is therefore supposed to proceed via a direct hydrogen abstraction from the α–carbon (pathway A–>C–>E in Figure 2.5). A summary of our experiments on the dealkylations and oxygenations of heteroatom containing compounds is given in Table 2.2. Table 2.2 Dealkylation and oxygenation of heteroatom containing compounds


a) N-dealkylation of aliphatic amines containing α-hydrogens Ex. CYP3A4 catalyzed N-deethylation of Lidocaine

NH

N

O

NH

NH

O

mw 234 mw 206

NH

NH

O

mw 206

HN

mw 73

b) N-oxidation of nitrogen containing compounds without hydrogens at the α-carbon. Ex. N-hydroxylation of 2-acetamidofluorene

N

O

H

N

O

OHmw 223 mw 239

N

O

Hmw 239

OH


28


c) Sulfur dealkylation Ex. S-demethylation of S-methylthiopurine

N

N

N

NH

S

N

N

N

NH

SH

mw 166 mw 152

N

N

N

NH

SO

mw 182 d) Sulfur oxidation Ex. S-oxidation of sulfides to sulfoxides

S

O

S

mw 146 mw 162

S

O

mw 162 e) Phosphor oxidation Ex. Phosphothionate oxidation of parathione

O2N P S

O

O

O2N P O

O

Omw 291 mw 275

O2N P O

O

Omw 275

f) Oxygen dealkylation Ex O-deethylation of 7-ethoxycoumarine

O OO O OHOmw 190 mw 162


N–dealkylation The CYP3A4 catalyzed N–deethylation of the local anesthetic lidocaine is described in Table 2.2a.52 The reaction was readily mimicked by the EC/MS system.

If the heteroatom (X) in Figure 2.5 is a nitrogen with neighboring α–hydrogens, the main pathway is supposed to be A–>B–>C–>E. The proposed mechanism for electrochemical oxidation of aliphatic amines is described in Figure 2.6.53

RN

HH

R

+e-- e-

- RN

H

R

+H+

-

- H+

O

H

+H2O

I II III IV

IV

RN

R

HRN

H

R

+ RN

H

R

OH

RN

HH

R

RN

R

H..

.. ..

....

Figure 2.6 Electrochemical oxidation of aliphatic amines.


29

The first two steps are the same for both mechanisms. A one electron oxidation gives an aminium cation radical, which, upon deprotonation, gives an α–carbon centered radical. In the P450 catalyzed mechanism, the electron and the proton are donated to the iron–oxygen intermediate of the enzyme and the hydroxyl group is introduced on the α–carbon by subsequent reaction with the formed [Fe–OH]3+ species (radical recombination). In the electrochemical oxidation, the electron is donated to the working electrode, and the proton is lost to any compound acting as a base in solution. Since there is no active iron–oxygen intermediate present in the electrochemical system, the neutral radical (III) is further oxidized at the working electrode to the iminium ion (IV). Finally, hydrolysis of the iminium ion (IV) will give the dealkylated amine and the corresponding aldehyde.

As discussed earlier, the stability of the initial radical cation is very important for the outcome of the enzyme catalyzed oxidation. The same is true for the electrochemical oxidation mechanism. In addition to the data presented in Table 2.2, several nitrogen containing compounds have been tested for dealkylation in the EC/MS system. For aliphatic, tertiary amines the oxidation potential for the formation of dialkylamines decreases as the size of the alkyl chain increases. In 10 mM ammonium acetate (pH 7), a series of trialkylamines was oxidized to the corresponding secondary amines as follows: triethylamine (650 mV), tripropylamine (330 mV), tributylamine (270 mV) and tripentylamine (170 mV).

N–oxides 2–Acetamidofluorene can be oxidized enzymatically at the amide nitrogen, as described in Table 2.2b.54 Electrochemical oxidation of 2–acetamidofluorene in 10 mM ammonium acetate gave a product at m/z 240 at 315 mV, corresponding to a gain of oxygen. The MS/MS fragmentation pattern suggests a hydroxylation at the aromatic ring system and not at the amide nitrogen. The N–oxidation was thus not mimicked by the EC/MS system. It should be noted that hydroxylation at the aromatic ring system is also a common metabolic reaction for 2–acetamidofluorene.55 Pyridine and 2,3–lutidine were also studied as model compounds for a possible electrochemical formation of N–oxides, but no electrochemical reaction was observed.

In the enzymatic mechanism, described in Figure 2.5, the radical cation intermediate (B) can react directly with the [Fe=O]2+ complex to give the N–oxide (D) instead of donating a proton from the α–carbon. In the electrochemical system, there is no such active iron–oxygen intermediate available, and as a consequence, no N–oxidation takes place.


30

Sulfur, S–dealkylation and sulfoxide formation The major products from cytochrome P450 catalyzed oxidations, as well as from electrochemical oxidations of sulfur containing compounds, are sulfoxides.20 However, in some cases cytochrome P450 gives the dealkylated product as well as the sulfoxide. One example is the S–demethylation of S–methylthiopurine, described in Table 2.2c.54 In 0.1 M acetic acid, the sulfur was electrochemically oxidized to the sulfoxide at 850 mV and dealkylation was not observed. Three other sulfides, dibutyl sulfide (Table 2.2d), diphenyl sulfide and thioanisole, were also oxidized in the electrochemical system. They were all oxidized to the corresponding sulfoxides at +250 to +450 mV in 0.1 M acetic acid. The position of the oxygen at the sulfur was confirmed by comparison of the MS/MS fragmentation pattern of the electrochemically generated sulfoxides with purchased standards of dibutyl sulfoxide and diphenyl sulfoxide. At increased potentials, the sulfoxides were further oxidized to sulfones. An acidic electrolyte clearly favored the sulfoxide formation.

In most cases, the electrochemical oxidation provides a good mimic of the enzymatic oxidation. A correlation between the rates of sulfoxide formation for a number of para–substituted thioanisoles and the electrochemical oxidation potentials has been reported by Watanabe et al.5

S–Methylthiopurine is easily protonated, but the other sulfides do not give [M+H]+ ions in solution, and are thus not detectable with an electrospray interface. Electrochemical oxidation of the sulfides gives sulfoxides that are easily protonated and can be detected by electrospray MS. For these experiments, both an electrospray interface and an atmospheric pressure chemical ionization (APCI) interface have been used. When APCI was used, the sulfides could be detected before oxidation, but the oxidized products gave signals with >80 times higher intensity. For analysis of sulfides, on–line electrochemical oxidation might thus be a way to reach lower detection limits for electrospray as well as for APCI.

Phosphor, phosphine oxide formation The P450 catalyzed phosphothionate oxidation of parathion, a widely used pesticide, is described in Table 2.2e.54 Electrochemical oxidation gave the same oxidation product at a potential of 600 mV in 0.1 M acetic acid. Compounds containing phosphorous, generally have an even lower oxidation potential than sulfur containing compounds, and cytochrome P450 catalyzed oxidation as well as electrochemical oxidation almost exclusively result in phosphine oxides.20


31

Oxygen, dealkylation of ethers Table 2.2f shows the O–deethylation of 7–ethoxycoumarin, a general test substrate for various P450 enzymes, e.g., human CYP1, CYP2 and CYP3 families.56 The O–deethylation could not be mimicked in the electrochemical cell. 7–Methoxycoumarine was also analyzed, but no demethylation products were observed.

Due to the electronegative character of oxygen, dealkylation of ethers by cytochrome P450 is thought to proceed via a direct hydrogen abstraction from the α–carbon (pathway A–>C–>E in Figure 2.5).20 No electrochemical oxidation was observed for any of the ethers tested here. The explanation for this is most likely the same as the reason for the direct hydrogen abstraction by cytochrome P450. The oxidation potential of ethers is too high for cytochrome P450 to extract an electron, and too high for electrochemical oxidation in aqueous solutions. It should be noted that some aromatic ethers have relatively low oxidation potential and are probably oxidized by cytochrome P450 via a one electron mechanism.42,57

Other types of cytochrome P450 catalyzed oxidations Aliphatic and aromatic hydroxylations and dealkylation/oxygenation of heteroatom containing compounds are the most common reactions catalyzed by cytochrome P450. There are however other types of cytochrome P450 catalyzed oxidations, and some of these are described in Table 2.3. Table 2.3 Other types of cytochrome P450 catalyzed oxidations


a) Oxidative deamination of primary amines Ex. Oxidative deamination of amphetamine

O

mw 135 mw 134

NH2

NH

mw 133 b) Alcohol and aldehyde oxidation

NOH

N

O

HN

O

OH

mw 109 mw 107 mw 123

N

O

H

mw 107 c) Oxidative dehalogenation

NH

O

mw 107mw 127

NCl

mw 93

N


32


d) Dehydrogenation

HO

N

O

H

O

N

O

mw 151 mw 149

O

N

O

mw 149 Oxidative deamination, oxidative dehalogenation and alcohol/aldehyde oxidation are basically oxygen transfer processes with insertion of a hydroxyl group at the α–carbon, as described earlier. The structure of the substrate and the physical features of the P450 active site will dictate the specific mechanistic pathway taken. The hydroxylated products are unstable and will nonenzymatically lose ammonia, HX or water to give the final products.20 The mechanisms for alcohol oxidation and dehydrogenation will be discussed in more detail. Oxidative deamination Amphetamine is oxidized by cytochrome P450 to the ketone, as described in Table 2.3a.54 Electrochemical oxidation of amphetamine in 0.1 M NH4OH gave about 5 % of a product at m/z 134 at 850 mV which was presumed to correspond to the imino structure presented in Table 2.3a. In 0.1 M NH4OH, hexylamine gave trace amounts of a compound at m/z 98 at high oxidation potentials, believed to originate from the corresponding nitrile.

The mechanism described for the electrochemical oxidation of primary amines is basically the same as the mechanism described for the oxidation of tertiary amines described in Figure 2.6. Presumably, hydrolysis of the intermediate imines (Figure 2.6 IV) gives ammonia and the corresponding aldehydes or ketones in the case of α–branched primary amines.53 Under basic conditions, further deprotonation and oxidation to nitriles has been reported.58,59 The oxidation potentials are much higher for primary than for tertiary amines, and under the conditions used in these experiments, no oxidation was observed in acidic or neutral solutions. By changing the electrolyte and/or the electrode material, a better mimic of the enzymatic oxidation might be obtained, but under the conditions used in these experiments, no formation of ketones or aldehydes was observed for any of the amines examined.

Alcohol and aldehyde oxidation Cytochrome P450 is known to catalyze the oxidation of alcohols to aldehydes that can be further oxidized to the corresponding carboxylic acids as illustrated in Table 2.3b. The electrochemical oxidation of 2–(hydroxymethyl)–pyridine gave about 3 % of pyridine–2–


33

aldehyde at 750 mV in 0.1 M NH4OH. The identity of the product was confirmed by comparison of the MS/MS fragmentation pattern with a purchased pyridine–2–aldehyde standard. Electrochemical oxidation of aldehydes to carboxylic acids was not observed in our experiments but has been reported to occur under different electrochemical reaction conditions.36

, H+--e-- e-

OHHH

H OOHH OHH +

H+-

R O

H2O-

+[Fe]3+

OHROH

rebound

Oxygen

Fe OH[ ] 3++

OHR

+Fe O[ ]3+

or -H

, H+--e-

OHRH

(a)

(b)

.

.

.

Figure 2.7 a) Cytochrome P450 catalyzed oxidation of benzylic alcohols to aldehydes via gem-diol as suggested by Vaz and Coon. R=H or CH3. b) Electrochemical oxidation of benzyl alcohol to benzaldehyde. Vaz and Coon have suggested the mechanism presented in Figure 2.7a for the oxidation of benzylic alcohols by cytochrome P450 2B4 and 2E1.60 There are alternative pathways possible from the radical to the aldehyde, but the intermediate gem–diol seems to be the most likely route. The mechanism for electrochemical oxidation of benzyl alcohol in Figure 2.7b is initiated in the same way as the enzymatic oxidation, but the aldehyde is probably formed by one electron oxidation of the radical followed by deprotonation.36

The oxidation of aldehydes by cytochrome P450 normally yields the corresponding carboxylic acid, either by hydrogen abstraction followed by “oxygen rebound” as described earlier, or by a mechanism involving nucleophilic attack by the iron–peroxo–porphyrin intermediate. This mechanism has been suggested for the P450 2B4 catalyzed oxidation of cyclohexane carboxaldehyde.25

In conclusion, the P450 catalyzed oxidation of alcohols to aldehydes was mimicked by the EC/MS system but further oxidation to the carboxylic acid was not observed.

Oxidative dehalogenation The oxidation of 2-chloromethylpyridine (2–picolinechloride) shown in Table 2.3c was the only halogenated compound tested in the electrochemical system. The oxidation product


34

was identified as 2–picoline by comparison of MS/MS fragmentation pattern with a 2–picoline standard. The dehalogenation was thus not mimicked by the EC/MS system. Further investigation is necessary to be able to draw any conclusions about halogenated compounds in general.

Dehydrogenation Acetaminophen is oxidized by several human cytochrome P450 enzymes, including CYP 2E1, 1A2, 2A6, 3A4 and 2D6, to its toxic metabolite N–acetyl–p–benzoquinoneimine.61 The dehydrogenation of acetaminophen in Table 2.3d was readily mimicked by the EC/MS system.

H+-,e-

-

OH

N OH

O

N OH

O

N

O

H+-,e-

-

OH

N OH

O

N

O+ +

O

N OH

Fe OH[ ]3 +

Fe OH[ ]3 +

[ 3 +

e-- H+-,

Fe O][

e-- H+-,

Fe O ]3 +

(a)

(b)

Figure 2.8 a) Cytochrome P450 catalyzed dehydrogenation of acetaminophen proposed by

Koymans et al.62 b) Electrochemical oxidation of acetaminophen. The mechanism proposed by Koymans et al. is presented in Figure 2.8a.62 The formation of the radical intermediate is believed to occur via electron transfer followed by proton abstraction. Another one electron oxidation followed by a second proton abstraction yields the final product. The mechanism for the electrochemical oxidation in Figure 2.8b is practically the same as the suggested enzymatic mechanism.


35

2.4 Conclusions

This study presents an overview of the different oxidations catalyzed by cytochrome P450 and how they correlate to electrochemical oxidations. The EC/MS system provided successful mimics in cases where the P450 catalyzed reactions are supposed to proceed via a mechanism initiated by a one electron oxidation, such as N–dealkylation, S–oxidation, P–oxidation, alcohol oxidation and dehydrogenation. The P450 catalyzed reactions thought to be initiated via direct hydrogen atom abstraction, such as O–dealkylation and hydroxylation of unsubstituted aromatic rings generally had a too high oxidation potential to be electrochemically oxidized within the potential limits of water, and were not mimicked by the EC/MS system. An electrochemical system, like the one described here, obviously has a number of drawbacks. Some reactions, as for example the important aliphatic hydroxylations, are not mimicked and the stereochemistry induced by different cytochrome P450 enzymes can not be obtained without modifications of the electrode surfaces. Despite these limitations, valuable information can be obtained concerning the sensitivity of the substrate towards oxidation and the regions of the molecule where oxidations are likely to take place. In order to draw any conclusions for specific classes of compounds, more specific work is needed. Analogy between cytochrome P450 oxidation rates and electrochemical oxidation potentials for oxidations proceeding via initial one electron oxidation has been reported by the group of Guengerich and others.5,7,63 In a large chemical library with similar compounds, comparison of the oxidation potentials might provide useful information about the relative metabolic stability.

Electrochemistry can be very useful for small–scale synthesis of metabolites, starting from a drug. The on–line EC/MS system is suitable for optimization of the electrochemical conditions. A potential sweep with the EC/MS system, as described in this chapter, will give a quick answer regarding what potentials should be used for optimum yield of the desired product. The simplicity of the system, and the ease and speed with which it can be applied to a large number of compounds, make it a useful tool in drug metabolism research.

Acknowledgement The authors wish to thank Ruben de Kanter for helpful discussions about cytochrome P450 metabolism.

2.5 References

1. Lund H, Baizer MM, In Organic electrochemistry, an introduction and a guide (3rd edn); Marcel Dekker: New York, 1991.


36

2. Shono T, Toda T, Oshino N, J. Am. Chem. Soc., 1982; 104: 2639-2641. 3. Shono T, Toda T, Oshino N, Drug Metab. Dispos., 1981; 9: 481-482. 4. Hall LR, Iwamoto T, Hanzlik RP, J. Org. Chem., 1989; 54: 2446-2451. 5. Watanabe Y, Iyanagi T, Oae S, Tetrahedron Lett., 1982; 23: 533-536. 6. Guengerich FP, Macdonald TL, Acc. Chem. Res., 1984; 17: 9-16. 7. Macdonald TL, Gutheim WG, Martin RB, Guengerich FP, Biochemistry, 1989; 28: 2071-

2077. 8. Hambitzer G, Heitbaum J, Anal. Chem., 1986; 58: 1067-1070. 9. Getek TA, The development of thermospray mass spectrometry combined with

electrochemical redox processes, 34th ASMS Conference on Mass Spectrometry and Allied Topics, Cincinnati, OH, 1986.

10. Getek TA, Korfmacher WA, McRae TA, Hinson JA, J. Chromatogr., 1989; 474: 245-256. 11. Volk KJ, Yost RA, Brajter-Toth A, Anal. Chem., 1992; 64: 21A-33A. 12. Deng H, Berkel GV, Electrochemistry On-Line with Electrospray: Insights into Oxidative

Metabolism of Tamoxifen, 47th ASMS Conf. Mass Spectrometry and Allied Topics, Dallas, TX, 1999.

13. Baczynskyj L, EC-LC-ESI MS: Electrochemistry Combined with Liquid Chromatography and Electrospray Mass Spectrometry, 44th ASMS Conf. Mass Spectrometry and Allied Topics, Portland, OR, 1996.

14. Jurva U, Bruins AP, Wikström HV, The ESA electrochemical cell coupled with electrospray ionization mass spectrometry, 47th ASMS Conf. Mass Spectrometry and Allied Topics, Dallas, TX, 1999.

15. Jurva U, Wikström HV, Bruins AP, Rapid commun. Mass spectrom., 2000; 14: 529-533. 16. Deng H, Berkel GJV, Electroanalysis, 1999; 11: 857-865. 17. Jurva U, Wikström HV, Bruins AP, In vitro mimicry of metabolic oxidation reactions by

electrochemistry/mass spectrometry, 15th IMS conference, Barcelona, Spain, 2000. 18. Iwahashi H, J. Chrom. B, 1999; 736: 237-245. 19. Phillips IR, Shephard EA (eds). Cytochrome P450 protocols ; Humana press, 1998. 20. Sono M, Roach MP, Coulter ED, Dawson JH, Chem. Rev., 1996; 96: 2841-2887. 21. Meunier B, Bernadou J, Structure and Bonding, 2000; 97: 1-35. 22. Montellano PROd (ed). Cytochrome P450, Structure, Mechanism, and Biochemistry (2nd

ed). ; Plenum Press, 1995. 23. Lewis DFV, Pratt JM, Drug Metab. Rev., 1998; 30: 739-786. 24. Groves JT, Han YZ, Models and Mechanisms of Cytochrome P450 Action, p. 3-48 in

ref.22. 25. Wertz DL, Valentine JS, Structure and Bonding, 2000; 97: 37-60. 26. Horn AS, Tepper P, Weide JVD, Watanabe M, Grigoriadis G, Seeman P, Pharm. Weekbl.

Sci. Ed., 1985; 7: 208.


37

27. Crespi CL, Chang TKH, Waxman DJ, Determination of CYP4A11-Catalyzed Lauric acid 12-Hydroxylation by High-Performance Liquid Chromatography with Radiometric Detection, p. 163-167 in ref. 19.

28. Eberson L, Utley JHP, Hammerich OHC, Anodic oxidation of hydrocarbons, p. 505-533 in ref. 1.

29. Newcomb M, Shen R, Choi SY, Toy PH, Hollenberg PF, Vaz ADN, Coon MJ, J. Am. Chem. Soc., 2000; 122: 2677-2686.

30. Swart PJ, Oelen WE, Bruins AP, Tepper PG, Zeeuw RAD, J. Anal. Toxicol., 1994; 18: 71. 31. Crespi CL, Chang TKH, Waxman DJ, CYP2C19-Mediated (S)-Mephenytoin 4'-

Hydroxylation Assayed by High-Performance Liquid Chromatography with Radiometric Detection, p. 135-139 in ref. 19.

32. Waxman DJ, Chang TKH, Spectrofluorometric Analysis of CYP2A6-Catalyzed Coumarin 7-Hydroxylation, p. 111-116 in ref. 19.

33. Chang TKH, Crespi CL, Waxman DJ, Spectrofluorometric Analysis of Human CYP2E1-Catalyzed p-nitrophenol Hydroxylation, p. 147-152 in ref. 19.

34. Crespi CL, Chang TKH, Waxman DJ, Determination of CYP2C9-catalyzed Diclofenac 4'-Hydroxylation by High-Performance Liquide Chromatography, p. 129-133 in ref. 19.

35. Yilmazer M, Stevens JF, Deinzer ML, Buhler DR, Drug Metab. Dispos., 2001; 29: 223-231.

36. Hammerich O, Svensmark B, Anodic Oxidation of Oxygen-Containing Compounds, p. 615-657 in ref. 1.

37. Guengerich FP, Macdonald TL, FASEB J., 1990; 4: 2453-2459. 38. Montellano PROd, Oxygen activation and reactivity, In ref. 22. 39. Schäfer HJ, Electrolytic Oxidative Coupling, p. 949-1027 in ref 1. 40. Jeftic L, Adams RN, j. Am. Chem. Soc., 1970; 92: 1332-1337. 41. Xu X, Lu W, Cole RB, Anal. Chem., 1996; 68: 4244-4253. 42. Sato H, Guengerich FP, J. Am. Chem. Soc., 2000; 122: 8099-8100. 43. Cavalieri EC, Rogan EG, Devanesan PD, Cremonesi P, Cerny RL, Gross ML, Bodell WJ,

Biochemistry, 1990; 29: 4820-4827. 44. Cremonesi P, Stack DE, Rogan EG, Cavalieri EL, J. Org. Chem., 1994; 59: 7683-7687. 45. Anzenbacher P, Niwa T, Tolbert LM, Sirimanne SR, Guengerich FP, Biochemistry, 1996;

35: 2512-2520. 46. Berkel GJV, Zhou F, Anal. Chem., 1995; 67: 2916-2923. 47. Zhou F, Berkel GJV, Electrochemically enhanced electrospray ionization mass

spectrometry, 42nd ASMS Conf. Mass Spectrometry and Allied Topics, Chicago, Illinois, 1994.

48. Zhou F, Berkel GJV, Anal. Chem., 1995; 67: 3643-3649. 49. Diehl G, Liesener A, Karst U, Analyst, 2001; 126: 288-290.


38

50. Deng H, Berkel GJV, Takano H, Gazda D, Porter MD, Anal. Chem., 2000; 72: 2641-2647. 51. Guengerich FP, Yun CH, Macdonald TL, J. Biol. Chem., 1996; 271: 27321-27329. 52. Olinga P, Merema M, Hof IH, Jong KPd, Slooff MJ, Meijer DKF, Groothuis GMM, Drug

Metab. Dispos., 1998; 26: 5. 53. Steckhan E, Nitrogen-Containing Compounds, p. 581-613 in ref. 1. 54. Gibson GG, Skett P (eds). Introduction to Drug Metabolism (2nd ed). ; Blackie Academic

and professional, 1994. 55. Meerman J, The role of sulfation in the chemical carcinogenesis by N-hydroxy-

arylacetamides, PhD Thesis, University of Groningen, 1982. 56. Waxman DJ, Chang TKH, Use of 7-Ethoxycoumarin to monitor the Human CYP1, CYP2,

and CYP3 Families, p. 175-179 in ref. 19. 57. Yun C, Miller GP, Guengerich FP, Biochemistry, 2000; 39: 11319-11329. 58. Fleischmann M, Korinek K, Pletcher D, J.Chem.Soc. Perkin Trans. II, 1972; 1393-1403. 59. Feldhues U, Schafer HJ, Synthesis, 1982; 145-146. 60. Vaz ADN, Coon MJ, Biochemistry, 1994; 33: 6442-6449. 61. Dong H, Haining RL, Thummel KE, Rettie AE, Nelson SD, Drug Metab. Dispos., 2000;

28: 1397-1400. 62. Koymans L, Kelder GMD-oD, Koppele JMT, Vermeulen NPE, Xenobiotica, 1993; 23:

633-648. 63. Galliani G, Rindone B, Dagnino G, Salomona M, Eur. J. Drug Metab., 1984; 9: 289-293.

Chapter 3 Electrochemically assisted Fenton reaction

39

Chapter 3

Electrochemically assisted Fenton reaction: The

reaction of hydroxyl radicals with xenobiotics

followed by on–line analysis with HPLC/MS/MS

Oxygen radicals are generated in vivo by various processes, often as toxic intermediates in different metabolic transformations, and have been shown to play an important role for a large number of diseases. In this chapter we introduce an electrochemical flow through system that allows generation of hydroxyl radicals for reaction with xenobiotics and subsequent detection of the oxidation products on–line with HPLC/MS/MS. The system is based on the Fenton reaction and is predominantly aimed at the generation of hydroxyl radicals, but by minor variations to the system, a broad range of other radicals can be produced. Optimization of the system was performed with the radical scavenger 5,5-dimethyl-1-pyrroline-N-oxide (DMPO). Under the same physical conditions, one injection through the electrochemical cell gave a higher yield of the oxidation product N-hydroxy-5,5-dimethyl-pyrrolidin-2-one than what was attained after 60 minutes with a chemical Fenton system catalyzed by ascorbic acid. Since the iron is added as Fe3+, the initial mixture is “inactive” until it reaches the electrochemical cell. This makes it very suitable for on–line analysis of the generated compounds, since the whole reaction mixture, including substrate, can be kept in a vial in an autosampler. The system described provides a useful tool for investigation of new radical scavengers and antioxidants. Since the hydroxyl radical adds readily to unsaturated π–systems, the technique is also suitable for on–line generation and characterization of potential drug metabolites resulting from hydroxylation of double bonds and aromatic systems.

3.1 Introduction

Oxygen radicals in the body are reactive, short lived and play a role in most major health problems.1 A number of radicals are formed in biological systems via a range of different processes. The oxygen containing radicals involved in “oxidative stress” together with some non radical species are collectively called Reactive Oxygen Species (ROS) and are often toxic intermediates in different metabolic processes. Examples of ROS known to induce damage in vivo are H2O2, organic hydroperoxides, HOCl, NO⋅, O2

.-, alkoxyl radicals and the hydroxyl radical. ROS can be derived from numerous sources in vivo


40

including normal respiration, photochemical reactions and enzymatic reactions. A large number of enzymes have been shown to be capable to generate ROS and include the cytochromes P450, various oxidases, peroxidases, lipogenases and dehydrogenases.2

One of the most hazardous radicals is the extremely reactive hydroxyl radical, with an almost diffusion limited half life of about one nanosecond. It has been shown to play a major role in a number of diseases. Oxidative stress induced by hydroxyl radicals is for example believed to be the main cause of damage to the neurons in Parkinson’s disease3,4. Oxidation of proteins and enzymes may contribute to the build–up of glutamate in stroke. The reaction of hydroxyl radicals with nucleic acids causes strand breaks and alters bases and thereby contributes to carcinogenesis.5 Accumulation of oxidized proteins and enzymes has also been observed during the process of aging.6

The main source of hydroxyl radicals in vivo is probably the so called Haber–Weiss reaction, where O2

.- reduces Fe3+ to Fe2+ and in that way initiates the Fenton reaction between Fe2+ and hydrogen peroxide.2

Fe 2+Fe3+ +

OH-

H2O2

+ O2

+ +Fe 2+ + Fe3+

O2

-.O2

-. -.

OH. OH.

Other metal ions such as Cu+, Ti3+, Cr2+ and Co2+ can replace Fe2+ in the reaction.7 In vivo, only iron and copper are present in high enough concentrations to be relevant in the Haber–Weiss reaction. Iron is the most abundant and consequently the most important transition state metal, but because of its high abundance in the mammalian cell nucleus copper is probably more important than iron for oxidative damage to DNA.8 The reaction can also be initiated by reduction of Fe3+ by other reducing agents present in the body, as for example ascorbic acid.9 Ironically, the antioxidant property of ascorbic acid also enables it to catalyze the Fenton reaction and thereby increase the oxidative damage to the surrounding tissue.

In order to gain information about how radicals interact with endogenous compounds as well as drugs and xenobiotics and to allow the study of new radical scavengers and antioxidants, the generation of radicals followed by accurate measurement of the oxidation products is of great importance.

In vitro, hydroxyl radicals can be generated in a number of ways, for example via radiolysis of water10, photolysis of hydrogen peroxide11, chemical systems based on Fenton chemistry7 and from enzymatic sources such as xanthine/xanthine oxidase.12


41

Fe2+ Fe3+

e-

H2O2+ + +OH-

OH.

Figure 3.1 Principle of the Fenton reaction. Regeneration of Fe2+ from Fe3+ is done by chemical reduction with L(+) ascorbic acid, or by electrochemical reduction at the working electrode. The overall Fenton reaction presented in Figure 3.1 is a simplification of a much more complicated mechanism still under debate, but a more thorough mechanistic discussion lies beyond the scope of this article.13,14 Hydroxyl radicals are generated as Fe2+ donates an electron to hydrogen peroxide. The iron is oxidized to Fe3+ and becomes inactive for further reaction. In a chemical Fenton system, a reducing agent, such as ascorbic acid, is usually added to regenerate Fe2+.15

The system presented in this article is based on electrochemically assisted Fenton chemistry, where the regeneration of Fe2+ is achieved by reduction of Fe3+ at the working electrode.16 Since the iron is added as Fe3+, the mixture will be “inactive” until it reaches the electrochemical cell. This makes it very suitable for on–line analysis of the generated compounds with a number of analysis techniques, such as ESR or UV–detection. In our experiments, the electrochemical cell has been connected on–line with an HPLC column connected to a triple quadrupole mass spectrometer. Although the focus has been on the generation of hydroxyl radicals, the system presented can be applied for the generation of a large range of radicals.

3.2 Experimental

Chemicals The following chemicals were obtained commercially: 5,5-dimethyl-1-pyrroline-N-oxide (DMPO), Sigma; Iron(III)chloride hexahydrate, 98 %, Aldrich; ethylenediaminetetraacetic acid disodium salt (EDTA), 99 %, Merck; L(+) ascorbic acid, 99.7 %, Merck; dimethylsulfoxide, 99.9 %, Aldrich; dibutylsulfoxide, 96 %, Aldrich; Methanol, gradient grade, Merck; hydrogen peroxide, 35 % solution in water, Aldrich. The water used in the experiments was purified in a Maxima Ultrapure water system (ELGA, High Wycombe, Bucks, U.K.) and was sonicated for about 15 minutes before use.


42

Description of the EC/LC/MS system

Mac Labcell control

and read-out

ESACoulochem5020 cell

Potentiostat

LoopInjector

LoopInjector

PumpSciex

API 3000LC/MS/MS

Make-up flow| | | | | |

500 µl

20 µlColumnHPLC

Two-pumpgradientsystem

Figure 3.2 The EC/LC/MS system

A schematic overview of the EC/LC/MS system used for the electrochemically assisted Fenton experiments is given in Figure 3.2. A Brownlee microgradient system (Brownlee labs, Santa Clara, CA, USA) delivered a flow of 1-50 µl/min. The analytes were injected with a 0.5 ml injection loop and passed through an ESA Coulochem 5020 guard cell (ESA Inc., Bedford, Massachusetts, USA) connected to a home made potentiostat. The ESA 5020 guard cell has a porous graphite working electrode and all reported cell potentials are vs. a palladium reference electrode. A MacLab system with Chart 3.5.7 software (AD Instruments, Castle Hill, NSW, Australia) was used to control the potentiostat and to apply the desired potential over the electrochemical cell. The cell current was continuously registered by Chart. To dilute the sample and to adjust the acetonitrile content to fit the start of the gradient, a make up liquid containing 10 % acetonitrile and 0.1 % formic acid was added at a flow rate of 50 µl/min from a KDS100 syringe pump (KD Scientific, Boston, Massachusetts, USA). The diluted sample was collected on–line in a 20 µl injection loop, and injected onto a reverse phase HPLC column (Alltech Alltima, C18, 5 µm, 150 x 2.1 mm, Alltech Associates, Deerfield, IL, USA). Two Series 200 micro LC pumps (Perkin Elmer, Norwalk, CT, USA) were programmed to deliver a gradient from water and acetonitrile at a total flow rate of 200 µl/min. Both phases contained 0.1 % formic acid. The first two minutes, the acetonitrile concentration was kept at 15 %, and was then linearly increased to 75 % over a period of 10 minutes. The mobile phase


43

composition was linearly brought back to 15 % acetonitrile in 2 minutes. The system was allowed to equilibrate for at least 10 minutes between the injections. Full–scan spectra or product ion spectra were acquired continuously with an API 3000 triple quadrupole mass spectrometer (MDS–Sciex, Concord, Ontario, Canada) equipped with a TurboIonSpray interface. The interface was operated at room temperature and a splitter was introduced between the HPLC–column and the mass spectrometer to reduce the flow to approximately 40 µl/min. The MS was operated at such a low orifice voltage that “up front” collision induced dissociation did not take place.

The electrochemically assisted Fenton system (EC-Fenton system), general

conditions Unless otherwise mentioned, the sample solutions injected through the electrochemical cell contained 1 mM substrate, 1 mM FeCl3, 1.02 mM EDTA (to prevent precipitation of ferric hydroxide) and 10 mM hydrogen peroxide. The final concentration of acetonitrile in the sample solution was 21 %. The hydrogen peroxide was kept in the refrigerator and was added just before the injection. The “inactive” Fenton mixture was injected at a flow of 50 µl/min. After 1 minute, the flow was lowered to 2 µl/min (the efficiency of the electrochemically assisted Fenton reaction increases at low flow rates). 20 Minutes after the first injection, the diluted sample was injected with the second injection loop onto the HPLC column. The potential over the electrochemical cell was kept at –0.5 V.

Comparison between the electrochemically assisted Fenton system and a

chemical Fenton system. To a glass vial was added: FeCl3 (200 µl, 10 mM in acetonitrile), EDTA (200 µl, 10.2 mM in water) and 1380 µl water. The solution was vortexed for 10 s in a MS1 Minishaker (IKA WORKS, Inc., Wilmington, NC, USA) and then DMPO (200 µl, 10 mM in acetonitrile) was added. After another 10 s vortex, the solution was split equally between two glass vials. To the vial destined for the chemical Fenton reaction, ascorbic acid (5 µl, 1 M in water) was added. At t=0, hydrogen peroxide (10 µl, 1 M in 90 % acetonitrile, 10 % water) was added to both vials. The sample without ascorbic acid was injected onto the EC/LC/MS system as described in the section above. The vial containing ascorbic acid was vortexed at 1000 rpm. Samples were taken at 1, 15, 30, 60 and 120 minutes and were kept in the freezer until they were analyzed the same day using identical HPLC/MS conditions as those used in the EC/LC/MS system. Both experiments were performed at room temperature.


44

Formation of methyl, butyl, and hydroxymethyl radicals The samples were prepared as described in the section “general conditions”. For the formation of methyl radicals, dimethylsulfoxide (2 %, 0.26 M) was added to the sample before injection. Butyl radicals were obtained by addition of dibutylsulfoxide (2 %, 0.12 M), and hydroxymethyl radicals were generated by addition of methanol (2 %, 0.63 M). 3.3 Results and discussion

N

O

H+

- -

+N

O

OHa)

N

O

OH

H

OH H-

N

OH

O

CH3OHOH

H2O + CH2OHb)

c)N

O

CH2OH

H

-

+N

O

CH2OH

-

+N

O

HCH2OH H-

Figure 3.3 a) Reaction of the hydroxyl radical with DMPO b) Formation of hydroxymethyl radical c) Reaction of the hydroxymethyl radical with DMPO Scavenging of the hydroxyl radical with the spin trap DMPO In Figure 3.3a, the scavenging of hydroxyl radical by DMPO is shown. The radical product is stabilized by delocalization of the unpaired electron spin density over nitrogen and oxygen in the aminoxy group. However, in the case of the hydroxyl radical, the stability of the spin adduct formed is rather low and the radical loses a hydrogen atom to give 2–hydroxy–DMPO. The product ion spectrum given in Figure 3.4 and the suggested fragmentation pathway described in Figure 3.5 suggest that the product has been rearranged to N-hydroxy-5,5-dimethyl-pyrrolidin-2-one.


45

QuickTime™ and aGraphics decompressorare needed to see this picture.

0

100

200

300

400

500

600

700

Cou

nts/

s

m/z

130

112

97

88

84

69

6055

41

Products of m/z 130

N

OH

O

Figure 3.4 Product ion spectrum of the DMPO adduct with m/z 130

m/z 130

- H2O

m/z 112 m/z 84

- HNCO

N C OH

+ +

m/z 69

C2H4-

+N C O

H

+ N C O

H

+N

HOH

O

+

N O+

m/z 41

m/z 112

- HNCO

m/z 130

N C O

H+

+

N O

OHH

H - H2NOH

m/z 97

+- CO

m/z 69

m/z 130

N O

OHH +

N O+

N O

OHH

H+

N OHH

HH2C=C=O- +

N

H

OHH

+

N O

OHH +

- H2NOH

m/z 88

+

m/z 55

N O+

m/z 112

NO

+

O+

H2C=C=O-

N+N +

m/z 70

Figure 3.5 Suggested fragmentation pathways for the protonated N-hydroxy-5,5-dimethyl-pyrrolidin-2-one


46

It has been debated that the DMPO–OH radical can be formed by a non–radical mechanism.13,17 A simple method to test if the adduct is formed via addition of hydroxyl radical or by a different mechanism is to add a second radical scavenger, such as methanol, ethanol or dimethyl sulfoxide (DMSO). Figure 3.3b shows the reaction between hydroxyl radical and methanol. The hydroxyl radical abstracts a hydrogen atom from methanol to yield a hydroxymethyl radical which then in turn reacts with DMPO (Figure 3.3c).18 Subsequent loss of a hydrogen atom results in a new nitrone. The detection of 2–hydroxymethyl–DMPO was taken as indirect evidence that hydroxyl radicals were formed in the electrochemical system. Other mechanisms can not be entirely excluded but hydroxyl radicals are clearly generated in the system. QuickTime™ and aGraphics decompressorare needed to see this picture.

a) -0.5 V

b) Cell turned off

-

+N

O

CH2OH

Figure 3.6 Mass chromatograms of m/z 144, corresponding to the protonated 2–hydroxymethyl–DMPO obtained by electrochemically assisted Fenton reaction a) The cell potential was kept at –0.5 V b) No potential was applied over the electrochemical cell. The mass chromatogram of the [M+H]+ ion of 2–hydroxymethyl–DMPO is shown in Figure 3.6. The same reaction mixture was injected in both experiments. In Figure 3.6a, the potential over the cell was kept at –0.5 V, a potential where the Fe3+ ions are readily reduced to Fe2+ and thereby initiating the production of hydroxyl radicals. In Figure 3.6b no potential was applied over the cell, and as a consequence, no 2–hydroxymethyl–DMPO was formed.

It is important to realize that the electrochemical reaction is limited to a thin layer next to the electrode surface and the ions in the bulk solution do not take part in the electrochemical reaction.19 In principle, direct electrochemical reduction of hydrogen peroxide could give one hydroxyl radical and one hydroxyl ion. Since this reaction takes place at the electrode surface, the produced hydroxyl radicals are immediately reduced to


47

hydroxyl ions at a reductive potential. In contrast, Fe2+ ions produced at the working electrode are free to diffuse into the solvent phase and become available for reaction with hydrogen peroxide. If the substrate used to trap the radicals is added in high enough concentration, a large part of the hydroxyl radicals will be trapped without a substantial loss of hydroxyl radicals to the working electrode.

Comparison between the electrochemically assisted Fenton system and a

chemical Fenton system. Under the conditions used in these experiments, one injection through the electrochemical cell gave a higher yield of N-hydroxy-5,5-dimethyl-pyrrolidin-2-one than what was obtained after 60 minutes with a chemical Fenton system catalyzed by ascorbic acid. The calculated cell volume of the ESA Coulochem 5020 guard cell is 40 µl. At a flow rate of 2 µl/min, the Fenton mixture stays 20 minutes in the electrochemical cell. The electrochemical regeneration of Fe2+ provided by the ESA 5020 guard cell is thus more efficient than the chemical regeneration by ascorbic acid.

Generation of radicals other than hydroxyl radicals in the EC Fenton system Hydroxyl radical generation was the main goal in these experiments, but the electrochemically assisted Fenton system can also be used to generate a range of other radicals. Because of the extreme reactivity of the hydroxyl radical, it will react with most compounds added to the system and generate new radicals. One example is the formation of hydroxymethyl radical by addition of methanol to the system described above. Another example of how radicals can be derived from the hydroxyl radical is the reaction with dialkylsulfoxides.


48

QuickTime™ and aGraphics decompressorare needed to see this picture.

0

500000

1000000

1500000

2000000

2500000

3000000

Cou

nts/

s

m/z

128

255

206

[M+H]+

[2M+H]+

[M+DMSO+H]+

-

+N

O

a)

0

500000

1000000

1500000

2000000

Cou

nts/

s

m/z

[M+H]+

[2M+H]+

170

339

N

O

+

-

b)

Figure 3.7 Products from reaction of DMPO with radicals derived from the hydroxyl radical. a) 2-methyl-DMPO from addition of dimethylsulfoxide. b) 2-butyl-DMPO from addition of dibutylsulfoxide. The hydroxyl radical adds to the sulfur and an alkyl radical is released. Figures 3.7a and 3.7b show the full–scan spectra of 2–methyl–DMPO and 2–butyl–DMPO generated by addition of dimethylsulfoxide and dibutylsulfoxide respectively. When DMSO was added to the system, DMPO was almost completely converted to 2–methyl–DMPO, suggesting that the initial formation of hydroxyl radicals is larger than indicated by scavenging and formation of N-hydroxy-5,5-dimethyl-pyrrolidin-2-one. DMPO has to compete with several other reactions of the hydroxyl radical, such as reduction to hydroxide ions at the electrode or by Fe2+, reaction with EDTA and reaction with another hydroxyl radical to form hydrogen peroxide. The relative stability of the methyl radical compared to the hydroxyl radical provides a longer time to encounter and react with DMPO.


49

Effect of Fe(III)/EDTA concentration on the formation of N-hydroxy-5,5-

dimethyl-pyrrolidin-2-one

0

10

20

30

40

50

60

70

80

90

100

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

DM

PO

add

uct w

ith m

/z 1

30 (

%)

Concentration Fe(III)/EDTA (mM)

Figure 3.8 Effect of the Fe(III)/EDTA concentration on the formation of N-hydroxy-5,5-dimethyl-pyrrolidin-2-one given as percent of the sum of the integrated peaks of the protonated DMPO and N-hydroxy-5,5-dimethyl-pyrrolidin-2-one in the mass chromatograms. The effect of Fe(III)/EDTA concentration on the formation of N-hydroxy-5,5-dimethyl-pyrrolidin-2-one is illustrated in Figure 3.8. The electrochemical regeneration of Fe2+, shown in Figure 3.1 theoretically allows Fe3+ to be added in catalytic amounts. In an electrochemical system where the ferric ions have constant access to the working electrode, as described by Oturan and Pinson, the ferrous ions will be continuously regenerated, and can thus be added in small amounts.16 In the flow through system described here, the Fenton reaction mixture will stay 20 minutes in the electrochemical cell at a flow rate of 2 µl/min. In this period, the Fenton chemistry does not proceed at a high enough rate to allow addition of catalytic amounts of Fe3+ for a high formation of N-hydroxy-5,5-dimethyl-pyrrolidin-2-one.


50

Effect of the flow rate on the formation of N-hydroxy-5,5-dimethyl-pyrrolidin-

2-one

0

5

10

15

20

25

30

35

40

0 1 2 3 4 5 6 7 8 9 10

DM

PO

add

uct w

ith m

/z 1

30 (

%)

Flow rate (µl/min)

Figure 3.9 Effect of the flow rate on the formation of N-hydroxy-5,5-dimethyl-pyrrolidin-2-one given as percent of the sum of the integrated peaks of the protonated DMPO and N-hydroxy-5,5-dimethyl-pyrrolidin-2-one in the mass chromatograms. The electrochemical regeneration of Fe2+, illustrated in Figure 3.1 is only active as long as the ferric ions can be reduced at the working electrode. The efficiency of the electrochemically assisted Fenton reaction is thus dependent on the time that the reaction mixture spends in the electrochemical cell. The cell volume and the flow rate therefore have a large influence on the radical formation. The ESA Coulochem 5020 guard cell has a calculated volume of 40 µl. At a flow of 1 µl/min, the reaction mixture spends 40 minutes in the electrochemical cell. In Figure 9, the effect of flow rate on the formation of N-hydroxy-5,5-dimethyl-pyrrolidin-2-one is shown. It is clear that the efficiency of the electrochemically assisted Fenton reaction goes down as the flow goes up. For obvious reasons, high flow rates would be preferred in order to save time when the system is connected on–line. A larger electrochemical cell would probably increase the efficiency and allow the use of higher flow rates. On the other hand, the larger the cell volume, the longer it takes to get through the electrochemical cell. 3.4 Conclusions

An electrochemically assisted Fenton system for generation of radicals on–line with HPLC/MS/MS has been developed. The fact that the “inactive” reagent is activated upon


51

passage through the electrochemical cell makes it very suitable for on–line analysis of the products, since the whole reaction mixture including substrate can be kept in a vial in an autosampler. The described system provides a useful tool for the investigation of new radical scavengers and antioxidants. In our previous work, we described how electrochemistry on–line with mass spectrometry can be used to mimic a number of cytochrome P450 catalyzed oxidations.20,21 The electrochemically assisted Fenton system provides a different type of oxidation and is able to insert an oxygen on organic molecules in cases where pure electrochemistry fails to mimic the enzymatic oxidation. It is important to realize that the EC–Fenton system requires the use of significantly lower flow rates than pure electrochemical oxidation.

Acknowledgements The authors whish to thank Dr. Lars Weidolf at AstraZeneca R&D Mölndal for helpful discussions about this chapter.

References 1. L. Packer "Oxidants and Antioxidants, Part A" ; Academic Press, 1999. 2. J. P. Kehrer, "The Haber-Weiss reaction and mechanism of toxicity", Toxicol., 2000, 149,

43-50. 3. P. Foley; P. Riederer, "Influence of neurotoxins and oxidative stress on the onset and

progression of Parkinson's disease.", J. Neurol, 2000, 247, 82-94. 4. W. Linert; G. N. L. Jameson, "Redox reactions of neurotransmitters possibly involved in

the progression of Parkinson's disease", J. Inorg. Biochem., 2000, 79, 319-326. 5. R. A. Floyd, "Role of oxygen free radicals in carcinogenesis and brain ischemia", FASEB

J., 1990, 4, 2587-2597. 6. B. S. Berlett; E. R. Stadtman, "Protein Oxidation in Aging, Disease, and Oxidative Stress",

J. Biol. Chem., 1997, 272, 20313-20316. 7. P. Wardman; L. P. Candeias, "Fenton Chemistry: An Introduction", Radiat. Res., 1996,

145, 523-531. 8. M. J. Burkitt, "Copper-DNA Adducts", In "Oxygen Radicals in Biological Systems, Part D"

; Packer, L., Ed.; Academic Press:, 1994; Vol. 234. 9. G. R. Buettner; B. A. Jurkiewitcz, "Catalytic Metals, Ascorbate and Free Radicals:

Combinations to avoid.", Radiat. Res., 1996, 145, 532-541. 10. K. P. Madden; H. Taniguchi, "The role of the DMPO-hydrated electron spin adduct in

DMPO-OH spin trapping", Free Rad. Biol. Med., 2001, 30, 1374-1380.


52

11. E. J. Feltham; M. J. Almond; G. Marston; K. S. Wiltshire; N. Goldberg, "Reactions of hydroxyl radicals with alkenes in low-temperature matrices", Spectrochim. Acta Part A, 2000, 56, 2589-2603.

12. B. Halliwell, "Superoxide-dependent formation of hydroxyl radicals in the presence of iron chelates", FEBS Lett., 1978, 92, 321-326.

13. S. Goldstein; D. Meyerstein; G. Czapski, "The Fenton Reagents", Free Rad. Biol. Med., 1993, 15, 435-445.

14. C. C. Winterbourn, "Toxicity of iron and hydrogen peroxide: the Fenton reaction", Toxicol. lett., 1995, 82/83, 969-974.

15. A. Ghiselli, "Aromatic Hydroxylation: Salicylic acid as a probe for measuring hydroxyl radical production", In "Free radical and antioxidant protocols" ; Armstrong, D., Ed.; Humana Press:, 1998; Vol. 108.

16. M. A. Oturan; J. Pinson, "Hydroxylation by Electrochemically Generated OH radicals. Mono- and Polyhydroxylation of Benzoic acid: Products and Isomers' Distribution", J. Phys. Chem., 1995, 99, 13948-13954.

17. D. A. Wink; R. W. Nims; J. E. Saavedra; W. E. U. JR; P. C. Ford, "The Fenton oxidation mechanism: Reactivities of biologically relevant substrates with two oxidizing intermediates differ from those predicted for the hydroxyl radical.", Proc. Natl. Acad. Sci., 1994, 91, 6604-6608.

18. H. Iwahashi; C. E. Parker; R. P. Mason; K. B. Tomer, "Radical identification by liquid chromatography/thermospray mass spectrometry", Rapid Commun. Mass Spec., 1990, 4, 352-354.

19. P. T. Kissinger; C. R. Preddy; R. E. Shoup; W. R. Heinman, "Fundamental concepts of analytical electrochemistry", In "Laboratory techniques in electroanalytical chemistry" ; Kissinger, P. T., Heinman, W. R., Eds.; Marcel Dekker:, 1984.

20. U. Jurva; H.V. Wikstrom, L. Weidolf, A.P. Bruins, “Comparison between electrochemistry/mass spectrometry and cytochrome P450 catalyzed oxidation reactions” Rapid Commun. Mass Spectrom. 2003, 17, 800-810.

21. U. Jurva; A. P. Bruins; H. V. Wikström, "In vitro mimicry of metabolic oxidation reactions by electrochemistry/mass spectrometry", Rapid Commun. Mass Spec., 2000, 14, 529-533.

Chapter 4 The electrochemically assisted Fenton system applied to the prenylated flavonoid Xanthohumol

53

Chapter 4

The electrochemically assisted Fenton system

applied to the prenylated flavonoid Xanthohumol

The reactivity of the prenylated flavonoid xanthohumol (XN) towards hydroxyl radicals has been investigated in an electrochemically assisted Fenton system connected on–line with LC/MS/MS. An experiment with rat liver slices was also performed and it was found that out of the 10 metabolites showing a mass gain of 16 Da (oxygen), 9 were also formed in the electrochemically assisted Fenton system. Three previously reported metabolites were identified as dehydrocycloxanthohumol (DX), dehydrocycloxanthohumol hydrate (DH) and its furano analogue DF. Based on the data derived from the electrochemically assisted Fenton system, seven additional metabolites from the liver slice experiments are suggested. The main metabolite, 1-(3,5-dihydroxy-7-methoxy-2,2-dimethyl-chroman-6-yl)-3-(4-hydroxy-phenyl)-propenone, was collected and identified by NMR. The main oxidation products formed in the electrochemically assisted Fenton system are believed to originate from addition of a hydroxyl radical at the prenyl function. The radical scavenging properties of the prenyl function might provide an explanation for the increased ability of prenylated flavonoids to inhibit lipid peroxidation as compared to their non–prenylated counterparts. The electrochemically assisted Fenton system provides a valuable tool for investigation of the reactivity of flavonoids towards oxygen radicals. In the case of xanthohumol, the system produces an array of oxidation products that resemble oxidative metabolism in biological systems. If the system is scaled up it can provide a means for synthesis of metabolites in larger quantities for further characterization with other techniques, such as NMR. 4.1 Introduction

Flavonoids is the collective name for more than 4000 different diphenylpropanes that commonly occur in plants, which are present in the human diet such as fruits, vegetables, nuts and seeds.1 The family members of flavonoids include flavones, flavonols, isoflavones, flavanones and chalcones. Chalcones are isomers of flavanones and both flavonoid types are key metabolites in the biosynthesis of flavonoids. The chalcone xanthohumol (XN), described in Figure 4.1, is the major prenylated flavonoid of the hop


54

plant (Humulus lupulus) and is subsequently present in beer, the major dietary source of prenylated chalcones for humans.2

A B

O

OHOMe

OH

HO2 4

62'

5'4'

6'

2"

1"

3"

4"

5"

α

β

3'

Figure 4.1 Xanthohumol

A number of in vitro studies suggest that xanthohumol is a promising agent for chemoprevention of cancer, an important issue in public health-related research aimed to halt or reverse the development and progression of cancer cells through the use of nontoxic nutrients and/or pharmacological agents. Experiments with human cell lines demonstrated that XN is an effective antiproliferative agent (prevents cell growth) in human breast cancer cells, colonal cancer cells and ovarian cancer cells.3 XN has also been shown to induce quinone reductase in mouse Hepa 1c1c7 cells, a model to asses potential anticarcinogenic activity.4 A number of cytochrome P450 enzymes responsible for metabolic activation of chemical carcinogens were effectively inhibited by XN.5 Inhibition by XN of the mutagenic activation of the potential human carcinogen 2-amino-3-methylimidazo[4,5]quinoline has also been demonstrated.6

Flavonoids generally show antioxidant properties that are related to their ability to chelate metal ions and to scavenge reactive oxygen species, such as peroxyl radicals and hydroxyl radicals.7

Rodriguez et al. demonstrated that xanthohumol and other prenylated flavonoids are better inhibitors of microsomal lipid peroxidation than their nonprenylated counterparts.7 In the same study, xanthohumol did not chelate Fe2+ or Fe3+, indicating that the antioxidant effect of the prenylated chalcones might originate from their ability to scavenge reactive oxygen species on the prenyl function.

Despite the high abundance of flavonoids in nature, little is known about their metabolism. In a study with rat liver microsomes, it was found that the major metabolites of XN originate from oxidation at the prenyl function (i.e. positions 1” to 5” in Figure 4.1) and that the CYP1A and CYP2B families might be involved in their biotransformation.8 A metabolism pathway initiated by cytochrome P450 catalyzed epoxidation of the prenyl double bond was proposed.


55

Xanthohumol was thus considered an interesting compound for the electrochemically assisted Fenton system described in chapter 3. The following questions were investigated: i) Can the on–line electrochemically assisted Fenton system be used to gain additional understanding of the antioxidant properties of xanthohumol? ii) Is it possible to mimic the biotransformation of xanthohumol in rat liver slices with the electrochemically assisted Fenton system? 4.2 Experimental

Chemicals Xanthohumol (XN), dehydrocycloxanthohumol (DX), dehydrocycloxanthohumol hydrate (DH) and its furano analogue 1-[4-hydroxy-2-(1-hydroxy-1-methyl-ethyl)-6-methoxy-2,3-dihydro-benzofuran-5-yl]-3-(4-hydroxy-phenyl)-propenone (DF) were generous gifts from Dr. J. F. Stevens at the Institute of Plant Biochemistry, Halle, Germany and Dr. C.L. Miranda at Oregon State University, Corvallis, Oregon, USA.

The following chemicals were obtained commercially: Iron(III)chloride hexahydrate, 98 %, Aldrich; ethylenediaminetetraacetic acid disodium salt (EDTA), 99 %, Merck; L-(+)-ascorbic acid, 99 %, Aldrich; hydrogen peroxide, 35 % solution in water, Aldrich; tert-butyl hydroperoxide, 70 % solution in water, Fluka; DMSO–d6, 99.95%, Merck. The water used in the experiments was purified in a Maxima Ultrapure water system (ELGA, High Wycombe, Bucks, U.K.) and was sonicated for about 15 minutes before use.

The chemicals used in the liver slice experiment were of analytical grade and were obtained from commercial sources. The EC/LC/MS setup The EC/LC/MS system was set up as described in chapter 3 with the following modifications.9 To dilute the sample and to adjust the acetonitrile content to fit the start of the gradient, a make up liquid containing 25 % acetonitrile and 0.1 % formic acid was added at a flow rate of 50 µl/min from a KDS100 syringe pump (KD Scientific, Boston, Massachusetts, USA). The diluted sample from the electrochemical cell was collected on–line in a 100 µl injection loop and injected onto a reverse phase HPLC column (Kromasil, C8, 5 µm, 250 x 4.6 mm, Eka Nobel AB, Bohus, Sweden). Two Series 200 micro LC pumps (Perkin Elmer, Norwalk, CT, USA) were programmed to deliver a gradient from water and acetonitrile at a total flow rate of 1 ml/min. Both phases contained 0.1 % formic


56

acid. During the first two minutes, the acetonitrile concentration was kept at 30 %, and was then linearly increased to 95 % over a period of 15 minutes. The mobile phase composition was then linearly brought back to 30 % acetonitrile in 2 minutes. The system was allowed to equilibrate for at least 10 minutes between the injections. A splitter was introduced between the HPLC–column and the mass spectrometer to reduce the flow to approximately 40 µl/min. The standards were injected on the same LC/MS system with the electrochemical cell disconnected. The ESA coulochem working electrode is porous graphite and all reported cell potentials are versus a palladium reference electrode. HPLC system for collection of fractions for NMR The following adjustments were made to the analytical system described above. The dissolved sample was injected with a 1000 µl injection loop onto a reverse phase HPLC column (Kromasil, C8, 5 µm, 250 x 10 mm, Eka Nobel AB, Bohus, Sweden). The two Series 200 micro LC pumps were programmed to deliver the same gradient as described for the analytical system but at a total flow rate of 4 ml/min. A splitter was introduced between the HPLC–column and the mass spectrometer to reduce the flow to the mass spectrometer to approximately 40 µl/min. The second outlet from the splitter was connected to a Perkin Elmer 785A UV-detector and the UV absorption was registered at 370 nm. The fractions were collected manually in 10 ml glass tubes and evaporated under a stream of nitrogen gas. NMR NMR spectra were run on a Varian UNITY instrument at 500 MHz (1H) and 125 MHz (13C) in DMSO-d6 at room temperature. The solvent resonances (δH 2.50 and δC 49.51) were used as internal shift references. 2-Dimensional experiments (1H-1H COSY, 1H-13C HSQC and HMBC) were carried out using standard Varian pulse sequences. The electrochemically assisted Fenton system (EC–Fenton system) The samples injected through the electrochemical cell contained 1 mM xanthohumol, 1 mM FeCl3, 1.02 mM EDTA and 10 mM hydrogen peroxide for generation of hydroxyl radicals or 10 mM tert-butyl hydroperoxide for generation of tert-butoxyl radicals. The final concentration of acetonitrile in the sample was 40 %. Hydrogen peroxide was kept in the refrigerator and was added just before the injection. The “inactive” Fenton mix was injected at a flow rate of 50 µl/min. After 1 minute, the flow rate was lowered to 2 µl/min. Twenty minutes after the first injection, the diluted sample was injected with the second injection loop onto the HPLC column. The cell potential was kept at –0.5 V.


57

The chemical Fenton system The following system was found to give the highest yield of the oxidation product eluting at RT 7.7 minutes. To a 20 ml glass vial was added: Xanthohumol (1 ml, 10.0 mM in 50% acetonitrile/50% water), FeCl3 / EDTA solution (1 ml, 10.0 mM FeCl3 , 10.2 mM EDTA in 50% acetonitrile/50% water), L(+)ascorbic acid (8.8 mg, 50 µmol), acetonitrile (4 ml) and water (4 ml). The reaction was started by addition of hydrogen peroxide (20 µl, 10M in water). The solution was stirred at room temperature for 20 h. The vial was put in the freezer and the acetonitrile layer was collected and evaporated under a stream of nitrogen gas. The residue was dissolved in a small volume of mobile phase and injected onto the HPLC system for collection of fractions for NMR. Rat liver slice experiment The procedure for preparation of precision–cut liver slices has been described in detail earlier.10 Liver slices of male Whistar rats were incubated in 3.2 ml William’s medium E that was pre–warmed and gassed with 95% O2 / 5% CO2 and supplemented with glucose (final concentration 25 mM) and gentamicin (50 µg/ml). Three slices were individually incubated in a six–well culture plate that was placed in a plastic container, continuously gassed with humidified 95% O2 / 5% CO2, and shaken back and forth (90 times/min) in a cabinet at 37°C. At t=0, xanthohumol (3.2 µl, 0.1M) was added. Medium samples (1 ml) were taken out after 1, 5 and 24h and kept in the freezer until the time of analysis. Prior to analysis by LC/MS as described above, the thawed samples were diluted 5 times in mobile phase (70% water, 30% Acetonitrile, 0.1% HCOOH).


58

4.3 Results and discussion

Introduction

O OHOMe

OH

HO

O O

OHOMe

OH

OHO

DH, mw 370 DF, mw 370

mw 370

O

OHOMe

OH

HO

Xanthohumol, mw 354

OHO

OHOMe

OH

HO

O OHOMe

OH O

DX, mw 352

mw 370O

OHOMe

OH

O

HO

a) b)

c) d)

e) f)

Figure 4.2 Molecular structure of xanthohumol and the metabolites found in experiments with rat liver microsomes and identified by means of HPLC, MS/MS and NMR. a) xanthohumol b) dehydrocycloxanthohumol (DX) c) dehydrocycloxanthohumol hydrate (DH) d) 1-[4-hydroxy-2-(1-hydroxy-1-methyl-ethyl)-6-methoxy-2,3-dihydro-benzofuran-5-yl]-3-(4-hydroxy-phenyl)-propenone (DF) e) metabolite with hydroxylation on the phenol function f) Metabolite suggested to be 5’’-isopropyl-5’’-hydroxydihydrofurano[2’’, 3’’:3’, 4’]-2’, 4-dihydroxy-6’-methoxychalcone. The metabolites e and f were suggested on basis of retention time and MS/MS fragmentation only.

The metabolites of xanthohumol reported from experiments with rat liver microsomes are given in Figure 4.2.8 The metabolites c and d, dehydrocycloxanthohumol hydrate (DH) and its furano analogue, 1-[4-hydroxy-2-(1-hydroxy-1-methyl-ethyl)-6-methoxy-2,3-dihydro-benzofuran-5-yl]-3-(4-hydroxy-phenyl)-propenone (DF) were proposed to originate from an initial epoxidation of the prenyl double bond (2", 3") catalyzed by cytochrome P450 epoxidase. Subsequent nucleophilic attack from the 4' hydroxyl group at C–3" or C–2" would give DH or DF. The metabolite dehydrocycloxanthohumol (DX) originates from dehydration of DH. Cytochrome P450 hydroxylase catalyzed hydroxylation of the phenol B–ring was proposed to give the metabolite hydroxylated at the phenol function (B–ring).


59

14.72.5 x 10

6

100 150 200 250 300 350

6.2 x 105179

235

299

355

a)

b)

14.72.5 x 10

6

100 150 200 250 300 350

6.2 x 105179

235

299

355

a)

b)

Figure 4.3 50 µM Xanthohumol standard. a) Mass chromatogram of m/z 355, corresponding to the [M+H]+ ion of xanthohumol. b) Product ion spectrum of m/z 355. Step size 0.1 was used. The number above the peak in chromatogram a represent the retention time in minutes. The signal intensity is indicated as counts/s in the upper right corner.

The mass chromatogram of m/z 355, corresponding to the [M+H]+ ion of xanthohumol is given in Figure 4.3 together with its corresponding product ion spectrum. The MS/MS fragmentation pathway for xanthohumol suggested by Stevens et al. is presented in Figure 4.4.2


60

O

OHOMe

OH

HO

m/z 179

m/z 299

m/z 355

OMe

OH

H

O

C

O

m/z 235

RDA

O

OHOMe

OH

O

H

H

m/z 355

+ +

+

OMe O

C

OHO+

m/z 179

m/z 179

OMe O

C

O

CH2

O

OMe

HO O

C

O

+

+O

O

OH

OMe

OH

H +

O

O

OH

OMe

HO

H2C

H2C

m/z 299

Figure 4.4 Fragmentation of the [M+H]+ ion of xanthohumol. The loss of the prenyl function is illustrated with a mechanism involving nucleophilic attack at the 4’–OH2

+ oxygen by C–1” as suggested by Stevens et. al.

The fragment at m/z 299 was explained with loss of the prenyl substituent through a nucleophilic attack at the 4’–OH2

+ oxygen by C–1”. The presence of this fragment in a product ion scan of m/z 371 would suggest that the mass gain of 16 Da is located at the prenyl function, leaving the chalcone skeleton intact. Isomerization to the flavanone structure followed by cleavage of the heterocyclic ring via a retro Diels–Alder (RDA) mechanism yields a fragment at m/z 235. Subsequent loss of the prenyl substituent gives a


61

fragment at m/z 179. A gain of 16 to this fragment giving a fragment at m/z 195 would suggest hydroxylation at the A–ring (Figure 4.1).

O

OHOMe

OH

HO

m/z 299

O

OHOMe

OH

O

H

H

m/z 355

+

O

OHOMe

OH

OH

+

+ O

OHOMe

OH

OHa)

b)

O

OHOMe

OH

OH

+

+

m/z 355

m/z 355

m/z 355

O

OHOMe

OH

O

H2C

H +

m/z 299

Figure 4.5 Alternative mechanism for loss of the prenyl function via initial proton transfer to the prenyl double bond.

Alternative mechanisms for loss of the prenyl function via initial proton transfer to the prenyl double bond are presented in Figure 4.5. We believe that these mechanisms are more likely than a mechanism proceeding via a nucleophilic attack at the 4’–OH2

+ oxygen by C–1” as described in Figure 4.4. The retro Diels–Alder mechanism described in Figure 4.4 provides a plausible explanation for the formation of the fragment at m/z 235 but this fragment can also be explained without rearrangement to the flavanone structure as described in Figure 4.6. For simplicity of presentation, the fragments described in Figure 4.4 are used in the interpretation of MS/MS data throughout this chapter.


62

O

OHOMe

O

HO

HH +

m/z 355

O

OHOMe

OH

HOH H

H

+

C

OMe

OH

HO

O

+C

OMe

OH

HO

OC

OMe

OH

HO

O

+

+

m/z 355

m/z 235m/z 235m/z 235 Figure 4.6 Alternative to the retro Diels–Alder mechanism

The two metabolites DH and DF (Figures 2c and 2d) were identified earlier and were available as standards.8 The mass chromatograms of protonated DH and DF are presented in Figure 4.7 together with the corresponding product ion spectra.


63

Figure 4.7 Standards of known metabolites of xanthohumol with a mass gain of 16 Da. Mass chromatograms of m/z 371 for a) 10µM DH standard, b) 50 µM DF standard. Products of m/z 371 for c) 10µM DH standard, d) 50 µM DF standard. The numbers above the peaks in chromatograms a and b represent retention times in minutes. The signal intensity is indicated as counts/s in the upper right corner of each chromatogram and spectrum.

11.7DF

12.4DHa) DH standard

b) DF standard

100 150 200 250 300 350

147179

233251

299

311

353

3711.8 x 10

4

100 150 200 250 300 350

147179

233

251

299311

353

371

193

2.9 x 105c) DH standard d) DF standard

2.0 x 10 5

1.4 x 10 611.7DF

12.4DH 12.4DHa) DH standard

b) DF standard

100 150 200 250 300 350

147179

233251

299

311

353

3711.8 x 10

4

a) DH standard

b) DF standard

100 150 200 250 300 350

147179

233251

299

311

353

3711.8 x 10

4

100 150 200 250 300 350

147179

233

251

299311

353

371

193

2.9 x 105c) DH standard d) DF standard

2.0 x 10 5

1.4 x 10 6


64

O OHOMe

HOH

HO

O+

OH

O+

OHOMe

OH O

H+O

O

OH

OMe

H+O

O

O

OH

OMe

H+O

O

OH

O OHOMe

OH O

H+

H+O O

C

OOMe

H+O O

C

OOMe

OH

O

C

OOMe

H+O

O OHOMe

OH

HO

O

H+

RDA

H2O-

H2O-

RDA

m/z 371 m/z 371 m/z 299

m/z 353 m/z 371m/z 147

m/z 353 m/z 233 m/z 251 m/z 179

H

OH-

H

OH-

Figure 4.8 Suggested fragmentation of the [M+H]+ ion of DH

The suggested MS/MS fragmentation pathway for DH is presented in Figure 4.8. Compared to the fragmentation pattern of xanthohumol, the fragment at m/z 235 has been replaced with m/z 251. The fragment at m/z 353 corresponds to dehydration of the [M+H]+ ion of DH and the fragment at m/z 233 is a result of dehydration of the fragment at m/z 251. The loss of a water molecule is favorable because the resulting double bond is in conjugation with the A–ring. The fragment at m/z 147 indicates that the B–ring and the α, β– double bond are intact. Results from the EC–Fenton system and the liver slice experiment Figure 4.9 a–d shows mass chromatograms of m/z 371, corresponding to a gain of 16 compared to the [M+H]+ ion of xanthohumol. The peaks with retention times 7.7, 8.7, 9.6, 10.7, 11.0, 11.7, 12.4, 12.5 and 13.7 minutes in the EC–Fenton system (Figure 4.9a) are also found in the experiment with liver slices (Figure 4.9b). The peaks with the retention times 11.2, 12.8 and 13.0 in the EC–Fenton experiment have no corresponding peaks in the liver slice experiments. There is also a peak present in the chromatogram from the liver slice experiments at RT 9.3 that is absent in the chromatogram from the EC–Fenton system. The yield of metabolites from the liver slice experiment was low (less than 3% after 24h), and useful product ion spectra were only obtained for the peaks with retention times 7.7 and 12.5 minutes. The fragmentation patterns for those two peaks were the same as for the corresponding peaks in the EC–Fenton system. The product ion spectra for the different peaks in mass chromatogram 4.9a are given in Figure 4.10a–l.

The mass chromatogram of m/z 387 in Figure 4.9e represent a gain in molecular weight of 32 Da, corresponding to an introduction of two oxygen atoms at the


65

xanthohumol structure. Since the intensity of the peaks are much lower compared to the mono hydroxylated products, no thorough investigation of the MS/MS fragmentation was performed.

a)

b)

c)

d)

7.7

8.7

9.6 10.7

11.0,

11.211.7

12.4

12.813.0

13.7

12.5

7.7

8.79.6 11.0

11.7

12.4,

12.59.3

12.4DH

11.7DF

13.7

10.7

e)

1.6 x 10 6

1.4 x 10 5

1.4 x 10 6

2.0 x 10 5

1.5 x 10 5

2.3

3.23.6

5.8

6.2

7.0

7.78.8

10.010.3

11.9 12.1

12.8 16.8

a )

b )

c)

d )

a)

b)

c)

d)

7.7

8.7

9.6 10.7

11.0,

11.211.7

12.4

12.813.0

13.7

12.5

7.7

8.79.6 11.0

11.7

12.4,

12.59.3

12.4DH

11.7DF

13.7

10.7

e)

1.6 x 10 6

1.4 x 10 5

1.4 x 10 6

2.0 x 10 5

1.5 x 10 5

2.3

3.23.6

5.8

6.2

7.0

7.78.8

10.010.3

11.9 12.1

12.8 16.8

a )

b )

c)

d )

Figure 4.9 Mass chromatograms of m/z 371 for a) the EC–Fenton system b) the liver slice experiment at 24 h c) 10 µM DH standard d) 50 µM DF standard. Chromatogram e represents the mass chromatogram of m/z 387 for the EC–Fenton system. The numbers above the peaks represent retention times in minutes. The signal intensity is indicated as counts/s in the upper right corner of each chromatogram.


66

100 150 200 250 300 350

147

179

233251

299

311353

371

1.3 x 105a) RT 7.7 min

100 150 200 250 300 350

179

233

311

353

371

195

315

b) RT 8.7 min 2 X 104

100 150 200 250 300 350

179

251 353371315

3.0 x 103c) RT 9.6 min

107 153 209

235

256

269

297325285

100 150 200 250 300 350

179

233

311353

1.6 x 104d) RT 10.7 min

100 150 200 250 300 350

147

233

183

303

1.2 x 104e) RT 11.0 min

100 150 200 250 300 350

179

353

235

269

f) RT 11.2 min1.6 x 10

4

[1]

[3]

[2]

[6][5]

[4]

100 150 200 250 350300100 150 200 250 350300

Figure 4.10 Product ion spectra of m/z 371. The retention times correspond to the peaks in chromatogram 4.9a. The signal intensity is indicated as counts/s in the upper right corner of each spectrum. The number within brackets is the compound number assigned to each peak and will be used throughout the text. Extracted ion chromatograms of the fragments (not presented in this Figure) clearly show which fragment belongs to a certain peak and for clarity of presentation, some fragments belonging to overlapping peaks have been removed from spectra 10i and 10k.


67

100 150 200 250 300 350

147

179233

251

299 353

371

193

g) RT 11.7 min (DF)8.0 x 10

3

100 150 200 250 300 350

147

179

233 251

299

311

353

371

h) RT 12.4 min (DH)4.0 x 10

4

100 150 200 250 300 350

179

233

353311

i) RT 12.5 min 6.0 x 104

100 150 200 250 300 350

179

233

251

353371

205

297

j) RT 12.8 min 2.5 x 104

100 150 200 250 300 350

179

353371

297

235

k) RT 13.0 min 1.2 x 104

100 150 200 250 300 350

1.4 x 104

179

233 251

299

353

371

l) RT 13.7 min

[7][8]

[9][10]

[11] [12]

100 150 200 250 350300

Figure 4.10 Continued

Identification of the products in chromatogram 4.9a and 4.9b

RT 7.7 minutes (Compound 1) Compound 1 was the main product with m/z 371 from both the EC–Fenton system and from the experiment with liver slices (Figure 4.9a and 4.9b). Both peaks have the same MS/MS fragmentation pattern that suggests hydroxylation at the prenyl group. In order to obtain further structural information about this peak, a larger batch was prepared in a chemical Fenton system and HPLC–fractions were collected for NMR. The combined fractions gave 0.8 mg pure product that was dissolved in DMSO–d6 and analyzed by


68

NMR. The results from the NMR analysis and the suggested structure are presented in Table 4.1.

Table 4.1 NMR data (DMSO-d6, 500 MHz) and suggested structure for compound 1 (RT

7.7 minutes).

O

OMeHO

O

OH

HO

1

23

4

56

α

β1'2'

3'

4'5'

6'

1"

2"

3"

4" 5" Atom no.a δ1H, multiplicity, (Jb in Hz) δ13Cc α 6.69 d (J = 16 Hz) 126.3 β 7.06 d (J = 16 Hz) 143.6 2/6 7.45 d (J = 8.6 Hz) 130.2 3/5 6.78 d (J = 8.6 Hz) 115.9 4 OH and 4’ OH 9.74 s and 10.01 s 6’ OCH3 3.59 s 55.3 5’ 6.11 s 91.2 1” a 2.73 dd (J = 16.6, 5.7 Hz) 25.9 1” b 2.32 dd (J = 16.7, 8.1 Hz) 2” 3.56 d‘t’ (J = 7.9, ca 5.4 Hz) 67.7 2” OH 5.10 d (J = 4.8 Hz) 4” 1.14 s 25.4 d 5” 1.04 s 19.7 d a Xanthohumol atom numbering used. b Coupling constants were directly determined from the 1H spectrum and may deviate from the theoretical values. c 13C shifts determined from the 1H –13C HSQC spectrum. d May be interchanged

The 1H and 13C shift values recorded for the cyclic prenyl moiety are in good agreement with those published for xanthohumol B (= DH) by Stevens et al.11 The 1H NMR spectrum of xanthohumol B differed from that of its isomer eluting at 7.7 min (compound 1) by the presence of a low-field hydrogen bonded 2’-hydroxy group (δH 14.8). This indicates that compound 1 does not have a hydrogen bond between the 2’-OH and the carbonyl function,


69

presumably because the original 2’-OH group in compound 1 has provided the heteroatom in the newly formed pyrano ring.

In addition to the data presented in Table 4.1, the NOESY spectrum (2–dimensional) shows interaction between the B–ring protons and the α and β protons, which demonstrate that the O=C–CH=CH–PhOH moiety is not modified. The doublet at 5.10 ppm showed a cross peak with the H2O signal, indicating that this proton (2” OH) is exchangeable.

The UV absorption maximum in acetonitrile was shifted from 370 nm for XN and DH to 320 nm, which is in agreement with the UV characteristics of other chalcones lacking 2’-hydroxy groups12

RT 8.7 minutes (Compound 2) Compound 2 was formed both in the EC–Fenton system and in the liver slice experiment. The product ion spectrum of compound 2 is presented in Figure 4.10b. The fragment at m/z 195 corresponds to a gain of 16 compared to the fragment at m/z 179 in the fragmentation of xanthohumol (Figure 4.3). A possible explanation for this fragmentation pattern is hydroxylation at the 5’–position. The strong peak at m/z 353 suggests that water is readily lost in the fragmentation. Aromatic hydroxyl groups are strongly bound to the aromatic ring and loss of water is generally not a major fragmentation pathway. However, with a hydroxylation at the 5’–position, ring A of compound 2 contains three hydroxyl groups and the loss of water from compound 2 can be explained by the formation of quinone fragments as illustrated in Figure 4.11.

m/z 371 m/z 353 m/z 353

O

OHO

OH

OH

+

O

OHO

OH

O

H

H

OH

+

O

OHO

OH

H

O

+

m/z 353

O

OHO

O

O

H H

H O

OHO

O

O

H+

+

O

O

OO

+

m/z 353m/z 233

5'

Figure 4.11 Compound 2, alternative 1: hydroxyl group located at the 5’–position. Possible explanation for the fragments at m/z 353 and m/z 233.


70

m/z 371

O

OHO

O

HO

HH

HO

+O

OHO

OH

HO

H

H

HO

+m/z 371

–

O

OHO

OH

HO

HO

H+

m/z 315

O

O

OH

O

HO

HO

H+

m/z 315

RDA

+

O

C

OO

HO

HO

H

m/z 195

1''

Figure 4.12 Compound 2, alternative 2: hydroxyl group located at the 1’’–position. Possible explanation for the fragments at m/z 315 and m/z 195.

Hydroxylation at the 5’–position is one explanation for the appearance of a fragment at m/z 195 but this fragment can also be explained from hydroxylation at the 1’’–position as illustrated in Figure 4.12. We have assigned this structure to compound 4 (see below) and believe that compound 2 is hydroxylated at the 5’–position but further characterization is needed to differentiate between these isomers. RT 9.6 minutes (Compound 3) Compound 3 was formed both in the EC–Fenton system and the liver slice experiment. The product ion spectrum of compound 3 is given in Figure 4.10c. A plausible explanation for this fragmentation pattern is a compound hydroxylated at the a-position of the a, ß–double bond. The suggested fragmentation is described in Figure 4.13.


71

O

OHOMe

OH

O

H

H OH+

mw 371mw 371

O

O

OH

OMeH

H

OH

OH+

m/z 353

m/z 315

m/z 297

- CO

O

OHH+

OMe

O

m/z 269

- CO

H+O

OMe O

C

O

RDA

- H2O

m/z 179

- H2C=C(CH3)2

- H2C=C(CH3)2

- H2O

O

O

OH

OMe

OH

H+

O

O

OH

OMe

OH+

m/z 325

O

OH

OMe

OH

H+

m/z 235

+

OMe

OH

H

O

C

O

O

O

OH

OMe

OH

H

H

H

OH

+

m/z 371 m/z 107

+OH

H2C

m/z 107

+H2C

OH

+

α

hydroxyl at α− carbon

H

OHO

OH

OMe OH

O

Figure 4.13 Suggested fragmentation of protonated Compound 3.

RT 10.7 and 12.5 minutes (Compounds 4 and 9) Compounds 4 and 9 were both formed in the EC–Fenton system as well as in the liver slice experiment. The product ion spectra are presented in Figure 4.10d and 4.10i. Both spectra contain mainly the fragments at m/z 353, 233 and 179. An explanation for this fragmentation pattern with a very intense fragment at m/z 233 would be the chalcone and the flavanone of a compound hydroxylated at position 1”. The suggested fragmentation of the chalcone is presented in Figure 4.14.


72

O

OHOMe

OH

HO

OH

H

+

m/z 371

1"

OCH3

O

HO

O

OHH

+

H2O_

OCH3

O

HO

O

OH

H

H

H+

OCH3

O

HO

O+

OH_

m/z 353

m/z 353

m/z 233

m/z 353

OCH3

OH

O

OH

O

H+

OCH3

OH

O

O

OH

H

+

OCH3

OH

O

O

OH

RDA

OCH3

OH

C

O

O

m/z 233

OH_

m/z 353m/z 353

Figure 4.14 Suggested fragmentation of a product hydroxylated at position 1”.

RT 11.0 minutes (Compound 5) Compound 5 was formed both in the EC–Fenton system and the liver slice experiment. The product ion spectrum (Figure 4.10e) contains two fragments at m/z 183 and m/z 303 that are not observed in any of the other spectra. A possible explanation for the formation of these two fragments would be the structure presented in Figure 4.15. Loss of the entire prenyl chain (loss of 68, product ion m/z 303) from the aromatic A–ring would not be expected. Through the introduction of an hydroxyl group at position 3’, the A–ring is no longer an aromatic ring and loss of the prenyl function becomes feasible.


73

O

OHOCH3

OH

OH

OHH

+

m/z 371

–

Proton transfer

O

OHOCH3HO

OH

HO

H+

m/z 303

RDA

OH_

O

C

O

OH

HO

OCH3

H+

m/z 183 m/z 303

H+

O

OCH3

HO

OHOH

O

3'

Figure 4.15 Suggested fragmentation of protonated Compound 5.

RT 11.2 minutes (Compound 6) Compound 6 was only detected in the EC–Fenton system. The presence of a fragment at m/z 235 in Figure 4.10f suggests that the prenyl group and the A ring are intact and that the hydroxylation has occurred at either the phenol function or at the α,β–double bond. Hydroxylation at the α,β–double bond has been assigned to compounds 3 (α–carbon) and 10 (β–carbon). Since hydroxylation at position 2 or 6 is assigned to the peak at RT 13.0 minutes (see below), we suggest that compound 6 corresponds to 3–hydroxy–xanthohumol. RT 11.7 (Compound 7, DF) and 12.4 minutes (Compound 8, DH) Compounds 7 and 8 (Figure 4.10g and 4.10h) are present in the EC–Fenton system as well as in the liver slice experiment and were identified as DF and its analogue DH on grounds of having the same retention times and product ion spectra as the corresponding DF and DH standards. RT 12.8 minutes (Compound 10) Compound 10 was only detected in the EC–Fenton system. The presence of the fragment at m/z 251 in Figure 4.10j indicates that the oxygen is located somewhere at the A–ring or at the prenyl function. The most intense fragment at m/z 205 is specific for this compound and the fragment at m/z 297 is only present in spectra 10c and 10k where the hydroxylation has been suggested on the α,β–double bond or the B-ring. Figure 4.16 illustrates how hydroxylation at the β-carbon of the α,β–double bond could account for this type of fragmentation pattern.


74

O

OHOMe

OH

HO

OH

m/z 371

- CO- H2O

α

hydroxyl at β−carbon

β

H +

O

OHOMe

OH

HO

O

H +

H +

O

OHOMe

OH

HO

O

O

OH

MeO

HO

OHHO

+ proton

transfer

hydrogenradical transfer

O

MeO

HO

HOO

H

H

+

m/z 251

OH

MeO

HO

HOO

H+

m/z 251 m/z 233

MeO

HO O

+

m/z 205

O

CH2

OH

MeO

HO

OHHO

+

O

CH3

OH

MeO

HO

HOO

+

OH

–O

CH2

OH

MeO

HO

HOO

H+

C

MeO

HO O

O+MeO

HO

HOO

OH

+

O

O

OH

OMe

HO

H

HO

+

Figure 4.16 Suggested fragmentation of protonated compound 10.

RT 13.0 minutes (Compound 11) Compound 11 was only detected in the EC–Fenton system. The fragment at m/z 235 in Figure 4.10k indicates that the A–ring and the prenyl function are intact. There is also a


75

very abundant fragment ion at m/z 297. The suggested fragmentation scheme in Figure 4.17 describes how hydroxylation at position 2 or 6 in the B–ring could account for this type of fragmentation pattern.

O

OHOMe

OH

O

H

H

HO

m/z 371+

OMe

OH

OOH

H

OH

HO

+

m/z 371

H2O–

m/z 353

OMe

OH

O OH

HO

+

OMe

OH

O OH

O+

H2C=C(CH3)2–

m/z 297 Figure 4.17 Possible fragmentation pathway to account for the stable fragment at m/z 297 in the product ion spectra of Compound 11.

RT 13.7 minutes (Compound 12) Compound 12 was detected in both the EC–Fenton system and in the liver slice experiment. The fragment at m/z 299 in the product ion spectrum in Figure 4.10l indicates that the oxygen is located at the prenyl function. An explanation for this compound would be hydroxylation at either position 4’’ or 5’’. The suggested fragmentation is given in Figure 4.18 where the hydroxyl group is located at the 5’’–position. It should be noted that several isomers, for example the flavanone of DF or DH, could give this type of fragmentation pattern and further characterization is needed to differentiate between the isomers.


76

O

OHOMe

OH

O

H

H

HO

+

O

O

OH

OMe

OH

H

OH

+

m/z 371m/z 371

RDA

+ O

C

OOMe

OH

H

OH

–

OH

O

OHOMe

OH

HO+

m/z 299

m/z 251

–

OH

m/z 179

O

C

OOMe

HO+

Figure 4.18 Possible fragmentation of protonated Compound 12.


77

Dehydration products QuickTime™ and aGraphics decompressorare needed to see this picture.

17.5

17.5

17.5

a)

b)

c) 3.0 x 105

1.2 x 104

1.7 x 105

8.2

100 150 200 250 300 350

191

215205

233d) RT 8.2 min [13] 1.7 x 103

353

100 150 200 250 300 350

f) DX standard 233

191 215205 311

1.4 x 105

353

100 150 200 250 300 350

4.0 x 104e) RT 17.5 min

[14]

353

233

191 215205

311

Figure 4.19 Mass chromatograms of m/z 353 for a) the EC–Fenton system b) the liver slice experiment at 24 h c) 10 µM DX standard. Product ion spectra of m/z 353 for d) compound 13 e) compound 14 and f) DX standard. The numbers above the peaks in chromatogram a–c represent retention times in minutes. The signal intensity is indicated as counts/s in the upper right corner of each chromatogram and spectrum.


78

Figure 4.19 a–c shows mass chromatograms of m/z 353, corresponding to a gain of 16 (oxygen) followed by a loss of 18 (water) compared to the [M+H]+ ion of xanthohumol. The EC–Fenton system (chromatogram 4.19a) gives two major products with m/z 353 at RT 8.2 (compound 13) and 17.5 minutes (compound 14). Compound 13 is present only in the EC–Fenton system. The product ion spectrum of compound 13 is presented in Figure 4.19d and is very similar to the spectrum of DX in Figure 4.19f. A plausible explanation for this product would the dehydrated equivalent of compound 1. Compound 14 was identified as dehydrocycloxanthohumol (DX) on basis of comparison of the retention time and product ion spectrum with a DX standard. A peak with m/z 353 and the same retention time was also found in the liver slice experiment, but it was not strong enough to perform a useful product ion scan. This peak does however most certainly also correspond to DX.

A summary of the suggested oxidation products from the EC–Fenton system with hydrogen peroxide is presented in Table 4.2.


79

Table 4.2 Suggested products in the EC–Fenton system

Compound (RT in minutes)a

Suggested structure Present in the liver slice experiment

Means of identification

1 (7.7)

O

OMeHO

O

OH

HO

Yes MS/MS and NMR

2 (8.7)

O

OHOMe

OH

HOOH

Yes MS/MS

3 (9.6)

O

OHOMe

OH

HOOH

Yes MS/MS

4b (10.7)

O

OHOMe

OH

HO

HO

Yes MS/MS

5 (11.0)

O

OHOMe

OH

O

HO

Yes MS/MS

6 (11.2)

O

OHOMe

OH

HOOH

No MS/MS

a The retention times for compounds 1-12 refer to chromatogram 9a. The retention times for compounds 13 and 14 refer to chromatogram 18a. b The chalcone structure is presented but the flavanone structure is also possible.


80

Table 4.2 Continued

Compound (RT in minutes)

Suggested structure Present in the liver slice experiment

Means of identification

7 (11.7, DF)

O

OHOMe

OH

OHO

Yes Comparison of RT and MS/MS with standard

8 (12.4, DH)

O OHOMe

HO

OOH


9c (12.5)

O

OH

OOMe

HO

OH

Yes MS/MS

10 (12.8)

O

OHOMe

OH

HO

OH

No MS/MS

11 (13.0)

O

OHOMe

OH

HO

OH

No MS/MS

12 (13.7)

O

OHOMe

OH

HO

HO

Yes MS/MS

13 (8.2)

O

OCH3HO

O

OH

No MS/MS

14 (17.5, DX)

O OHOMe

OOH


c The flavanone structure is presented but the chalcone is also possible.

Mechanisms of product formation The results in Table 4.2 show that the major products from the EC–Fenton system originate from insertion of an oxygen atom at the prenyl function. The hydroxyl radical is an electrophile and adds readily to aromatic rings and double bonds.13 The double bond in


81

the prenyl group has a higher electron density than the conjugated chalcone skeleton and is therefore more prone towards an attack by the hydroxyl radical. The hydroxyl radical can also abstract a hydrogen atom from various organic compounds and thereby generate a new radical. In the section below we provide some possible explanations for how the many oxidation products are formed. Addition to the prenyl double bond

HO OCH3

OH O

OHOH

H_

OCH3

OH O

OHO

HO

OCH3

OH O

OHOHO

H2O–

OCH3

OH O

OHO

HO OCH3

OH O

OH

OHH

HO OCH3

OH O

OH

OH

4'

2'

4'

2'

2"3"

O

OCH3HO

O

OH

HO

O

OCH3HO

O

OH

H2O–

13

1 8 (DH) 7 (DF)

14 (DX)

2' 2' 2'

2' 2'

4' 4' 4'

4'4'

4'

2'

Figure 4.20 Suggested reaction pathway for the formation of compounds 1, 7 (DF), 8 (DH), 13 and 14 (DX).

Addition of a hydroxyl radical on the 2” position of the prenyl group gives a more stable radical than addition on the 3” position, and must therefore be assumed to be the main reaction pathway. The resulting tertiary carbon radical can undergo a number of subsequent reactions. Formation of an epoxide over the 2”, 3” double bond as illustrated in Figure 4.20 would provide an explanation for the formation of compounds 1, 7 (DF) and 8 (DH). A nucleophilic attack by the 2’–hydroxyl group on C–3” gives compound 1. A nucleophilic attack by the 4’–hydroxyl group on C–2” and C–3” gives compounds 7 (DF) and 8 (DH). Subsequent loss of water from compounds 1 and 8 (DH) would also explain the presence of compounds 13 and 14 (DX).


82

In analogy with the formation of compound 7 (DF), a nucleophilic attack by the 2’-hydroxyl group on C–2” to give the furano analogue of compound 1 would be expected but no such compound was detected. Compound 7 (DF) is a minor product compared to compound 8 (DH). They elute close to each other in the chromatogram and have similar product ion spectra (Figures 9a, 10g and 10h). If a furano analogue of compound 1 was formed it might be hidden under the large peak of compound 1.

The formation of an epoxide from the tertiary carbon radical is illustrated by the loss of a hydrogen radical in Figure 4.20. This is obviously a simplified view of a more complex reaction. In Figure 4.21, three possible routes are described for the epoxide formation.

R

H

OH H R

HO

R

HOH

R

HOH

R

HOH

R

HOH

1)

2)

3)

OH

H2O

R

HO

Fe 3+

2+Fe

R

HO

H+

HO–OH

R

HOH

+

+

R

HOH

OHOH_

+

R

HO

H+.

.

..

.

.

.

.

Figure 4.21 Possible mechanisms for epoxidation of the prenyl function.

In the first alternative, the hydrogen radical is taken up by a second hydroxyl radical to give the epoxide and a water molecule. This reaction is likely to occur if the two reactants encounter each other but unlikely on grounds of kinetics. Because of its high reactivity, the half–life of the hydroxyl radical is very short. Besides reaction with xanthohumol, the hydroxyl radical can be consumed by a number of competing reactions, such as reduction to hydroxide ion at the electrode or by ferrous ion. This means that the hydroxyl radical concentration will be low. The tertiary carbon radical is much more stable than the


83

hydroxyl radical but it will still have a low concentration. Thus, the chance is low that the tertiary carbon radical and the hydroxyl radical will encounter each other in the solution which suggests that this route contributes very little to the epoxide formation.

In the second alternative, the tertiary carbon radical is oxidized by ferric ion to give a tertiary carbocation and ferrous ion. A subsequent nucleophilic attack by the hydroxyl oxygen gives the protonated epoxide. In our experiments, the initial ferric ion concentration as well as the substrate concentration are 1 mM. Upon passage through the electrochemical cell a large part of the ferric ions will be reduced to ferrous ions but a substantial amount of ferric ions will still be expected in the solution since the electrochemical conversion is probably not 100%. In addition, Ferrous ions are constantly oxidized to ferric ions in the solution by hydrogen peroxide (Fenton reaction) and by hydroxyl radicals. Thus, the ferric ion concentration is expected to be relatively high which makes a ferric ion much more likely to encounter and react with the tertiary carbon radical than a hydroxyl radical. We therefore believe that ferric ion is a more likely candidate as the second oxidant in the reaction than the hydroxyl radical.

In the third alternative, the tertiary carbon radical is oxidized by hydrogen peroxide to give a tertiary carbocation, a hydroxide ion and a hydroxyl radical. In this case, the second oxidation generates a new hydroxyl radical and does not lead to termination of the radical reaction. Since the tertiary carbon radical is substantially more stable than the hydroxyl radical, this reaction might not be favored. However, the hydrogen peroxide concentration is 10 mM in our experiments which makes hydrogen peroxide the most concentrated oxidant in the system.

In conclusion, addition of hydroxyl radical to the prenyl double bond generates a tertiary carbon radical that is oxidized to a tertiary carbocation by ferric ion or possibly hydrogen peroxide. A subsequent nucleophilic attack by the hydroxyl oxygen gives the protonated epoxide. Hydrogen atom abstraction The mechanism described above with initial addition of a hydroxyl radical can also be applied to most of the products with a hydroxyl group introduced on the chalcone skeleton. However, a mechanism involving hydrogen abstraction from the phenolic hydroxyl groups is probably the main pathway since the resulting radical would be stabilized over the chalcone skeleton. This generates a relatively stable radical that could perhaps survive long enough to encounter a second hydroxyl radical and be hydroxylated through a radical–radical reaction. Even with this prolonged lifetime of the radical, the chance of encountering a ferric ion or hydrogen peroxide is far greater than to encounter a hydroxyl radical. Subsequent reaction with hydroxide ion (or water) would result in the same hydroxylated species as illustrated for the formation of compounds 2 and 5 in Figure 4.22.


84

HO OCH3

OH

OH

O

OH H2OO OCH3

OH

OH

O

O OCH3

OH

R

O OCH3

OH

RHO

Compound 5 (RT 11.0 min)

a)

b)

O OCH3

OH

R

H

O OCH3

OH

R

Compound 2 (RT 8.7 min)

O OCH3

OH

R

H OHHO OCH3

OH

R

OH

Fe3+ Fe2+O OCH3

OH

R

H

O OCH3

OH

R

H

+

OH–

1)Fe3+ Fe2+

2) OH–

..

.

..

.

Figure 4.22 Hydrogen atom abstraction followed by a second oxidation by Fe3+ illustrated for the formation of Compound 2 (RT 8.7 min) and Compound 5 (RT 11.0 min).

The intermediate epoxide over the prenyl double bond provides a plausible explanation for the formation of the compounds 1, 7 (DF), 8 (DH), 13 and 14 (DX). However, the formation of these compounds can also be explained by a mechanism involving initial hydrogen atom abstraction as illustrated for the formation of compound 8 (DH) in Figure 4.23.


85

HO OCH3

OH

OH

O

OH H2OO OCH3

OH

OH

O

OCH3

OH

OH

O

OOCH3

OH

OH

O

O

HO OH–2)

Fe3+ Fe2+

1)

.

..

Figure 4.23 Alternative mechanism for the formation of compound 8 (DH) via hydrogen atom abstraction.

Due to the increased electron density of the prenyl double bond compared to the conjugated double bonds in the chalcone skeleton, we believe that addition of the electrophilic hydroxyl radical to the prenyl double bond is the most likely explanation for the formation of compounds 1, 7 (DF), 8 (DH), 13 and 14 (DX), but the mechanism initiated by hydrogen atom abstraction cannot be ruled out.

Compounds 3 and 10 have been assigned to hydroxylation of the α– and β–carbon of the α, β–double bond. Figure 4.24 illustrates how both products can be accounted for by initial hydrogen abstraction from the hydroxyl group at position 4. The introduction of a hydroxyl group at the β–carbon can be achieved by protonation of the quinone followed by addition of a hydroxide ion. In order to regain the conjugated system the generated diol would be easily dehydrated to give either of the compounds 3 or 10. The same products would be expected from hydrolysis of an epoxide formed via an addition mechanism as described above. However, the electron density of the α, β–double bond is relatively low and the mechanism initiated by hydrogen abstraction seems more likely.


86

OH H2O

OH–

OH

O

Rα

β

4O

O

Rα

β

O

O

R

H

α

β

Fe3+

Fe2+

O

O

R

H

α

β

+O

O

R

H

H

α

β+

H+

O

O

R

H

HHOα

β

O

O

R

H

HHO

H

αβ

+OH

O

R

OHα

β

OH

O

RHHO

OHH

α βH2O–

H2O–OH

O

R

OH

α

β

Compound 3

Compound 10

...

Figure 4.24 Possible mechanism for the formation of compounds 3 and 10.

Tert-butoxyl radicals As described in chapter 3, the EC–Fenton system can be modified to generate a large range of radicals. In these experiments, hydrogen peroxide was replaced with tert-butyl hydroperoxide in order to generate tert-butoxyl radicals. Hydroxyl radicals can also be formed from such a system, but the relative stability of the tert-butoxyl radical compared to the hydroxyl radical makes the tert-butoxyl radical the major product of the reaction.


87

12.2

13.2

14.9

20.2

20.7

22.3

2.5 x 105

Figure 4.25 Extracted ion chromatograms of m/z 427 corresponding to a gain of 72 compared to the [M+H]+ ion of xanthohumol, equivalent with the mass gain from addition of (CH3)3C-O combined with loss of a hydrogen atom. The numbers above the peaks represent retention times in minutes. The signal intensity is indicated as counts/s in the upper right corner.


88

50 100 150 200 250 300 350 400 450

a) RT 12.2 min179

233

299

353371

3.5 x 103

50 100 150 200 250 300 350 400 450

b) RT 13.2 min8.0 x 10

3179

233

299

353

371207

50 100 150 200 250 300 350 400 450

c) RT 14.9 min1.6 x 10

3

121

101 207

234

354

50 100 150 200 250 300 350 400 450

d) RT 20.2 min 233

353

427311

50 100 150 200 250 300 350 400 450

e) RT 20.7 min

233

299

371

102

251

353

324 427195

1.6 x 104

1.2 x 103

50 100 150 200 250 300 350 400 450

f) RT 22.3 min2.5 x 10

3

179

299353

371427

410

50 100 150 200 300250 350 400 450

Figure 4.26 Product ion spectra of m/z 427. The retention times correspond to the peaks in the chromatogram in Figure 4.25. The signal intensity is indicated as counts/s in the upper right corner of each spectrum.

Figure 4.25 shows the mass chromatogram of m/z 427, corresponding to a gain of 72 compared to the [M+H]+ ion of xanthohumol, equivalent with the mass gain from addition of (CH3)3C–O radical combined with the loss of a hydrogen atom. The corresponding product ion spectra are presented in Figure 4.26. The fragment at m/z 299, indicating that the gain of 72 is located on the prenyl function, is present in the spectra for the oxidation products eluting at 12.2, 13.2 and 22.3 minutes. The 20.7 minute product also contains a fragment at m/z 299 but the tert-butoxyl is probably located at an aromatic carbon as discussed below. The product eluting at 14.9 minutes contains two even fragments at m/z


89

354 and m/z 234 that correspond to the loss of a (CH3)3C–O radical followed by loss of the prenyl function. In order to form stable radical fragments the tert-butoxyl group is probably located at the chalcone skeleton. Based on the assumption that the fragment at m/z 121 corresponds to O=C–ph–OH the tert-butoxyl group is suggested to be located at the β–carbon. The main oxidation product eluting at 20.2 minutes has a very similar fragmentation pattern to compounds 4 and 9 in the system with hydrogen peroxide (Figure 4.10d and 4.10i). A possible explanation for this oxidation product would be a compound with the tert-butoxyl group located at position 1” on xanthohumol. The spectrum of the 20.7 minute product has a very different distribution of fragments compared to the other spectra and contains a large fragment at m/z 371. Figure 4.27 illustrates how a product with the tert-butoxyl group located at position 5’ could explain this type of fragmentation pattern.

m/z 427

O

OHO

OH

HO

O HH

5'+

_

O

OHO

OH

HO

OH

H+

m/z 371

_ from prenyl function

O

OHO

OH

O

OH 5'+

from tert-butoxyl function

m/z 371

H2O-proton transfer

O

OHO

O

O

H

+

m/z 353

rearrangementto flavanone RDA

m/z 233

+

O

O

OO

O O

C

OO

O

H +

m/z 251

proton transfer

_

O O

C

OO

HO

H +

m/z 195 Figure 4.27 Suggested fragmentation of the 20.7 minute compound with the tert-butoxyl group at position 5’. Formation of the fragments at m/z 353 and 233 is explained in detail in Figure 4.11


90

These results indicate that the prenyl function is not only reactive towards the hydroxyl radical, but can also contribute to the scavenging of other radicals, which might explain why xanthohumol and other prenylated flavonoids are better inhibitors of microsomal lipid peroxidation than their nonprenylated counterparts. Possible applications in drug research In the mimicry of cytochrome P450, the reactions that were possible to mimic with the EC/MS system took place via a one electron oxidation (Chapter 2). The substrates that were supposed to be oxidized via direct hydrogen atom abstraction, such as aliphatic hydroxylation, O-dealkylation and epoxidation or hydroxylation of aromatic rings or double bonds without electron donating substituents, were not electrochemically oxidized within the potential limits of water. In addition to the obvious use of the on–line system in the study of antioxidants and radical scavengers, the EC–Fenton system provides a different type of oxidation, and certain enzymatic oxidations that cannot be mimicked by pure electrochemical oxidations can be mimicked by the EC–Fenton system.

The experiments with xanthohumol presented here demonstrate the possibilities of the EC–Fenton system to generate hydroxylated products that cannot be obtained via direct electrochemical oxidation. The results also bring up one of the drawbacks of this system: the hydroxyl radical is not a selective reagent and will often give a broad range of products. However, although the range of hydroxylated products makes specific generation of one desired product difficult, the yield of the oxidation products is generally high enough for comparison of retention times with metabolites generated in other in vitro systems or in vivo and allows for the generation of product ion spectra that may provide sufficient information for the characterization of an unknown metabolite. For further characterization of the products with other techniques, such as NMR, the system needs to be scaled up.

For investigation of the reactivity of antioxidants towards hydroxyl radicals the interpretation of the MS/MS spectra often allows identification of a specific part of the molecule in which the oxygen is located and thus which part of the molecule is sensitive towards a radical attack. 4.4 Conclusions

The reactivity of the prenylated flavonoid xanthohumol (XN) towards hydroxyl radicals and tert-butoxyl radicals has been investigated in an EC–Fenton system connected on–line with a triple quadrupole mass spectrometer equipped with an electrospray interface. The majority of the products originate from addition of the radicals on the prenyl function. The ability of the prenyl group to scavenge radicals might explain the increased ability of


91

prenylated flavonoids to inhibit lipid peroxidation compared to their non–prenylated counterparts.

A majority of the metabolites formed in an experiment with rat liver slices were also generated in the EC–Fenton system, indicating that in the case of xanthohumol the system provides a good way of mimicking cytochrome P450-type oxidations. The results presented demonstrate how the EC–Fenton system can be used for generation of metabolites for easy comparison of retention times with metabolites generated in other in vitro systems or in vivo . In the present setup, the yields obtained are well above the requirements for LC/MS and sufficient for generation of product ion spectra in LC/MS/MS. If the system is scaled up, for example by the use of a larger electrochemical cell, additional techniques such as NMR can be applied for further characterization of the oxidation products. 4.5 References

1. G. Cao; E. Sofic; R. L. Prior, "Antioxidant and prooxidant behavior of flavonoids: structure-activity relationships", Free Rad. Biol. Med., 1997, 22, 749-760.

2. J. F. Stevens; M. Ivancic; V. L. Hsu; M.L.Deinzer, "Prenylflavonoids from Humulus Lupus", Phytochemistry, 1997, 44, 1575-1585.

3. C. L. Miranda; J. F. Stevens; A. Helmrich; M. C. Henderson; R. J. Rodriguez; Y. H. Yang; M. L. Deinzer; D. W. Barnes; D. R. Buhler, "Antiproliferative and cytotoxic effects of prenylated flavonoids from hops (Humulus Lupus) in human cancer cell lines", Food Chem. Toxicol., 1999, 37, 271-285.

4. C. L. Miranda; G. L. M. Asponso; J. F. Stevens; M. L. Deinzer; D. R. Buhler, "Prenylated chalcones and flavanones as inducers of quinone reductase in mouse Hepa 1c1c7 cells", Cancer Lett., 2000, 149, 21-29.

5. M. C. Henderson; C. L. Miranda; J. F. Stevens; M. L. Deinzer; D. R. Buhler, "In vitro inhibition of human P450 enzymes by prenylated flavonoids from hops, Humulus lupus", Xenobiotica, 2000, 30, 235-251.

6. C. L. Miranda; Y. H. Yang; M. C. Henderson; J. F. Stevens; G. Santana-Rios; M. L. Deinzer; D. R. Buhler, "Prenylflavonoids from hops inhibit the metabolic activation of the carcinogenic heterocyclic amine 2-amino-3-methylimidazo[4,5-F]quinoline, mediated by CDNA-expressed human CYP1A2", Drug Metab. and Dispos., 2000, 28, 1297-1302.

7. R. J. Rodriguez; C. L. Miranda; J. F. Stevens; M. L. Deinzer; D. R. Buhler, "Influence of prenylated and non-prenylated flavonoids on liver microsomal lipid peroxidation and oxidative injury in rat hepatocytes", Food Chem. Toxicol., 2001, 39, 437-445.


92

8. M. Yilmazer; J. F. Stevens; M. L. Deinzer; D. R. Buhler, "In vitro biotransformation of xanthohumol, a flavonoid from hops (Humulus Lupus), by rat liver microsomes", Drug Metab. Dispos., 2001, 29, 223-231.

9. U. Jurva; H. V. Wikström; A. P. Bruins, "Electrochemically assisted Fenton reaction: reaction of hydroxyl radicals with xenobiotics followed by on-line analysis with HPLC/MS/MS", Rapid Commun. Mass Spectrom., 2002, 16, 1934-1940.

10. R. de Kanter; M. H. de Jager; A. L. Draaisma; J. U. Jurva; P. Olinga; D. K. F. Meijer; G. M. M. Groothuis, "Drug metabolizing activity of human and rat liver, lung, kidney and intestine slices", Xenobiotica, 2002, 32, 349-362.

11. J. F. Stevens; A. W. Taylor; G. B. Nickerson; M. Ivancic; J. Henning; A. Haunold; M. L. Deinzer, "Prenylflavonoid variation in Humulus Lupus: distribution and taxonomic significance of xanthogalenol and 4'-O-methylxanthohumol", Phytochemistry, 2000, 53, 759-775.

12. T. J. Mabry; K. R. Markham; M. B. Thomas "The Systematic Identification of Flavonoids" ; Springer: New York, 1970.

13. C. v. Sonntag; H.-P. Schuchmann, "Pulse radiolysis", In "Oxygen Radicals in Biological Systems, Part C"; L. Packer (Ed.); Academic Press:, 1994; Vol. 233.

Chapter 5 The oxidative metabolism of the dopamine agonist N–0923 mimicked by EC/MS and electrochemically assisted Fenton reaction.

93

Chapter 5

The oxidative metabolism of the dopamine agonist

N–0923 mimicked by EC/MS and electrochemically

assisted Fenton reaction.

The two on–line oxidation systems described earlier, the EC/MS system and the EC-Fenton system, have been applied to the dopamine agonist S(–)-(N-propyl-N-2-thienylethylamine)-5-hydroxytetralin (N–0923). The oxidative metabolism previously reported from rat liver perfusion experiments was partly mimicked by both methods. The EC/MS system gave a mimic of the N–dealkylation while the oxidation of the phenol function was not fully mimicked since the catechols (or p–hydroquinones) formed were immediately oxidized to the corresponding quinones. By collecting the oxidized sample solution and injecting it again at a reductive potential, the corresponding catechols (or p–hydroquinones) could be generated.

Oxidation in the EC–Fenton system gave rise to five different products, all showing a gain of oxygen. Interpretation of the product ion spectra suggests that three of these products are hydroxylated at the phenolic ring or at the benzylic carbons 1 or 4. Two other compounds were assigned as the N–oxide and a compound with the oxygen located at the thienylethyl group. The results give a clear indication that the EC/MS system and the EC–Fenton system give different products and that they can be complementary to each other. 5.1 Introduction

The EC/MS system described if chapter 2 and the EC–Fenton system described in chapter 3 are both useful tools for the generation of oxidative metabolites. Both systems have their limitations. Since the mechanism of oxidation is different in these two systems, it is interesting to investigate their individual limitations and the possibility that they may be complementary to each other. We have therefore chosen to investigate to what extent a combination of the two systems can be used to generate metabolites of a drug with a known metabolism pattern.

The drug 2-(N-propyl-N-2-thienylethylamine)-5-hydroxytetralin (N–0437) was first synthesized by Horn et al.1 In vitro as well as in vivo studies have proven it to be a potent and selective D–2 receptor agonist.2,3 N–0437 is a racemic mixture and it was


94

shown that the R(+)–enantiomer, N–0924, is inactive in reducing dopamine release following i.p. injection while the S(–)–enantiomer, N–0923 (Figure 5.1), acts as an agonist at both pre– and postsynaptic receptors and might be of therapeutic use in Parkinson's disease.4,5 The metabolism of N–0437 has been studied extensively.6-8 The S(–)–enantiomer, N–0923, was available in our laboratory and was considered a suitable candidate for a test in the EC/MS system as well as the EC-Fenton system.

N S

H

OH

12

345

4a6

7

88a

9

10

11

1213

14

15 16

17

Figure 5.1 2S(-)-(N-propyl-N-2-thienylethylamino)-5-hydroxy-tetralin, N-0923

5.2 Experimental

Chemicals N–0923 is the S(–)–enantiomer of the dopamine agonist N–0437 [2-(N-propyl-N-2-thienylethylamino)-5-hydroxytetralin] and has been synthesized earlier.1 The following chemicals were obtained commercially: Iron(III)chloride hexahydrate, 98 %, Aldrich; ethylenediaminetetraacetic acid disodium salt (EDTA), 99 %, Merck; hydrogen peroxide, 35 % solution in water, Aldrich; ammonium acetate, p.a., MERCK; glacial acetic acid, p.a., MERCK; lithium trifluoromethane sulfonate, 99.995%, Aldrich; trifluoromethane sulfonic acid, 98%, Aldrich.

The water used in the experiments was purified in a Maxima Ultrapure water system (ELGA, High Wycombe, Bucks, U.K.) and was sonicated for about 15 minutes before use.

Instrumentation The EC/MS system was set up as described in chapter 2 and the EC–Fenton system was used in the same way as described in chapter 3. All reported cell potentials are versus a palladium reference electrode. In all experiments where an HPLC–column was used to separate the oxidation products a splitter was introduced between the HPLC–column and the mass spectrometer to reduce the flow to approximately 40 µl/min. The voltammograms


95

and mass chromatograms presented in Figures 5.4, 5.6, 5.7, 5.9 and 5.11 have been smoothed by "Kalman smooth" in Multiview 1.4 (MDS–Sciex, Concord, Ontario, Canada) prior to reconstruction in DeltaGraph 4.5 (SPSS Inc., Chicago, Illinois, USA)

Electrochemical reaction conditions in the EC/MS system (method 1) N–0923 was tested in pure water and in the following supporting electrolytes: acetic acid, ammonium acetate, lithium trifluoromethane sulfonate, and trifluoromethane sulfonic acid. The supporting electrolytes were used in 0.1, 1 and 10 mM concentrations. The electrolytes did not affect the oxidation reaction pathway for N–0923. A high electrolyte concentration gave a better conversion of the sample but also gave increased background ion signals and decreased the sample ion signal.

Electrochemical reaction conditions in the EC–Fenton system (method 2) The sample solution injected through the electrochemical cell contained 1 mM N–0923, 1 mM FeCl3, 1.02 mM EDTA (to prevent precipitation of ferric hydroxide) and 10 mM hydrogen peroxide. The final concentration of acetonitrile in the sample solution was 21 %. The hydrogen peroxide was kept in the refrigerator and was added just before the injection. The “inactive” Fenton mixture was injected at a flow of 50 µl/min. After 1 minute, the flow was lowered to 2 µl/min. A make up liquid containing 10 % acetonitrile and 0.1 % formic acid was added to the effluent from the electrochemical cell at a flow rate of 50 µl/min to dilute the sample solution and to adjust the acetonitrile content to fit the start of the HPLC–gradient. 20 Minutes after the first injection, the diluted sample solution was injected with the second injection loop onto the HPLC column. The reductive potential over the electrochemical cell was kept at –0.5 V. 5.3 Results and discussion

N–0923 contains three major sites for oxidation, a tertiary amine, a thienyl group and a phenol function. The main metabolites reported from in vitro rat liver experiments are presented in Figure 5.2. The parent drug (1) (R=H) undergoes either hydroxylation or N–dealkylation to give the metabolites (2), (3), and (4), while metabolite (5) is the result of both hydroxylation and N–dealkylation. Glucuronidation and sulfation take place at the hydroxyl function of (1), (3), and (4). One of the hydroxyl groups in the metabolites (2) and (5) may also undergo glucuronidation and sulfation.


96

OR

N S

H

OR

NH

OR

N

RO

S

OR

N S

OR

NH

RO

(1) (2)

(3) (4)

(5) Figure 5.2 The major metabolites of N-0923 observed in rat liver perfusion experiments. R=H, glucuronide or sulfate as explained in the Results and Discussion section.

Method 1, electrochemical oxidation The products of electrochemical reactions of the parent drug are given in Figure 5.3. Voltammograms with detection of m/z values of protonated products are presented in Figure 5.4. As shown in Figures 5.3 and 5.4, the phenolic ring is oxidized to a catechol (or a p–hydroquinone) at about +0.2 V, but the products are immediately oxidized to the corresponding quinones. As the potential is increased to about +0.3 V, the tertiary amine function is also oxidized. Since the groups at the nitrogen are different, three N–dealkylation products are formed.


97

SHN

mw 169

N-0923

N S

H

O

O mw 287 mw 219

N

O

O

H

mv 331

mw 329

mw 315

N

O

O

S

OH

N

HO

S

OH

N S

Figure 5.3 N–0923, oxidation reaction pathways observed in the electrochemical cell. For simplicity, ortho–hydroxylation is shown, in analogy with the distribution observed in perfused rat liver experiments. para–Hydroxylation followed by oxidation to para–quinone can not be ruled out.

5.0 x 106

Cell potential (mV)

m/z 316

Counts/s

0 200 400 600 800 1000 1200 1400

m/z 330

m/z 220

m/z 170 m/z 288

9.0 x 105

Counts/s

Cell potential (mV)0 200 400 600 800 1000 1200 14000 200 400 600 800 1000 1200 1400

Figure 5.4 Voltammograms with detection of m/z values of protonated products from the electrochemical oxidation of N–0923


98

Compared to the reactions that take place in the rat liver experiments, the N–dealkylation is mimicked by the electrochemical cell. The electrochemical oxidation of the phenol function does not give a complete simulation of enzymatic oxidation, since the products are further oxidized to quinones. Enzymatic oxidation is obviously more selective than electrochemical oxidation. Furthermore, in rat liver experiments glucuronidation and sulfation protect the hydroxyl function and therefore oxidation to the quinone structure does not take place. It also appears in the rat liver experiments that dealkylation can take place prior to hydroxylation of the aromatic ring. In the EC/MS experiments, oxidation of the phenol function occurs at a lower cell potential than dealkylation and metabolites (3) and (4) in Figure 5.2 can therefore not be formed in the absence of protection of the hydroxyl group.

Electrochemical oxidation followed by reduction Upon electrochemical oxidation (method 1), the phenol function of N–0923 is oxidized at a lower potential than the tertiary amine function. By selecting the appropriate electrochemical conditions, dealkylated or non–dealkylated electrochemical products, can thus be produced selectively. Unlike the enzymatic oxidation, the electrochemical oxidation of the phenol function does not give catechols as final oxidation products, but the catechols are further oxidized to quinones. With an electrochemical system like the one described above, a change of the electrochemical conditions from oxidation to reduction is easily performed by switching the electrode potentials. By collecting the sample solution and injecting it at a reductive potential, the quinones are reduced back to catechols as described in Figure 5.5.


99

OO

NR2

R1

NR2

R1

HO

OH

e-2 , H+2

Figure 5.5 Electrochemical reduction of quinones

0 100 200 300 400 500 600 700 800 900 1000

m/z 220

m/z 222

Counts/s

1.5 x 105

Reductive Cell potential (mV, negative)100 200 300 400 500 600 700 800 900 10000

0 100 200 300 400 500 600 700 800 900 1000

m/z 220

m/z 222

Counts/s

1.5 x 105

Reductive Cell potential (mV, negative)100 200 300 400 500 600 700 800 900 10000

Figure 5.6 Extracted ion voltammograms with detection of m/z values from the electrochemical reduction of the quinone product with mw 219. The quinone was generated by injecting N-0923 at an oxidative potential of + 1.0 V The oxidation products were collected after oxidation of 20 µM N–0923 at +1.0 V, a potential where both the phenol function and the tertiary amine are oxidized. Figure 5.6 shows the extracted ion voltammograms with detection of m/z values from the electrochemical reduction of the quinone product with mw 219 (Figure 5.3). As the reductive potential is increased, the [M+H]+ ion of the quinone is replaced by the [M+H]+ ion of the corresponding catechol (or p–hydroquinone). Metabolites (2) and (5) in Figure 5.2, can thus be generated in the EC/MS system.

Method 2, the electrochemically assisted Fenton system The oxidation products of the EC–Fenton system were separated by gradient elution in an on–line LC/MS system. The mass chromatogram of the [M+H]+ ion of N–0923 is


100

presented in Figure 5.7a. The product ion spectra of the [M+H]+ ion of N–0923 is shown in Figure 5.7b with the suggested fragmentation given in Figure 5.8. The major fragment at m/z 147 corresponds to a loss of propyl–thienylethylamine, leaving the positive charge on the 5–hydroxytetralin structure. A gain of 16 Da to this fragment giving m/z 163 would thus suggest hydroxylation somewhere at the 5–hydroxytetralin structure. Following the same line of reasoning, the presence of a fragment at m/z 186, through a m/z 16 increase from m/z 170, would suggest that the oxygen is located somewhere at the propyl–thienylethylamine function. A fragment at m/z 127 instead of m/z 111 would then suggest hydroxylation on the thienylethyl group and a fragment at m/z 135 instead of m/z 119 would correspond to hydroxylation at either the phenolic ring or at the benzyl carbons 1 or 4.

9.9

a) N-0923 (m/z 316) 3.5 x 106

50 100 150 200 250 300 350

316

1.5 x 106b) N-0923

218170

147

119111

9172132

9.9

Figure 5.7 a) Mass chromatogram of m/z 316, the [M+H]+ ion of N–0923. b) Product ion spectrum of the [M+H]+ ion of N–0923.


101

m/z 316

N-0923

m/z 218OH

N CH2+

OH

N S

H

+

m/z 147OH

+

N SH

H+

Proton

transfer

m/z 170

OH

N SH+

m/z 316

H2CS+

m/z 111

N S

H

H +

m/z 170

NCH2

H+

m/z 72

m/z 91

m/z 119

C7H7+ - CO H2C=CH2-

OH

H+

m/z 147

CH2

O

H

+ CH2

O

H

+

C

O HHH+

CH2

CH2

O

CH2

+

CH2

O

CH2

+

Figure 5.8 Suggested fragmentation of the [M+H]+ ion of N-0923.

When the EC–Fenton system was applied, at least five different products with a gain of oxygen were obtained. Figure 5.9a shows the mass chromatogram of m/z 332, corresponding to an m/z 16 increase from m/z 316, the [M+H]+ ion of N–0923. About 20 % of the N–0923 was oxidized. The product ion spectra given in Figures 5.9b–5.9f correspond to the peaks in the mass chromatogram in Figure 5.9a.


102

6.97.2 8.3

8.79.1 5.0 x 105a) m/z 332

50 100 150 200 250 300 350

1.4 x 104b) RT 6.9 min

332234218

208

147

135

117111

10772

50 100 150 200 250 300 350

1.9 x 104c) RT 7.2 min

332304234

208

170

163

145,

147135

111

72

50 100 150 200 250 300 350

4.5 x 104d) RT 8.3 min

332186

147

127119

10772

50 100 150 200 250 300 350

1.6 x 105e) RT 8.7 min

332

314234

170

163

145

135111

72 107

50 100 150 200 250 300 350

9.0 x 104f) RT 9.1 min

332314

234186

168,

170

163145,

147

127

111

72206

6.9

Figure 5.9 Mass chromatogram and product ion spectra of the oxidation products of N–0923 from the EC-Fenton system. The signal intensities in counts/s are given in the upper right corner of each chromatogram and spectrum. The numbers above the peaks in chromatogram a represent retention times in minutes. a) Mass chromatogram of m/z 332, corresponding to a gain of oxygen to the [M+H]+ ion of N-0923. b-f) Product ion spectra of the peaks with RT 6.9, 7.2, 8.3, 8.7 and 9.1 minutes in chromatogram a.

The spectra for the peaks at RT 6.9, 7.2 and 8.7 minutes (Figure 5.9b, 5.9c and 5.9e) contain the fragments 135 and 163, suggesting that the gain of oxygen is located at the


103

phenolic ring or at the benzyl carbons 1 or 4. Based on the product ion spectra alone, we were not able to fully characterize these products. These products are further discussed below.

The spectrum for the peak at RT 8.3 minutes in Figure 5.9d is similar to the spectrum of N–0923 with the exceptions that the fragment at m/z 170 has been replaced with m/z 186, and the fragment at m/z 111 has been replaced with m/z 127. This is consistent with a gain of oxygen on the thienylethyl group.

The spectrum for the peak with RT 9.1 minutes in Figure 5.9f contains fragments at m/z 127, 163 and 186, suggesting a gain of oxygen at the 5–hydroxytetralin structure as well as on the thienylethyl group. The absence of a fragment at m/z 135 suggests that the oxygen is not located at the phenolic ring or at the benzyl carbons. Since the product ion scan is performed on m/z 332, hydroxylation in two different positions is not possible. This could be a case of overlapping peaks, although extraction of the different fragment ions did not show any difference in retention times. In order to get fragments at both m/z 127 and 163 from a single molecule, the oxygen has to be located in such a position of the molecule that it can be transferred to either the 5–hydroxytetralin structure or to the thienylethyl function during the fragmentation. Following the discussion about the mechanism for the electrochemical dealkylation in chapter 2, hydroxylation at any of the positions next to the nitrogen (positions 2, 15 or 9) would have resulted in N–dealkylation. No N–dealkylation products were observed in the EC–Fenton experiments. The fragment at m/z 168 is only present in this spectrum and would be consistent with a gain of 16 Da to the propyl–thienylethylamine followed by a loss of water. An explanation for this fragmentation pattern would be the N–oxide. Some possible explanations for oxygen atom transfer during the fragmentation are given in Figure 5.10.


104

OH

N S

HO

+

OH

N S

HO

OH

N S

OHH

+

OH

N S

HOH

+

+

OH

N S

OH

H

H

+

OH

N S

OH

H

H+

+O

OH

H

m/z 163

m/z 332

Figure 5.10 Possible explanations for oxygen transfer during the fragmentation of N–0923 N–oxide.

Comparison of the oxidation products from the EC–Fenton system with the

electrochemically generated catechol (or p-hydroquinone) from the EC/MS

system Interpretation of the product ion spectra for the peaks at RT 6.9, 7.2 and 8.7 minutes (Figures 5.9b, 5.9c and 5.9e) suggest that the gain of oxygen is located at the phenolic ring or at the benzyl carbons 1 or 4. Interpretation of the product ion spectra alone did not allow a more precise determination of the structures. The hydroxyl radical acts as an electrophile and reacts mainly by addition to unsaturated systems, but can also abstract a hydrogen atom.9 The main products would thus be expected to be hydroxylation on the phenolic ring. Hydroxylation at positions 1 and 4 is also possible by a hydrogen abstraction mechanism as described in chapter 4. In order to gain additional information about the products with RT 6.9, 7.2 and 8.7 minutes, a comparison of retention times and product ion spectra was made with the presumed catechol obtained from reduction of the quinone product with mw 329 (Method 1, Figure 5.3). Since the quinones can only be formed electrochemically if the initial hydroxylation occurs at ortho or para position, the reduced product with mw 331 is bound to be either the catechol or the para–hydroquinone. A comparison between the two experiments is presented in Figure 5.11. The retention times in Figure 5.11b differ from those given in Figure 5.9a because the experiments were performed on separate occasions and the split between the HPLC–column and the mass spectrometer was different.


105

a) Oxidized - reduced

1.6 x 106

50 100 150 200 250 300 350

1.3 x 105c) Ox - red

RT 9.7 min

332

314234

170

163

145

135111

72107

9.7

b) EC–Fenton

8.2

8.5

9.4

9.710.1 2.0 x 10

6

50 100 150 200 250 300 350

1.6 x 105

332314234

170

163

145

135111

72107

d) EC-Fenton

RT 9.7 min

a) Oxidized - reduced

Figure 5.11 Mass chromatograms and product ion spectra of the oxidation products of N–0923. The signal intensities in counts/s are given in the upper right corner of each chromatogram and spectrum. The numbers above the peaks in chromatogram a and b represent retention times in minutes. a) Mass chromatogram of m/z 332, corresponding to a gain of oxygen to the [M+H]+ ion of N-0923, from reduction of the electrochemically generated quinone. b) Mass chromatogram of m/z 332 from the electrochemically assisted Fenton system. The retention times differ from Figure 5.9 because the experiments were performed on separate occasions. c) Product ion spectrum of the peak with RT 9.7 minutes in chromatogram a. d) Product ion spectrum of the peak with RT 9.7 minutes in chromatogram b.

Figures 5.11a and 5.11b show that the electrochemically generated catechol (or para–hydroquinone) elutes at the same retention time as the fourth peak from the EC–Fenton system (9.7 minutes in Figure 5.11b, 8.7 minutes in Figure 5.9a). The product ion spectra given in Figures 5.11c and 5.11d are also very similar. However, in the spectrum belonging to the product from the EC–Fenton system (5.11d), the fragment at m/z 145 has a much higher intensity than in the spectrum from the electrochemically generated catechol


106

(5.11c). This fragment results from a loss of water from the fragment at m/z 163. The MS/MS conditions were the same for both spectra and if the compounds were exactly the same, such a substantial difference would be highly unlikely between two injections of identical products. It is possible that the location of the hydroxyl group does not sufficiently change the polarity of the compound to allow separation of the isomers on the HPLC–column (C–18). In that case, the peak at RT 9.7 minutes could correspond to a mixture of N–0923 hydroxylated at ortho, meta and para positions. This would also mean, that the compound generated from the EC/MS system could consist of both the catechol and the p–hydroquinone.


107

H H

OH

HO

+

m/z 163

b) ortho

+

OH

O

H

HO

OHH

+

OH

HO

H

H

c) meta

hydrogen

transfer

+

transfer

hydrogen

OH

O

H

H

+

O

OH

H +

m/z 163

H+

OH

+

OH H

H

H

– H2O

+

OH

OH H

+

OH

O

H

HH

hydrogen transfer

m/z 145OH

+

H2O–

+O

OH

HH

m/z 163

+

HHO

OH

H

a) para

123

4

432

1

Figure 5.12 Possible fragmentation and stabilization pathways for the fragment at m/z 163 when the hydroxyl group is located at the a) para, b) ortho or c) meta positions

The tendency to lose a water molecule from the fragment at m/z 163 is thus the main difference between the products with RT 9.7 minutes from the two systems. Figure 5.12 describes the suggested continued fragmentation of the fragment at m/z 163 when the hydroxyl group is located at the ortho, meta or para positions. When the hydroxyl group is


108

located at the para position (Figure 5.12a) a proton can be transferred to the “para hydroxyl group” to make water a suitable leaving group. Hydrogen atom transfer from C3 to C2 provides a second route to the fragment at m/z 145.

When the hydroxyl group is located at the ortho or meta positions (Figures 5.12b and 5.12c) an initial hydrogen atom transfer is required to make the benzylic hydrogens at position 4 available for attack by the hydroxyl group. A hydrogen atom transfer can also lead to a stable structure of the ion at m/z 163 and thus prevent it from losing a water molecule.

Based on the proposed fragmentation of m/z 163 in Figure 5.12, we suggest that the product from reduction of the quinone in the EC/MS system is mainly the catechol (the hydroxyl group located at the ortho position) and therefore gives a low yield of the fragment at m/z 145. The product from the EC–Fenton system with the same retention time (9.7 minutes) could correspond to a mixture of N–0923 hydroxylated at the ortho, meta or para positions, with the para conformation responsible for the fragment at m/z 145.

If these assumptions are correct, the two compounds eluting at RT 8.2 and 8.5 minutes (RT 6.9 and 7.2 minutes in Figure 5.9a) would correspond to two compounds hydroxylated at position 1 or position 4.

In order to obtain further information about the ortho, meta, para distribution, the application of further techniques, such as NMR, is required. Because of the low flow rates (2 µl/min) required for good conversion efficiency in the EC–Fenton system, generation of amounts sufficient for analysis with proton NMR is a very time consuming process. At a flow rate of 2 µl/minute and 20 % yield of the desired oxidation product, it would take about 5 days (and nights) to synthesize 1 mg of a compound weighing 300 Da. The EC–Fenton system presented here is highly suitable for on–line generation of oxidation products in concentrations well above the required concentrations for analysis with electrospray mass spectrometry. For generation of large quantities of oxidation products, an off–line electrochemical system or a flow–through electrochemical cell with a larger surface area than the ESA Coulochem 5020 guard cell is needed.

Summary The structures suggested for the final products from both systems are given in Figure 5.13, where a–c illustrates the compounds from the EC/MS system and d–f the products from the EC–Fenton system. Compared to the metabolites reported from rat liver experiments (Figure 5.2), the metabolites (2) and (5) can be generated in the EC/MS system by reduction of the quinone products. In the rat liver experiments, two dealkylation products (3) and (4) were detected, but only one dealkylation product of the catechol was observed (5). It is possible that compound b in Figure 5.9 was not formed in the liver experiment. On the other hand, it might have been formed in concentrations too low to be detected.


109

mw 331

mw 289mw 221 OH

N

HO

S

H

OH

NH

HO

OH

N

HO

S

OH

N S

OH

OH

N S

HO

OH

N S

O

+

-

Suggested oxidation products of N-0923 from the EC-Fenton system

Suggested products of N-0923 from the EC/MS system after reduction of the quinones

mw 331 RT 6.9 or 7.2 min

mw 331 mw 331

a) b)

c)

d) e)

f)

OH

N S

OH

OH

N S

OH

g)

h)

mw 331

mw 331 RT 6.9 or 7.2 min

RT 8.3 min RT 8.7 min

RT 9.1 min Figure 5.13 Suggested oxidation products of the EC/MS system (a–c) and the EC-Fenton system (d–h). The compounds a–c are presented as catechols although the p–hydroquinone conformation can not be entirely excluded. The retention times in minutes given for compounds d–h are taken from the mass chromatogram in Figure 5.9a.


110

Metabolite 2 in Figure 5.2 was also formed in the EC–Fenton system. None of the dealkylation products were observed in the EC–Fenton experiments. Four additional compounds were tentatively identified from the interpretation of the product ion scans: two compounds hydroxylated at positions 1 or 4, a compound with the oxygen located at the thienylethyl group, and the N–oxide. None of these products were reported from the rat liver experiments. Dealkylation is the main pathway for oxidation of tertiary amines by cytochrome P450 but N–oxides are also formed.10 Even though N–oxides of N–0923 were not formed in the rat liver experiment, the formation of N–oxides in the EC–Fenton system can be useful for the characterization of metabolites in general. Metabolites 3 and 4 can not be generated from any of the systems without the introduction of a protective group on the 5–hydroxyl group prior to the electrochemical oxidation.

5.4 Conclusions

The results presented here demonstrate that the two methods are well suited to complement each other. Even though the EC–Fenton system does not give much additional information in the case of N–0923, it is clear that some metabolites that can not be generated in the EC/MS system can be generated in the EC–Fenton system. In this case two compounds hydroxylated at the benzylic carbons, a presumed N–oxide, a compound with the oxygen located at the thienylethyl group and a compound hydroxylated at the phenolic ring without further oxidation to the quinone structure. Even though the mechanism of hydroxylation is different from the P450 mechanism, the large range of hydroxylated compounds usually obtained from the EC–Fenton system allows comparison of HPLC retention times with real metabolites and characterization with MS/MS.

The electrochemically assisted Fenton system described here makes use of the separation of the oxidation products by means of HPLC connected on–line with the mass spectrometer, a setup that was also used in this chapter for characterization of the products from the direct electrochemical oxidation. In a setup comprising an HPLC column the very informative electrochemical potential sweeps can not be performed, but a quick insight is obtained into the composition of a mixture of isomeric oxidation products.

If larger quantities are needed for further characterization with for example NMR, both methods can be scaled up by the use of a larger electrochemical cell.

5.5 References

1. A. S. Horn; P. Tepper; J. v. d. Weide; M. Watanabe; G. Grigoriadis; P. Seeman, "Synthesis and radioreceptor binding of N-0437, a new extremely potent and selective D2 dopamine receptor agonist.", Pharm. Weekbl. Sci. Ed., 1985, 7, 208-211.


111

2. J. V. d. Weide; J. B. d. Vries; P. G. Tepper; A. S. Horn, "In vitro binding of the very potent and selective D2 dopamine agonist, [3H] N-0437 to calf caudate membranes.", Eur. J. Pharm., 1987, 134, 211-219.

3. J. V. d. Weide; J. B. d. Vries; P. G. Tepper; D. N. Krause; M. L. Dubocovich; A. S. Horn, "N-0437: a selective D-2 dopamine receptor agonist in in vitro and in vivo models.", Eur. J. Pharm., 1988, 147, 249-258.

4. W. Timmerman; B. H. C. Westrink; J. B. d. Vries; P. G. Tepper; A. S. Horn, Eur. J. Pharm., 1989, 162, 143-150.

5. W. Timmerman; P. G. Tepper; B. G. J. Bohus; A. S. Horn, Eur. J. Pharm., 1990, 181, 253-260.

6. P. J. Swart; W. E. Oelen; A. P. Bruins; P. G. Tepper; R. A. d. Zeeuw, "Determination of the dopamine agonist N-0923 and its major metabolites in perfused ratlivers by HPLC-UV-atmospheric pressure ionization mass spectrometry.", J. Anal. Toxicol., 1994, 18, 71-77.

7. T. K. Gerding; B. F. H. Drenth; R. A. d. Zeeuw; P. G. Tepper; A. S. Horn, "The metabolic fate of the dopamine agonist 2-(N-propyl-N-thienylamino)-5-hydroxytetralin in rats after intravenous and oral administration. II. Isolation and identification of metabolites.", Xenobiotica, 1990, 20, 525-536.

8. T. K. Gerding; B. F. H. Drenth; H. J. Roosenstein; R. A. d. Zeeuw; P. G. Tepper; A. S. Horn, "The metabolic fate of the dopamine agonist 2-(N-propyl-N-thienylamino)-5-hydroxytetralin in rats after intravenous and oral administration. I. Disposition and metabolic profiling", Xenobiotica, 1990, 20, 515-524.

9. C. v. Sonntag; H.-P. Schuchmann, "Pulse radiolysis", In "Oxygen Radicals in Biological Systems, Part C" ; Packer, L., Ed.; Academic Press:, 1994; Vol. 233.


Chapter 6 Practical aspects of electrochemistry on–line with ESI/MS

113

Chapter 6

Practical aspects of electrochemistry on–line with

ESI/MS

6.1 Introduction

In 1994, Zhou and Van Berkel connected a number of different electrochemical cells on–line with electrospray ionization mass spectrometry (ESI/MS).1,2 Since then, the number of publications in the field of EC/ESI/MS has been steadily increasing. The ease of use and the many advantages that can be gained from adding an electrochemical cell in front of the electrospray ionization mass spectrometer (ESI/MS) will most certainly lead to an extended use in the future. However, the coupling of electrochemistry on–line with ESI/MS is a fairly new technique and many practical aspects are yet to be explored. In order to make the two techniques compatible, there are a number of things that have to be considered, such as currents between the electrochemical cell and the electrospray needle, signal suppression from the supporting electrolyte and retention of the analytes in the electrochemical flow cell. The aim of this chapter is to provide some solutions to the most common problems that can be encountered in an EC/ESI/MS system.

6.2 Currents between the electrochemical cell and the electrospray

needle

An electrochemical cell in an on–line setup close to the electrospray needle will suffer from disturbing currents originating from the applied electrospray potential. A high positive or negative potential is applied at the electrospray needle, and a current will inevitably be flowing between the tip of the electrospray needle and the electrochemical cell if it is kept at ground potential. A number of ways to circumvent this problem were described by Zhou and Van Berkel as they presented their on–line electrochemical cells and will not be discussed in detail here.1,2 One setup recommend by Zhou and Van Berkel is to float the electrochemical cell at the same potential as the potential as applied at the electrospray needle. They also described how the introduction of a metal connector at ground between the electrochemical cell and the electrospray needle can solve the problem of interfering currents.


114

Most efforts to combine electrochemistry with mass spectrometry have been aimed at characterization of short–lived intermediates in electrochemical reactions or to make ions of compounds with low polarity for signal enhancement in electrospray mass spectrometry. In our experiments, we were primarily interested in the end–products of the electrochemical oxidations and were therefore not dependent of a short response time between the electrochemical cell and the ESI/MS. Due to the relatively large distance between the electrochemical cell and the electrospray interface, currents between the electrospray needle and the electrochemical cell did not turn out to be a major problem in our EC/ESI/MS–setup and the electrochemical cell was connected on–line with the ESI/MS without further modifications. Before we came to this conclusion, we performed a number of experiments with the electrospray needle kept at ground potential in order to avoid currents interfering with the electrochemical reaction. A description of the setup is presented in Figure 6.1.

Pump EC flowcell

+500V

+70V

loop injector

-3.5 kV

Figure 6.1 Experimental setup for EC/ESI/MS with the electrospray needle at ground. The ring around the electrospray needle represents a large stainless steel nut. A potential of –3.5 kV was applied at the nut and the electrospray needle was connected to ground. Positive ions in the solution will be drawn to the surface of the droplet in the same way as when the needle is kept at a high positive voltage with a surrounding at ground potential. When the positive ions finally end up in the gas phase, they will be drawn towards the negatively charged nut because of the direction of the electric field. However, by applying a high enough flow of nebulizing gas, a sufficient amount of positive ions will be physically pushed against the electric field and end up at the front of the mass spectrometer where they are "sucked in" by the next electric field. Although a part of the ions end up on the nut, the amount of ions arriving at the detector was about 60 % of what was observed in normal electrospray mode. When the mass spectrometer was operated in the negative mode, a potential of +3.5 kV was applied to the nut. In this type of setup, the outlet from the electrochemical cell can be directly connected to the electrospray emitter


115

and the delay time between the electrochemical reaction and the detection can be minimized.

6.3 Signal suppression from the supporting electrolyte

For most organic solvents used in electrochemistry, the addition of supporting electrolytes is necessary for any electrochemical reaction to occur. Aqueous solutions of organic molecules have a moderate conductivity and some electrochemical reactions will take place even without addition of supporting electrolytes. However, in order to achieve a good electrochemical conversion in the electrochemical cell it is essential to add some kind of supporting electrolyte to increase the conductivity.

When choosing the suitable electrolyte for an electrochemical reaction it is of great importance to consider what pH is most favorable for the desired electrochemical reaction. Changing the pH of the solution can completely alter the outcome of the electrochemical reaction. Most tertiary amines are for example not oxidized at all in 0.1 M acetic acid, while they are easily oxidized in neutral or basic media as illustrated for tributylamine in Figure 6.2.


116

0x100

1x106

2x106

3x106

4x106

5x106

6x106

0 200 400 600 800 1000 1200

pH 9 pH 7

0.1 M HAc

ions/s

cell potential (mV)

Figure 6.2 Extracted ion voltammograms of m/z 130, the [M+H]+ ion of dibutylamine from electrochemical oxidation of tributylamine in different solvents.

The problem when it comes to coupling electrochemistry on–line with ESI/MS is that most traditional supporting electrolytes for electrochemistry are not suitable for electrospray mass spectrometry. Some examples of common supporting electrolytes are silver perchlorate, tetraalkylammonium perchlorates, tetraalkylammonium p–toluenesulfonates, phosphate buffers and borate buffers. All of these supporting electrolytes will give rise to strong background peaks in the mass spectra or lead to crystallization on the electrospray needle or at the curtain plate. In order to avoid these types of problems it is often necessary to consider other supporting electrolytes than the salts traditionally used.

As a general rule, it is always recommended to start with a supporting electrolyte that is also suitable for the ESI/MS. For aqueous solutions, we recommend either ammonium acetate or ammonium formate. For lower pH:s, diluted acetic acid or formic acid can be used alone or added to the ammonium acetate/formate solutions. Ammonia can be added for electrochemical reactions requiring higher pH:s. The disadvantage of electrolytes based on ammonium acetate/formate is that ammonia will sometimes react as a nucleophile with the electrochemically generated intermediates. If such reactions occur, a different supporting electrolyte has to be chosen. We have testes a number of different salts as supporting electrolytes for the electrochemical oxidation of lidocaine (see chapter 2). The results are presented in Table 6.1.


117

Table 6.1 Effects of different electrolytes on the oxidation of lidocaine.

Salt

Lidocaine onset of

oxidation (mV)

Background noise

(pos. ions)

Comments e

ammonium

acetate

230

low

100 % conversion. pH may be modified with acetic acid and

ammonia.

LiCF3SO3

320

very high d

100 % conversion. pH may be

modified with LiOH and HCF3SO3

lithium chloride

350

mediuma

100 % conversion. pH may be modified with LiOH and HCl.

lithium fluoride

470

mediuma

100 % conversion. pH may be modified with LiOH and HF.

lithium formateb

260

high


modified with LiOH and HCOOH.

lithium perchlorate

320

high d


modified with LiOH and HClO4.

sodium chloride

320

highc 100 % conversion. pH may be modified with NaOH and HCl.

Crystallizes on the curtain plate.

sodium formate

350

very high


modified with NaOH and HCOOH.

sodium dihydrogen

citrate

410

very high d


modified with NaOH and citric acid.

All salts were 10 mM in 80 % water and 20 % acetonitrile. 10 µM lidocaine was used to test the conversion and ion suppression. The flow rate was kept at 25 µl/min. aLithium forms clusters with acetonitrile. Otherwise rather clean spectra. bpH modified to 7 with LiOH and formic acid. cSodium forms clusters with acetonitrile. Otherwise rather clean spectra. dThe background noise in the negative mode is also very high. e The experiments were performed with the dissolved salts only and pH was not adjusted. Lidocaine was oxidized to 100 % in the presence of all supporting electrolytes tested, but the background noise and signal suppression in the positive mode was higher than for ammonium acetate for all salts in Table 6.1. Some negative ions, such as trifluoromethane sulfonate, perchlorate and dihydrogencitrate also give rise to high background noise and


118

signal suppression in the negative mode. The background noise in the positive mode comes from positively charged clusters of neutral molecules with Li+ or Na+. If acetonitrile is added as organic modifier to the aqueous solution, clusters of acetonitrile and Li+ or Na+ will give an increased background noise when the mass spectrometer is operated in positive mode. The cluster formation is generally more pronounced with sodium than with lithium. The use of a different organic modifier, such as methanol, might decrease these effects. It should be noted that methanol, like ammonia, can react as a nucleophile with electrochemically generated intermediates.

All of the salts presented in Table 6.1 can thus be used as supporting electrolytes. If the compound under investigation is readily oxidized in ammonium acetate/formate but undesired reactions occur with ammonia, the best choice is generally to use a combination of acetic acid/formic acid and sodium or lithium acetate/formate. For electrochemical conversions requiring high pH, sodium or lithium hydroxide can be added. If acetonitrile is added as organic modifier, it is preferable to use lithium acetate/formate because in general the cluster formation with acetonitrile is less pronounced for lithium than sodium.

In our experiments, we always worked with water based solutions. For certain electrochemical reactions, the presence of water is undesirable and dry organic solvents are preferred. Zhou and Van Berkel suggested lithiumtrifluorosulphonate as a suitable electrolyte for organic solvents like acetonitrile or dichloromethane.2

In conclusion, the choice of supporting electrolyte will always be a trade–off between good electrochemical conversion and good conditions for the ESI/MS. If the oxidized/reduced mixture is separated on an HPLC–column, most salts will elute within the unretained front and get separated from the reaction products. Therefore, the use of HPLC–separation between the electrochemical cell and the ESI/MS allows the use of a broader range of supporting electrolytes.

6.4 Retention in the electrochemical flow cell

Many species dissolved in solution have a tendency to adsorb on the electrode surface. In an electrochemical flow cell this will lead to retention of the analyte and thus a delayed arrival at the mass spectrometer. The adsorption is dependent on the type of compound investigated, the solvent composition, the electrode material and the potential applied to the working electrode. Due to electrostatic interactions, negatively charged species will adsorb at positive potentials while positively charged species will adsorb on the electrode at negative potentials. The pre–concentration of tamoxifen described by Pretty, Deng, Goeringer and Van Berkel relies on the fact that positively charge ions are accumulated on the working electrode at a negative potential and then rapidly released as a more positive potential is applied.3


119

In aqueous solution, neutral organic molecules interact with the electrode primarily because of hydrophobic interactions and adsorb to the greatest extent on the uncharged working electrode.4,5 This phenomenon is particularly pronounced for organic compounds with non–polar groups and thus with low solubility in water. As a positive or negative potential is applied, the dipole charges of the surrounding water molecules are attracted towards the electrode and displace the neutral organic molecules on the surface. In pure aqueous solution, adsorption on the working electrode can be a great cause of retention in an electrochemical flow cell. The addition of an organic modifier, such as acetonitrile, to the solution makes the neutral molecules less prone towards adsorption and will often solve the problem.

6.5 The electrochemical cell

Several home–made electrochemical cells have been presented in the literature over the last decade. They have all been constructed to fit the purpose of each investigation and have indeed served their purpose at the time. We have also experimented with some different cell design, but have come to the conclusion that the flow–through cells provided by ESA are best suited for our applications.

Working electrodes

Counter and reference electrodes

Figure 6.3 The ESA Coulochem cell 5011

6

The ESA Coulochem cell 5011, presented in Figure 6.3, was used for the majority of our experiments. It contains two flow–through porous graphite working electrodes connected in series, each with independent palladium reference and counter electrodes. In our experiments, we connected the cell to a home made potentiostat and used both working electrodes for oxidation. The large surface area allows a high electrochemical conversion at moderate flow rates (<100 µl/min). At higher flow rates, the conversion efficiency goes down. It should be noted that our electrochemical cells had all been used extensively in


120

analytical applications before we received them and that the conversion in a new cell might be good also at higher flow rates. The ESA Coulochem 5020 guard cell was used for the EC–Fenton experiments (chapter 3). It is constructed in the same way as the 5011 cell but contains only one large porous working electrode and is constructed to withstand high pressure. This can come in handy for certain applications since it can be connected prior to an HPLC column. The ESA Coulochem cell 5011 is not constructed to withstand high pressure and can only be used post–column or in a setup like the one described in chapter 4, where the electrochemical products are collected in an injection loop at low pressure.

In conclusion, the most important consideration for setting up an EC/ESI/MS system is a definition of the purpose of the system. If the main aim of the setup is to detect short–lived electrochemical intermediates, the delay time between the EC cell and the ESI/MS has to be minimized. It will then be of great importance to set up the system in such a way that currents between the electrochemical cell and the electrospray needle can be avoided. If, on the other hand, the main aim is to characterize the end–products of an electrochemical reaction, a larger distance can be allowed between the electrochemical cell and the ESI/MS. The installation of an HPLC column between the electrochemical cell and the ESI/MS allows the use of a broad range of supporting electrolytes. 6.6 References

1. F. Zhou; G. J. V. Berkel, "Electrochemically enhanced electrospray ionization mass spectrometry", Proceedings of the 42nd ASMS Conference on Mass Spectrometry and Allied Topics, Chicago, Illinois, 1994.

2. F. Zhou; G. J. V. Berkel, "Electrochemistry Combined On-Line with Electrospray Mass Spectrometry", Anal. Chem., 1995, 67, 3643-3649.

3. J. R. Pretty; H. Deng; D. E. Goeringer; G. J. v. Berkel, "Electrochemically modulated preconcentration and matrix elimination for organic analytes coupled on-line with electrospray mass spectrometry", Anal. Chem., 2000, 72, 2066-2074.

4. F. C. Anson, "Patterns of ionic and molecular adsorbtion at electrodes", Acc. Chem. Res., 1975, 8, 400-407.

5. P. T. Kissinger; C. R. Preddy; R. E. Shoup; W. R. Heinman, "Fundamental concepts of analytical electrochemistry", In "Laboratory techniques in electroanalytical chemistry" ; Kissinger, P. T., Heinman, W. R., Eds.; Marcel Dekker:, 1984.

6. http://www.esainc.com/products/HPLC/EC_Detectors/esa_coulochem2.html

Chapter 7 Bioactivation of the catecholamine pro–drugs PD217015, (–)-GMC6650 and (+)-GMC6650 in male Wistar rats

121

Chapter 7

Bioactivation of the catecholamine pro–drugs PD217015,

(–)-GMC6650 and (+)-GMC6650 in male Wistar rats

The bioactivation of the potential catecholamine pro–drugs PD217015, (–)-GMC6650 and (+)-GMC6650 has been studied by LC/MS/MS analysis of blood plasma and brain samples from male Wistar rats following oral administration. The catechol metabolite of PD217015, 5,6-dihydroxy-DPAT, was identified in the brain by means of comparison of retention time and product ion spectrum with a racemic 5,6-dihydroxy-DPAT standard. From interpretation of the product ion spectra, three additional metabolites of PD217015 are suggested. The N–dealkylated metabolite 6-propylamino-3,4,5,6,7,8-hexahydro-2H-naphtalene-1-one and two hydroxylated metabolites suggested to be 2–hydroxy–PD217015 and 4–hydroxy–PD217015. Hydroxylation at position 2 is proposed to constitute an intermediate step in the formation of the active principle 5,6-dihydroxy-DPAT.

For GMC6650, a metabolite believed to correspond to the catechol active principle 1–propyl-1,2,3,4,4a,5,10,10a-octahydro-benzo[g]quinoline-6,7-diol was only detected in the brain of the rat given the (–)–enantiomer.* A compound hydroxylated somewhere at the positions 4 through 10a of GMC6650 was detected for both enantiomers in plasma as well as in the brain. 7.1 Introduction

Parkinson’s disease is a progressive neurodegenerative disease involving neurodegeneration of dopaminergic neurons of the substantia nigra, a part of the midbrain.1 Loss of these neurons results in a hypokinetic disorder marked by akinesia (impaired initiation of movement), bradykinesia (reduction of voluntary movement), muscular rigidity and tremor.2 The pathogenesis of Parkinson’s disease is most likely multi-functional and various processes have been shown to play a role, such as mitochondrial effects, glutamate toxicity, genetic factors and oxidative stress.3

Although the exact cause of Parkinson’s disease is not known, much indicates that oxidative stress is one of the important factors in the degeneration of dopaminergic neurons. Increased concentrations of iron has been observed in the substantia nigra of patients with Parkinson’s disease, something that could lead to increased generation of

* This structure has now been confirmed by comparison of retention time and product ion spectrum with a synthesized reference of 6,7-di-OH-PBGQ.


122

hydroxyl radicals through the Fenton reaction.4 Glutathione is a major antioxidant in the brain with a crucial role in the primary cellular defense against oxidative stress. A reduction of the glutathione levels in the substantia nigra has been observed during Parkinson’s disease, suggesting an increased formation in reactive oxygen species that contributes to the death of neuronal cells.1 CYP2E1, an isoform of cytochrome P450 that is associated with free radical production and formation of endogenous toxins, is selectively localized in nigral dopamine containing cells and might also play a role in the development of Parkinson’s disease.5

The control of normal motor behavior is under influence of the so-called reinforcing basal ganglia-thalamocortical motor circuit.3 The two major subtypes of dopamine receptors, dopamine D1 and D2 receptors, both play a crucial role in this motor circuit. Very simplified, the dopamine D1 receptors are involved in the “direct pathway” that leads to stimulation of cortically generated movement, while dopamine D2 receptors are involved in the “indirect pathway” that leads to inhibition of cortically generated movement. Dopamine itself is an agonist at both receptor subtypes and under normal circumstances, there is a balance between the direct and the indirect pathways. In a moving disorder like Parkinson’s disease, the absence of dopamine in the substantia nigra (due to deterioration of dopaminergic neurons) disables the direct pathway leading to a reduction of cortically generated movement. To compensate for the decreased number of useful dopaminergic neurons and to minimize the symptoms of Parkinson’s disease a continuous stimulation of the dopamine receptors is needed. Since dopamine itself acts as an agonist at both receptor subtypes it seems reasonable that a mixed dopamine D1/D2 agonist would be most effective. It has been demonstrated that mixed D1/D2 agonists give rise to additional abnormal behavior in rats compared to when selective dopamine D1 or D2 agonists are administrated.6 Consequently, there must be synergistic dopamine D1/D2 receptor interactions, which would be lost upon administration of a selective dopamine D1 or D2 agonist.

Oral administration of catecholamines for the treatment of central nervous system (CNS) disorders like Parkinson's disease is generally not successful because of their poor oral bioavailability.6,7 In order to reach the bloodstream and enable transport to the brain, the largest portion of the drug has to pass through the gastrointestinal tract and the liver, where most catecholamines are subjected to rapid biotransformation. Glucuronidation, sulfation and auto-oxidation are common metabolic fates for the catechol function while the amino function often is subjected to N-dealkylation or oxidative deamination.8-11 By introducing protecting groups at the hydroxy and amino functions, the oral bioavailability can be increased. Ideally, this results in a pharmacologically inactive compound with high oral bioavailability that is slowly transformed in vivo into the pharmacologically active catecholamine. The rapid biotransformation of the catechol moiety has successfully been


123

slowed down by protection of the hydroxy groups by pivaloyl or benzoyl esters.12,13 N-alkylation of the amino function of dopamine increases the lipophilicity and improves transfer through the blood-brain barrier.14 The dipivaloyl ester of dopamine in combination with the introduction of a dihydropyridine derivative on the amino function gave an increased transfer through the blood-brain barrier and an extended dopaminergic activity.15

A new type of catecholamine pro–drug has recently been developed in our laboratory in a collaborative project.6,16,17 In principle, the catechol function is replaced with a cyclohexenone function that is biotransformed into a catechol in vivo. We have used high-performance liquid chromatography / tandem mass spectrometry (LC/MS/MS) to examine the metabolism of two such enone pro–drugs, PD217015 (S-enantiomer of PD148903, Figure 7.1a) and GMC6650 (Figure 7.1b), in male Wistar rats following oral administration.

a) PD148903 (racemic)PD 217015 (S-enantiomer)

b) GMC6650

N

O

12

3

44a

567

89

9a10 10aN

O1

2

3

4 56

78

4a

8a

N

OH

HO

c) 5,6-di-OH-DPAT d) R-(-)-10,11-dipivaloylapomorphine

N

O

O

H

O

O

mw 249 mw 247

mw 263 mw 435

*

Figure 7.1 a) 6-(N,N-dipropylamino)-3,4,5,6,7,8-hexahydro-2H-naphtalene-1-one (PD148903). PD217015 is the S-enantiomer of PD148903. b) 1-Propyl-trans-2,3,4,4a,5,7,8,9,10,10a-decahydro-1H-benzo[g]quinolin-6-one (GMC6650). c) 5,6-dihydroxy-2-(N,N-dipropylamino)-tetralin (5,6-di-OH-DPAT). d) R-(–)-10,11-dipivaloylapomorphine (DPA, used as internal standard).

In vitro, the racemic pro–drug PD148903 (Figure 7.1a) has no binding affinity for the dopamine (DA) receptor family. However, biochemical and behavioral experiments in vivo proved that the pro-drug is converted to a DA agonist.18-20 The S-enantiomer of PD148903 (PD217015) was found to induce potential anti–Parkinson effects in animal models of PD.16 Based on structural similarities, it was hypothesized that the pro-drug is


124

converted to one or more dopaminergic, hydroxylated aminotetralins. Subcutaneous administration of 10 mg/kg PD217015 to male Wistar rats in a striatal microdialysis experiment revealed that the enone was biotransformed into the mixed dopamine D1/D2 agonist (S)-5,6-dihydroxy-2-(N,N-dipropylamino)tetralin ((S)-5,6-di-OH-DPAT, Figure 7.1c). The metabolite was identified by comparison of retention times with a racemic 5,6-di-OH-DPAT standard in an HPLC system equipped with an electrochemical detector.6,16,17

The other catecholamine pro–drug tested, 1-propyl-trans-2,3,4,4a,5,7,8,9,10,10a-decahydrobenzo[g]quinoline-6-one (GMC6650, Figure 7.1b), is an analog of PD148903 and has been synthesized and evaluated in our laboratory.6 Neither of the enantiomers has in vitro binding for the DA receptor family. Striatal microdialysis experiments revealed the (–)–enantiomer as a highly potent and long lasting DA agonist while the (+)-enantiomer was pharmacologically inactive at the DA receptors. It was assumed that (–)-GMC6650 follows the same bioactivation pathway as PD217015 and is converted in vivo to trans-1-propyl-6,7-di-OH-2,3,4,4a,5,10,10a-octahydro-benzo[g]quinoline.

To obtain further evidence for the biotransformation of these enone pro–drugs, we have given relatively high doses of PD217015, (–)-GMC6650 and (+)-GMC6650 orally to male Wistar rats that had been operated to insert a jugularis valve for blood sampling. An analytical method was developed for analysis of blood plasma and of brain tissue by means of LC/MS/MS.

The main aim of these studies was to provide further evidence for the in vivo conversion of the enones into their catechol counterparts. The second objective was to look for additional metabolites in brain and plasma, in order to better understand the biotransformations of these pro-drugs in vivo.

7.2 Experimental

In vivo rat experiments

PD217015 Test 1: Three male Wistar rats were operated to insert a jugularis valve for blood sampling. The test animals were given 40 µmol/kg PD217016 p.o. One of the rats died 15 minutes after the oral administration. The two surviving rats showed dopamine agonist stereotyped behavior throughout the experiment. Blood (0.3 ml) was sampled at 0, 15, 30 and 60 minutes. Mercaptoethanol (10 µl, 0.35 %) and EDTA (10 µl, 10 %) were added to the blood and the samples were then centrifuged for 10 minutes at 3000 rpm at in a Chilspin MSE centrifuge (Chilspin, MSE, England). The supernatants were transferred to 1.5 ml plastic eppendorf vials and were kept in the freezer at –18°C until the day of analysis.


125

PD217015 Test 2: One male Wistar rat was given 10µmol/kg PD217016 p.o. 65 Minutes following the oral administration, the test animal was killed with an overdose of chloral hydrate. The whole brain was taken out, weighed, dissected, frozen on dry ice and homogenized in 10 ml 60 % acetonitrile / 40 % water containing 0.1 % formic acid and 0.01 % mercaptoethanol. The homogenized brain was centrifuged for 10 minutes at 3000 rpm in a Chilspin MSE centrifuge (Chilspin, MSE, England). The supernatant was transferred to a glass tube and was kept in the freezer at –18°C until the day of analysis. GMC6650 Two male Wistar rats were operated to insert a jugularis valve for blood sampling. One of the rats was given 10 µmol/kg (+)-GMC6650 p.o. and the other rat was given 10 µmol/kg (–)-GMC6650 p.o. The rat given (–)-GMC6650 showed dopamine agonist stereotyped behavior in the form of sniffing and chewing 10 minutes following the oral administration. The rat given (+)-GMC6650 showed no stereotyped behavior throughout the experiment. Blood (0.3 ml) was sampled at 0, 15, 30 and 60 min. Mercaptoethanol (20 µl, 0.35 %) and EDTA (20 µl, 10 %) were added to the blood and the samples were then centrifuged for 10 minutes at 3000 rpm in a Chilspin MSE centrifuge (Chilspin, MSE, England). The supernatants were transferred to 1.5 ml plastic eppendorf vials and were kept in the freezer at –18°C until the day of analysis. At 65 min, the two test animals were killed with an overdose of chloral hydrate. A heart punction was performed and a large volume of blood was collected in EDTA vacutainer tubes (Becton Dickinson, France). This large volume of blood was worked up in the same way as the smaller samples with the exception that no mercaptoethanol was added to the tubes. The heart punction of the rat given (+)-GMC6650 failed. Therefore, no data are reported in the experimental section for (+)-GMC6650 at 65 minutes. In addition, the whole brains were taken out, weighed, dissected, frozen on dry ice and homogenized in 10 ml 60 % acetonitrile / 40 % water containing 0.1 % formic acid and 0.01 % mercaptoethanol. The homogenized brain was centrifuged for 10 minutes at 3000 rpm in a Chilspin MSE centrifuge (Chilspin, MSE, England). The supernatant was transferred to a glass tube and was kept in the freezer at –18°C until the day of analysis. Work-up procedure of the plasma samples To a 1.5 ml plastic eppendorf vial was added R-(–)-dipivaloyl apomorphine (DPA, Figure 7.1d) (20 µl, 10 µM) as internal standard. 200 µl Plasma was then added to the vial and after about 5 seconds of vortexing, 600 µl acetonitrile was added to promote precipitation of plasma proteins. The mixture was vortexed for about 10 seconds and centrifuged at 9000 RPM for about 10 minutes in a Sigma 110 microcentrifuge (SIGMA Laborzentrifugen, Osterode, Germany). The liquid phase was transferred to 10 ml glass


126

tubes an evaporated under a flow of nitrogen gas. The samples were dissolved in 150 µl mobile phase and transferred to 200 µl glass vials for injection onto the LC/MS/MS system. Work-up procedure of the brain samples 1 ml of the homogenized brain samples was transferred to a second glass tube and was spiked with DPA to a final concentration of 1 µmol/kg brain tissue. The samples were evaporated under a flow of nitrogen gas, dissolved in 150 µl mobile phase and transferred to 200 µl glass vials for injection onto the LC/MS/MS system. LC/MS/MS conditions The worked up samples were transferred to a series 200 autosampler (Perkin Elmer, Norwalk, CT, USA) equipped with a 20 µl injection loop, and injected onto a reverse phase HPLC column (Alltech Alltima, C18, 5 µm, 150 x 2.1 mm, Alltech Associates, Deerfield, IL, USA). Two Series 200 micro LC pumps (Perkin Elmer, Norwalk, CT, USA) were programmed to deliver a gradient from water and acetonitrile at a total flow rate of 200 µl/min. Both phases contained 0.1 % formic acid. For the first two minutes, the acetonitrile concentration was kept at 15 %, and was then linearly increased to 95 % over a period of 15 minutes. The mobile phase composition was kept at 95 % acetonitrile for another 5 minutes and was then linearly brought back to 15 % acetonitrile in 2 minutes. The system was allowed to equilibrate for at least 10 minutes between the injections. Full-scan spectra, product ion spectra or selected reaction monitoring (SRM) traces were acquired continuously with an API 3000 triple quadrupole mass spectrometer (MDS-Sciex, Concord, Ontario, Canada) equipped with a TurboIonspray interface. A splitter was introduced between the HPLC–column and the mass spectrometer to reduce the flow to approximately 40 µl/min. The MS was operated at such a low orifice voltage that “up front” collision induced dissociation did not take place. Quantification of the integrated peaks For all samples analyzed, the internal standard (dipivaloylapomorphine, Figure 7.1d) was added to represent a concentration of 1 µmol/l plasma in the plasma samples and 1 µmol/kg brain in the brain samples. The concentrations mentioned in the results and discussion section are calculated from the integrated SRM traces. For quantification of the compounds PD217015, 5,6-di-OH-DPAT, (–)-GMC6650 and (+)-GMC6650, blank plasma from each individual rat was spiked with 1 µM of the respective compound and 1 µM DPA. The relation between the integrated peaks of each respective compound and DPA in this blank plasma sample was used to compensate for signal suppression in the collected samples. The plasma concentrations given for these compounds have thus been corrected


127

to account for signal suppression from endogenous compounds in each individual plasma. Signal suppression from endogenous compounds in the brain samples has not been taken into account and the given concentrations must be regarded as approximate determinations. The concentration of metabolites for which there was no standard available has been calculated with the same corrections as was used for the corresponding pro–drug and must also be regarded as approximate determinations. 7.3 Results and Discussion

PD217015 The mass chromatogram of m/z 250 corresponding to the [M+H]+ ion of PD217015 (2 µM standard) is given in Figure 7.2 together with its corresponding product ion spectrum. The suggested collision induced dissociation (CID) of m/z 250 is presented in Figure 7.3. The fragment at m/z 102 corresponds to protonated dipropylamine. A metabolite showing a gain of 16 to this fragment would suggest that an oxygen atom has been introduced somewhere at the dipropylamine structure. Following the same line of reasoning, a gain of 16 to the fragment at m/z 149 would suggest a hydroxylation somewhere at the 3,4,5,6,7,8-hexahydro-2H-naphtalen-1-one structure.


128

a) 2 µM PD217015 std

6

b) RT 6.4 min, 2 µM PD217015 std, product ions of m/z 250

50 100 150 200 250

5560

7991

102

121

131

149

208

25093

5.8 x 10 5

6.4a) 2 µM PD217015 std

Figure 7.2 2 µM PD217015 standard. The signal intensities in counts/s are given in the upper right corner of each chromatogram and spectrum. a) Mass chromatogram of m/z 250, corresponding to the [M+H]+ ion of PD217015. b) Product ion spectrum of the peak with m/z 250 at RT 6.4 minutes in chromatogram 7.2a.


129

N

O

H

H

PD217015

m/z 250

+N H-

NH

H+

m/z 102

O

H+

m/z 149CH2

O

+

CH2

CH2

O

+C2H4-

+

- CO

CH2

O

CH2+

m/z 121m/z 93

CH2

O

H

H +

+CH2

CH2

O

H

CH2

O

H

+

C

O

H

HH

+

C2H4-CH2

CH2

O

+

m/z 149

m/z 121

m/z 60

NH3+

m/z 131

+HH2O-

m/z 149

+

O

H

CH2

O

+

m/z 149

CH2+

m/z 131

ProtonTransfer

ProtonTransfer

Figure 7.3 Suggested CID for the [M+H]+ ion of PD217015

PD217015, metabolites For identification of (S)-5,6-di-OH-DPAT, a racemic 5,6-di-OH-DPAT standard was used. The mass chromatogram of m/z 264 corresponding to the [M+H]+ ion of 5,6-di-OH-DPAT (2 µM standard) is given in Figure 7.4a. The corresponding product ion spectrum is presented in Figure 7.5a.


130

a) 2 µM (S)-5,6-di-OH-DPAT

m/z 264

1.1 x 106

6.8b) brain, SRM 264 –> 163 2.2 x 103

4.0

c) plasma, Product ion chromatogram of m/z 266 7.2 x 105

3.75.2

3.9d) plasma, Product ion chromatogram of m/z 208 9.7 x 104

8.53.0

6.8a) 2 µM (S)-5,6-di-OH-DPATm/z 264

Figure 7.4 Chromatograms for metabolites of PD217015. The signal intensities in counts/s are given in the upper right corner of each chromatogram. The numbers above the peaks represent retention times in minutes. a) Mass chromatogram of the [M+H]+ ion of 2 µM 5,6-dihydroxy-DPAT standard. b) SRM chromatogram (264 → 163) from rat brain dissected 60 minutes after oral distribution of 10 µmol/kg PD217015. c) Product ion chromatogram of m/z 266 from rat plasma taken 60 minutes after oral distribution of 10 µmol/kg PD217015. d) Product ion chromatogram of m/z 208 from the same rat plasma as in chromatogram c.


131

a) 2 µM 5,6-di-OH-DPAT, pr. 264

50 100 150 200 250

102

1.0 x 106

117

145

163

264

b) RT 6.8 min, brain, pr. 264

50 100 150 200 250

102

117

145

163

264

5.0 x 102

c) RT 3.7 min, plasma, pr. 266

50 100 150 200 250

60

121

91

2.6 x 104 266

165

102119

147129137

d) RT 4.0 min, plasma, pr. 266

50 100 150 200 250

91

60

1.3 x 105 266

165

102

119

147

248

129137

50 100 150 200 250

131

149

93

1.8 x 103

266

24858

100

118

144

e) RT 5.2 min, plasma, pr. 266

50 100 150 200

5560

79

131

149208

7.0 x 103

67

91

9395

86

105107

121123

f) RT 3.9 min, plasma, pr. 208

a) 2 µM (S)-5,6-di-OH-DPAT

Figure 7.5 Product ion spectra for metabolites of PD217015. The signal intensities in counts/s are given in the upper right corner of each spectrum. a) Product ion spectrum of m/z 264 from 2 µM 5,6-dihydroxy-DPAT standard. The step size 0.1 was used. For all the other spectra, the step size 1 was used. b) Product ion spectrum of m/z 264 taken at 6.8 minutes in the chromatogram from rat brain dissected 60 minutes after oral distribution administration of 10 µmol/kg PD217015. Because of the low signal intensity, the scan range was limited to m/z 100-200. c-e) Product ion spectra of m/z 266 for the peaks at RT 3.7, 4.0 and 5.2 minutes in chromatogram 7.4c. f) Product ion spectrum of m/z 208 for the peak at RT 3.9 minutes in chromatogram 7.4d.


132

The selected reaction monitoring (SRM) transition 264→163, representing a loss of dipropylamine from the [M+H]+ ion of 5,6-di-OH-DPAT, was chosen to quantify 5,6-di-OH-DPAT in the dissected rat brain. The SRM chromatogram given in Figure 7.4b shows a peak at 6.8 minutes, the same retention time as observed for the racemic 5,6-di-OH-DPAT standard (Figure 7.4a). Despite the rather low concentration, it was possible to generate a useful product ion spectrum of m/z 264 in the brain sample. The distribution of fragments in the resulting spectrum presented in Figure 7.5b is the same as in the product ion spectrum of the 5,6-di-OH-DPAT standard in Figure 7.5a. Based on comparison of retention time and product ion spectrum with the 5,6-di-OH-DPAT standard it was concluded that 5,6-di-OH-DPAT is formed in vivo in rat following oral administration of PD217015.

To account for the formation of (S)-5,6-di-OH-DPAT in vivo, it is reasonable to assume an initial hydroxylation at position 2 on PD217015 (Figure 7.1a). A product ion scan of m/z 266, corresponding to a mass gain of 16 Da to the [M+H]+ ion of PD217015, revealed three peaks with retention times 3.7, 4.0 and 5.2 minutes in the plasma sample taken at 60 minutes (Figure 7.4c). The corresponding product ion spectra are given in Figures 7.5c, 7.5d and 7.5e.

The first two peaks in Figure 7.4c (RT 3.7 and 4.0 minutes) show similar product ion spectra and the only difference is the presence of a fragment at m/z 121 in the peak with RT 3.7 minutes. The suggested CID for a compound with a hydroxyl group introduced at position 2 on PD217015 is presented if Figure 7.6. The fragments found at m/z 119 and m/z 91 are supposedly formed via loss of a water molecule followed by rearrangement to the phenol structure. This is only possible if the hydroxyl group is located either at positions 2,3 or 4 on PD217015. With the hydroxyl group located at position 2, the fragment at m/z 121 can be generated through two different routes. We suggest that the peak at RT 3.7 minutes corresponds to PD217015 hydroxylated at position 2 and represents an intermediate step in the formation of 5,6-di-OH-DPAT.


133

N

O

H

HHO

HH

m/z 266

+ - N H

OH

HO

HH H

+

+N

O

H

H2O-

N H-

H2O-

m/z 165

m/z 248 O

H+

m/z 147

m/z 147O

HH H

+

O

H

H

+

m/z 91

m/z 119

C7H7+ - CO H2C=CH2-

OH

H+

m/z 147

OH

H+

m/z 147

CH2

O

H

+ CH2

O

H

+

C

O HHH+

CH2

CH2

O

CH2

+

CH2

O

CH2

+

N

O

HO

H+

N_

Retro Diels Alder O

HO

H+ H+O

H2O-

[M+H] = 128+

m/z 266 m/z 139 m/z 121

+CH2

CH2

O

HO

+CH2

O

HO

m/z 121

CH2

O+

OH

–

Figure 7.6 Suggested CID for the [M+H]+ ion of 2-hydroxy-PD217015

In the spectrum of the compound eluting at 4.0 minutes, there is only a trace of m/z 121 and an intense fragment at m/z 119. The suggested CID presented in Figure 7.7 illustrates how a compound hydroxylated at position 4 on PD217015 can give rise to an intense fragment at m/z 119. We suggest that the metabolite eluting 4.0 minutes corresponds to 4-hydroxy-PD217015.


134

m/z 266

4

O

O

H

HH

H

+N

O

OH4

m/z 165

H2O_

m/z 147

O

+

O

H+

C

O

H

+C

OH

+

H +m/z 119

_ CO

+

_ CO

m/z 119

H+

m/z 266

4

O

O

H

HH

H

+N

O

OH4

m/z 165

H2O_

m/z 147

O

+

O

H+

C

O

H

+C

OH

+

H +m/z 119

_ CO

+

_ CO

m/z 119

H+

Figure 7.7 Suggested CID for the [M+H]+ ion of 4-hydroxy-PD217015 The peak with RT 5.2 minutes in Figure 7.4c is weak but nevertheless a useful product ion spectrum was obtained. The presence of a fragment at m/z 149 suggests that the 3,4,5,6,7,8-hexahydro-2H-naphtalen-1-one structure is intact. The fragment at m/z 118 suggests a gain of an oxygen atom at the dipropylamino group. The fragment at m/z 100 would then be consistent with a loss of water from m/z 118. The N–oxide of PD217015 or a product hydroxylated at one of the propyl chains would be consistent with this fragmentation pattern.

A common metabolic route for tertiary amines is N-dealkylation.21 The product ion chromatogram of m/z 208, presumably the [M+H]+ ion of monopropyl PD217015, is presented in Figure 7.4d. Three peaks were observed at retention times 3.0, 3.9 and 8.5 minutes. The product ion spectra for the two peaks at RT 3.0 and 8.5 showed no similarities to the product ion spectrum of PD217015 and most likely represent endogenous compounds with mw 207. However, the product ion spectra for the peak at RT 3.9 given in Figure 7.5f shows a similar fragmentation pattern to PD217015. The fragment found at m/z 149 suggests that the 3,4,5,6,7,8-hexahydro-2H-naphtalen-1-one structure is intact. This peak is believed to correspond to the [M+H]+ ion of the N–depropylated metabolite of PD217015.


135

GMC6650 The mass chromatogram of m/z 248 corresponding to the [M+H]+ ion of (–)-GMC6650 (1 µM standard) is given in Figure 7.8 together with its corresponding product ion spectrum. We were not able to account for all the fragments in the spectra, but the suggested CID of m/z 248 with plausible explanations for most of the fragments is presented in Figure 7.9.

RT 6.3 1.2 x 10

6a) (-)GMC6650, m/z 248

b) Products of m/z 248

50 100 150 200 250

4.5 x 104

248

206

189171

161119105

91

7972

55

147145143

129

135133

RT 6.3 1.2 x 106

Figure 7.8 1 µM (–)-GMC6650 standard. The signal intensities in counts/s are given in the upper right corner of each chromatogram and spectrum. a) Mass chromatogram of m/z 248, corresponding to the [M+H]+ ion of GMC6650. b) Product ion spectrum of the peak with m/z 248 at RT 6.3 minutes in chromatogram a.


136

N

O

HGMC 6650

m/z 248

+

O

HHN H

H+

O

N HH+

O

CH2

+

m/z 189

C2H4-

O

CH2

+

m/z 161

CH2

O+

CH2

O

+

O+ m/z 133

C2H4- - CO

+

m/z 105

+

O

CH2

H

HCH2

+

m/z 171 C2H4-

CH2+

m/z 143

+

m/z 171

a) m/z 248 --> 189 --> 161 --> 133 --> 105

b) m/z 189 --> 171 --> 143

m/z 189

CH2

- H2O

- H2O

+

C2H4-

CH2+

m/z 143

m/z 171

+

m/z 171

Figure 7.9 Suggested CID for the [M+H]+ ion of GMC6650


137

+

m/z 248

GMC 6650

N

O

H

Proton

Transfer

N

O

HH +

m/z 206

c) m/z 248 --> 206

d) m/z 189 --> 147 --> 119 --> 91

O

CH2

H

+

m/z 189 O

H

CH3 +

O

H

_

+

m/z 147

m/z 147

+

O

H

H H O H+

C2H4-

RDA C

O H+m/z 119

- COC7H7

m/z 91

H2O-

H+

m/z 129

vI

129

+

Hydrogen

transfer

Figure 7.9 Suggested CID for the [M+H]+ ion of GMC6650, continued

GMC6650, metabolites Figure 7.10a shows the SRM chromatogram of the presumed catechol detected in the brain 60 minutes following oral administration of 10 µmol/kg (–)-GMC6650. The two SRM transitions were chosen based on the assumption that the catechol product would lose propylamine in the same way as GMC6650 and a water molecule. Both SRM transitions gave a peak at 6.4 minutes. The presumed catecholamine metabolite elutes just after the GMC6650 peak in the chromatogram. The same distribution was observed in the chromatogram for PD217015 and its identified catecholamine metabolite 5,6-di-OH-DPAT. We find it very reasonable to assume that the peak at 6.4 minutes in Figure 7.10a corresponds to the catechol metabolite 1–propyl-1,2,3,4,4a,5,10,10a-octahydro-


138

benzo[g]quinoline-6,7-diol (6,7-di-OH-PBGQ).* The corresponding catechol for the (+)–enantiomer of GMC6650 was not detected in brain or plasma. QuickTime™ and aGraphics decompressorare needed to see this picture.

RT 6.4 6.8 x 102a) brain, MRM 262 –>203

and 262 –>185

RT 4.2

2.6 x 104

6.7 x 103

b) plasma, product ion scan of m/z 264, TIC

c) RT 4.2 product ion scan of m/z 264

50 100 150 200 250

246187169

159145989172

264

131133135

120117

141

6.8 x 102

Figure 10 Chromatograms and product ion spectrum for metabolites of (–)-GMC6650. The signal intensities in counts/s are given in the upper right corner of each chromatogram and spectrum. The numbers above the peaks in chromatograms a and b represent retention times in minutes. a) MRM chromatogram (262→203 and 262→185) of the suggested catechol metabolite 6,7-di-OH-PBGQ detected in brain 65 minutes following oral distribution administration of (–)-GMC6650. b) Product ion chromatogram of m/z 264 detected in plasma 65 minutes following oral administration distribution of (–)-GMC6650. c) Product ion spectrum of the peak with m/z 264 at RT 4.2 minutes in chromatogram b. In order to account for the formation of the catechol metabolite, it is reasonable to assume an initial hydroxylation at position 7 on GMC6650. In contrast to PD217015, there was only one metabolite showing a gain of 16 Da to the GMC6650 structure found in the brain

* This structure has now been confirmed by comparison of retention time and product ion spectrum with a synthesized reference of 6,7-di-OH-PBGQ.


139

as well as in the plasma. This metabolite eluted at RT 4.2 minutes and was found for both enantiomers, although the concentration of this metabolite was much higher for (–)-GMC6650 than for (+)-GMC6650. The possibility that this metabolite would correspond to the N-oxide was excluded since a synthetic standard of the GMC 6650 N-oxide was injected and eluted several minutes after the metabolite. The product ion chromatogram of m/z 264, corresponding to a mass gain of 16 Da to the [M+H]+ ion of (–)-GMC6650, is given in Figure 7.10b. The corresponding product ion spectrum is given in Figure 7.10c. The fragments at m/z 187, 169, 159 and 141 are consistent with the introduction of a hydroxyl group followed by loss of a water molecule from the fragments at m/z 189, 171, 161 and 143 found in the spectrum of GMC6650 (Figure 7.8b). If the suggested CID of the [M+H]+ ion of GMC6650 is correct, this would suggest hydroxylation somewhere at the positions 4 to 10a on GMC6650. We were not able to pinpoint the location of the oxygen from the interpretation of the product ion spectrum. Summary of the pro–drug biotransformation The concentrations of PD217015 and its suggested metabolites found in plasma and brain following oral administration of PD217015 to male Wistar rats are given in Table 1, with the corresponding structures presented in Figure 7.11a. The identification of 5,6-di-OH-DPAT is based on comparison of retention time and product ion spectrum with a racemic 5,6-di-OH-DPAT standard. The metabolites hydroxylated at positions 2 and 4 and the dealkylated metabolite are suggested on basis of the interpretation the respective product ion spectra.

The concentrations of (+)– and (-)–GMC6650 and there suggested metabolites found in plasma and brain following oral administration of 10 µmol/kg (+)– or (-)–GMC6650 to male Wistar rats are presented in Table 2 with the corresponding structures given in Figure 7.11b. The two metabolites are suggested on basis of the interpretation the respective product ion spectra and on their appearance in the chromatograms. There was one pronounced difference in the metabolism between the two GMC6650 enantiomers: Only the (–)–enantiomer gave detectable levels of the catechol product in the brain. Since the (4aR, 10aS)–enantiomer of 6,7-di-OH-PBGQ is the potent enantiomer of the catechol, the (–)–enantiomer of GMC6650 probably corresponds to the (4aR, 10aS) conformation. Only trace amounts of the catecholamine product were found in the plasma for both enantiomers. The catecholamine product is thus either formed only for the (–)–enantiomer or is enantioselectively transferred to the brain. The rat given the (–)–enantiomer of GMC6650 showed dopamine agonist stereotyped behavior 10 minutes after oral administration, while the rat given the (+)–enantiomer did not show any stereotyped behavior during the experiment.


140

Table 7.1 PD217015 and metabolites in brain and plasma Conditions

PD217015

5,6-di-OH-DPAT

hydroxylated metabolite

dealkylated metabolite

Brain at 65 min from rat given 10µmol/kg PD217015

130a

70

200

trace

Plasma at 60 min from rats given 40µmol/kg PD217015

3560 b

(2 rats)

trace

940

70

aConcentrations in the brain are given as nmol/kg brain tissue. bConcentrations in the plasma are given as nmol/l plasma. Table 7.2 GMC6650 and metabolites detected in brain and plasma following oral administration of 10µmol/kg (-)-GMC6650 or (+)-GMC6650.

Compound

GMC6650

catechol metabolite

hydroxylated metabolite

(–)-GMC6650 Brain at 65 min Plasma at 65 min

260 a 220 b

100 0

100 160

(+)-GMC6650 Brain at 65 min Plasma at 60 min

560 20

0 0

200 26

aConcentrations in the brain are given as nmol/kg brain tissue. bConcentrations in the plasma are given as nmol/l plasma


141

mw 249

a) PD217015

N

O

H

N

OH

HO

H

mw 263

mw 207

NH

O

H

N

O

H

HO

mw 265

N

O

b) GMC6650

N

O

HO

N

OH

HO

mw 247 mw 263

mw 261

(S)-5,6-di-OH-DPAT

6,7-di-OH-PBGQ

N

O

H

OH

mw 265

Figure 7.11 a) Suggested metabolism of PD217015. b) Suggested metabolism of GMC6650. Only the (–)-enantiomer of GMC6650 gave detectable levels of a compound believed to be 6,7-di-OH-PBGQ.

Pharmacokinetics The plasma concentration of the three tested pro–drugs during 1 h following oral administration is presented in Figure 7.12. Although the number of test animals is too low to allow any certain conclusions from these experiments, some general trends can be recognized. The rats given 40 µmol/kg PD217015 (Figure 7.12a) showed dopamine agonist stereotyped behavior throughout the experiment. This is consistent with the high concentration of PD217015 in the plasma that allowed continued biotransformation into the pharmacologically active 5,6-di-OH-DPAT.

The rat given (–)–GMC6650 also showed dopamine agonist stereotyped behavior throughout the experiment. The plasma concentration of (–)–GMC6650 was high during the experiment and although the concentration goes down over time, the high


142

concentration in the plasma presumably allowed continued biotransformation into the pharmacologically active 6,7-di-OH-PBGQ. The rat given (+)–GMC6650 showed no stereotyped behavior throughout the experiment. The plasma concentration was very low considering that the brain contained a relatively high amount (+)–GMC6650 (Table 2). The reason for this distribution is not clear at this point.

0

1000

2000

3000

4000

0 20 40 60

plas

ma

conc

entr

atio

n (n

M)

Time (minutes)

a) PD217015

0

200

400

600

0 20 40 60

plas

ma

conc

entr

atio

n (n

M)

Time (minutes)

b) (–)GMC6650

0

50

100

150

0 20 40 60

plas

ma

conc

entr

atio

n (n

M)

Time (minutes)

c) (+)GMC6650

Figure 12 Plasma concentrations of the catecholamine pro–drugs following oral administration distribution of: a) 40 µmol/kg PD217015; b) 10 µmol/kg (–)-GMC6650 and c) 10 µmol/kg (+)-GMC6650

7.4 References

1. S. Bharath; M. Hsu; D. Kaur; S. Rajagopalan; J. K. Andersen, "Glutathione, iron and Parkinson's disease", Biochem. Pharmacol., 2002, 64, 1037-1048.


143

2. K. D. Sethi, "Clinical aspects of Parkinson's diesease", Curr. Opin. Neurol., 2002, 15, 457-460.

3. F. Blandini; G. Nappi; C. Tassorelli; E. Martignoni, "Functional changes in the basal ganglia circuitory in Parkinson's disease", Prog. Neurol., 2000, 62, 63-88.

4. E. C. Hirsch; B. A. Faucheux, "Iron metabolism and Parkinson's disease", Movement Disorders, 1998, 13, 39-45.

5. P. Jenner, "Oxidative mechanisms in nigral cell death in Parkinson's disease", Movement Disorders, 1998, 13, 24-34.

6. B. J. Venhuis, PhD thesis, "Enone pro–drugs of catecholamines", Department of Pharmacy, University of Groningen, 2002.

7. I. R. Yode; M. J. Raxworthy; P. A. Gulliver; D. Dijkstra; A. S. Horn, "The metabolism of dopamine, N,N-dialkylated dopamines and derivatives of the dopamine agonist 2-amino-dihydroxy-1,2,3,4-tetrahydronaphtalene (ADTN) by catechol-O-methyltransferase", J. Pharm. Parmacol., 1983, 36, 309-313.

8. P. J. Swart; W. E. Oelen; A. P. Bruins; P. G. Tepper; R. A. de Zeeuw, "Determination of the dopamine agonist N-0923 and its major metabolites in perfused rat livers by HPLC-UV-atmospheric pressure ionization mass spectrometry.", J. Anal. Toxicol., 1994, 18, 71-77.

9. T. K. Gerding; B. F. H. Drenth; R. A. de Zeeuw; P. G. Tepper; A. S. Horn, "The metabolic fate of the dopamine agonist 2-(N-propyl-N-thienylamino)-5-hydroxytetralin in rats after intravenous and oral administration. II. Isolation and identification of metabolites.", Xenobiotica, 1990, 20, 525-536.

10. T. K. Gerding; B. F. H. Drenth; H. J. Roosenstein; R. A. de Zeeuw; P. G. Tepper; A. S. Horn, "The metabolic fate of the dopamine agonist 2-(N-propyl-N-thienylamino)-5-hydroxytetralin in rats after intravenous and oral administration. I. Disposition and metabolic profiling", Xenobiotica, 1990, 20, 515-524.

11. C. E. Tedford; V. B. Ruperto; A. Barnett, "Characterization of Rat Liver Glucuronyltransferase that Glucuronidates the selective D1 Antagonist, SCH 23390, and other benzazepines", Drug Metab. Dispos., 1991, 19, 1152-1159.

12. N. Bodor; K. B. Sloan; T. Higuchi; K. Sasahara, "Improved delivery through Biological Membranes. 4. Prodrugs of L-Dopa", J. Med. Chem., 1977, 20, 1435-1445.

13. H. V. Wikström; P. Lindberg; P. Martinson; S. Hjorth; A. Carlsson, "Pivaloyl Esters of N,N-dialkylated Dopamine Congeners. Central Dopamine-Receptor Stimulating Activity", J. Med. Chem., 1978, 21, 864-867.

14. R. J. Borgman; J. J. McPhillips; R. E. Stitzel; I. J. Goodman, "Synthesis and Pharmacology of Centrally Acting Dopamine Derivatives and Analogs in Relation to Parkinson's Disease", J. Med. Chem., 1973, 16, 630-633.


144

15. N. Bodor; H. H. Farag, "Improved Delivery through Biological Membranes. 13. Brain-Specific Delivery of Dopamine with a Dihydropyridine - Pyridinium Salt Type Redox Delivery System", J. Med. Chem., 1983, 26, 528-534.

16. N. Rodenhuis, PhD Thesis, "New centrally acting dopaminergic agents with an improved oral availability: synthesis and pharmacological evaluation", Department of Pharmacy, University of Groningen, 2000.

17. B. J. Venhuis; H. V. Wikström; N. Rodenhuis; S. Sundell; D. Dijkstra, "A new type of prodrug of catecholamines: An opportunety to improve the treatment of Parkinson's disease", J. Med. Chem., 2002, 45, 2349-2351.

18. S. J. Jonson; T. G. Heffner; L. T. Meltzer; T. A. Pugsley; L. D. Wise, "Dihydroanalogues of 5- and 7-hydroxy-2-aminotetralins: Synthesis and dopaminergic activity", Abstracts of papers, 208th National Meeting of the American Chemical Society Meeting, Washington DC, USA, 1994.

19. B. J. Venhuis; H. V. Wikström; D. Wustrow; L. T. Meltzer; L. D. Wise; S. J. Johnson; D. Dijkstra, "PD148903, derivatives and analogs", Abstracts of the XVIth International Symposium on Medicinal Chemistry, Bologna, Italy, 2000.

20. D. Dijkstra; B. J. Venhuis; H. V. Wikström; S. J. Johnson; L. D. Wise; D. J. Wustrow; L. T. Meltzer In Patent, WO0128977 A1, 2001.


145

Overall conclusions and future prospects

In these studies, two electrochemical techniques have been developed for oxidation of drugs and xenobiotics on-line with electrospray mass spectrometry. The first technique, the EC/MS system, provided successful mimics of cytochrome P450 catalyzed oxidations in cases where the enzymatic oxidations are supposed to proceed via a mechanism initiated by a one electron oxidation, such as N–dealkylation, S–oxidation, P–oxidation, alcohol oxidation and dehydrogenation. The P450 catalyzed reactions initiated via direct hydrogen atom abstraction, such as O–dealkylation and hydroxylation of aromatic rings without electron donating groups were not mimicked by the EC/MS system.1 This investigation demonstrated what types of biotransformations that can be mimicked by EC/MS and provides a basis for how the EC/MS technique can be implemented into biotransformation studies. The second technique, the EC-Fenton system, allows generation of hydroxyl radicals for reaction with xenobiotics and subsequent detection of the oxidation products on–line with LC/MS/MS.2 The technique holds great promise as a tool for investigation of interactions between radicals and biologically relevant molecules, in particular radical scavengers and antioxidants. In addition, the technique is complementary to the EC/MS system and is suitable for generation of metabolites resulting from hydroxylation and epoxidation.

Both electrochemical techniques are under further investigation at AstraZeneca R&D Mölndal, Sweden, with the aim to implement them as regular tools in biotransformation studies.

In the present setup of the EC/MS system, the electrochemical oxidations are limited to oxidations that are possible at the porous graphite electrode. By chemical modification of the electrode surfaces, we expect that an even broader range of metabolic oxidations can be mimicked by the EC/MS system and consequently make the technique even more suitable in biotransformation studies.3-

6 This work has also opened up a new avenue in protein research through the specific cleavage of peptides.7 Both of these matters will be investigated further. The universities in Groningen and Twente are working on a grant application for a cooperative project that is financially supported by the pharmaceutical industry, instrument companies and STW.

Electrochemical techniques coupled on-line with electrospray mass spectrometry has received increased attention over the last couple of years and we expect to see a continued interest and much progress in this field in the future. 1. U. Jurva; H.V. Wikström, L. Weidolf, A.P. Bruins “Comparison between

electrochemistry/mass spectrometry and cytochrome P450 catalyzed oxidation reactions” Rapid Commun. Mass Spectrom. 2003, 17, 800-810.

146

2. U. Jurva, H.V. Wikström, A.P. Bruins, “Electrochemically assisted Fenton reaction:

reaction of hydroxyl radicals with xenobiotics followed by on-line analysis with high-performance liquid chromatography/tandem mass spectrometry” Rapid Commun. Mass Spectrom. 2002, 16, 1934-1940.

3. A.O. Solak, L.R. Eichorst, W.J. Clark, R.L. McCreery, “Modified carbon surfaces as

"organic electrodes" that exhibit conductance switching”. Anal. Chem. 2003, 75, 296-305. 4. M.A. Rahman, M. Won, Y. Shim, “Characterization of an EDTA bonded conducting

polymer modified electrode: Its application for the simultaneous determination of heavy metal ions” Anal. Chem. 2003, 75, 1123-1129.

5. H. Ma, N. Hu, J.F. Rusling, “Electroactive Myoglobin films grown layer-by-layer with

poly(styrenesulfonate) on pyrolytic graphite electrodes”. Langmuir 2000, 16, 4969-4975. 6. E.I. Iwuoha, M.R. Smyth, “Reactivities of organic phase biosensors: 6. Square-wave and

differential pulse studies of genetically engineered cytochrome P450(cam) (CYP101) bioelectrodes in selected solvents” Biosensors & Bioelectronics 2003, 18, 237-244.

7. H.P. Permentier, U. Jurva, B. Barroso, A.P. Bruins, “Electrochemical oxidation and

cleavage of peptides analyzed with on-line mass spectrometric detection” Rapid Commun. Mass Spectrom. 2003, 17, 1585-1592.

147

Summary

The main goal of these studies was the development of new techniques on-line with electrospray mass spectrometry for in vitro generation and characterization of drug metabolites. In order to achieve this goal two electrochemical techniques have been developed for oxidation of drugs and xenobiotics on-line with electrospray mass spectrometry.

In chapter 1, a short theoretical introduction is provided to the main concepts discussed in this thesis. The first section gives a historic overview of the developments of electrochemistry coupled on-line with mass spectrometry. The second section gives a short, theoretical description of cytochrome P450, the most important enzyme family in oxidative drug metabolism. The third section provides an introduction to reactive oxygen species (ROS) with the focus on the hydroxyl radical. The ROS section is followed by a brief overview of antioxidants and the chapter is finished with a definition of the scope of the thesis.

In chapter 2, the extent to what electrochemistry on–line with electrospray mass spectrometry (EC/MS) can be used to mimic cytochrome P450 catalyzed oxidations has been investigated. Comparisons at the mechanistic level have been made for most reactions in an effort to explain why certain reactions can, and some cannot, be mimicked by electrochemical oxidations. The EC/MS system provided successful mimics in cases where the P450 catalyzed reactions are supposed to proceed via a mechanism initiated by a one electron oxidation, such as N–dealkylation, S–oxidation, P–oxidation, alcohol oxidation and dehydrogenation. The P450 catalyzed reactions initiated via direct hydrogen atom abstraction, such as O–dealkylation and hydroxylation of aromatic rings without electron donating groups generally had a too high oxidation potential to be electrochemically oxidized below the oxidation potential limit of water, and were not mimicked by the EC/MS system.

In chapter 3, an electrochemical flow through system was introduced that allows the generation of hydroxyl radicals for reaction with xenobiotics and subsequent detection of the oxidation products on–line with LC/MS/MS. The system is based on the Fenton reaction and is predominantly aimed at the generation of hydroxyl radicals, but by minor variations to the system, a broad range of other radicals can be produced. Since the iron is added as Fe3+, the initial mixture is “inactive” until it reaches the electrochemical cell. This makes it very suitable for on–line analysis of the compounds generated, since the whole reaction mixture, including substrate, can be kept in a vial in an autosampler. The system described provides a useful tool for investigation of new radical scavengers and antioxidants. Since the hydroxyl radical adds readily to unsaturated π–systems, the technique is also suitable for on–line generation and characterization of potential drug metabolites resulting from hydroxylation of double bonds and aromatic systems.

148

In chapter 4, the reactivity of the prenylated flavonoid xanthohumol (XN) towards hydroxyl radicals has been investigated in the electrochemically assisted Fenton system (EC-Fenton system). An experiment with human liver slices was performed and it was found that out of the 10 metabolites showing a mass gain of 16 Da (oxygen), 9 were also formed in the EC-Fenton system. The electrochemically assisted Fenton system provides a valuable tool for investigation of the reactivity of flavonoids towards oxygen radicals. In the case of xanthohumol, the system produces an array of oxidation products that resembles oxidative metabolism in biological systems.

In chapter 5, the two on–line oxidation systems, the EC/MS system and the EC–Fenton system, have been applied to the dopamine agonist S(–)-(N-propyl-N-2-thienylethylamine)-5-hydroxytetralin (N–0923). The oxidative metabolism previously reported from rat liver perfusion experiments was partly mimicked by both methods. The results give an indication that the EC/MS system and the EC–Fenton system give different products and that they can be complementary to each other.

In chapter 6, some of the different problems that one might encounter when coupling an electrochemical cell on-line with an electrospray ionization mass spectrometer are discussed. Some possible solutions are provided for each situation.

In chapter 7, the bioactivation of the potential catecholamine pro–drugs PD217015, (–)-GMC6650 and (+)-GMC6650 has been studied by LC/MS/MS analysis of blood plasma and brain samples from male Wistar rats following oral administration. The catechol metabolite of PD217015, 5,6-dihydroxy-DPAT, was identified in the brain. From interpretation of the product ion spectra, three additional metabolites of PD217015 are suggested. The N–dealkylated metabolite 6-propylamino-3,4,5,6,7,8-hexahydro-2H-naphtalene-1-one and two hydroxylated metabolites suggested to be 2–hydroxy–PD217015 and 4–hydroxy–PD217015. Hydroxylation at position 2 is proposed to constitute an intermediate step in the formation of the active principle 5,6-dihydroxy-DPAT.

For GMC6650, a metabolite believed to correspond to the catechol active principle 1–propyl-1,2,3,4,4a,5,10,10a-octahydro-benzo[g]quinoline-6,7-diol was detected in the brain of the rat given the (–)–enantiomer.

Although MS/MS data often provides useful information about where on a molecule an oxidation has taken place, additional information is sometimes needed to provide sufficient characterization of an unknown metabolite. Both of the electrochemical techniques described above can be scaled up to generate amounts sufficient for analysis with other techniques. If a compound generated with any of the techniques shows the same retention time and product ion spectrum as an unknown metabolite found in a biological sample, a scaled up system could be used to generate this metabolite in amounts sufficient for characterization by IR and NMR.

149

The simplicity of the techniques, and the ease and speed with which they can be applied to a large number of compounds make them valuable tools in drug metabolism research.

150

Samenvatting

De doelstelling van dit onderzoek was het ontwikkelen van nieuwe technieken gekoppeld aan electrospray massaspectrometrie, om geneesmiddel-metabolieten in vitro te genereren en te karakteriseren. Twee electrochemische technieken zijn ontwikkeld die geschikt zijn voor oxidatie van geneesmiddelen en xenobiotica on-line met electrospray massaspectrometrie.

In het inleidende hoofdstuk is een overzicht gegeven van de belangrijkste begrippen van dit onderzoek. Een historisch overzicht van de ontwikkelingen van electrochemie on-line met massaspectrometrie wordt gevolgd door een beschrijving van cytochroom P450, het belangrijkste enzymsysteem voor oxidatief metabolisme. Tevens wordt Reactive Oxygen Species (ROS) met fokus op het hydroxyl-radikaal beschreven. De ROS sectie wordt gevolgt door een korte inleiding over antioxidanten en het hoofdstuk wordt afgerond met een definitie van het overgrijpende doel van dit proefschrift.

In het tweede hoofdstuk wordt het onderzocht in hoeverre het mogelijk is om cytochroom P450 gekatalyseerde oxidaties na te bootsen met een electrochemische cel gekoppeld on-line met electrospray massaspectrometrie (EC/MS). Vergelijkingen zijn gemaakt op grond van het mechanisme van het merendeel van de reacties om een verklaring te geven, waarom sommige reacties wel en sommige reacties niet worden nagebootst door de electrochemische reacties. Het EC/MS systeem bootst cytochroom P450 gekatalyseerde reacties na, welke geïnitiëerd worden door en 1-elektron oxidatie, zoals bij voorbeld N–dealkylatie, S–oxidatie, P–oxidatie, alcohol oxidatie en dehydrogenatie. De P450 gekatalyseerde oxidaties die via abstractie van een waterstof molecuul geïnitiëerd zijn, zoals bij voorbeld O–dealkylatie en hydroxylatie van aromatische systemen zonder elektron donerende groepen, worden meestal niet geoxideerd bij de potentialen die mogelijk zijn in water en werden dus niet nagebootst door het EC/MS systeem.

In hoofdstuk 3 wordt een electrochemisch flow-through systeem geïntroduceerd (het EC-Fenton systeem) dat geschikt is om hydroxyl-radikalen te genereren welke met xenobiotica kunnen reageren. De oxidatieprodukten van deze reactie kunnen vervolgens on-line met LC/MS/MS gedetecteerd worden. Het systeem is gebaseerd op de Fenton reactie en is vooral gericht op het genereren van hydroxyl-radikalen, maar door kleine veranderingen in het systeem kunnen tevens een grote hoeveelheid andere radikalen gegenereerd worden. Omdat ijzer als Fe3+ aan het monster toegevoegd is, blijft het mengsel inaktief totdat het in de elektrochemische cel komt. Dit maakt het systeem bij uitstek geschikt voor on-line analyse van de produkten omdat het hele mengsel, inclusief het substraat, in een autosampler gezet kan worden. Het systeem is geschikt voor onderzoek aan nieuwe radicaal-vangers en antioxidanten. Omdat het hydroxyl-radicaal reageert met

151

π-systemen is het ook geschikt voor on-line productie en detectie van metabolieten die geoxideerd zijn aan dubbel-bindingen en aromatische systemen.

In hoofdstuk 4 wordt de reaktiviteit van het geprenyleerde flavonoid xanthohumol (XN) met hydroxyl-radicalen onderzocht in het EC-Fenton systeem. Negen van tien metabolieten, die in een experiment met ratten-lever plakken gevormd waren, kon ook met het EC-Fenton systeem geproduceerd worden. Hieruit kan worden geconcludeerd dat het EC-Fenton systeem gescikt is om de reaktiviteit van flavanoiden tegen radikalen te onderzoeken. In het geval van xanthohumol genereert het EC-Fenton systeem een reeks van oxidatie-produkten die aantonen dat de EC-Fenton reactie inderdaad lijkt op oxidatief metabolisme in een biologisch systeem.

In hoofdstuk 5 zijn de twee systemen, namelijk het EC/MS systeem en het EC-Fenton systeem, gebruikt voor de dopamine agonist, S(–)-(N-propyl-N-2-thienylethylamine)-5-hydroxytetralin (N–0923). Het oxidatief metabolisme, dat eerder werd waargenomen in experimenten met perfusie van ratte-lever, was gedeeltelijk nagebootst door beide metoden. De resultaten geven een indicatie dat het EC/MS systeem en het EC-Fenton systeem verschillende produkten geven en dat ze elkaar kunnen aanvullen.

In hoofdstuk 6 worden verschillende problemen behandeld die kunnen ontstaan als een electrochemische cel gekoppeld wordt aan electrospray massaspectrometrie. Verschillende oplossingen zijn beschreven voor elke situatie.

In hoofdstuk 7 wordt het metabolisme van de potentiële catecholamine pro-drugs PD217015, (–)-GMC6650 en (+)-GMC6650 bestudeerd met LC/MS/MS analyse van bloed-plasma en hersen-monsters van mannelijke Wistar ratten na orale administratie. De catechol-metaboliet van PD217015, 5,6-dihydroxy-DPAT, werd geïdentificeerd in de hersenen. Gebaseerd op interpretatie van de produkt ion spektra, zijn drie verdere metabolieten van PD217015 voorgesteld. De N-dealkyleerde metaboliet 6-propylamino-3,4,5,6,7,8-hexahydro-2H-naphtalene-1-one en de twee gehydroxyleerde metabolieten 2–hydroxy–PD217015 en 4–hydroxy–PD217015. Het is aannemelijk dat hydroxylatie op positie 2 is een intermediair product in de formatie van de actieve metaboliet 5,6-dihydroxy-DPAT.

Voor GMC6650, werd in de hersenen van de rat die (-)-GMC6650 had gekregen een metaboliet gedetecteerd die waarschijnlijk overeenstemt met de actieve catechol metaboliet 1–propyl-1,2,3,4,4a,5,10,10a-octahydro-benzo[g]quinoline-6,7-diol.

Data van MS/MS geeft vaak nuttige informatie over welk deel van een molecuul geoxideerd is, maar om een onbekende metaboliet volledig te karakteriseren is vaak meer informatie nodig. Beide hier beschreven electrochemische technieken kunnen gemodificeerd worden zodat een grotere hoevelheid produkt gevormd kan worden. Een electrochemisch oxidatieprodukt kan als een metaboliet geïdentificeerd worden door

152

vergelijking van de retentie tijd en het produkt ion spektrum met een metaboliet gevormd in een biologisch systeem. Het electrochemische systeem kan dan gebruikt worden om voldoende hoeveelheid stof te genereren voor het karakteriseren met IR en NMR.

Het gemak en de snelheid waarmee deze electrochemische systemen toegepast kunnen worden voor een groote aantal monsters, maken ze waardevolle hulpmiddelen in het ondezoek van metabolisme.

153

List of publications

U. Jurva; A. P. Bruins; H. V. Wikström, "In vitro mimicry of metabolic oxidation reactions by electrochemistry/mass spectrometry", Rapid Commun. Mass Spectrom., 2000, 14, 529-533.

U. Jurva; H.V. Wikström; A.P. Bruins, “Electrochemically assisted Fenton reaction: reaction of hydroxyl radicals with xenobiotics followed by on-line analysis with high-performance liquid chromatography/tandem mass spectrometry” Rapid Commun. Mass Spectrom., 2002, 16, 1934-1940. U. Jurva; H.V. Wikström; L. Weidolf; A.P. Bruins “Comparison between electrochemistry/mass spectrometry and cytochrome P450 catalyzed oxidation reactions” Rapid Commun. Mass Spectrom. 2003, 17, 800-810. R. de Kanter; M. H. de Jager; A. L. Draaisma; U. Jurva; P. Olinga; D. K. F. Meijer; G. M. M. Groothuis, "Drug metabolizing activity of human and rat liver, lung, kidney and intestine slices", Xenobiotica, 2002, 32, 349-362. H.P. Permentier; U. Jurva; B. Barroso; A.P. Bruins “Electrochemical oxidation and cleavage of peptides analyzed with on-line mass spectrometric detection”, Rapid Commun. Mass Spectrom. 2003, 17, 1585-1592.

154

Acknowledgements

First of all, thank you Håkan and Andries for four great years of science! Thank you Håkan for giving me this opportunity by calling me up that day in the autumn of 1997. Many thanks for your enthusiasm and for your never ending stream of ideas. Thank you Andries for teaching me everything I know about mass spectrometry. It’s a true privilege to have learned MS/MS interpretations from one of the very best. Thank you both for always being there when I needed you but also for giving me the freedom to do what I wanted to do. If Peter Stjärnlöf was still with us I would also thank him for recommending me and for his support when I made the big decision to go to Groningen. Thank you Eva for always making me feel welcome in your home (in particular during the ice hockey tournaments) and for all the nice dinners. Thank you Ala for teaching me how real Indonesian food is supposed to taste like.

Thank you Albert, Alessandro, Annie, Damon, Grielof, Hjalmar and Margot for all the good times we had in and outside the MS lab. To Annie and Margot, thanks for all the laughs and for taking care of me! Thank you Margot for letting me chop wood and use the chain saw. I felt almost like I was back at the farm preparing for the winter. Thank you Annie for a great game of football with you and your daughters. Too bad that I was too big for the trampoline though. Thank you Hjalmar for helping me look for my glasses at Storm! I wonder who’s wearing them now... Thank you Damon and Tina for all the good times. I hope that we can make that fishing trip someday!

To all the guys on the forth floor, Pieter, Erik, Bas-Jan, Sander, Evert, Marguérite, Sandrine, Nienke, Cor, Durk, Yi and Danyang, thank you for all the good times and for teaching me that Grolsch is a gift from the gods and that it’s best enjoyed together with your colleagues on Fridays after work in de Toeter!

Thomas, thanks for teaching me how to catch a paling and how to stay up all night and then go fishing in the morning. Yuki, thank you for all the good parties! Olga, thank you for trying to bring modern theatre into my life. I’m afraid that I’m too much of a cultural idiot to understand the beauty of it but I had a good time anyway.

Thank you Jonas for everything. Thank you Fred for coming up with the great idea to use the EC-Fenton system on

flavonoids and for all the help with chapter 4. Thank you Ruben for the liver slice incubations and for the discussions about cytochrome P450. I promise to submit the manuscript soon! Thank you Neal for all the useful comments. Jan, many thanks for all the in vivo experiments. Thank you Danyang for the very useful standards.

Thank you Janita, Janneke and Astrid for all the help you gave me when I first came to Groningen and had to deal with Dutch laws and regulations. Thank you Margriet

155

for helping me with all the practical things that had to be fixed for the promotie. I would have been completely lost without you!

Thank you Lasse Weidolf for believing in the EC/MS project from the start and for the many useful discussions about metabolism. I hope we can finish what we set out to do in Mölndal as well!

I would like to thank all the members of the reading committee, professor Nibbering, professor Castagnoli and professor Franke for taking the time and effort to read my thesis.

I would also like to thank my sister Sofia for being a joy in my life and in particular for helping her sloppy brother to clean his apartment when he left Groningen two years ago. I would probably still have been busy moving things from one place to another if you hadn’t been there.

Thanks to those of you that I have forgotten to mention for your forgiveness of my bad memory.

Jaha, det var det det... Det tog ett tag och har kostat många kvällar och helger men till slut blev det inte så pjåkigt ändå. Fast jag visste egentligen redan från början att det skulle fixa sig på ett eller annat sätt. Det brukar nämligen alltid göra det, fixa sig alltså. Livet blir som sagt så mycket lättare när man är säker på att saker och ting ska fixa sig. Man slipper bekymra sig över ödesdigra konsekvenser av ett eventuellt misslyckande och då blir det mycket lättare att klara av det man föresatt sig. Sedan är det bara att som vi Tidaholmare säger ”ge sig fan påt” så brukar saker och ting gå vägen. Sedan gör det givetvis livet lättare att lära sig att titta på resultatet av det man gjort och i sitt glada hjärta känna att ”det där blev bra” även vid de tillfällen då mer nogräknade själar kanske skulle betrakta det som halvbra och göra mer än vad som krävs. Den som inte är lite lat kommer sällan på några uppfinningar som förenklar tillvaron.

När farsan kunde prata brukade han alltid säga att ”det fixar jag” och var i det ögonblick han sa det fullständigt övertygad om att saker och ting skulle lösa sig även om han inte riktigt visste hur just då. Nästan alltid hade han rätt även om det ofta slutligen var morsan som såg till att det egentligen hände något.

Så slutligen vill jag tacka min gamle fader och min ömma moder för att ni lärde mig att saker och ting alltid fixar sig samt att det inte är så farligt om det mot förmodan skulle gå åt skogen i alla fall.

university of groningen electrochemistry on-line with mass spectrometry … · 2016-03-09 ·...

Documents