PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRC/DateDue.p65-p. 15

EXAMINATION OF THE MEMBRANE BINDING DOMAIN OF HUMAN
CYCLOOXYGENASE -2 ENZYME USING SITE-DIRECTED SPIN LABELING
AND ELECTRON PARAMAGNETIC RESONANCE SPECTROSCOPY

By

Zahra MirAfzali

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry
&
Department of Biochemistry and Molecular Biology

2003

ABSTRACT

EXAMINATION OF THE MEMBRANE BINDING DOMAIN OF HUMAN
CYCLOOXYGENASE —2 ENZYME USING SITE-DIRECTED SPIN LABELING
AND ELECTRON PARAMAGNETIC RESONANCE SPECTROSCOPY

By

Zahra MirAfzali

The widespread use of aspirin, ibuprofen, Celebrex®, Vioxx® and other non-

steroidal anti-inflammatory drugs (NSAIDs) has called for a deeper understanding of the
physiological systems in which they work. NSAIDs exerts their anti-inflammatory and
analgesic effects through the inhibition of cyclooxygenase —1 and —2 (COX-1 and —2).
The biological necessity for two COX isozymes is not fully understood as these enzymes
have nearly identical catalytic properties. Several evidences support that the COX
isozymes have different biological functions. The basis for the differential biological
activities is not known, but may relate to differences in their membrane binding domains,
which are poorly conserved.

Site-directed spin labeling (SDSL) is a powerful tool for determination of the
membrane protein structure and the topology of protein binding to the lipid bilayer. 25
mutant enzymes that contained single reactive cysteines substituted for amino acids
within the human COX-2 membrane-binding domain were constructed and spin labeled.
The accessibility of each protein-bound spin label with freely diffusing oxygen as a non-

polar paramagnetic reagent, and NiEDDA as a polar paramagnetic reagent, was measured

using power saturation EPR spectroscopy. Our results show the accessibility parameter
(II) for both oxygen and NiEDDA, has a periodicity of 3.6 consistent with an (It-helical
configuration. Furthermore, I'onygen and nNiEDDA both have the same period and phase.
Our results indicate that both polar and non-polar paramagnetic relaxers do not access the
amino acids located in the membrane-binding domain due to the very low values of

TIM),gen and l'INiEDDA. This indicates that the membrane binding domain of hCOX-2 is

located in the interfacial region of the membrane and is sandwiched between the main
body of the 144 KDa COX dimer and the lipid bilayer. We have identified three amino
acids that participate in anchoring the protein to the lipid bilayer. Human COX-2 is the
first peripheral membrane protein mapped using site-directed spin labeling.

Another aspect of this dissertation was to study the conformational changes in the
mouth of the cyclooxygenase channel. The purpose of this research was to determine
whether the COX substrate and inhibitor binding sites are flexible, e. g., can be opened
and closed. Ten cysteine double mutants located in the helices that form the opening of
the cyclooxygenase active site and membrane binding of human COX-2 were
constructed. The dipolar broadening in the presence and absence of NSAIDs, arachidonic
acid and heme were compared in detergent. The same experiments were also performed
in liposomes. The interspin distance measurements were done using Fourier
convolution/deconvolution technique.

Our results indicated that binding of heme, arachidonic acid, and NSAIDs had no
significant effect on the conformation of hCOX-2. However, a slight conformational
change was observed in the structure of the enzyme in the presence and absence of the _

lipid bilayer.

DEDICATION

In the Loving Memory of My Father

Professor
Abolghassem MirAfzali

iv

ACKNOWLEDGEMENTS

First of all, I would like to thank the person who was the mastermind behind this
work and made all of this research possible: Professor David DeWitt, also known as “The
Big Boss.”

I had a very rough time as a graduate student because my research was based on
the work that was supposedly done by the previous graduate student in our lab. It took me
almost a year to realize that the previous work was “science fraud” and I found myself in
a situation to start the research from scratch. Professor DeWitt helped me tremendously
during all those tough and dark days of my life and encouraged me constantly. He not
only taught me biochemistry but also taught me many life lessons. He taught me to
tolerate difficult people and situations and to concentrate on my goals. He is the best
example of a dedicated scientist who spends all his time including weekends, holidays,
Christmas and first day of New Year in his lab. He supervised my research EVERY DAY
including the weekends. I was always welcome to ask any questions of him and he
guided me step by step through this work. I also thank him for supporting me financially
to attend various scientific meetings. I have really great memories from the activities that
we had outside his lab, such as group lunches, the dinner invitation, the pizza party and
the “end of summer” party at his house, the bowling game and our trip to Tahoe City in
California.

I would like to thank my chemistry research advisor Professor John McCracken
for giving me the opportunity to be a dual major student and work in Professor DeWitt’s

lab in the Biochemistry Department. Professor McCracken is not only a great scientist but

also a great lecturer. Before joining Professor McCracken’s lab I was his teaching
assistant in a Physical Chemistry course and I learned a lot from attending his classes. I
also enjoyed all the amusing stories that he told us in our group meetings. Professor
McCracken accepted me as his student when my former advisor Professor Babcock was
seriously ill and gave me the opportunity to start a new project. He also supported me as a
research assistant all these years so I could do research without being concerned about
teaching duties. When I look back I realize that joining Professor McCracken’s lab was
the best and the most professional decision that I have ever made in my scientific life.

I would like to Professors Lynmarie A. Posey, Ned Jackson and Honggao Yan for
serving in my defense committee.

I would like to thank the members of three cyclooxygenase research laboratories:

In Professor DeWitt’s lab I would like to thank Jeffrey Leipprandt for growing
bioreactor after bioreactor of cells for me and for tolerating all the frustrating moments in
our lab. This research was absolutely impossible without Jeff.

I thank Christi Hemming for being so entertaining and full of life and making me
laugh all the time. Christi is one of the most pleasant people that I have ever met in my
life.

I thank my dear friend Prashanti Franklin for her friendship and the outstanding
undergraduate student Matthew Sekedat for helping me with the liposome incorporation
studies of Cox enzyme.

In Professor Smith’s lab I would like to thank Professor William Smith for his
encouragement and his faith in me. I also thank Jill Rieke, Dr. Cindy DeLong, Jiayan Liu,

Dr. Gilad Rimon (Ben Gurion University at Negev, Israel), Dr. Sigal Fleisher (Ben

vi

Gurion University at Negev, Israel), Dr. Byron Wingerd and Dr Dimitry Kuklev for our
scientific discussions and also for their friendship. I am still in shock about their lab
relocation to the other school down the road, also known as University of Michigan.

In Professor Garavito’s lab I would like to thank Professor Mike Garavito for his
scientific discussions about lipid bilayer interface and X-ray crystallography and also for
introducing the Nebulizer to me which made my life much easier.

I am in debt to Dr. Anne Mulichak who is an outstanding scientist for discussing
the crystal structure of COX-2 with me and making the membrane binding domain
pictures of COX—2. I would also like to thank Amy Scharmen, Karen Poster-Verrill and
Nicole Webb for their friendship.

I would like to thank all the scientists in the MSU Center for Advanced
Microscopy.

Thank to Dr. Alicia Pastor Lecha for helping me with the transmission electron
microscopy and patiently teaching me this technique. The electron microscopy work was
impossible without her. I will never forget all the fun that we had in the dark room and in
the microscope room and also the wonderful dinner party at Alicia’s new house.

I would like to thank Robert Pcionek for his very genuine friendship and for
teaching me how to use the complicated electron microscope instrument. I can’t thank
him enough.

Another big thanks goes to Ewa Danielewicz for the scanning electron
microscopy work on liposomes. I really enjoyed working with Ewa and listening to her

amazing stories about her sons.

vii

I would like to acknowledge Dr. Shirley Owens for her impressive confocal laser
scanning microscopy work on fluorescence liposomes and also for her kindness and
friendship.

I would like to thank Kermit Johnson who is an exceptional technician for
teaching me to use the loop-gap resonator. Kermit is my hero.

I thank Dr. Andrew Ichimura for teaching me to use the EPR spectrometer at the
beginning of my work in the EPR lab. I really enjoyed my late night talks with Andrew.

I also would like to thank Dr. Joseph Leykam in Macromolecular Structure
Sequencing and Synthesis Facility who taught me various techniques including peptide
synthesis. I enjoyed talking to Joe and listening to his amusing stories.

I would like to thank the members of the fifth floor labs in Biochemistry building
for their wonderful friendship and day-to-day interaction.

I would like to thank my various friends in the Chemistry Department especially
Montserrat Rabago-Smith and Chrysoula Vasileiou for their friendship. I will always
remember the night that Montserrat made me incredibly drunk with several Tequila shots.

I would like to thank my family especially my mother Dr. Guity Mehr MirAfzali
for her encouragement during all these years and my love Michael for making my life so
bright especially during the last several months that I was writing my dissertation. I really
enjoyed all our “geeky” scientific discussions about magnetic resonance and C
programming. Your love and encouragement saved my life. I love you.v v

I thank MSU Center for Biological Modeling for the graduate student award for

Spring 2001 and also for the travel award.

viii

At last I would like to thank the National Institutes of Health for supporting this

“VERY EXPENSIVE” project.

ix

TABLE OF CONTENTS

List of Tables XIII
List of Figures XIV
List of Abbreviations XXI

Chapter 1- Literature Review
Prostaglandins, Prostaglandin Endoperoxide H2 Synthases Isozymes and

Nonsteroidal Anti-Inflammatory Drugs 1
1.1 . Introduction 2
1.2. Prostaglandins 2
1.2.1. Functions of Prostaglandins 3
1.2.2. Prostanoid Biosynthesis 5
1.3. Cyclooxygenases—l and —2 7
1.4. Nonsteroidal Anti-Inflammatory Drugs (NSAIDs) 14
1.4.1. Historical background 14
1.4.2. The Problem of Pain 15
1.4.3. The Cost of Pain 16
1.4.4. Inhibition of Cyclooxygenase with NSAIDs 17
1.4.5. COX-2 Selective Inhibition 18
1.5. Statement of Problem 21
1.6. References 25
Chapter 2

Incorporation of Cyclooxygenase Isozymes into Preformed Liposomes and
Electron Microscopy Studies of Liposomes and Proteoliposomes 29
2.1. Abstract 30
2.2. Reconstitution of Membrane Proteins into Lipid Vesicles 31
2.2.1. Factors Affecting Incorporation 31
2.3. Incorporation Techniques 33
2.3.1. Organic Solvent-Mediated Reconstitution 34
2.3.2. Reconstitution by Mechanical Means 35
2.3.3. Reconstitution by Using Detergent 35
2.3.4. Direct Incorporation of Proteins into Preformed Liposomes 36
2.4. Electron Microscopy 36
2.4.1. Transmission Electron Microscopy 36
2.4.2. Anatomy of Transmission Electron Microscopy 38
2.4.3. Staining 39

2.4.4. Scanning Electron Microscopy 4O

 

 

2.5. Materials and Methods 41

2.6. Results and Discussions 51
2.7. References 62
Chapter 3

Electron Paramagnetic Resonance Spectroscopy and Site-Directed Spin
Labeling 64
3.1. Abstract 65
3.2. Structural Studies of Biomolecules 66
3.3. Principles of Electron Paramagnetic Resonance Spectroscopy 70
3.3.1. Zeeman Interaction 70
3.3.2. The Resonance Condition 72
3.4. The Electron Paramagnetic Resonance Experiment 74
3.4.1. The EPR Spectrometer 74
3.4.2. The Loop-Gap Resonator 76
3.4.3. Rectangular Box Cavity (TE102) 78
3.5. Nitroxide Free Radicals 78
3.5.1. Nitroxide Radical Spectrum 83
3.5.2. The Spectrum of the Non-Oriented Systems (Solid Phase) 85
3.5.3. Spectral Anisotropy 89
3.5.4. The Effect of Nitroxide Motion on EPR Lineshapes 91
3.6. Accessibility Determination Using Power Saturation 94
3.6.1. Continuous Wave Saturation EPR (CWS-EPR) 95
3.6.2. Site-Directed Spin Labeling 97
3.6.3. The Theory of Site-Directed Spin Labeling 100
3.7. Materials and Methods 103
3.8. Results 106
3.9. Discussions 132
3.10. References 137
Chapter 4

Double Site Directed Spin Labeling Studies 140
4.1. Abstract 141
4.2. Dipole—Dipole Interaction 142
4.3. Determination of Electron- Electron Distance from Dipolar Interaction between
Two Spin Labels Measured by EPR in Immobilized State 145
4.3.1. Dipolar Splitting is Significant Compared with Linewidths of Spectra for
Corresponding Monoradical 145
4.3.1.1. Analysis of Spectral Lineshapes by Computer Simulation 147
4.3.1.2. Lineshape Deconvolution 147
4.3.1.3. Relative Intensity of the Half-Field Transition 147
4.3.1.4. Lineshapes Distortion Determined by the Ratio of peak Heights 148

xi

 

 

 

 

4.3.2. Dipolar Interaction is Small Compared with Linewidths of Spectra for

Corresponding Monoradical 148
4.3.2.1. 3-Pulse ELDOR (Electron-Electron Double Resonance) 149
4.3.2.2. 2+1 Sequence 149
4.3.2.3. 4-Pulse DEER (Double Electron-Electron Resonance) 149
4.3.2.4. Double-Quantum Coherence 150
4.4. Fourier Convolution Deconvolution of the Dipolar Coupling in the Spin-Labeled
EPR Spectra 150
4.4. 1. Theoretical Background 150
4.4.1.1 . Convolution 150
4.4.1.2. Fourier Transform 151
4.4.1.3. Convolution Theorem 154
4.4.2. Fourier Convolution Deconvolution of the EPR Dipolar Coupling 155
4.4.3. Spectral Analysis 157
4.4.4. Monoradical Impurities 162
4.5. Materials and Methods 163
4.6. Results 165
4.7. Discussion 181
4.8. References 184
Chapter 5

Conclusion and Future Direction 187
Appendix A

The Data Analysis Procedures for the Experiments Described in Chapter
Three 191
Appendix B

Application of MATLAB® Scripts for Measuring the Distance between
Two Nitroxide Spin Labels 202
Appendix C

EPR Spectra of Ten hCOX-2 Double Mutants 214

xii

 

 

LIST OF TABLES

Table 1.1. Generic and brand names of common NSAIDs 19

Table 2.1. The effect of oleic acid concentration on incorporation of hCOX-2 into
preformed liposomes 53

Table 2.2. The effect of oleic acid concentration on incorporation of oCOX-l into
preformed liposomes 5 3

Table 3.1. Relative cyclooxygenase activities of the single substituted hCOX-2

cysteine mutants 108
Table 3.2. Collision parameters for the membrane binding domain mutants 115
Table 4.1. Summary of methods for determining electron-electron distances 146

Table 4.2. EPR determined inter-nitroxide for mutants in Helix C and I98C. 183

xiii

LIST OF FIGURES

Figure 1.1. Nomenclature of prostaglandins 4
Figure 1.2. Biosynthetic pathway for prostanoid synthesis 6

Figure 1.3. A ribbon representation of the oCOX-l monomer with AA bound in the

cyclooxygenase channel 10
Figure 1.4. Mechanistic sequence for converting AA to PGGz. 11
Figure 1.5. Side view of the COX-2 membrane binding domain helices 13
Figure 1.6. Isoform- selective inhibitor binding 20
Figure 2.1. 50

A) Stack-plot of EPR spectra of H75C spin labeled proteoliposmes. Each EPR
spectrum was acquired using a different modulation amplitude

B) The plot of intensity of each spectra in Figure 2.1 .A versus the modulation
amplitude .

Figure 2.2. Representative negative stain electron micrograph of multilamellar lipid
membranes. 56
Figure 2.3. 58
A) Representative negative stain electron micrograph of proteoliposme
B) Representative negative stain electron micrograph of plain liposome
Figure 2.4. 60
A) Scanning electron micrograph of a liposome

B) Low magnification scanning electron micrograph of liposomes

xiv

 

 

Figure 2.5. Representative negative stain electron micrograph of gold labeled proteins

incorporated into a liposome 61
Figure 3.1. The EPR spectrometer 75
Figure 3.2. The two loop one gap resonator 77
Figure 3.3. The structure of TE102 rectangular box cavity 79

.Figure 3.4. The methanethiosulfonate spin label (MTSSL) reacts exclusively with the
sulfliydryl group on the introduced cysteine residue. 80
Figure 3.5. Two resonance structures of nitroxide 82
Figure 3.6. The energy level diagram for a system with one electron and one l4N

atom 86
Figure 3.7. Principle axes of the g6 matrix of the nitroxide spin label 87
Figure 3.8. Schematic diagram of the rigid limit EPR powder pattern of MTSSL 88

Figure 3.9. Orientation dependence of the EPR spectrum of a nitroxide. Simulated

spectra of a monocrystal sample with the external magnetic field B directed along

the principle axes g and A tensors. 90
Figure 3.10. Motion dependence of the EPR spectrum of a nitroxide 92
Figure 3.11. A typical power saturation curve 98

Figure 3.12. Model helices and idealized behavior of the accessibility parameter 105
Figure 3.13. CW-EPR spectra of COX-2 helix A spin label mutants in 30% w/v
sucrose 118
Figure 3.14. CW-EPR spectra of COX-2 helix B spin label mutants in 30% w/v

sucrose 120

XV

Figure 3.15. CW-EPR spectra of COX-2 helix C spin label mutants in 30% w/v
sucrose 122
Figure 3.16. CW-EPR spectra of Cox-2 helix A spin label mutants incorporated into
liposome 124
Figure 3.17. CW—EPR spectra of Cox-2 helix B spin label mutants incorporated into
liposomes 126
Figure 3.18. CW-EPR spectra of Cox-2 helix C spin label mutants incorporated into
liposomes 128
Figure 3.19. 129
A) Plot of relative accessibility of helix A side chains
B) Plot of mobility of helix A side chains
Figure 3.20. 130
A) Plot of relative accessibility of helix B side chains
B) Plot of mobility of helix B side chains
Figure 3.21. 131
A) Plot of relative accessibility of helix C side chains
B) Plot of mobility of helix C side chains
Figure 3.22. The crystal structure of the membrane-binding domain of hCOX-2 135

Figure 3.23. The structure of the membrane-binding domain of hCOX-2 that was

proposed by site-directed spin labeling 136
Figure 4.1. Dipolar coupling between two spin labels 143
Figure 4.2. Convolution of f(x) ® g(x) 152
Figure 4.3. Fourier transformation of a signal 153

xvi

 

Figure 4.4. A flow chart of the algorithm for distance analysis from dipolar

spin-spin interaction 159
Figure 4.5. The resulting spectra at each step of analysis outlined in Figure 4.4. 161
Figure 4.6. Analysis of Monoradical impurities 164
Figure 4.7. The X-ray structure of murine holo-COX-2 (5COX) with two cysteine
substitution mutations at residues F84 and 198 170
Figure 4.8. The X-ray structure of murine holo-COX-Z (5COX) with two cysteine
substitution mutations at residues W85 and I98 171
Figure 4.9. The X-ray structure of murine holo-COX-2 (5COX) with two cysteine
substitution mutations at residues N86 and 198 172
Figure 4.10. The X-ray structure of murine holo-COX-2 (5COX) with two cysteine
substitution mutations at residues V87 and 198 173
Figure 4.11. The X-ray structure of murine holo-COX-2 (5COX) with two cysteine
substitution mutations at residues V88 and 198 174
Figure 4.12. The X-ray structure of murine holo-COX-2 (5COX) with two cysteine
substitution mutations at residues T73 and 198 175
Figure 4.13. The X-ray structure of murine holo-COX-Z (5COX) with two cysteine
substitution mutations at residues V74 and 198 176
Figure 4.14. The X-ray structure of murine holo-COX—2 (5COX) with two cysteine
substitution mutations at residues H75 and 198 177
Figure 4.15. The X-ray structure of murine holo-COX-2 (5COX) with two cysteine

substitution mutations at residues Y76 and 198 178

xvii

 

 

 

Figure 4.16. The X-ray structure of murine holo-COX-2 (5COX) with two cysteine
substitution mutations at residues I77 and 198 179
Figure 4.17. Overlay of absorbance EPR spectra of H75C in liposome (red) and
H75C in Tween 20 (green) 180

Figure 1A. Step one: To open the spectrum with Bruker’s SimFonia sofiware 192

Figure 2A. Step two: To export the spectrum to WinEPR software 193
Figure 3A. Step three: To filter the noise in the spectrum 194
Figure 4A. Step four: To measure the peak to peak amplitude 195

Figure 5A. The stack plot of EPR spectra. Each spectrum was acquired in a

different microwave power 196
Figure 6A. Step five: Peak to peak amplitude m1=0 of a first derivative EPR signal is a
function of the square root of the incident microwave power 197
Figure 7A. Step six: To plot the peak-to-peak amplitude of spectra as a function of the
square root of the incident microwave power. 198
Figure 8A. Step seven: Curve fitting the curve to the following equation:
Y=(P1*x)/((l+((2"(1/P2)-1)*((x"2)/P3)))"(P2)) 199
Figure 9A. Step eight: To obtain the unknown values 200
Figure 10A. A new software was recently written in C programming language in
collaboration with Professor Michael Romalis in Princeton University. 201
Figure 18. Overlaid absorbance spectra of singly and doubly labeled proteins 209
Figure 2B. Fourier transform of I'I*(m), the doubly labeled spectrum. Only the

ends of this curve contain useful information. A close-up of this is shown in the

next figure. 210

xviii

 

 

 

Figure 3B. Zoom of Fourier transform of doubly labeled spectrum 211
Figure 4B. Fourier transform using calculated tau components 212

Figure SB. Spectra resulting from FCD analysis of dipolar coupling 213
Figure 1C. Absorbance EPR spectra of T73C/I98C double mutant in detergent

(Tween 20) and liposome and in the presence of Co-PPIX, arachidonic acid and
various COX inhibitors 216-220

Figure 2C. Absorbance EPR spectra of V74C/I98C double mutant in detergent

(Tween 20) and liposome and in the presence of Co-PPIX, arachidonic acid and
various COX inhibitors. 222-226

Figure 3C. Absorbance EPR spectra of H75C/I98C double mutant in detergent

(Tween 20) and liposome and in the presence of Co-PPIX, arachidonic acid and
various COX inhibitors. 228-232

Figure 4C. Absorbance EPR spectra of Y76C/I98C double mutant in detergent

(Tween 20) and liposome and in the presence of Co—PPIX, arachidonic acid and
various COX inhibitors. 234-238

Figure SC. Absorbance EPR spectra of I77C/I98C double mutant in detergent

(Tween 20) and liposome and in the presence of Co-PPIX, arachidonic acid and

various COX inhibitors. 240-244
Figure 6C. Absorbance EPR spectra of F84C/I98C double mutant in detergent

(Tween 20) and liposome and in the presence of Co-PPIX, arachidonic acid and

various COX inhibitors. 246-250

xix

 

 

Figure 7C. Absorbance EPR spectra of W85C/I98C double mutant in detergent
(Tween 20) and liposome and in the presence of Co-PPIX, arachidonic acid and
various COX inhibitors. 252-256
Figure 8C. Absorbance EPR spectra of N86C/I98C double mutant in detergent
(Tween 20) and liposome and in the presence of Co-PPIX, arachidonic acid and
various COX inhibitors. 258-262
Figure 9C. Absorbance EPR spectra of V87C/I98C double mutant in detergent
(Tween 20) and liposome and in the presence of Co-PPIX, arachidonic acid and
various COX inhibitors. 264-268
Figure 10C. Absorbance EPR spectra of V88C/I98C double mutant in detergent
(Tween 20) and liposome and in the presence of Co-PPIX, arachidonic acid and

various COX inhibitors. 270-274

 

 

CLSM

COX

CW

DEER

DPPH

DOPC

DOPS

DSDSL

EDTA

EGF

ELDOR

EPR

ER

FCD

FFT

FORTRAN

FT

MBD

MOI

MTSSL

MLV

LIST OF ABBREVIATIONS

Arachidonic Acid

Confocal laser scanning microscopy
Cyclooxygenase

Continuous wave

Double Electron-Electron Resonance

0t, a’-diphenyl-B- picrylhydrazyl
1,2-Dioleoyl-sn-Glycero-3-Phosphocholine
l,2-Dioleoyl-sn-Glycero-3-[Phospho-L-Serine]
Double Site Directed Spin Labeling
Ethylene Diarnine Tetraacetic Acid
Epidermal Growth Factor
ELectron-electron DOuble Resonance
Electron Paramagnetic Resonance
Endoplasmic Reticulum

Fourier Convolution Deconvolution

Fast Fourier Transformation

F ORmula TRANslation (A programming language)
Fourier Transformation

Membrane Binding Domain

Multiplicity of Infection
MethaneThioSulfonate Spin Label

MultiLamellar Vesicle

xxi

NiAA
NiEDDA
NMR
NSAID
PBS
PDB
PG
PGGz
PGHz
PGHS
POX
SDSL
SEM
Sf-2 1
TEM

ULV

Nickel Acetyl Acetonate

Nickel Ethylen Diarnine Diacetic Acid
Nuclear Magnetic Resonance

Non Steroidal Anti Inflammatory Drugs
Phosphate Buffer Saline

Protein Data Bank

Prostaglandin

Prostaglandin G2

Prostaglandin H2

Prostaglandin Endoperoxide H2 Synthase
Peroxidase

Site-Directed Spin Labeling

Scanning Electron Microscopy
Spodoptera F rugiperda

Transmission Electron Microscopy

UniLamellar Vesicle

xxii

Chapter 1- Literature Review

Prostaglandins, Prostaglandin Endoperoxide H2 Synthase Isozymes and
Nonsteroidal Anti-Inflammatory Drugs

 

1.1. Introduction

Pain is the most significant and costly problem confronting human beings and has
a profound impact on the work productivity and the quality of life.

Prostaglandins are a group of physiologically active substances that play a major
role in pain, inflammation and chronic inflammatory diseases, such as asthma and
rheumatoid arthritis. Cyclooxygenase (COX) isozymes are membrane proteins
responsible for the synthesis of prostaglandins. Inhibition of COX with nonsteroidal anti-
inflammatory drugs (NSAIDs) acutely reduces inflammation and pain.

In the following review, prostaglandins and their functions in body, the structure
and the catalytic mechanism of cyclooxygenase and also different classes of nonsteroidal

anti-inflammatory drugs will be discussed.

1.2. Prostaglandins

Prostaglandins are a group of physiologically active substances having diverse
hormone-like effects in animals. Prostaglandins act in a manner similar to that of
hormones, by stimulating target cells into action. However, they differ from hormones in
that they act locally, near their site of synthesis, and they are metabolized very rapidly. In
terms of chemical structure, prostaglandins are 20—carbon fatty acid derivatives
containing a 5-carbon ring (Figure 1.1). They were discovered in human semen in 1935
by Swedish physiologist Ulf Von Euler (1), who named them thinking they were secreted

by the prostate gland. Prostaglandins are now known to be of widespread occurrence in

 

 

 

animal tissues, where they are formed from polyunsaturated fatty acids and are rapidly
metabolized (2).

Most prostaglandins are synthesized from arachidonic acid (20:4 AS‘S'H‘M). These
are called “series 2” products, because most have two double bonds. However, the triene
fatty acid 20:3 A3,11.14 can also be used; the products have one fewer double bond than the
arachidonic acid derivatives and are called the “series 1” products. Two (1)6 fatty acids,
20:3 Am"14 and arachidonic acid (20:4 As‘s‘n’”), are substrates for most prostaglandin
biosynthesis, producing the “series 1” and “series 2” products respectively. In addition,
the 20:5 A5,8.ll.l4,l7 fatty acid, an (1)3 fatty acid, can also be used for prostaglandin
biosynthesis. Humans can synthesize (09 fatty acids, such as oleic acid and its 20:3 A5’8‘”
derivative. However, this is ordinarily a minor pathway, and the 20:3 AS'B‘” cannot be

used to make functional prostaglandins (2).

1.2.1. Functions of Prostaglandins

Prostaglandins (PG) play critical roles in normal physiological processes. Platelet—
derived TXA2 is an important mediator of platelet aggregation and thus hemostasis.
During periods of stress, PGs of the E and I series are important regulators of renal blood
flow. PGs likewise are important in modulating many aspects of reproductive biology
including ovulation, fertilization, fetal development, and parturition. The opposing
actions of different PG classes help to maintain bronchial tone. The processes of bone
formation and resorption are also subject to regulation by PGs. Macrophage

differentiation is also modulated by PGs. PGs are vital to the maintenance of mucosal

 

 

A O
"NW/(3021'!
Ov Type 1 o
W
(‘11
"NW 0
-' — 0021-1
x Type 2 -
6H
H9
.t"'l\=//\/\002H
Type 3
g _ O
OH
31
H5
H9
H6

Figure 1.1. Nomenclature of prostaglandins

PGA

PGB

PGC

PGD

PGE

PGF

A) The number of carbon-carbon double bonds determines the type of prostaglandin.

B) Letters A to F are designated as the substituents on the cyclopentane ring.

f

 

integrity in the gastrointestinal tract and also play a role in the regulation of motility and
secretion. PGs also affect immune function in a number of ways (3).

PGs have been implicated in a wide variety of disease processes. The huge annual
market for nonsteroidal anti-inflammatory drugs (NSAIDs) which inhibit COX activity is
a testament to the role of PGs in acute inflammation and chronic inflammatory diseases,
such as asthma, rheumatoid/osteo arthritis, and inflammatory bowel disease (IBD).
Epidemiological and animal studies indicate that inhibition of PG synthesis is efficacious
in the prevention of coronary artery thrombosis, Alzheimer’s disease, gastrointestinal
cancer and breast cancer. The properties of PGs that contribute to disease progression
include their thrombotic activity, ability to modulate cellular apoptosis and other cell

cycle parameters, angiogenic activity and other functions yet to be identified (3).

1.2.2. Prostanoid Biosynthesis

Prostaglandin formation is a three-step process, which includes:
a) Arachidonic acid release from membrane phospholipids.
b) Conversion of arachidonic acid to PGHz by cyclooxygenase enzyme.
c) Conversion of PGH2 to specific prostaglandins, followed by their release from the
cell.

The overall process in shown in Figure 1.2.

202

 

 

  
     
 

    
  

 

 

 

— — COO- \ A O‘m- .-“\___/\/\COO
— — PGHS o--- / ’
AA cyclooxygenase H 60H
PGGZ
PGHS
peroxidase 2 AH
2A'+ H20
Hg PGD Thromboxane
. "“WCOO- A sanhase (1)... INWCOO' Sanhase L
W f 0....
o H b” H bH
PGDZ PGH2
I Fr’nifase Prostacyclin
so Synthase
PGF COO-
Synthase
/
° 11 ‘
JWCOO‘ HQ . '9
/ ‘ WCOO /
H 5 q .
O HO H OH OH
PGE PGI
2 PGFZQ 2

Figure 1.2. Biosynthetic pathway for Prostanoid synthesis. Taken from (4). Copyright

permission was obtained.

 

Initially, arachidonic acid is cleaved from membrane phospholipids following
hormonal activation of one of the phospholipases A2s (PLA2). Two classes of PLA2 have
been shown to be involved in the mobilization of arachidonic acid from phospholipids,
cytoplasmic PLA2 (cPLA2) and several secretory PLA2 (sPLA2) (5). The cPLA2 is
believed to be the major mediator of agonist-induced arachidonic acid released in cells
because of its specificity in cleaving. Arachidonic acid is hydrolyzed from the sn-2
position of membrane phospholipids. Several sPLA2 have also been shown to mobilize
arachidonic acid, and subsequently increase prostanoid production, although specificity
for arachidonate cleavage from phospholipids has not been shown as clearly as with

cPLA2 (6). Release of free arachidonic acid is increased upon activation of cPLA2.

1.3. Prostaglandin Endoperoxide H2 Synthases—l and —2

Cyclooxygenases are N-glycosylated membrane proteins responsible for the
synthesis of prostaglandins involved in important physiological processes, such as
smooth muscle contraction, inflammation, parturition and platelet aggregation (7). Two
isoforrns have been identified: COX-1 and —2, which are also referred to as Prostaglandin
endoperoxide H2 synthases —l and —-2 (PGHS-1 and —2) (7). COX-1 and -2 are of
particular interest because they are the major targets of nonsteroidal anti-inflammatory
drugs (NSAIDs) including aspirin, flurbiprofen, naproxin and new COX-2 inhibitors such
as Celebrex®, Vioxx®, Bextra® and Arcoxia®. Inhibition of the COXs with NSAle
acutely reduces inflammation, pain, and fever. Long term use of these drugs reduces

fatal thrombotic events, as well as the development of colon cancer and Alzheimer’s

 

* v-

 

 

disease. The isozymes are peripheral membrane proteins that catalyze the synthesis of
prostaglandins in the arachidonic acid cascade.

The primary structures of both isoforrns share a 60% amino acid sequence identity
within the species. The major differences in sequence between COX-1 and —2 are in the
N-terrninal and C-terrninal regions of the isozymes. In N—terminal, COX-1 has a sequence
of hydrophobic residues within the signal peptide that are lacking completely in COX-2.
Conversely, COX-2 contains a highly conserved l8-amino acid cassette at the C-
terrninus, whose function has not yet been determined.

The theoretical molecular masses of COX-1 and —2 are 66 kDa and 67 kDa
respectively. COX-1 has a mass of 72 kDa and COX-2 has both masses of 74 and 72
kDa. The 72 kDa masses can be attributed to the addition of carbohydrate groups at three
N-glycosylation sites in both COX-l and-2 (Asn 68, Asn 144, Asn 410). Glycosylation of
all three sites is essential for enzyme activity in both isozymes (8). The 74 kDa mass of
COX-2 is attributed to an inefficient N-glycosylation at a unique fourth site, Asn580,
whose glycosylation has no effect on the enzymatic activity of COX-2 (8). Otherwise, the
enzymes share a remarkable similarity in sequence within a species, with an epidermal
growth factor (EGF) domain, three N-glycosylation sites, and conservation of all
catalytically important residues in both COX-1 and COX-2.

The three dimensional structures of ovine 567-amino acid COX-1 complexed with
the NSAID flurbiprofen (9), with 2-bromoacetoxybenzoic acid (a potent aspirin
analogue) (10), as well as with iodinated indomethacin and suprofen (11), have been
determined by X-ray crystallography. X-ray structures of unligated murine COX-2 and

COX-2 complexed with flurbiprofen, indomethacin and a selective inhibitor (SC-558)

 

 

have also been obtained (12). The structure of human COX-2, determined in the presence
of a selective inhibitor is similar to that of COX-1 (13). The crystal structures revealed
the presence of three folding units: an EGF homology domain, a catalytic domain
containing both cyclooxygenase and peroxidase catalytic sites and a membrane—binding
domain (Figure 1.3). From solubilization and crystallization studies of COXs, it has been
shown that the isozymes function as homodimers, forming an interface between EGF
homology domain of each monomer.

The crystal structure of COX-1 complexed with flurbiprofen shows the drug
localized in a long, hydrophobic channel within the cyclooxygenase active site. The
opening of this site begins with the membrane—binding domain and extends into the
catalytic domain, with one active site in each monomer. Presumably, the arachidonic acid
substrate would be cleaved from the membrane lipids and could directly access the
hydrophobic cyclooxygenase active site channel. The crystal structure of Co+3-heme
ovine COX-1 complexed with arachidonic acid was solved (14). This structure revealed
specific substrate-enzyme interactions within the cyclooxygenase active site. In the
cyclooxygenase reaction, arachidonic acid is positioned in the hydrophobic pocket in an
extended L-shaped conformation with the carboxylate group positioned at the base of the
channel by virtue of an electrostatic interaction with arginine 120 (I5, 16). C13 of the
arachidonic acid is positioned near tyrosine 385 at the top of the channel (Figure 1.4).

The porphyrin radical of the oxidized heme group, generated by reduction of a
hydroperoxide substrate in the peroxidase reaction, abstracts a hydrogen from tyrosine
385 creating a tyrosyl radical (I7, 18). Next this tyrosyl radical abstracts the 13 pro-S

hydrogen from arachidonate creating an arachidonyl radical, which then reacts with a

 

    
    
   
 
  

Peroxidase
A ctive Site

 

Catalytic
Domain

 

 

 

Domain

Membrane
Binding
Domain

Figure 1.3. A ribbon representation of the oCOX-l monomer with AA bound in the
cyclooxygenase channel. The EGF domain, MBD, and catalytic domain are shown in
green, orange, and blue, respectively. (Inset) The native dimer with the twofold axis
running vertical. Taken from (14). Copyright permission was obtained.

Images in this dissertation are presented in color.

 

 

 

 

 

 

 

Figure 1.4. Mechanistic sequence for converting AA to PGG2.Taken from (14).
Copyright permission was obtained. Images in this dissertation are presented in color.

 

molecule of molecular oxygen at C11. The resulting peroxyl radical next attracts C9 of
the fatty acid resulting in an oxygen bridge between C9 and C11, followed by the
formation of a cyclopentane ring connecting C8 and C12 (Figure 1.4). Addition of
molecular oxygen at C15 and the transfer of the radical back to tyrosine 385 results in the
formation of the hydroperoxide PGG2 which exits the cyclooxygenase active site, leaving
it primed for another round of catalysis. PGG2 next diffuses to the peroxidase active site,
which is a shallow channel containing a heme group and on the opposite face of the
enzyme, to undergo the two-electron reduction in the peroxidase active site converting
the lS-hydroperoxy group to the 15-alkyl group of PGH2.

X-ray crystallographic studies have demonstrated that neither isozyme contains
amino acid sequences of sufficient length and hydrophobicity to function as classical
transmembrane domains. Instead, COX-1 and —2 are thought to associate with cellular
membranes through a novel, monotopic membrane binding domain that transverses a
single leaflet of the lipid bilayer (9). The X-ray crystallography shows that projecting
from the globular catalytic domain in each monomer are four amphipathic alpha helices
(labeled A, B, C and D) (Figure 1.5). Hydrophobic and aromatic residues protrude from
these helices and away from the hydrophobic surface of the catalytic domain to create a
hydrophobic patch. Chimeric constructs of green fluorescent protein linked to the N-
terminal sequences containing the MBD domain from either COX-1 or COX-2 are
targeted to the proper membrane location in cells, providing further evidence that these
sequences are responsible for membrane localization of the isozymes (19). These helices
also form the opening to the cyclooxygenase active site and were predicted to facilitate

interaction with the lipid bilayer. The substrate of COXs, arachidonic acid, is released

12

 

 

{ K64

 

Figure 1.5. Side view of the COX-2 membrane binding domain helices
This picture was made by Dr. Anne Mulichak.
Images in this dissertation are presented in color.

13

 

from the bilayer and presumably travels through the core of the helices A-D to the
cyclooxygenase active site. The membrane-binding domain of COX-1 and COX-2 has

the least sequence similarity to each other, sharing only 38% sequence similarity (7).

1.4. Nonsteroidal Anti-Inflammatory Drugs (NSAIDs)

1.4.1. Historical background

The use of medicinal substances to relieve pain and fever dates back to ancient
Egypt, where a decoction of dried leaves of myrtle applied to the back and abdomen of
patients, was used to expel pains from the womb. Later, in Greece, the bitter extracts
from the bark of the poplar tree were used in patients with eye diseases. Extracts of the
willow bark were chewed to relieve the pain of childbirth and to reduce fever. These
beneficial effects remained largely unknown until the first published report of willow
bark was done in England by Reverend Edward Stone in 1763. Later on the active
component of willow bark was identified as salicin, which is metabolized to salicylate.

Salicylic acid was synthesized in Germany in 1860 by Kolbe and Lautemann,
leading to an extended use as an external antiseptic, an antipyretic, and an antirheumatic
drug which caused dyspepsia and tasted bitter when given orally. Salicylic acid has a
tendency to cause upset stomach in frequent users, and it was this side effect that inspired
Felix Hoffman, a German chemist employed by the Bayer Corporation whose arthritic
father used salicylic acid, to synthesize aspirin from the original willow bark compound

in 1897. In 1899, Heinrich Dreser named the compound "Aspirin", the "a" referring to the

acetyl grouping and the "spirin" recalling the botanical genus spiraea, from which
salicylates could be extracted.

During the beginning of the 20th century, aspirin was recognized as an
antipyretic, anti—inflammatory, and an analgesic drug. Later on, the development of other
drugs with similar clinical effects, like Phenylbutazone in 1949 and Indomethacin in
1963, inspired pharrnacologists and biochemists to search for a common mode of action.
Little was known except that they produced an anti—inflammatory effect both

quantitatively and qualitatively different from the more potent glucocorticoids (20-23).

1.4.2. The Problem of Pain

Pain is one of the most significant and costly healthcare problems confronting
society today. Pain impacts not only the lives of patients and their families, but also
places great strains upon providers as well as the healthcare systems that bear the cost of
treatment. Indeed, pain is the single most common reason patients seek medical care,
accounting for one-half of all physician office visits. Yet, nearly two-thirds (64%) of pain
sufferers report that they only seek medical care when they can endure the pain no longer.
A recent survey reports that four of ten adults (42%) experience pain daily, while nine out
of ten (89%) experience pain every month. Of the latter, more than 26 million individuals
(15%) suffer severe debilitating pain. Forty-three percent of adults - a projected 83

million individuals - report that pain frequently affects their participation in activities. In

15

 

 

the workplace, a significant portion of workdays lost (25%) is due to pain. In
hospitalized patients, pain is associated with longer recovery times and poorer patient
outcomes. Pain, therefore, has a profound impact on the quality of life, work productivity,

and the cost of healthcare (24).

1.4.3. The Cost of Pain

In US the annual cost for healthcare and lost productivity due to the pain is more
than 100 billion dollars. The annual cost attributed to back pain, migraines, and arthritis
is about 40 billion dollars and the annual cost of the physician visits due to the pain is 40
million dollars. The prescription analgesic market is currently estimated at over $9 billion
in the US. and $26 billion globally. In the US. alone, pain accounts for over 40 million
clinical visits per year and 4 billion lost workdays ($65 billion in productivity) (24, 25).
Despite the large sales numbers, it is widely agreed by practitioners and patients that the
market is poorly served by existing therapeutic approaches due mostly to their poor
efficacy or debilitating side effects. Because of this unmet need, it is projected that sales
of drugs related to pain management will grow by as much as 20% each year for the
foreseeable future. One need only look to the short timeframe required for two recently

launched pain drugs, OxyContin® (Purdue Pharma) and Celebrex® (Pfizer), to surpass

the $1 billion annual sales level, 36 months and 10 months, respectively, to appreciate the

market’s appetite for new products (26).

16

 

1.4.4. Inhibition of Cyclooxygenase with NSAID

The principal pharmacological effect of NSAIDs is due to their ability to inhibit
prostaglandin synthesis by blocking the cyclooxygenase activity of both COX-1 and
COX-2. Their therapeutic effectiveness as analgesics, antipyretics, anti-inflammatories
and anti-thrombogenics is due to their inhibition of prostanoid synthesis, which also
accounts for their side-effect profile.

Traditional NSAIDs can be grouped into four classes based on their modes of inhibition
of COX:

Class 1: Simple, competitive, reversible inhibitors that compete with arachidonic
acid for binding to the COX active site.

Included in this class are: Ibuprofen (Advil®, Mortin®); Piroxican (Feldene®); Sulindac
sulfide (Clinoril®); Flufenamate; Mefenamic acid (Ponstel®) and Naproxen
(Naprosyn®).

Class II: Competitive, time-dependent, reversible inhibitors that bind to the COX
active site in a first phase, to form reversible enzyme-inhibitor complexes.

Included in this class are: Indomethacin (Indocin®); Flurbiprofen (Ansaid®),
Meclofenamic acid (Meclomen®) and Diclofenac (Voltarol®).

Class III: Competitive, time-dependent, irreversible inhibitors that form an
enzyme inhibitor complex after a covalent conformational change in the protein.
Included in this class is: Aspirin.
The acetylation of COX-1 by aspirin is an irreversible process that inhibits COX activity

but not the peroxidase activity.

17

 

Class IV: Time-dependent COX-2 inhibitors. These drugs have larger side groups
to occupy extra side pocket in COX-2.
Included in this class: Celecoxib (Celebrex®); Rofecoxib (Vioxx®); Valdecoxib
(Bextra®), Etoricoxib (Arcoxia®) and SC58125.

Table 1.] lists the generic and brand names of common NSAIDs.

1.4.5. COX-2 Selective Inhibition

The original NSAle have long been known for their lethal side-effects. As
reported in the American Journal of Medicine (27), over 100,000 people are hospitalized
each year and at least 16,500 die each year from gastrointestinal (GD bleeding in United
States. These figures are considered conservative, and they are only figures for NSAIDs
being used to treat arthritis. It also doesn’t include the number of people who die of other
complications from NSAID use such as congestive heart failure. A true analysis of the
number of deaths world wide from NSAIDs would be overwhelming.

The discovery that COX-2 is involved in inflammation, fever, and pain has led to
the development of a new class of inhibitors that acts selectively on this isozyme. COX-2
has a larger active site and a side pocket into which the new specific bulkier inhibitors fit.
Crystal structures of COX-2/inhibitor complexes (13) and (12) reveal that selective COX-
2 inhibitors utilize the side pocket created by IS23V and I434V substitutions in COX-1

(Figure 1.6).

Table 1.1. Generic and brand names of common NSAIDs

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

GENETIC NAME BRAND NAME

Aceclofenac Preservex

Acematacin Emflex

Azapropazone Rheumox

Benoxaprofen Oraflex, Operen

Bromfenac Duract

Celecoxib Celebrex

Diclofenac Voltaren, Rhumalgan, Diclomax, Motifene 75

Diflunisal Dolobid

Etodolac Lodine

Etoricoxib Arcoxia

Fenbufen Lederfen

Fenoprofen Naflon, Fenopron, Progesic

Floctafenine Idarac

Flurbiprofen Ocufen, Froben, Ansaid

Ibuprofen Advil, Haltran, Trendar, Nuprin, Motrin, Aches-N-Pain, Brufen,
Nurofen, Actiprofen, ACT-3, Rafen

Indomethacin Indocid, Indocin, Indoflex, Indameth, Rheumacin LA, Flexin,
Nisiad

Ketoprofen Orudis, Oruvail, Ketovail, Ketocid, Alrheumat

Ketorolac Toradol, Acular

 

Meclofenamate

Meclomen

 

Mefenamic Acid

Ponstan, Mefic, Ponstel, Meflam, Dysman, Opustan

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Meloxicam Mobic

N abumetone Relafen, Relifex

Naproxen Naprosyn, Anaprox, Naprelan, Synflex, Nycopren, Arthroxen,
Laraflex

Naproxen sodium Aleve

Oxaprozin Daypro

Oxyphenbutazone Oxalid, Tandearil

Phenylbutazone Butazolidin, Azolid, Butacote, Butatab

Piroxicam Feldene, Larapam, Piroflam, Flamatrol

Rofecoxib Vioxx

Sulindac Clinoril

Suprofen Suprol

Tenoxicam Tilcotil, Mobiflex

Tiaprofenic acid Surgam

Tolfenamic Clotem

Tolmetin Tolectin

Valdecoxib Bextra

 

l9

 

Flurbiprofen

 

Figure 1.6. Isoform-selective inhibitor binding.

The COX-1 and COX-2 active sites are shown superimposed (COX-1, yellow; COX-2
pink). Two inhibitors are seen; flurbiprofen (orange), a nonselective inhibitor, and SC-
558 (blue), a COX-2 -sclective inhibitor, NSAIDs achieve COX inhibition by occupying
the upper portion of the active site channel, preventing the fatty acid substrate fi'om
gaining access to the active site tyrosine seen at the upper right. The COX-2 selective
inhibitor projects lefiward into a side pocket that is not exploited by the nonselective
inhibitor. Taken from (28).

Images in this dissertation are presented in color. Copyright permission was obtained.

These “COX-2” inhibitors are able to selectively inhibit COX-2, alleviating pain,
fever and inflammation without causing the gastrointestinal problems associated with

non-selective NSAIDs (22).

1.5. Statement of Problem

The main goal of this dissertation is to characterize the structural interactions of
the peripheral membrane protein hCOX-2 with lipid bilayers. COX isozymes do not
contain transmembrane sequences. Instead, both isozymes contain four short contiguous
amphipathic (Jr-helices that constitute a membrane binding domain thought to interact
with a single leaflet of the lipid bilayer (9). However, this model has never been
confirmed experimentally. Amphipathic or-helices are common structural motifs that
previously have been grouped into 7 classes according to their function (29). Examples of
well characterized roles for these motifs include; cell membrane disruption (mellittin and
other transmembrane domains), transportation of lipids (apolipoproteins), membrane
insertion (integral proteins containing transmembrane domains), and solvent/protein
interfaces (most globular proteins). In COX isozymes amphipathic helices are though to
be used for stable association to the membrane.

The COX isozymes have been studied extensively because of their essential and
regulatory role in prostaglandin synthesis and also because they are the sites of action of
nonsteroidal anti-inflammatory drugs (NSAIDs) such as aspirin and ibuprofen. However
very little is known about how the membrane binding domains of COX isozymes interact
with the lipid bilayers. The crystal structures of these enzymes can be used to predict how

the COXs may interact with membrane, but more direct methods must be used to confirm

21

or refute these predictions. Furthermore, it is not possible to determine the COX
isozymes orientation in a membrane, or to determine lipid binding sites in the COXs from
a crystal structure, as few lipid or detergent molecules are visible using X-ray analysis.

The experiments outlined in chapter 3 of this dissertation were designed to answer
these important basic questions. The results of the experiments in this chapter might have
general applicability to a new class of monotopically-associated peripheral membrane
proteins. Sequalene hepone cyclase is another monotopic membrane protein, the only
other structurally characterized protein of this new class.

The use of site-directed spin labeling (SDSL) to examine the protein -membrane
interaction is a technique that has been pioneered by Wayne L. Hubbell (30, 31). In this
technique, site directed mutagenesis is used to incorporate single cysteine substitutions at
various positions into a series of mutant proteins. These cysteines are then covalently
modified with a spin-labeled reagent. The differential collision rates of nitroxides side
chains in these proteins with polar and non-polar paramagnetic reagents can be used to
determine which of these residues are solvent-exposed or membrane-imbedded. These
collision rates are measured using power saturation electron paramagnetic resonance
(EPR) spectroscopy. One of the best examples of the use of SDSL has been in the
identification of membrane-embedded and solvent accessible amino acids in
bacteriorhodopsin. While the crystal structure of bacteriorhodopsin, like the COXs, was
known beforehand, these experiments were necessary to determine the orientation of this
structure within the membrane. SDSL combined with EPR spectroscopy has provided
essential information about the membrane interactions of several other proteins including;

diphtheria toxin (32-34), colicin E1 (35, 36) lactose perrnease (37-39), KcsA potassium

22

channel (40), aA-Crystallin (41-43), phospholipase A2 (44, 45), T4 lysozyme (46), tear
lipocalin (47, 48) and FepA (49-51).

In the research described in chapter three, 25 mutant enzymes that contained
single reactive cysteines substituted for amino acids within the hCOX-2 membrane-
binding domain were constructed and spin labeled. The accessibility of each spin label
with freely diffusing oxygen, a non-polar paramagnetic reagent, and NiEDDA, a polar
paramagnetic reagent was measured using power saturation EPR spectroscopy.

These studies require the incorporation of purified COX enzyme into artificial cell
membranes or liposomes. An increasing number of techniques for preparation of
proteoliposomes have been published, because the approach to the incorporation of a new
protein into liposomes is still an empirical one. It has been impossible until now to
predict which conditions are necessary for a particular protein to be incorporated into the
liposomes.

In chapter 2 of this dissertation the incorporation of hCOX-2 into preformed
liposome is discussed. The goal of this research is to optimize the method for
reconstitution of active hCOX-2 in liposomes and increase the protein content of these
proteoliposomes. The shape and size of the liposomes and proteoliposomes have been
extensively studied by electron microscopy. To directly detect the hCOX-2 protein
molecules, gold labeled anti bodies were attached to the proteins and the gold particles
were directly observed by transmission electron microscopy.

In chapter 4 of this dissertation, the conformational changes in the mouth of the
cyclooxygenase channel are studied by double site directed spin labeling technique. The

purpose of this research was to determine whether the COX substrate and inhibitor

23

 

 

binding sites are flexible, e.g., can be opened and closed, and whether binding of
substrate and/or inhibitor locks these enzymes in a static position. One unexplained
feature of the enzymology of the COX isozymes is the mechanism whereby certain
nonsteroidal anti-inflammatory drugs (NSAIDs) inhibit these enzymes in a time-
dependent manner. Upon incubation, NSAIDs such as indomethacin and flurbiprofen
form very tight non-covalent, but essentially irreversible complexes with COX isozymes.
Presumably, these complexes result from time-dependent conformational changes in the
enzymes upon inhibitor binding. NSAID-induced conformational changes in doubly
labeled COXs would be observable as shifts in the spin coupled nitroxide lineshape. The
mechanism for time-dependent inhibition of the COXs is of considerable practical
importance because newly developed NSAIDs that are selective for COX-2 owe their
selectivity to the fact that they are time-dependent inhibitors of COX-2 and only simple
competitive inhibitors of COX-1 (52). Thus, being able to measure conformational
changes in the enzyme structure may increase our understanding of the dynamics of
enzyme catalysis as well as provide information about mechanism of drug inhibition.

Ten cysteine double mutants located in the helices that form the opening of the
cyclooxygenase active site and membrane-binding domain of hCOX-2 were constructed.
The dipolar broadening in the presence and absence of NSAIDs, in the presence and
absence of heme and in the presence and absence of arachidonic acid were compared in
detergent. The same experiments were also performed in liposomes. The distances
between the two spin labels were also measured in various samples of hCOX-2. The
interspin distance measurements were done using the Fourier convolution/deconvolution

technique (53).

24

 

1.6. References

10.

ll.

l2.

l3.

14.

von Euler, U. S. (1936) Journal of Physiology 88, 213-34.

Rocca, B., and FitzGerald, G. A. (2002) International Immunopharmacology 2,
603-630.

Funk, C. D. (2001) Science 294, 1871-1875.

Goodwin, D. C., Landino, L. M., and Mamett, L. J. (1999) Faseb Journal 13,
1121-1136.

Clark, J. D., Schievella, A. R., Nalefski, E. A., and Lin, L. L. (1995) Journal of
Lipid Mediators and Cell Signalling 12, 83-117.

Murakami, M., Shimbara, S., Kambe, T., Kuwata, H., Winstead, M. V.,
Tischfield, J. A., and Kudo, I. (1998) Journal of Biological Chemistry 273,
14411-14423.

Smith, W. L., DeWitt, D. L., and Garavito, R. M. (2000) Annual Review of
Biochemistry 69, 145-182.

Otto, J. C., Dewitt, D. L., and Smith, W. L. (1993) Journal of Biological
Chemistry 268, 18234-18242.

Picot, D., Loll, P. J., and Garavito, R. M. (1994) Nature 367, 243-249.

Loll, P. J., Picot, D., and Garavito, R. M. (1995) Nature Structural Biology 2,
637-643.

Loll, P. J., Picot, D., Ekabo, O., and Garavito, R. M. (1996) Biochemistry 35,
7330-7340.

Kurumbail, R. G., Stevens, A. M., Gierse, J. K., McDonald, J. J., Stegeman, R.
A., Pak, J. Y., Gildehaus, D., Miyashiro, J. M., Penning, T. D., Seibert, K.,
Isakson, P. C., and Stallings, W. C. (1996) Nature 384, 644-648.

Luong, C., Miller, A., Barnett, J., Chow, J., Ramesha, C., and Browner, M. F.
(1996) Nature Structural Biology 3, 927-933.

Malkowski, M. G., Ginell, S. L., Smith, W. L., and Garavito, R. M. (2000)
Science 289, 1933-1937.

25

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

Bhattacharyya, D. K., Lecomte, M., Rieke, C. J ., Garavito, R. M., and Smith, W.
L. (1996) Journal of Biological Chemistry 271, 2179-2184.

Rieke, C. J., Mulichak, A. M., Garavito, R. M., and Smith, W. L. (1999) Journal
of Biological Chemistry 274, 17109-171 l4.

Hsi, L. C., Hoganson, C. W., Babcock, G. T., Garavito, R. M., and Smith, W. L.
( 1995) Biochemical and Biophysical Research Communications 207, 652-660.

Tsai, A. L., Hsi, L. C., Kulmacz, R. J., Palmer, G., and Smith, W. L. (1994)
Journal of Biological Chemistry 269, 5085-5091.

Li, Y., Smith, T., Grabski, S., and DeWitt, D. L. (1998) Journal of Biological
Chemistry 273, 29830-29837.

Donahue, K. E. S. (1996) The Watermark 19, 80-83.

Gartrell, D. (1997), UCLA multimedia website.

Liebeskind, J. C., and Meldrum, Marcia L. (1997) in Proceedings of the Eighth
World Congress on Pain (Progress in Pain Research and Management) (T. S.
Jensen, J. A. T., and Z. Wiesenfeld-Hallin, Ed.), IASP Press, Seattle.

Meldrum, M. L. (1999) American Pain Society Bulletin 9.
http://wwwampainsocogl.

Chase, M. (1998) Wall Street Journal.

http://www.ims-global.coml.

Singh, G. (1998) The American Journal of Medicine July 27, 3 IS.

FitzGerald, G. A., and Loll, P. (2001) Journal of Clinical Investigation 107, 1335-
1337.

Segrest, J. P., Deloof, H., Dohlman, J. G., Brouillette, C. G., and Anantharamaiah,
G. M. (1990) Proteins-Structure Function and Genetics 8, 103-117.

Hubbell, W. L., and Altenbach, C. (1994) Current Opinion in Structural Biology
4, 566-573.

Hubbell, W. L., Gross, A., Langen, R., and Lietzow, M. A. (1998) Current
Opinion in Structural Biology 8, 649-656.

26

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

47.

Oh, K. J., Zhan, H. J., Cui, C., Altenbach, C., Hubbell, W. L., and Collier, R. J.
(1999) Biochemistry 38, 10336-10343.

Oh, K. J., Zhan, H. J., Cui, C., Hideg, K., Collier, R. J., and Hubbell, W. L. (1996)
Science 273, 810-812.

Zhan, H. J., Oh, K. J., Shin, Y. K., Hubbell, W. L., and Collier, R. J. (1995)
Biochemistry 34, 4856-4863.

Salwinski, L., and Hubbell, W. L. (1999) Protein Science 8, 562-572.

Salwinski, L., Levinthal, C., Levinthal, F., and Hubbell, W. L. (1993) Biophysical
Journal 64, A183-A183.

Sun, J. 2., Voss, J., Hubbell, W. L., and Kaback, H. R. (1999) Biochemistry 38,
3100-3105.

Voss, J., Hubbell, W. L., HemandezBorrell, J., and Kaback, H. R. (1997)
Biochemistry 36, 15055-15061 .

He, M. M., Voss, J., Hubbell, W. L., and Kaback, H. R. (1997) Biochemistry 36,
13682- 13687.

Gross, A., Columbus, L., Hideg, K., Altenbach, C., and Hubbell, W. L. (1999)
Biochemistry 38, 10324- 10335.

Koteiche, H. A., and McHaourab, H. S. (2002) F ebs Letters 519, 16-22.

Koteiche, H. A., and McHaourab, H. S. (1999) Journal of Molecular Biology 294,
561-577.

Berengian, A. R., Parfenova, M., and McHaourab, H. S. (1999) Journal of
Biological Chemistry 274, 6305-6314.

Frazier, A. A., Roller, C. R., Havelka, J. J., Hinderliter, A., and Cafiso, D. S.
(2003) Biochemistry 42, 96-105.

Frazier, A. A., Wisner, M. A., Malmberg, N. J., Victor, K. G., Fanucci, G. E.,
Nalefski, E. A., Falke, J. J ., and Cafiso, D. S. (2002) Biochemistry 41, 6282-6292.

McHaourab, H. S., Kalai, T., Hideg, K., and Hubbell, W. L. (1999) Biochemistry
38, 2947-2955.

Gasymov, O. K., Abduragimov, A. R., Yusifov, T. N., and Glasgow, B. J. (1997)
Biochemical and Biophysical Research Communications 239, 191-196.

27

 

 

48.

49.

50.

51.

52.

53.

Glasgow, B. J ., Gasymov, O. K., Abduragimov, A. R., Yusifov, T. N ., Altenbach,
C., and Hubbell, W. L. (1999) Biochemistry 38, 13707-13716.

Klug, C. S., Eaton, S. S., Eaton, G. R., and Feix, J. B. (1998) Biochemistry 37,
9016-9023.

Klug, C. S., and Feix, J. B. (1998) Protein Science 7, 1469-1476.
Klug, C. S., Su, W. Y., and Feix, J. B. (1997) Biochemistry 36, 13027-13033.
DeWitt, D. L. (1999) Molecular Pharmacology 55, 625-631.

Rabenstein, M. D., and Shin, Y. K. (1995) Proceedings of the National Academy
of Sciences of the United States of America 92, 8239-8243.

28

Chapter 2
Incorporation of Cyclooxygenase Isozymes into Preformed Liposomes and Electron

Microscopy Studies of Liposomes and Proteoliposomes

29

2.1. Abstract

The site-directed spin labeling technique requires the incorporation of the purified
COX enzyme into the liposomes. Therefore the goal of the research described in this
chapter is to optimize the method for reconstitution of active hCOX-2 into liposomes and
increase the protein content of these proteoliposomes.

In this research unilamellar vesicle liposome (ULV) is used as the artificial cell
membrane model and direct incorporation technique is used to incorporate hCOX-2 into
the liposomes. The size and shape of the liposomes including ULV and multilamellar
vesicles (MLV) and proteoliposomes were extensively studied by negative staining
transmission electron microscopy. In negative staining technique, the background and
surroundings are stained to increase the contrast in the sample. In order to directly detect
the hCOX-2 protein molecules, gold-labeled antibodies were attached to the proteins and
the gold particles were directly observed by transmission electron microscopy. Scanning
electron microscopy technique was also used to observe the 3—dimensional images of
liposomes and proteoliposomes.

In section 2.2 of this chapter factors effecting incorporation of an isolated protein
into the liposomes are discussed. In section 2.3 an overview of various incorporation
techniques will be given. Section 2.4 briefly describes various electron microscopy
techniques. Section 2.5 provides all the protocols that were used for the experiments in

this chapter and finally in section 2.6 the results will be discussed.

30

 

 

2.2. Reconstitution of Membrane Proteins into Lipid Vesicles

Six years after the description of liposome formation by Alec Bangham in 1965
(1), the reconstitution of a membrane protein into liposomes was first reported by
Kagawa and Racker in 1971 (2). Since then a large number of membrane proteins have
been isolated and re-incorporated into the artificial cell membranes, where they often
exhibit their normal biological activity.

An increasing number of techniques for preparation of such proteoliposomes has
been published (3, 4), and the reason for this is that the approach to the incorporation of a
new protein into the liposomes is still an empirical one. It has been impossible until now
to predict which conditions are necessary for a particular protein to be incorporated into
the liposomes. So one just has to try to find which method works best. According to E.
Racker’s comment in his book “ Reconstitutions of Transporters, Receptors and
Pathological State” (page 38) “ if nothing works, invent a new reconstitution

procedure. .

2.2.1. Factors Affecting Incorporation

One of the most important factors is the incorporation of an isolated protein into
the liposome is the physical state of the lipids. Natural membranes are normally fluid, i. e.
in the liquid crystalline phase above the phase transition temperature Tm. Many proteins
are known to be incorporated more readily into lipid bilayers above Tm and are even

sometimes found to be excluded from gel-state lipids (5). Other proteins are incorporated

31

into membranes only over a narrow range around Tm e.g. glucagon (6). A third group of
proteins interacts with phospholipids over a wide range of temperature exhibiting
preferential interaction with lipids in the gel phase (7).

The molecular basis for interaction of proteins with the lipid membranes near the
phase transition or with gel-state lipids is thought to be due to point or line defects
facilitating protein-lipid interactions as well as protein-protein interactions along such
defects. The spontaneous incorporation of membrane proteins into vesicles by freeze
thawing probably relies on the conditions that induce packing defects. The importance of
the physical state for the incorporation of proteins into liposomes becomes obvious from
the observation that M13 virus coat protein shows a maximum incorporation near the
phase transition. The protein assembles into the exterior liposome surface well below the
phase transition temperature Tm above this temperature there is almost no incorporation.

In general three types of lipid-protein interactions are distinguished:

1) Electrostatic interaction
2) A mixture of electrostatic and hydrophobic interaction

3) Hydrophobic interaction

The second important factor in the incorporation of an isolated protein into the
liposome is the composition of the lipids chosen for the incorporation of a protein. As an
example, charged lipids often have a marked effect on the incorporation. The use of
charged lipids certainly reinforces electrostatic interactions, but this is not the only effect.

It has been shown that negatively charged lipids (phosphatidic acid, phosphatidylinositol,

32

cardiolipin) also improve the incorporation of the acidic proteins e.g. cytochrome oxidase
(8).

The ionic strength of the buffer is the third important factor in the incorporation of
the proteins into the lipid vesicles. The ionic strength of the buffer has an influence on
lipid-protein interactions, especially when electrostatic interactions are important. Cations
such as (MgZI, Ca2+) enhance incorporation at certain concentrations, higher
concentrations of these cations often being inhibitory (9). Also the pH of the buffer can
have a high influence on the incorporation.

Other factors that may influence the incorporation are the type and size of the
liposomes. Some proteins e.g. cytochrome b5 favor small vesicles (20 nm) over large
vesicles (100 nm) by a factor of about 200 probably due to the looser packing of the polar
head groups of phospholipids in small vesicles (10). On the other hand, cytochrome c
oxidase incorporates readily into multilamellar liposomes, but less or not at all into small

unilamellar liposomes.

2.3. Incorporation Techniques

There are essentially four presently known mechanisms for incorporating, i. e., and

reconstituting, proteins into liposome:

1) Methods involving the use of an organic solvent.
2) Methods involving the use of the mechanical means.
3) Methods involving the use of detergents.

4) Direct incorporation of the protein into the preformed liposomes.

33

These techniques will be described briefly in the following sections.

2.3.1. Organic Solvent-Mediated Reconstitution

Organic solvents have been used widely to prepare large liposomes in procedures
including ethanol injection, ether infusion and reverse-phase evaporation. However, the
usefulness of these techniques is limited because exposure of membrane proteins to the
organic solvents, which often denature them.

The only suitable method reported to date for organic solvent-mediated
reconstitution of membrane proteins is the reverse phase evaporation. Szoka and
Papahadjopoulos (11) developed a technique for the incorporation of membrane proteins
into large unilamellar liposomes after extraction into a hydrocarbon solvent together with
phospholipids. The membrane protein was added to a suspension of lipids in a suitable
buffer then a solvent e. g. hexane, pentane, diisopropyl ether, and diethyl ether was added
and the mixture was sonicated for a few minutes under argon. After the removal of the
organic solvent in a rotary evaporator, the large unilamellar liposomes of about 1 um are
formed. Critical aspects of this technique are the selection of the organic solvent as well

as the volume ratio of aqueous to non-aqueous phase (12).

34

2.3.2. Reconstitution by Mechanical Means

Second method uses mechanical means to produce large and small unilamellar
vesicles (ULV) from multilamellar vesicles (MLV) by swelling of the dry phospholipid
films in excess buffer. Such mechanical means include sonication of MLVs and forcing
multilamellar lipid vesicles through a French press. Sonication and freeze-thawing of a
mixed suspension of lipids and isolated proteins has been widely used in earlier stages of
membrane protein reconstitution to demonstrate the function of these purified proteins for

which detergent dialysis was insufficient (13).

2.3.3. Reconstitution by Using Detergents

Of the several different methods used to remove detergent from a mixture of
detergent, phospholipid and protein, the dialysis procedure has proved highly successful.
In such a method, the proteins and phospholipids are cosolubilized in a detergent to form
micelles. The detergent is then removed, resulting in the spontaneous formation of bilayer
vesicles with the protein incorporated therein. The detergent is incorporated into
liposome as well the protein and thus, these methods require removal of the detergent by
methods such as dialysis, gel exclusion chromatography or absorption on hydrophobic
resins. The methods that use detergent are very slow because detergent removal must be
as complete as possible. Also a phase change that take place during this process slows
detergent removal even further. Another disadvantage is that one cannot control the

orientation of protein incorporated into the liposomes using the detergent methods (14).

35

2.3.4. Direct Incorporation of Proteins into Preformed Liposomes

The forth process involves the direct incorporation of proteins into preformed
liposomes. One of the key features for successful incorporation of the delipidated
proteins into preformed lipid bilayers appear to be the state of organization of the bilayer.
Bilayer conducive to spontaneous incorporation of large membrane proteins are achieved
by incorporating ‘impurities’, such as fatty acids, lysophospholipids, mixtures of
structurally different phospholipids, cholesterol, detergent and membrane bound proteins.
The putative effect of impurities is the formation of ‘organizational defects’ that act as

sites for fusion of vesicles, not only with other vesicles, but also with proteins (15) .

2.4. Electron Microscopy

2.4.1. Transmission Electron Microscopy

The 1986 Nobel Prize in Physics was awarded jointly to Ernst Rusko, Gerd
Binnig, and Heinrich Rohrer. Rusko was recognized for his work on electron optics and
the original design of the transmission electron microscope in the 19305. Binnig and
Rohrer were acknowledged for their design of the scanning tunneling microscope a half
century later in the 19805. In presenting the award, the Nobel Prize committee noted the
significance of the electron microscopes in all areas of science, calling them “ one of the

most important inventions of this century (l6).”

36

The transmission electron microscope (TEM) was developed after the effective
role of wavelength on theoretical resolution was apparent. Green light, which is used for
light microscope, has a wavelength of 500 nm and therefore has a theoretical resolution
of about 200 nm. As you move down the electromagnetic spectrum to shorter
wavelengths, you will pass blue and violet into the range of the ultraviolet (<400 nm). An
ultraviolet microscope with a theoretical resolution of 50 nm could be used. But
ultraviolet light is absorbed by glass, so the lenses would have to be made of quartz. This
would make the microscope extremely expensive and fragile for only a moderate increase
in resolution. X-ray microscopes would have a vastly superior increase in resolution, but
x-rays can not easily be refracted to form an image. Electron waves offered the best
alternative. The electron being a charged particle could be easily refracted in a magnetic
field. Also, it could be accelerated by an electrical potential. The stronger the potential
the faster the electron will move, and according to the De Broglie relationship the shorter
the wavelength therefore the better resolution. In fact a typical electron microscope at an
accelerating voltage of 75000 V would have a wavelength less than 5 picometer. This
makes the theoretical resolution about 100,000 times better than that of light, well worth
undertaking the engineering effort. Unfortunately, this theoretical resolution has never
come even close to being attained. The basic draw back is that magnetic fields can not be
manipulated, shaped and grouped the way an optical engineer does with glass lenses.

Electron microscopy is used in a variety of ways in biology, medicine, and the
material sciences. Examples in the biological sciences include diagnosis of human,
animal, and plant diseases; development of disease pathology; study of morphological

and developmental aspects of organisms, tissues, and cells; identification of pollen,

37

 

viruses, bacteriophage, and diatoms; visualization of subcellular components and
structures such as DNA; subcellular localization of elements, enzymes, and proteins; in
situ hybridization of gene products; and confirmation of biological, physiological, and

biochemical data (17).

2.4.2. Anatomy of Transmission Electron Microscopy

The ray of electrons is produced by pin-shaped cathode heated up with current.
The electrons are vacuumed up by a high voltage at the anode. The acceleration voltage is
between 50000 and 150000 V. The higher it is the shorter the electron wavelength and
the higher the resolution. The accelerated ray of electrons passes a drill-hole at the
bottom of the anode. Its following was in analogous to that of a ray of light in a light
microscope. The lens systems consist of electronic coils generating an electromagnetic
field. The condenser first focuses the ray. It then passes through the object, where it is
partially deflected. The degree of deflection is dependent on the electron density of the
object. The greater the mass of the atoms, the greater is the degree of deflection.
Biological objects have only weak contrast since they consist mainly of atoms with low
atomic numbers (C, H, N, and 0). Consequently it is necessary to treat the preparation
with special contrast enhancing chemicals (heavy atoms) to get at least some contrast.
Additionally they should not be thicker than 100 nm. After the electron beam passes
through the object, the scattered electrons are collected by an objective electromagnet
lens to form an image. This image is subsequently enlarged by an additional lens system

call projective. The formed image is made visible on fluorescent screen or is documented

38

on photographic material. Photos taken with electron microscope are always black and
white. Color itself is a function of visible light, and no visible light is used to generate

images in an electron microscope (I 7).

2.4.3. Staining

In light microscopy the selective absorption of colored dyes by biological
structures enables the detail of a thin section to be seen more clearly. In electron
microscopy the equivalent technique uses the absorption of heavy metal atoms to produce
differential electron scattering and hence increased contrast in samples. Organic materials
consist mainly of elements of low atomic numbers, e.g., hydrogen (Z=1), carbon (Z=6),
nitrogen (Z=7), oxygen (Z=8), phosphorus (Z=15), sulfur (2:16), which have a low
electron-scattering power. Hence the contrast in an image of untreated material is low.
Staining by soaking in solutions containing tungsten (Z=74), osmium (Z=76), lead
(2:82) or uranium (Z=92) results in sites of high electron-scattering power, and
consequently images with more clearly visible specimen detail.

If the specimen is particulate, e.g., viruses, macromolecules or cellular fragments,
there is an alternative staining technique, negative staining. In positive staining the heavy
metal is absorbed by the tissue or particle itself, whereas in negative staining it is the
surroundings and background which are stained, leaving the particle itself unstained and
clearly visible against them. This is a very good technique to reveal fine surface
structures on particles. Negative staining transmission electron microscopy was used on

liposomes and proteoliposomes in this chapter (18).

39

 

 

2.4.4. Scanning Electron Microscopy

The Scanning Electron Microscope (SEM) is one of the most versatile and widely
used tools of modern science as it allows the study of both morphology and composition

of biological and physical materials.

By scanning an electron probe across a specimen, 3-D high-resolution images of
the morphology or topography of a specimen, with great depth of field, at very low or
very high magnifications can be obtained. Size, shape, and distribution as well as
statistical analyses of these parameters, may be performed.

In scanning electron microscope, the electron gun, produces a stream of
monochromatic electrons. The stream is condensed by the first condenser lens. This lens
is used to both form the beam and limit the amount of current in the beam. It works in
conjunction with the condenser aperture to eliminate the high-angle electrons from the
beam. The beam is then constricted by the condenser aperture, eliminating some high-
angle electrons. The second condenser lens forms the electrons into a thin, tight,
coherent beam. An objective aperture further eliminates high-angle electrons from the
beam. A set of coils then "scan" or "sweep" the beam in a grid fashion, dwelling on
points for a period of time determined by the scan speed (usually in the microsecond
range). The final lens, the objective, focuses the scanning beam onto the part of the
specimen desired. When the beam strikes the sample (and dwells for a few microseconds)
interactions occur inside the sample and are detected with various instruments. Before

the beam moves to its next dwell point these instruments count the number of interactions

40

and display a pixel on a cathode ray tube (CRT), whose intensity is determined by this
number (the more reactions the brighter the pixel). This process is repeated until the grid

scan is finished and then repeated, the entire pattern can be scanned 30 times per second

(17).

2.5. Materials and Methods

Materials. 1,2-Dioleoyl-sn-Glycero-3-Phosphocholine (DOPC) and 1,2- Dioleoyl -sn—
Glycero-3-[Phospho-L-Serine] (Sodium Salt) (DOPS) were obtained from Avanti Polar
Lipids Inc. (Alabaster, Alabama). Oleic acid 2 99% was purchased from Fluka. All
restriction enzymes and DNA modifying enzymes were purchased from New England
Biolabs (Beverly, Massachusetts). Ni-NTA was purchased from Qiagen Inc. (Valencia,
California). Specially purified aqueous detergent solution (10% solution) of Tween® 20
was obtained from Pierce (Rockford, Illinois). Arachidonic acid was obtained from
Cayman Chemicals (Ann Arbor, Michigan). Nanogold® - antirabbit (NFR) was
purchased from Nanoprobes Inc. (Yaphank, New York). Uranyl acetate was from Fluka.

Other chemicals were purchased from Sigma.

Site-directed mutagenesis. A histidine-tagged version of the human COX-2 protein with
cysteines 299 and 526 mutated to serines was subcloned into the pFastbac plasmid (Life
technologies Inc.). All of the COX-2 mutants were made using Quick Change
Mutagenesis method (Stratagene) using complementary 30-40 nucleotide long primers
containing the desired mutations. The mutations were confirmed by the

dideoxynucleotide sequencing method. Recombinant baculovirus expressing the mutant

41

COX-2 protein was generated using the Bac-to-Bac® baculovirus expression system

(Life technologies, Inc.).

Expression of recombinant COX protein in SF21 insect cells. Spodoptera frugiperda
(Sf21) insect cells were purchased from Invitrogen Corporation (Carlsbad, California)
and cultured at 27°C in HyQ-SFX-Insect serum —free insect cell media from HyClone
(Logan, Utah) supplemented with 0.1% Pluronic F—68 (Life Technologies, Inc.), 1x Lipid
concentrate (Sigma), and 0.2% glucose in four- 2800 ml Fembach flasks (Bellco
Biotechnology Inc., Vineland, NJ) each containing one liter of culture. The shakers were
rotated at 120 rpm using an Innova® model 4000 benchtop gyrotory incubator shaker
(New Brunswick Scientific Co. Inc,. Edison, NJ).

When cells reached a density of 1.5-2 million cells per milliliter, recombinant baculovirus
were added at an MOI of 001-0.] and the infection was allowed to proceed for 72 to 96

hours. Cells were harvested, washed with PBS, and then stored as pellets at -80°C.

Purification of His-tagged COX from baculovirus-infected SF21 insect cells. SF21
cell pellets were resuspended in 3 ml/g wet weight of Nickel-homogenization buffer (25
mM NaPO4, 100 mM NaCl, 20 mM Irnidazole, pH=7.4). The Tween® 20 detergent was
added (1% v/v) to the suspended cells. The cells were disrupted by Bioneb® cell
disruption system (Glas-Col, Terre Haute, Indiana). After low speed (10,000 x g for 1
hour) centrifugation to pellet cell debris, the supemant was poured into cramp-seal
ultracentrifuge tubes. The insoluble material was removed by ultracentrifugation at

100,000 x g for 2 hour. The solubilized extract was next incubated with Fast-flow Ni-

42

 

NTA resin for overnight at 4°C with rocking. The resin was poured into a column and
washed with 5 volume of Nickel-homogenization buffer containing 0.1% Tween®. Next
the resin was washed with 3 volume of Nickel-wash buffer (25 mM NaPO4, 300 mM
NaCl, 20 mM Irnidazole, pH=7.4) containing 0.1% Tween® 20. The protein was eluted
with Nickel-elusion buffer (25 mM NaPO4, 100 mM NaCl, 200 mM Irnidazole, pH=7.4)
containing 0.2% Tween® 20. Fractions with high specific cyclooxygenase activity were

pooled and concentrated in Amicon® Ultra centrifugal filter device with 30000

molecular weight cutoff (Millipore Co., Bedford, MA).

Cyclooxygenase assay. Cyclooxygenase activities were measured at 37°C by observing

the initial rate of oxygen consumption with an O2 electrode (YSI Inc., Yellow Springs,

OH) (19).

Protein determinations. Protein concentrations were measured using Bicinchoninic

acid method (Pierce, Rockford, Illinois) with bovine serum albumin as standard.

Preparation of large unilamellar vesicles (20). Large unilamellar vesicles (LUVs) were
prepared with a DOPC: DOPS ratio of 3:7 and various amounts of oleic acid (0%, 1.25%,
2.47%, 4.82%, 9.2%, 16.8%, 28.8% and 44.7% (w/w)) by mixing the appropriate aliquots
of lipids and fatty acid in chloroform and drying the solution under a stream of nitrogen.
The dried film of the phospholipids was dissolved in 0.5 ml of hexane and the organic
solvent was removed again under a stream of nitrogen. The resulting film was hydrated in

a buffer of 25 mM Tris-Cl and 50 mM KCl pH 7.4 for 1 h, vortexed thoroughly and

43

extruded through a polycarbonate filter with a 0.1 um pore diameter using a hand-held
Mini—Extruder (Avanti Polar Lipids, Birmingham, AL). Lipids were extruded at a total
concentration of approximately 38 mM. This process resulted into a transparent milky

suspension.

Incorporation of COX into the preformed liposomes (20). The COX protein was
added to the preformed liposomes. Three different protein: lipid molar radio was used

(1:250, 1:500 and 121000). The solution contained 25 mM Tris-Cl and 50 mM KCl pH

7.4. The mixture was incubated for 20 minutes at 37 C.

Separation of unincorporated proteins from the proteoliposome (21). 0.5 ml of
liposome suspension was mixed with 1 ml of 30% (w/v) Ficoll in 25 mM Tris-Cl and 50
mM KCl pH 7.4 to give a final concentration of 20% (w/v) Ficoll. The liposome
suspension was transferred to an ultracentrifuge tube. 3 ml of 10% (w/v) Ficoll was
gently layered on top of the liposome suspension. The upper Ficoll layer was covered
with a solution of in 25 mM Tris-Cl and 50 mM KCl pH 7.4 (0% Ficoll). The swing-out
rotor (Beckman SW 50.1) was used for 30 min at 100,000 g at 20 C. The liposomes and
proteoliposomes are separated at the interface between the 0% Ficoll and 10% Ficoll

layers. The unincorporated protein will remain in the 30% Ficoll layer.

Determination of total phosphorus: The total amount of phosphorus was measured by

ascorbic acid method of Ammon and Hinsberg, (22) modified by Lowry et al (23).

44

Gold Labeling. About 100 pl of proteoliposomes are placed into Eppendorf tube and 10
pl of primary antibody is added to them. The mixture is incubated for 30 minutes in room
temperature with occasional shaking. The non-attached primary antibodies are separated
from the complex of proteoliposomes-primary antibodies using the Ficoll gradient
technique, which was described previously. The liposomes and the complex of
proteoliposomes-primary antibodies are separated at the interface between the 0% Ficoll
and 10% Ficoll layers. The unattached antibody will remain in the 30% Ficoll layer. The
complex of proteoliposomes-primary antibodies are placed into a Eppendorf tube and
about 100 pl of gold immunoprobe NANOGOLD®-anti rabbit (NRF) (1.4 nm particle
attached to affinity-purified Fab’ fragment, raised in goat, against rabbit IgG (whole
molecule» is added to the tube and is incubated for 30 minutes in room temperature with
occasional shaking. The non-attached gold-labeled secondary antibodies are separated
from the complex of proteoliposomes-primary antibodies-gold labeled secondary
antibodies using the Ficoll gradient technique. The liposomes and the complex of
proteoliposomes-primary antibodies-gold labeled secondary antibodies are separated at
the interface between the 0% Ficoll and 10% Ficoll layers. The unattached gold-labeled
secondary antibody will remain in the 30% Ficoll layer. The transmission electron
microscopy was done on the sample as described using a diluted stain (0.1-0.2 % solution

of uranyl acetate). Because the NANOGOLD® particles are small, over staining with

uranyl acetate may tend to obscure direct visualization of individual NANOGOLD®

particles. Therefore a diluted stain was used.

45

 

Transmission electron microscopy of liposomes and proteoliposomes. A J EOL
transmission electron microscope (model IOOCXII, Tokyo, Japan) operating at 100 kV
under the vacuum was used. Forrnvar (polyvinylformaldehyde) - carbon coated copper
grids G-300 mesh (Electron microscopy sciences, Inc., Fort Washington, PA), were
employed. A drOp of liposome solution (concentration about 5 mM of lipid) was placed
on the grid and after 1 minute the excess was removed with a Whatman filter paper #1. A
thin film was left on the grid and allowed to air dry. Negative staining with one drop of a
1% solution of uranyl acetate was performed. After 1 minute this dr0p was again

removed with the filter paper, and the resulting stained film was dried in a dust free place.

Scanning electron microscopy of liposomes and proteoliposomes. The liposomes and
proteoliposomes with total lipid concentration between 15 mM to 38 mM were mixed
with an equal volume of 4% gluteraldehyde in 0.1 M sodium phosphate buffer, pH 7.0,
for 1 h at room temperature. After fixation one drop of the suspension was placed on the
coverslip previously coated with poly —L-lysine (Sigma) and allowed to stand for 5
minutes. The coverslip was then carefully washed with several drops of distilled water,
followed by dehydration in ethanol. The coverslips were dried using a Balzers critical
point dryer (Balzers CPD, FL-9496, Balzers, Liechtenstein) with liquid carbon dioxide as
the transitional fluid. Coverslips were mounted on metal stubs and coated with a 25-30
nm gold layer in an Emscope sputter coater (Emscope Laboratories Ltd., Ashford, Kent,
U.K.). The liposomal structures were observed with a JEOL scanning microscope (model
JSM-6300F, Tokyo, Japan) at 7 mm working distance using an accelerating voltage of 13

kV.

46

 

Solid Phase EPR measurements. All EPR measurements were recorded on a Bruker
ESP 300E X-band spectrometer equipped with a TE102 Bruker rectangular cavity. 200 ill
of samples were loaded into a 4-mm OD clear fused quartz sample tubes (Wilmad-
labglass, New Jersey). The first derivative absorption spectra of the samples at 183 1' 2K
were obtained by averaging 4 scans at a scan width of 110 Gauss and each scan time of 4
minutes. The temperature was controlled with Bruker BVT 2000 liquid nitrogen
temperature controller. The microwave power was kept at 1 mW to avoid the saturation

of the EPR lines.

Quantitation of EPR signal and the measurement of incorporation yield (24). In
order to measure the incorporation yield of COX enzyme into the preformed liposome,
single cysteine spin labeled hCOX-2 (H75C) was used. The incorporation yield can be
measured by dividing the total number of moles of spin labeled protein sample added to
the liposomes to total number of moles of the isolated spin labeled proteoliposome band
after Ficoll gradient. In spin counting experiment Fremy’s salt (K2 (803) 2) NO) is used
as a standard and is quantified optically (8243:1690 and 8545: 20.8). Because spin
quantitation by EPR is somewhat prone to error, we have used a detailed protocol for our
quantitation procedure. First, we have shown that both Fremy’s salt and the nitroxide spin
label signal are linear in square root of applied microwave power to at least 5 mW.
Consequently, we choose 1 mW power for our quantitation studies. Second, the linear
relationship between H1 modulation amplitude and the area under the EPR absorption

curve, even under condition of mild over modulation, provides a convenient means by

47

which to increase the accuracy of the determinations. Thus, we carried out a double
integration of the first derivative EPR signal of the Fremy’s salt and the spin labeled
protein samples. By plotting the area vs. the modulation amplitude we were able to
extract a least-squares value for area/gauss of modulation amplitude for the known and
unknown species. Four different modulation amplitudes were used: 1.0080 G, 1,424 G,
1.793 G and 2.013 G. Typical data for nitroxide spin label are shown in Figure 2.1.
Finally, from this value and the known concentration of (K2 (S03) 2) NO), we can

determine the nitroxide spin label concentration using the following equation.

= Ax (BM )Std

AStd (Bm )x

[X]

where A is the measured area under the absorption curve, Bm is the modulation amplitude

in mT.

48

   

 

 

Figure 2.1.

A) Stack-plot of EPR spectra of H75C spin labeled proteoliposomes. Each EPR spectrum
was acquired using a different modulation amplitude.
B) The plot of intensity of each spectra in Figure 2.1 .A versus the modulation amplitude.

The slope of this line is used in equation 2.1.

49

 

 

2.013 G

1.793 G

1.424 G

1.008 G

 

 

1 W750 isolated band (1 :500) protein to lipid rand
8.mE+008 -

   

7.00E+008 -

6.0054008 4

lntensrty

 

 

rY=-1.2144257+4.0916£ax
5.005% -

 

“Ila-ON -

 

 

 

11.0 1 1'2 ' 1i4 ' 1:6 ' 1TB ' 2.0
Modulation Amplitude [Gauss]

50

 

2.6. Results and Discussions

To study the effect of oleic acid concentration on the incorporation yield of COX
enzymes into the preformed liposome different amounts of oleic acid (0, 1.25, 2.47, 4.82,
9.2, 16.8, 28.8 and 44.7 % (w/w)) were added to DOPC: DOPS (7:3) molar ratio aliquots
of lipids in chloroform and dried under the stream of nitrogen and hydrated in Tris buffer
(pH=7.4). COX enzyme was added to the various samples of preformed liposomes. The
molar ratio of protein to lipid was about 1/500 and the mixture was incubated for 20
minutes at 37 °C. The unincorporated protein was separated from the proteoliposomes
using a F icoll density gradient. The activity of the isolated proteoliposome band and the
activity of the mixture of protein and liposomes before separation by density gradient was
measured at 37 °C by observing the initial rate of oxygen consumption with an O2
electrode. The concentration of lipid was determined by measuring the concentration of
phosphorous. The incorporation yield was determined by dividing the total activity of the
isolated band to the total activity of the protein and liposome mixture before separation
by density gradient. Our results show that there is a significant increase in the
incorporation yield upon adding the oleic acid. The optimum amount of oleic acid to
yield the highest incorporation is between 4.8 and 9.2 % (w/w) (See table 2.1 and 2.2).

To examine the effect of the ratio of the protein to lipid on the incorporation yield
of COX enzyme into preformed liposomes different protein to lipid ratios were used
(1:250, 1:500 and 1:1000). 1:500 protein to lipid ratio gave the highest incorporation

yield (data is not shown).

51

The incorporation efficiency and the specific activity of the incorporated hCOX-2
was studied using spin counting and electron paramagnetic resonance spectroscopy. In
this experiment a single Cys mutant (H75C) of hCOX-2 was purified, spin labeled and
incorporated into the liposomes. The total number of moles of spin labeled protein were
counted in a sample before separation by density gradient and in an isolated
proteoliposome band after density gradient. The incorporation yield was determined by
dividing the total number of moles of protein in an isolated band to the total number of
moles of protein in a protein and liposome mixture before separation by the density
gradient. In another experiment the incorporation yield for H75C mutant into preformed
liposomes was measured using an oxygen electrode as previously described. From the
spin counting experiments the total amount of protein before Ficoll gradient and the
amount of the protein in the isolated proteoliposome band after gradient were found to be
1.48E-7 mole and 8.037E-8 mole respectively. From the specific activity measurements
the total amount of protein before Ficoll gradient and the amount of the protein in the
isolated proteoliposome band afier gradient were found to be 1.50E-7 mole and 8.369E-8
mole respectively. The spin counting experiment indicated 54% incorporation and
oxygen assay experiment indicate 55% incorporation. These experiments show a good
correlation between the two techniques.

These experiments also indicate that there is no significant change in the specific
activity of the protein in the detergent versus the specific activity of the incorporated

protein in the liposomes.

52

Table 2.1. The effect of oleic acid concentration on incorporation of hCOX-2 into
preformed liposomes.

% Oleic Acid in DOPCzDOPS % Protein Molar Ratio
:3 ' 'Protein

1.25
2.47
4.

9

 

Table 2.2. The effect of oleic acid concentration on incorporation of oCOX-l into
preformed liposomes.

% Oleic acid in DOPCzDOPS % Protein olar Ratio
:3 “Protein

1.25

2.4
4.
9

 

53

The liposomes (unilamellar vesicles and unilamellar vesicles) and
proteoliposomes were extensively studied by transmission electron microscopy (TEM)
and scanning electron microscopy (SEM).

Multilamellar and unilamellar vesicles can be generated by a variety of
techniques, which lead to systems with differing lamellarity, size, trapped volume and
solute distribution. The straightforward hydration of lipid to produce multilamellar
vesicles (MLVs) results in systems, which exhibit low-trapped volumes. Figure 2.2
represents the negative stain electron micrograph of multilamellar vesicles. The onion
layer structure of MLVs is clearly shown in this micrograph.

Unilamellar vesicles (ULVs) can be produced directly from MLVs by extrusion
or sonication. Unilamellar vesicles were used as the artificial cell membrane models for
protein incorporation in this dissertation. Figure 2.3 (B) represents the negative stain
electron micrograph of unilamellar vesicles (ULVs). Figure 2.3 (A) shows the negative
stain electron micrograph of isolated proteoliposome band and Figure 2.4 (B) is a
scanning electron micrograph of isolated proteoliposomes band. Distributions of various
liposome sizes were observed. The diameter of about 90 liposomes were measured. The
average diameter was 100 i 15 nm. Scanning electron microscopy reveals the 3-D
structure of liposomes. Figure 2.4 (A) shows the 3-D spherical shape of the liposome.

In order to directly observe proteins in a proteoliposome, the hCOX-2 enzyme
was gold labeled using a gold immunoprobe. Figure 2.5. shows the gold labeled proteins
incorporated into a liposome. COX isozymes exist as dimer. Therefore there will be two

gold labeled antibody per each molecule. However because of the resolution limitation it

54

is not possible to observe two particles per molecule. We believe that each black dot on

the electron micrograph represents one protein molecule.

55

 

 

Figure 2.2 Representative negative stain electron micrograph of multilamellar lipid
membranes. (3:7 molar ratio of DOPS:DOPC and 9.2 % (w/w) oleic acid liposome was
used).

Bar=48.8 nm

(Magnification =205,000)

56

Figure 2.3

A) Representative negative stain electron micrograph of proteoliposome (3:7 molar
ratio of DOPS:DOPC and 9.2 % (w/w) oleic acid liposome was used).
Bar=150 nm

(Magnification = 67,000)

B) Representative negative stain electron micrograph of plain liposome (3:7 molar ratio
of DOPS:DOPC and 9.2 % (w/w) oleic acid)
Bar=200 nm

(Magnification = 50,000)

57

 

 

Figure 2.4.

A) Scanning electron micrograph of a liposome. The background of the
picture is the gold plated cover slips.

B) Low magnification scanning electron micrograph of liposomes.

Bar: 250 nm

59

 

..,

' m in

 

 

Figure 2.5. Representative negative stain electron micrograph of gold labeled proteins

incorporated into a liposome (3:7 molar ratio of DOPS:DOPC and 9.2 % (w/w) oleic acid
liposome was used).
BamSO nm

(Magnification = 200,000)

61

2.7 . References

10.

ll.

12.

13.

14.

15.

16.

Bangham A.D., S., M.M, and Watkins, J.C. (1965) Journal of Molecular Biology
13, 238-52.

Kagawa, Y., and Racker, E. (1971) Journal of Biological Chemistry 246, 5477—&.

Rigaud, J. L., Levy, D., Mosser, G., and Lambert, O. (1998) European Biophysics
Journal with Biophysics Letters 27, 305-319.

Rigaud, J. L., Pitard, B., and Levy, D. (1995) Biochimica Et Biophysica Acta-
Bioenergetics 1231, 223-246.

Mateu, L., Caron, F., Luzzati, V., and Billecocq, A. (1978) Biochimica Et
Biophysica Acta 508, 109-121.

Epand, R. M., Boni, L. T., and Hui, S. W. (1982) Biochimica Et Biophysica Acta
692, 330-338.

Epand, R. M., and Sturtevant, J. M. (1984) Biophysical Chemistry 19, 355-362.

Eytan, G. D., and Racker, E. (1977) Journal of Biological Chemistry 252, 3208-
3213.

Navarro, J., Chabot, J., Sherrill, K., Aneja, R., Zahler, S. A., and Racker, E.
(1985) Biochemistry 24, 4645-4650.

Greenhut, S. F., Bourgeois, V. R., and Roseman, M. A. (1986) Journal of
Biological Chemistry 261, 3670-3675.

Szoka, F., and Papahadjopoulos, D. (1978) Proceedings of the National Academy
of Sciences of the United States of America 75, 4194—4198.

Batzri, S., and Korn, E. D. (1973) Biochimica Et Biophysica Acta 298, 1015-
1019.

Racker, E. (1979) Methods of Enzymology 55 , 699-71 1.
Holloway, E. T., and Bohr, D. F. (1973) Circulation Research 33, 678-685.
Jain, M. K., and Zakim, D. (1987) Biochimica Et Biophysica Acta 906, 33-68.

http://www.nobe1.se/.

62

17.

18.

19.

20.

21.

22.

23.

24.

Flegler S.L., H. J. W., and Klomparens KL. (1993) Scanning and transmission
electron microscopy: an introduction, W.H. Freeman, New York.

Hayat MA. (2000) Principles and Techniques of Electron Microscopy,
Cambridge University Press, Cambridge.

Laneuville, O., Breuer, D. K., Dewitt, D. L., Hla, T., Funk, C. D., and Smith, W.
L. (1994) Journal of Pharmacology and Experimental Therapeutics 271, 927-934.
Singh, P. (2001) in PCT Int. App. WO 2001041740, USA.

New, R. R. C. (1990) Liposomes a Practical Approach, Oxford University Press,
New York.

Ammon, R. H., K. (1936) Z. physiol. Chem. 239, 207—216.
Lowry, O. H. R., N.R.; Leiner, K. Y.; Wu, M.L.; Farr, A. L.,. (1954) Ibid 207.

Weil J.A., B. J. R., Wertz J.E. (1994) Electron Paramagnetic Resonance, John
Wiley & Sons Inc., New York.

63

 

Chapter 3

Electron Paramagnetic Resonance Spectroscopy and Site Directed Spin Labeling

3.1. Abstract

Although the catalytic domain of COX isozymes has been extensively studied,
very little is known about how the putative membrane binding domain of COX isozymes
interacts with the lipid bilayers. X-ray crystallographic studies (I) have demonstrated that
neither isozymes contain amino acid sequences of sufficient length and hydrophobicity to
function as classical transmembrane domains. Instead, COX-1 and —2 are thought to
associate with cellular membrane through a novel, monotopic membrane-binding domain
that associates with a single leaflet of the lipid bilayer, which would allow orientation of
residues in the COX-1 helices paralleled to the lipid/water interface of the membrane.
The crystal structures of these isozymes can be used to predict how COXs interact with
membranes, but more direct methods must be used to confirm or refute these predictions.
Furthermore, it is not possible to orient the COXs in a membrane, or to determine lipid
binding sites in the COXs from a crystal structure, as few lipid or detergent molecules are
visible using X-ray analysis. The experiments outlined in this chapter were designed to
begin mapping the topology of COX-membrane interactions. The results of the
experiments in this chapter shall have general applicability to a new class of
monotopically-associated peripheral membrane proteins. Squalene hopene cyclase is
another monotopic membrane protein, the only other structurally characterized protein of
this type (2).

Site-directed spin labeling (SDSL) is a powerful tool for determination of
membrane protein structure and the topology of protein binding to the lipid bilayer (3, 4).

The strategy of SDSL involves the introduction of a nitroxide side-chain at a selected site

65

in a protein. This is usually accomplished by cysteine substitution mutagenesis, followed
by modification of the unique sulfhydryl group with a selective nitroxide reagent.
Nitroxide scanning experiments provide sequence-correlated data that identify regular
secondary structures and their orientation by virtue of periodicity in sequence position.

In this work, 25 mutant enzymes that contained single reactive cysteines
substituted for amino acids within the hCOX-2 membrane-binding domain were
constructed and spin labeled. The accessibility of each spin label to freely diffusing
oxygen as a non-polar paramagnetic reagent and NiEDDA as a polar paramagnetic
reagent was measured using power saturation EPR spectroscopy and the orientation of
the membrane binding domain with respect to the lipid bilayer surface was determined.

Section 3.2 of this chapter will give an overview of various structural biology
methods. In sections 3.3, 3.4 and 3.5, principles of EPR spectroscopy, the structure of the
EPR spectrometer and resonators and the spectral properties of nitroxide free radicals will
be discussed respectively. Section 3.6 will give a detailed description of the site directed
spin labeling technique and finally in the last two sections of this chapter, 3.7 and 3.8, the

experiments and the results will be discussed.

3.2. Structural Studies of Biomolecules

Structural biology is the science of biological phenomena at the molecular level.
The task of explaining biological function at a molecular level requires knowledge of the
composition and structure of the biological system. As genome research rapidly increases

the number of proteins known, the requirement for obtaining structural information will

66

 

increase dramatically (5). The techniques used to gain structural information are physical
chemistry methods that can be applied to biological problems. These techniques include
X-ray diffraction, neutron diffraction, electron diffraction, nuclear magnetic resonance
spectroscopy (NMR), electron paramagnetic resonance spectroscopy (EPR), fluorescence
spectroscopy, Raman spectroscopy, infrared spectroscopy, ultraviolet spectroscopy,
circular dichroism spectroscopy and X-ray absorption spectroscopy. Each physical
method provides a different perspective of the system, and multiple methods are needed
to achieve a full view.

In this present discussion, the following classes of techniques will be discussed:
High (atomic) resolution techniques, low resolution techniques, and techniques using
probes. High-resolution techniques use either X-ray crystallography or NMR and yield
molecular structures at atomic resolution. Low resolution techniques, such as electron
microscopy, atomic force microscopy and electron diffraction yields the overall shape of
the biomolecule at less than atomic resolution. Techniques involving probes are
fluorescence spectroscopy, spin label electron paramagnetic resonance spectroscopy and
certain NMR isotopic labeling experiments. These techniques do not result in atomic
resolution structure, but instead they yield information about specific parts of the
biomolecule. For example, probes can be used to measure the distance between two parts
of a complex, or to learn about the dynamics of specific parts of biomolecules.

X-ray crystallography and NMR spectroscopy are excellent techniques for
providing high resolution structural information (6, 7). X-ray diffraction can yield
interatomic distances for the form of the protein that is present in the crystal. When the

sample is cryo- cooled, resolution for proteins may approach 0.9 A. NMR can yield an

67

average solution structure or the structure in solid state based on multiple measurements
of short (<5-6 A) interproton distances. However they both have limitations. Both
methods are limited by sample requirements. Many samples do not crystallize and even if
they do, the crystal may not diffract well or their phases may not be solvable. NMR
requires relatively concentrated samples of molecules that are small in molecular weight.
Consequently, membrane proteins are difficult to study by either technique. Only very
few membrane proteins structures have been solved by X-ray crystallography or solid
state NMR (8, 9).

Because membrane proteins do not easily yield to crystallization or NMR, they
have been studied using lower resolution methods. Electron diffraction, a technique that
requires two-dimensional crystals, yields low resolution information in which helices can
often be resolved. Furthermore, membrane proteins are well suited to study by electron
diffraction (10, 11).

Electron microscopy provides a lower resolution structure without the need for
two-dimensional crystals. High-resolution electron micrographs show the overall shape
of molecules or complexes, but do not give atomic detail. Often, low-resolution data is
complementary to high-resolution data. For instance, GroEL-GroES complexes were
observed by electron microscopy before their crystal structures were solved. When the
crystal structures of GroEL and GroES (12-15) were solved several years later, the
electron micrographs were useful in defining how the two proteins interact.

The third class of biophysical techniques is those involving probes. For these
methods, the biomolecules are labeled with an unnatural probe that slightly changes the

molecular structure, and the probes are studied using spectroscopic methods. The

68

information that the probe yields is then used to make inferences about the structure,
dynamics, or environment of the molecule. Advantages of probes are that they can
simplify complicated systems, they do not require crystals, and often the sample quantity
and concentration requirements are much lower than for other techniques.

In fluorescence labeling the distance between two fluorophores can be measured
using Fluorescence Resonance Energy Transfer (FRET). This method is based on Foster
energy transfer theory. One fluorophore, the donor, transfers the energy to the acceptor
without emission of a photon. Fluorescence energy transfer can yield distances between
naturally fluorescent sites or added labels if all of the angles that define the relative
orientations of the axes of the donor and the acceptor are known or average values are
assumed. Distances of up to 80 A can be measured if suitable donors and acceptors are
chosen (16).

Site directed spin labeling has emerged as a powerful technique for determining
the structure of the biomolecules. By incorporating two spin labeled side chains,
distances between elements of secondary structure can be determined by quantitation of
spin-spin interaction between the nitroxide probes using EPR spectroscopy (17).

Exogenous spin labels have been routinely criticized as “ perturbing” the “real”
structure. Although researchers always have to be sensitive to the relationship between
reality in living systems and their observations, spin labeling is not more perturbing than
other physical methodologies. Any method that modifies the biomolecules with a probe is
subject to the concern of whether the studied features of the biomolecule are significantly
perturbed by the added probe. This is an important question also for X-ray

crystallography, for example, when a heavy metal is added to aid in scattering, in

69

fluorescence labeling when a fluorescent label is attached and in spin labeling when a
nitroxide spin label is attached. For the specific case of nitroxide spin labels, there have
been several inquiries into the extent to which the label perturbs the structure of the
proteins. Spin labels attached to three of the amino acids in myoglobin (His-12, Tyr-103,
and Tyr-151) were shown not to affect the properties of heme or the conformation and
stability of he protein (18). Similarly, studies of spin labeled T4 lysozyme have revealed

minimal perturbation in the structure (19).

3.3. Principles of Electron Paramagnetic Resonance Spectroscopy

3.3.1. Zeeman Interaction

Electron paramagnetic resonance spectroscopy (EPR) is very similar in concept to
nuclear magnetic resonance spectroscopy (NMR). Both deal with the interaction of the
electromagnetic radiation with the magnetic moment. In the case of EPR, the magnetic
moments arise from electrons, but in NMR the magnetic moments arise from nuclei. EPR
spectroscopy is applicable to systems in a paramagnetic state. The paramagnetic state is a
state having net electron spin angular momentum. The origin of EPR spectroscopy lies in
the spin of the electron and its associated magnetic moment (u). The component of

electron magnetic moment along the direction of the external magnetic field (z-axis) is:

”Z :_gefleMs 3'1

7O

eh

4am e

 

and, 'Be = where me is the mass of electron and g, is the free electron g-factor and

is equal to 2.0023 and Ms = :t

va—

In the absence of an external magnetic field, the energy levels of the magnetic
moment are degenerate. In the presence of an external magnetic field, there is an
interaction between the external magnetic field (Bo) and the magnetic moment (,u), which
is called the Zeeman interaction, and the energy levels will be separated. The Zeeman

interaction energy for the electron is:
1
= — = : + —
E (,uZ )eB0 gefl BOMs _ 2 gefieBO 3.2

e

and 1n the srmilar manner for nuclei spin 2 , e.g. 1H, the Zeeman interaction energy is:

_ __ _ l
E_ (#Z)nBO_ gnfleBOMI _i§gnfln30 3'3

 

and fl = eh , where mp is the mass of proton and M I = i%.
m
P

The ratio of the electron magnetic moment to the 1H magnetic moment is :

P = 659 3.4

 

71

 

where for 1H, g" = 5.5857. Therefore the magnetic moment of the electron is 659 times

larger than the magnetic moment of 1H and also the gap between the energy levels for
electron is 659 times larger than 1H. The transition energy between electron Zeeman
energy levels is in the microwave frequency (109 Hz) range and for nuclear Zeeman

energy level is in the radio frequency range (106 Hz)(20).

3.3.2. The Resonance Condition

In the presence of an external magnetic field the energy levels for an electron spin
are separated. This separation increases linearly with Bo. It means that if the magnetic
field is stronger, the separation between the two energy levels is more. If the energy of
the microwave matches the difference between the two energy states, a transition in the
population between the two energy levels happens and the spectra is obtained. The
frequency for the microwave radiation is in the range of 1-100 GHz. The difference

between the energy levels is expressed as:

AB = hv = gefleBO 3.5

where h is Planck’s constant and v is the resonance frequency. This equation describes

the “resonance” condition for electron paramagnetic resonance spectroscopy.
Each molecule has a thermal energy equal to kT. At ambient temperature this

energy is 208 cm". The energy difference between the electron spin states is typically 0.3

72

 

 

cm". Since the thermal energy of a molecule is more than the difference between the two
energy levels, at ambient temperature the electron spins populate both energy levels.

Due to Boltzmann law:

Ne — E —E - B
J=expwzexpflgfl=expfl=0.9936z1 3.6
Nle kT kT 208

where N: is the number of electrons in the upper state and N; is the number of

electrons in the lower state, k is the Boltzmann constant and T is the temperature in
Kelvin.

After irradiating the molecule with microwave energy, there is a transition
between the population of spins in the two energy levels. The net transition is from lower
state to upper state. The spins do not remain in the excited state and after a brief time the
spins emit their energy and return to the ground state. This is a spontaneous emission and
the spins lose their energy, in other words, they exchange their energy with the
surroundings, which is called lattice. The time that takes for the spins to transfer their
energy to lattice and relax to the ground state is called T1 or spin-lattice relaxation time.

In addition to spin-lattice relaxation, there is another phenomenon that describes
an interaction between spins, called the spin-spin relaxation, T2. The more efficient
relaxation pathway will predominate over the other and will have the greatest affect on
the EPR linewidth of the EPR absorbance. In solution, for an organic free radical such as
the nitroxide free radical, the spin-lattice relaxation time is much greater than the spin-

spin relaxation time. The reciprocal of the spin-lattice relaxation time, 1/T., is very small

73

in comparison to the spin-spin relaxation time, 1/1‘ 2. So the spin-spin relaxation time
determines the linewidth of the EPR spectrum. In these cases, the spin-spin relaxation

time is defined by the central linewidth, AB, in the first derivative spectrum (20).

)AB 3.7

3.4. The Electron Paramagnetic Resonance Experiment

Most EPR spectrometers operate at a fixed frequency, usually in the microwave
region, and the resonance condition is met by sweeping the applied magnetic field, B2.

The electron paramagnetic resonance spectrometer used throughout this study was
operated with the microwave field at approximately 9.3E9 Hz, designated X-band, with

the resonant field being approximately 3300 Gauss.

3.4.1. The EPR Spectrometer

A simplified diagram of an EPR spectrometer is shown in Figure 3.1. During
operation, the microwave energy from the klystron (A) travels down the wave conduit
(B) into a resonator (C), containing the sample in a special cavity (D). The electromagnet

(E) applies an external magnetic field (I) to the sample in order to create the excited and

74

 

 

 

 

 

 

B
E ._ _____ "3-)";
———— ———>
__.._ E®§ —->
.__ 'C _:
__F;_ 5:.{_ 12.)

 

 

 

 

Figure 3.1. The EPR spectrometer

A: Klystron B: Wave guide conduit C: Resonator D: Sample cavity E: Electromagnet
F: Helmholtz coils G: Detector H: Computer 1: Magnetic field.
Taken from (21)

75

ground spin states for the free radical electron. The microwave energy interacts with the
sample and returns into the wave-guide. The microwave signal returning from the sample
is routed by a circulator to a detector that measures its incident power. To enhance
sensitivity, the magnetic field at the resonator is modulated by Helmholtz coils (F) to
encode a new frequency of 100 kHz on any EPR absorption signals that may occur. The
modulation amplitude must be kept below the linewidth of the EPR absorption in order to
avoid linewidth distortion of these resonances in the derivative spectrum. The EPR signal
is fed from the detector (G) to a 100 kHz lock in amplifier detection of the modulated
signal. The modulated signal appears as the first derivative of the absorption spectrum

and the signal is output to a computer or recording device (H).

3.4.2. The Loop-Gap Resonator

The resonator cavity is critical for determining EPR sample conditions and signal
strength. The loop-gap resonator is shown in Figure 3.2. The microwave energy from the
wave guide conduit (not pictured) reaches the resonator through the transmission line
(A). The capacitance of the wire coupling loop (C) is adjusted with the coupling
adjustment (B) in order to match the transmission line impedance to the impedance of the
resonator circuit containing the sample tube in the sample cavity (D). The central core of
the resonator is made of a ceramic material and coated with a thin layer of silver. The
flux return loop (E) and the sample loop (G) are actually cylindrical holes bored in this
ceramic material and connected by a narrow gap (F). The configuration of these 100ps

and the thin gap between them is designed to optimize the magnetic component of the

76

1w

 

 

 

 

 

 

 

 

 

 

I
I
| G
l
I

W

C)

 

 

 

 

 

 

 

 

 

Figure 3.2. The two loop one gap resonator
A: Transmission line B: Coupling adjustment C: Wire coupling loop D: Sample cavity

E: Flux return loop G: Sample loop F: Gap.
Taken form (21)

77

microwave signal in the sample loop and minimize the electric component. This provides

the highest possible microwave magnetic field (I) density across the sample cavity (22).

3.4.3. Rectangular Box Cavity (TE102)

TE102 resonator is the most commonly used EPR cavity. This resonator was used
for the double site directed spin labeling studies on frozen samples (Chapter 4). Figure
3.3 shows the structure of this resonator. The sample is placed in the middle of the
resonator where the magnetic field density is maximum and the electric filed density is

minimum.

3.5. Nitroxide Free Radicals

Paramagnetic species are rarely found naturally, therefore to investigate biological
phenomena using electron paramagnetic resonance, spectroscopists incorporated stable
free radical spin probes into the biological system of interest. Site directed mutagenesis is
used to substitute cysteine residues at specific amino acids of the interest and a sulfhydryl
specific nitroxide spin label is attached to the mutated cysteine (Figure 3.4). A standard
nitroxide free radical containing an l4N nucleus will give a characteristic three line
spectrum as seen with the l-oxy-2,2,5,5-tetramethylpyrrolinyl-3-methyl)-
methanethiosulfonate spin label (MTSSL). This structure is stabilized by the four
adjacent methyl groups which exert a protective effect on the free radical and make this

spin label, and ones like it, stable and inert.

78

 

Figure 3.3. The structure of TE102 rectangular box cavity. The Picture was taken from

ESP 300e EPR spectrometer manual.

 

oooooooooooooooooooooo

s-g- s
__ o E __
N + PROTEIN .su ——- ; N

i - I
O 3 O

reagent in :siaé‘asaa'm'

Figure 3.4. The methanethiosulfonate spin label (MTSSL) reacts exclusively with the
free sulfliydryl group on the introduced cysteine residue. Taken form (23). Copyright
permission was obtained.

80

 

Depending on the questions the investigator wishes to answer (s)he may choose
one of several sulfliydryl-reactive label reagents that have been synthesized. Each of
these reagents is usually composed of two functional parts: 1) a sulfhydryl reactive group
that provides the means of attachment to a cysteine residue and 2) a nitroxide moiety.
Variation in the means of attachment or in the geometry of the nitroxide group provides
the researcher with many options to explore scientific questions.

The typical free radical spin probe has two resonance structures that contribute to
the overall nitroxide structure (Figure 3.5). The resonance contributing form A that
localizes the radical electron on the nitrogen also places a negative charge on the oxygen.
Environments which stabilize the negative charge on the oxygen in this contributing
form, such as polar solvents, would increase the unpaired electron density at the nitrogen
and consequently increase the hyperfine coupling constant a0 between the resonance
induced by the nitrogen spin. This trend is observed in nitroxide spectra in solvents of
varying polarity.

In addition, measurement of the three dimensional components of the hyperfine
tensor indicate that the radical electron is located primarily along the molecular axis
perpendicular to the plane of the three nitroxide bonds. The hyperfine tensor components
are determined by examining the spectra of a nitroxide in a solid crystal under different
magnetic field orientations. These values indicate that the electron primarily occupies a p

orbital on the nitrogen atom oriented perpendicular to the nitrogen bond plane (24).

81

 

Figure 3.5. Two resonance structures of nitroxide

82

N—Ci

3.5.1. Nitroxide Radical Spectrum

The total Hamiltonian for a system with one electron and one l4N atom is:

H=He+Hn+HHF+HQ 3.8

where He, Hn, Hm: and HQ are electron Zeeman, nuclear Zeeman, hyperfine interaction

and nuclear quadrupole interaction Harniltonians respectively. For simplicity the effect of

the nuclear quadrupole interaction in neglected.

H = gefl BOS

e — gnflnBOIZ +hASZIZ 3.9

Z

If the hyperfine interaction (A) is positive then the energy levels are:

_ 1 1
El —( 2,1lgefleBOSZ —gn/3nBOIZ +hASzIz l—E'l)
1 1
El=_§gefleBO—gnflnBO-§M 3.10
E =—lhv -hv —lhA
1 2 e n 2

83

1 1
£2 _ (——,01 gefleBOSz — gnflnBOIz +hASlel 2 ,0)

2
I
1
E2 ——§hUe

I 1
E3 — {-5,—1 I gefleBOSz -gnflnBOIz +hASZIZ |-§,-l)
l I

E =—lhv +hv +lhA
3 2 e n 2

1 1
154 45711228433052 —gn,6nBOIz +hASzIZ 15,—1)

l I

E zlhv +hv —-lhA
4 2 e n 2

1 1
E5 _ (5,01 gefieaosz — gnflnBOIz +hASzIz 15,0)
1
Es 781231330 3-14
1

E5 =§hve

84

 

I I
= — — —31
E6 (2’l'ge’6eBOSz gnflnBOIZ+hASZIZ12 )

l l
E zlhv —hv +lhA
6 2 e n 2

The hyperfine interaction is larger than the nuclear Zeeman interaction. The energy level

diagram is shown in Figure 3.6. The EPR transitions are governed by the selection rules

Ams = i1 and Am; = 0.

3.5.2. The Spectrum of the Non-Oriented Systems (Solid Phase)

The EPR spectra of spin labels in solid phase can be described by this

Hamiltonian:

H=lflelBO-ge-S—gnlflnlBO-I+S-A-I 3.16

ge is the electronic Zeeman interaction matrix, g" the nuclear g-factor, B0 is the external
magnetic field, S and I the electron and nuclear spin operators, respectively, and A the
electron nucleus hyperfine interaction matrix. The principle axes of the g6 matrix defining
field orientation are shown in Figure 3.7, with the x-axes lying along the N-O bond, 2-

axis parallel to the nitrogen, 2pz orbital and the y-axis perpendicular to the x,z-plane.

85

 

 

 

 

 

 

 

’I T—
I”
I”
,'—T—‘::““""_—— """""
I \
I \
[I ‘\‘\ ‘7
‘\ a",’
X .2 ,,,,,
[I —_"
I
h
‘ v.
\\ c ””””
\ ‘,—'
a—-—'"
/\\\ ’1’
\\ ”1’
\\ I”
\P A,’
‘\
\“
\“
\m‘
Electron Zeeman Hyperfine Nuclear Zeeman

 

I1/2,l>
|l/2,0>

ll/2,-l>

l-l/2,-l>

l- l/2,0>

l-1/2,1>

Figure 3.6. The energy level diagram for a system with one electron and one l4N atom.

86

 

Figure 3.7. Principle axes of the g6 matrix of nitroxide spin label. Taken from (25)

87

 

 

 

 

 

 

 

"11:0
rm
:1 y z
m':+l m|="I
l I l I l 7
z x y x y z
1 l l
3330 3370 3410
Magnetic Field (G)

Figure 3.8. Schematic diagram of the rigid limit EPR powder pattern of MTSSL.

Taken from (26). Copyright permission was obtained.

88

In Figure 3.8, the EPR spectrum of a spin labeled sample at low temperature (solid phase)
is shown. The EPR spectrum is equivalent to a powder in the rigid limit of polycrystalline

sample. The spectrum is composite of spectra of molecules differing by their projection

of the 14N nuclear moment into the external magnetic field Bo designated by the values of

m = +1, 0 and —1. 14N hyperfine contribution to the spectrum produce three step

 

hv—m A hv-m A

shoulders at magnetic fields B =———I—Z~ and six divergences at B = I x,y
z 8 .5 x,y g ,8
Z 3 x,y e

(24).

3.5.3. Spectral Anisotropy

In a nitroxide radical, the unpaired electron is localized to a p-type orbital that is
not spherically symmetric. In this situation, interactions between the unpaired electron
and the nitrogen nucleus and between the unpaired electron and the applied magnetic
field are sensitive to the orientation of the molecule relative to the applied field. In other
words, the position of the resonant field (g-value) and the magnitude of the hyperfine
splitting (A-value) are dependent on the orientation of the magnetic field relative to the
nitroxide molecular axes. Using a Cartesian coordinate system, the principle axes for the

nitroxide are diagramed in Figure 3.7. The z-axis is defined by the nitrogen 2p1t-orbital

that is associated with the unpaired electron and the x-axis is defined by the NO bond.
The y-axis is orthogonal to the x and z-axes. The molecular coordinate system for the

nitroxide is used to define the direction of the applied magnetic field relative to the

89

 

 

 

B||X V 20G
Jill

Ve-
Bllz . An Ji

Figure 3.9. Orientation dependence of the EPR spectrum of a nitroxide. Simulated
spectra of a monocrystal sample with the external magnetic field B directed along the
principle axes of g and A tensors. Taken from (25)

90

nitroxide. The orientation-dependent effect of the nitroxide g- and A- values relative to
the applied magnetic field is diagramed in Figure 3.9, which shows the spectra for a
nitroxide in a highly ordered crystal that is oriented such that the applied magnetic field is
parallel to each of the principle molecular axes, as noted in the figure. To described the
anisotropic nature of g and a, they may be characterized by second-rank tensors, g and A.
These tensors are diagonal in the nitroxide molecular axis system with principle values:
gxx, gyy’ 822, and A“, Ayy, All. The magnetic interactions between an applied magnetic
field and the unpaired electron for any other nitroxide orientation can be described by
arithmetic combinations of the principle values. As can be seen in Figure 3.9, the nuclear

hyperfine coupling is maximal when the 2p1t orbital is parallel to the applied magnetic
field, leading to All having the largest splitting values. As will be seen in section 3.5.4,

the above anisotropies may be partially or totally averaged if molecular motion is rapid

relative to the EPR timescale.

3.5.4. The Effect of Nitroxide Motion on EPR Lineshapes

Extraction of motional data from the EPR spectra of nitroxides is arguably the
most widespread use for spin labels. An EPR spectrometer has an effective time interval
for collecting data (1/v). If molecular motion is much faster than 1/v, all of the spectral
anisotropies will be averaged and single g- and A- values will be observed. Nitroxide
attached to macromolecules are usually characterized by longer rotational correlation

times where averaging of the g and A anisotropies is incomplete. These effects are

91

1: [ns]

01

 

 

 

 

 

ZOG

Figure 3.10. Motion dependence of the EPR spectrum of a nitroxide. Simulated spectra
of a nitroxide tumbling in solution with the rotational correlation time 1:. No significant
changes in the spectral line shape are observed at 'c exceeding 50 ns. Taken from (25).

92

illustrated in Figure 3.10, which demonstrates the correlation between the rate of
isotropic nitroxide tumbling and the resulting EPR spectra. For an ensemble of nitroxides
with very fast reorientation rates, all of the spectral anisotropies will be averaged and
single g- and A- values will be observed. This is the case for the spectrum at the top of
Figure 3.10, where the correlation time for the nitroxide is less than 0.1 ns. However, if
molecular motion is slow relative to the above criteria (bottom spectrum in Figure 3.10),
the nitroxide will appear to be motionless and the resulting spectrum is what one would
expect from the superposition of spectra from a sample of nitroxides frozen in all possible
orientations, often referred to as a ‘powder spectrum’. For intermediate amounts of
motion, various degrees of averaging will be observed (middle spectra in Figure 3.10).
The spectra in Figure 3.10 represent simple isotropic tumbling. However, in biological
systems, nitroxide motion is often anisotropic. In these situations, the principle g and A
tensors are not averaged equally. Nitroxide motion at any particular site is directly related
to its local environment and is reflected in its EPR spectrum.

The mobility of a nitroxide side chain depends on the structural features of its
local environment. One of the goals of SDSL is to relate the EPR spectrum for a nitroxide
side chain to its motion, and thus, to the structure of its local environment. This

information is readily achievable at a semi-quantitative level. For example, a convenient
measure of motion is the peak-to-peak first derivative width of the central (m1=0)
resonance (ABO also known as AHO in Figure 3.10). The numerical value for ABo is
primarily dependent on the degree of averaging of the anisotropic g tensor. Structure-

dependent variations in ABO

93

have been employed to map topology in many proteins by simple distinction between
buried and surface sites.

The ability to correlate a nitroxide’s EPR lineshape with its motion and then to
correlate that motion to a specific structural environment is one of the primary goals in

SDSL studies. A nitroxide side chain motion is the result of three contribution factors: 1)

TR, the rotational correlation time for the entire protein; 2) 1:13, the effective correlation

time due to local backbone flexibility; 3) 1:5, the effective correlation time due to

isomerizations about the bonds that link the nitroxide to the backbone. The resulting EPR
spectrum reflects the sum of each of these contributions. Therefore, how local structure

contributes to each of these is a subject of great interest for spin labeling studies (27, 28).

3.6. Accessibility Determination Using Power Saturation

A significant amount of information about a protein’s structure can be derived
from the degree of solvent exposure of its side chains. A nitroxide side chain introduced
at any region of interest can report on solvent exposure using the power saturation
technique. The technique exploits changes in T1 relaxation pathways by Heisenberg spin
exchange with diffusible, fast-relaxing paramagnetic species. The rate of exchange is
proportional to the frequency of encounters and thus the degree of side chain exposure. A
brief introduction to the physical nature of this phenomenon and its experiment utility in

spin-labeling studies is described below.

94

 

3.6.1. Continuous Wave Saturation EPR (CWS-EPR)

CWS-EPR is one of the methods to measure the T1. CW S-EPR can be performed
on a conventional EPR spectrometer, preferably fitted with a loop-gap resonator. The
loop-gap resonator is a recent development and is important in that it provides a sufficient
microwave energy density to saturate the nitroxide resonance.

A convenient experimental parameter determined from CW S-EPR is Pu2. P1/2 is

the microwave power required to saturate the signal to one-half the amplitude it would

have if it did not saturate at all. Changes in the Pm are related to the exchange rate (Wx):

 

_ 0 ~_ ____ x
AID1/2"P1/2 P1/2°c T1 " ( ‘ 3'17

As indicated, this relationship is approximate and assumes that T2 is independent
of the exchange rate. This is generally the case for nitroxides on membrane-bound

proteins where T2 is much less than T. and also the concentration of the exchange reagent

is such that W <1: i.
x 12

The amplitude of an EPR signal is proportional to the square root of the
microwave power. Also, the amplitude of the signal is a linear function of the square root

of the power:

A=NAPU2 3.18

95

where A is the peak to peak amplitude, N is the normalization factor, A is the
instrumental calibration factor. So, by increasing the power, a signal with larger
amplitude is obtained.

For higher microwave powers the amplitude of the signal is not a linear function

of the square root of the power and the equation in modified to:

NAP“2
A=———- 3.19
53/2

272'

where S =1+A2P72T1T2 and yis the magnetogyric ratio given by 7: ——ghQ&.

1/2 1/2
The value of P1/2 is found from the intersection of A=—I%i—and A=1—VA:——

S
(Figure 3.11):
2/3_
P - (2 1) 3.20

1/2“ 2
.A yZfiTé

and substituting this equation in 3.19 gives:

96

NAPl/2

 

A = 3.21
2/3 3/2
[1+(2 —1)P/P1/2]
Equation 3.21 can be modified as:
1/2
A IP 3.22

 

[1+(21/8—1)P/P“2]£

where I=NAand 8:3/2.

8 is an adjustable parameter and measures the homogeneity of the saturation of the
resonance line. For the homogeneous and inhomogeneous saturation limits, 8 = 1.5 and
8:05 respectively. Under the nitrogen (in the absence of any other paramagnetic

relaxers) e is always very close to 1.5 and indicates the homogeneous saturation (29,

30).

3.6.2. Site-Directed Spin Labeling

Eduard G. Rozantsev and Harden M. McConnell are the two pioneers in the spin
labeling technique who contributed seminally to this branch of science. Eduard
Rozantsev developed a broad range of nitroxyl compounds and reagents and Harden

McConnell’s foresight into the areas of biological structure and function brought the

97

 

Amplitude "11.0

 

A=N1.P""’

A:N}.P"’/2

A=NxP"°/s°’

 

I ' I r ' l ‘ l 1

o 2 4’ e 8

P112 SORT Power

Figure 3.11. A typical power saturation curve

98

10

world the broad leadership that guided this technique to where it is today. McConnell’s
lab was located in the Physical Chemistry Laboratories of the Department of Chemistry
at Stanford University, where his research prior to that point had been heavily involved in
areas of chemical physics, solid-state physics, and theoretical aspects of spin physics.
Rozantsev’s lab, then at the Institute of Chemical Physics, USSR Academy of sciences,
Moscow, was publishing extensively on the chemistry, synthesis and physical chemistry
of broad range of nitroxyl compounds (coined iminoxyl at the time). Rozantsev’s
research was published mainly in Russian academy journals. Since it was difficult to
obtain samples from “ behind the iron curtain”, a major synthetic effort was launched in
the McConnell lab to reproduce compounds synthesized in the Rozantsev’s lab and to
develop analogues of protein modification reagents.
Site directed spin labeling (SDSL) consists of
a) Manipulating the system of interest so that it contains a single reactive amino acid
residue or in the case of double site directed spin labeling (DSDSL) two reactive
amino acid residues, usually a cysteine, that may be selectively modified with an
appropriate spin label usually a sulfhydryl specific nitroxide reagent.
b) Selectively labeling the site
c) Examining the functional viability of the mutants, preferably before and after spin
labeling.
d) Characterizing the spin-labeled site by a variety of EPR techniques.
This approached was first introduced by Wayne Hubbell (a former student of Harden
McConnell), who recognized that the rapid advancement in the site-directed mutagenesis

could solve a problem that had plagued spin labeling since its inception- the lack of

99

proteins containing unique reactive cysteines. Indeed its relatively low abundance makes

cysteine the residue of choice in SDSL studies (24).

3.6.3. The Theory of Site-Directed Spin Labeling

The small molecules in a membrane/water system are partitioned between the
water and the fluid hydrophobic phase of the bilayer according to their polarity. Polar
molecules preferentially partition into the aqueous phase, and the non—polar molecules
preferentially partition into the membrane phase. The bilayer interior is non-uniforrn with
gradients of the polarity and fluidity along the direction of the bilayer normal. Thus, there
are expected to be gradients in both concentration and diffusion coefficient of small
molecules in equilibrium with the aqueous phase.

According to Henry’s law:

 

0 0
C (x)-C ex ( i’w)ex (:flfl'l) 323
m " i,w p RT p RT '

where Ci m(x)is the concentration of species i in the bilayer at the distance x from the

interface, Ci W(x) is the uniform concentration of neutral species i inside the water and

membrane respectively.
The nitroxide attached to the protein are expected to have a depth-dependent

collision rate with a paramagnetic reagent in the membrane.

100

The Heisenberg exchange rate Wx with a hydrophobic reagent that freely diffuses

into the lipid phase is proportional to collision rate and is expressed as:

Wx = 47tpngm(x)Cm(x) 3.24

where p is the exchange probability and is equal to l, g is the steric factor, d is the

collision diameter, Dm (x) is the position-dependent relative diffusion coefficient and

Cm (x) is the position-dependent concentration given by equation 3.23.

According to equation 3.17,

P —P0 =AP oc—W—x 325
1/2 1/2 1/2 22 '

where P1/2 and P1322“ the value in the presence and absence of the paramagnetic

reagent.
From equations 3.23-25, it follows that the logarithm of the ratio of saturation is
directly related to the difference in standard state chemical potentials of reagents at any

depth, since T2 cancels out.

 

AP 1 0 _ 0
1/2( ) = ‘Ifll (x) #2 (xn-l-COItS

<D=ln[
AP1129) RT

3.26

101

Usually compound 1 is the paramagnetic molecule, which can easily diffuse inside the
membrane (e.g. O2) and compound 2 is the paramagnetic molecule which can easily

diffuse in the water and not inside the membrane (e. g. NiEDDA, NiAA, CrOx'3).

To obtain equation 3.26, it is assumed that T2 is independent of the rate of

exchange, an approximation and it is desirable to make AP] /2 independent of T2. Since

9

the peak to peak first derivative line width AB is proportional 5‘1— then a quantity AP 1 / 2

2

is defined as:
—— oc W 3.27

AP’1 lzis independent of T2, but it depends on the Q factor of the resonator used in the

EPR spectrometer. To obtain a quantity independent of this factor, dimensionless

accessibility parameter, II is defined

AP’1/2
P1/2(DPPH)

3.28

 

where P ’1/2(DPPH) is the value determined for a sample of crystalline 2,2-diphenyl-l-

picrylhydrazyl.
In Figure 3.12 model helices with different topographies and corresponding

idealized behavior of accessibility parameter TI, is shown.

102

3.7. Materials and Methods

Material. The nitroxide spin label (1-oxyl-2,2,5,5-tetramethylpyrroline-3-methyl)-
methanethiosulfonate (MTSSL) was purchased from Reanal Fine Chemical (Budapest,
Hungary). NiEDDA was prepared as follows: equimolar amounts of Ni(OH)2 (464 mg)
and EDDA (881 mg) were mixed in 100 ml of 50% methanol in water. The mixture was
stirred for 24 hr at room temperature and an additional 24 hr at 60°C. The resulting blue
solution was filtered, and the solvent was evaporated to give 1.12 g of a blue powder.

Mass spectroscopy showed 1 peak at 231 Da (molecular ion [NiEDDA —1H]').

Sample preparation. Protein expression, purification and reconstitution into the

liposomes were described in chapter 2.

Spin labeling. The purified single cysteine mutant of COX-2 (4 mg/ml) was incubated
in a 20-fold molar excess of MTSSL overnight at 4°C. Unreacted spin labels were
removed by dialysis. Dialysis was performed at 4°C against 4 liter of liposome buffer (25
mM Tris-Cl and 50 mM KCl pH=7.4) using Slide-A-Lyzer® with 10,000 molecular
weight cut off (Pierce, Rockford, Illinois). The dialysis buffer was changed 4 times every

6 h. The spin labeled protein was concentrated to 15-20 mg/ ml.

CW EPR power saturation measurements: Continuos wave (CW) EPR spectroscopy

was performed on a Bruker ESP 300E spectrometer at room temperature (23:2 C)

103

 

equipped with a two loop one gap resonator (Medical Advances, Milwaukee, Wisconsin).
Samples were loaded into a gas permeable TPX capillary (J agmar Ltd., Krakpw, Poland)
with total volume of 4 ul. Power saturation experiments were carried out using 2 Gauss
modulation amplitude and varying microwave power (typically 0.1-36 mW). The scan
range of all spectra was 100 Gauss and the final spectra was obtained by signal averaging
8, 16 or 32 scans using the digitizer. The peak to peak amplitude of the first derivative

mizo resonance line (A) was measured and plotted against the square root of the incident

microwave power. The data points were then fit to the expression:

[Pl/2

8
1+ 2‘8—1 P J
I ( )/P1/2

 

Appm) =

using Microcal Origin ® (Microcal Software, Inc., Northampton, Massachusetts). Here I
is a scaling factor, P122 is the microwave power required to reduce the resonance
amplitude to half of its unsaturated value, and e is a measure of the homogeneity of the
saturation of the resonance. In this fit, I, e and PW are adjustable parameters and yield a
characteristic P122. Values for Pm are then generated for each sample under three
different conditions: (1) equilibrated with nitrogen; (2) equilibrated with 20% oxygen
(air); and (3) equilibrated with nitrogen in the presence of 50 mM NiEDDA. The
proteoliposome samples were added to a freeze-dried aliquot of NiEDDA. From these

values of Pin, a collision parameter for O2 (1'I°"’) was determined according to

104

”.2‘ v... u .- .-

 

 

Currant Opnon n Snuclmll 8m.

Figure 3.12. Model helices and idealized behavior of the accessibility parameter. Taken

from (23). Copyright permission was obtained.

105

’1,2(02) =[Pl/2(02)_Pl/2(N2)]/A3pp
P ’1/2(DPPH) Pl [2(DPPH)/ABpp(DPPH)

 

r10” =

ABpp is the peak-to-peak line width for the central resonance of the EPR spectra. A
similar expression can be written for HN’EDDA, the collision parameter for NiEDDA.
P1/2(DPPH) and ABpp(DPPH) were obtained for a crystal sample of DPPH (0t, 01’-

diphenyl-B-picrylhydrazyl).

3.8. Results

The technique of site-directed spin labeling requires a template cDNA that does
not contain reactive cysteines. A “cysteine-less” template was prepared by mutating
cysteins at positions 229 and 526 in the human histidine-tagged enzyme COX-2. The
cysteine-less template retained 70% of the cyclooxygenase and peroxidase activities of
the native Histidine-tagged COX-2, indicating that cysteins 299 and 526 are not required
for catalysis and that these substitutions only minimally distort the structure of the
protein.

Four helices (A—D) form the membrane-binding domain in the mouth of the
cyclooxygenase active sites. Helices A, B and C were studied by site directed spin
labeling in this chapter. By using the cysteine-less template, 25 single cysteine mutants
were constructed (Table 3.1). The histidine-tagged mutant proteins were expressed in sf-

21 cells using baculovirus constructs and purified using nickel affinity chromatography

as described in detail in chapter 2. All of the single cysteine mutants retained significant

106

cyclooxygenase activities (IO-70% of the native). Conjugation of the cysteines with the
spin label MTSSL had no effect on cyclooxygenase activity. These results demonstrated
that the cysteine substitutions and subsequent conjugation with the spin label did not
grossly disrupt the native structure of the enzyme.

Each of the 25 single mutant proteins were purified, spin labeled and concentrated to =20
mg/ml (z280 uM) and incorporated into preformed liposomes. The unincorporated
protein was separated from the proteoliposomes using a Ficoll density gradient. The
proteoliposomes were concentrated using ultracentrifugation. The proteoliposomes were
loaded into a gas permeable TPX capillary with total volume of 4 11.1. Continuous wave
(CW) EPR spectroscopy was performed on a Bruker ESP 300E spectrometer at room
temperature (23i2 °C) equipped with a two loop one gap resonator. Power saturation
experiments were carried out using 2 Gauss modulation amplitude and varying
microwave power (typically 0.1-36 mW). The scan-range of all spectra was 100 Gauss.
The noise level of each spectrum was filtered using Bruker’s WinEPR software. (See
Appendix A). The peak to peak amplitude of the first derivative m1=o resonance line (A)

was measured and plotted against the square root of the incident microwave power. The

data points then were fit to the following equation:

1/2

“'8
_ -8_ P
pp(0).],113 [1+(2 1)—P ]

Values for P ”2 are then generated for three different samples of each mutant: 1) The spin

labeled proteoliposome equilibrated with nitrogen; 2) The spin labeled proteoliposome

107

Table 3.1. Relative cyclooxygenase activities of the single substituted hCOX-2 cysteine

 

mutants.
Samples Activity before labeling “ Activity after labeling
Native COX-2 100 NAc
Cys'b 70 NA
E58C 50 50
F59C 50 50
L6OC 70 70
T6lC 60 60
R62C 50 50
I63C 70 70
K64C 7O 70
L65C 70 70
F66C 50 50
L67C 40 4O
N72C 60 60
T73C 40 40
V74C 60 60
H75C 70 70
Y76C 70 70
I77C 60 60
L78C 30 30

G83C 35 35

108

 

F84C 50 50

W85C 10 10
N 86C 50 50
V87C 30 30
V88C 30 30
N89C 30 30
N90C 20 20

 

a Cyclooxygenase activity determined by oxygen electrode as described previously on
purified enzyme before and after conjugation with MTSSL. All activities are normalized
to the specific activity of the His-Tagged native COX-2.

b Cys' : human COX-2 mutant in which cysteines at positions 299 and 526 were
substituted with serines.

c NA: Not Applicable

109

equilibrated with 20% oxygen (air); 3) The spin labeled proteoliposome in the presence
of 50 mM NiEDDA equilibrated with nitrogen. In order to equilibrate the samples with
the gas, the loop-gap resonator was connected to nitrogen and air cylinders. The sample
was purged with the gas for at least 20 minutes before performing the measurements. The

resonator remained connected to the gas during the measurements.

Power Saturation Experiments:

In the absence of any relaxer, the energy levels are saturated and homogeneity
factor (8) is close to 1.5. (See section 3.6.1 for detail theory discussion about power
saturation curves). The presence of other paramagnetic species can affect P1/2. For
example, the presence of a nearby fast-relaxing transition metal ion will enhance the rate
at which MTSSL radical electrons relax from the high-energy to low-energy state, thus
increasing the microwave power that can be absorbed before the transition saturates. On
the other hand very high microwave powers distort the line shapes and the data for the
peak to peak amplitude of the signals obtained at very high powers are not useful.
Therefore in the presence of relaxers the power saturation curves don’t roll over
significantly. The homogeneity factor (8) for power saturation curves in the presence of
relaxers, especially the metal relaxer is about 0.5. However the value of Pu2 is not
significantly affected by the homogeneity of the saturation curves because what

determines the Pm value of the curve is the initial slope of the curve.

110

EPR spectra: Mobility of the nitroxide side chains

Figures 3.13-15 show the room temperature EPR spectra of nitroxide spin label
side chain (R1) at sites 58—67, 72-78 and 83-90 in 30% sucrose. 30% w/v sucrose is used
in order to reduce the rotary diffusion rate of the protein. Sucrose at this concentration
does not affect the rotational mobility of the side chain.

The shape of spin label EPR spectrum reflects the degree of reorientational
motion of the nitroxide side-chain. This residual motion depends on the secondary and
tertiary structure of the spin label binding site and its vicinity. Three distinct types of
structural classes can be assigned based on mobility. The first class consists of sites
where the spectral line shape indicates rapid motions. These sites occur generally in loops
or at the solvent-exposed surfaces of helices where R1 is not in contact with protein
tertiary structure. Weak interaction between the nitroxide and the rest of the protein
results in the high degree of mobility. In this case, the apparent hyperfine splitting and the
line width are small, as found for the spectra of I63R1 and F66Rl in helix A, V87R1 in
helix C and E58R1 in the loop which connects the EGF domain to helix A. In turn, if
residual motion is restricted due to strong interaction of the nitroxide group with
neighboring side-chains or backbone atoms, the apparent hyperfine splitting and the line
width are increased. This is obvious for the spectra of V74R1 and I77Rl in helix B and
F84Rl, W85R1 and N86R1 in helix C, which clearly show the restricted motion of
nitroxide spin label. The line shapes in this second class are characterized by the presence
of resolved outer splittings indicating partial averaging of the A tensor. The broad central
line is also consistent with partial averaging of the g tensor. None of the residues in helix

A shows a restricted nitroxide motion. This reveals a loosely packed environment around

111

helix A, which is in agreement with the X-ray crystallographic structure (I). The third
class shows an intermediate degree of averaging which occurs at sites: F59R1, L60R1,
T61R1, R62Rl, K64R1, L65R1 and L67R1 in helix A, T73Rl, H75R1, Y76R1 and
I78Rl in helix B and G83R1, V88R1, N89R1 and N90R1 in Helix C.

In order to extract dynamics data from an EPR spectrum, a spectral parameter that

is sensitive to changes in mobility is needed. The inverse central line widths (AB'l also
known as AH") are widely used as indicators of mobility. However AB'l measurements
can be erroneous due to the sharp features of the central peak (m1=0) in proteins which

have fast rotational motion in solutions. 30% w/v sucrose is used in order to reduce the

rotary diffusion rate of the protein. But even in 30% w/v sucrose solution the AB"1

measurement is not very accurate. The peak to peak amplitude of the mi=o of a
normalized spectra (normalized to the total number of spins) is used as a more sensitive
indicator of mobility. In a normalized spectrum if spectral lineshape is narrow then the
peak to peak amplitude is large and if the lineshape is broad then the peak to peak
amplitude is small.

For this technique the noise level in spectrum is filtered using Bruker’s WinEPR
software and the spectrum is normalized to the total number of spin (area underneath the
spectrum) using Origin software. The peak to peak amplitudes of the normalized spectra
are measured and plotted as a function of the amino acid sequence (Figures 3.19-21 B). A

periodicity of 3.6 is observed in these plots which is consistent with an (ii-helical

configuration of the backbone.

112

Accessibilities to Oxygen and Nickel Ethylene Diarnine Diacetic Acid (NiEDDA)

The solvent exposure of a nitroxide spin label attached to a cysteine residue in a
protein can be measured by examining its collision frequency with another paramagnetic
reagent such as NiEDDA in solution. The collision frequency of a nitroxide probe with
paramagnetic relaxers depends on the product of its translational diffusion coefficient and
its local concentration. Molecular oxygen and water-soluble N iEDDA have been found to
be ideally suited because of the sizes and solubility properties. Molecular oxygen has a
low concentration in the tightly packed protein interior, whereas both the solubility and
the diffusion coefficient are high in the bilayer. Therefore, the collision frequency
between oxygen and spin labels in the interior of the protein is low, whereas it is high, if
the spin label side-chain faces the bilayer. The relative collision frequencies were
determined by the method of continuos wave power saturation. The increase in Pm that
occurs in the presence of NiEDDA or air is converted into a dimensionless accessibility
parameter 1'1 (31 ). II is proportional to the collision frequency of the nitroxide with the
respective reagent.

The accessibility parameter values in the presence of NiEDDA or with the sample
in equilibrium with air were plotted as a function of sequence position (See Table 3.2 and
Figures 3.19-21 A). The highest values of l'INiEDDA and hence the highest collision
frequency with NiEDDA is revealed for the side-chains attached to positions L60R1,
T61R1, K64Rl in helix A (Figure 3.19), H75R1, Y76Rl in helix B (Figure 3.20) and
V87Rl in helix C (Figure 3.21). The residues showing the largest collision frequency

mu st be oriented towards the aqueous phase, while those with low values of Hun-mm face

113

the protein or the bilayer. The accessibility parameter Hoxygen and the width of the

central absorption peak (or the normalized amplitude of the center peak (m1=0)) allow
discrimination between these cases. Low values of normalized amplitude of the center

peak (m1=0) and low oxygen accessibility is typical of the location of the nitroxide side-

chain buried inside the protein. The oxygen accessibility is lowest for positions F59R1,
K62Rl in helix A, V74R1 in helix B and G83R1, F84R1, W85R1, V88R1, N89R1 and

N90R1 in helix C and all these positions have a relative low values of normalized

amplitude of the center peak (mI=0)- On the other hand, high motional freedom and thus

narrow EPR lines combined with high oxygen accessibility are characteristic of the
nitroxide in contact with the lipid bilayer. This is the case for E58Rl, F66Rl and L67R1,
which show high accessibility to air and lower accessibility to NiEDDA and have narrow

EPR lineshapes.
As it is shown in Figures 3.19-21 A and B, there is a qualitative agreement in the
patterns of II and the normalized amplitude. The solvent —accessible nitroxides, which
display high 1'1 values, are likely to enjoy high mobility, which in turn will lead to high
values of the normalized amplitude of m1=o peak. This is obvious for L60R1, T61R1,
I63Rl and K64R1 in Helix A, H75R1 and Y76R1 in Helix B and V87R1 in Helix C.

The buried sites such as F59R1, R62Rl and L65R1 in helix A; V74R1 and I77R1

in helix B and W85R1 and N89R1 in helix C display low 1'1 values and low values of the

normalized amplitude of m1=0 peak.

114

 

 

 

Table 3.2. Collisional parameters for the membrane binding domain mutants.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Mutants P1/2(N2) P1/2(Air) P112 (50 mM NiEDDA) ABM, 1'I(Air) 11(NiEDDA)

E580 4.69 13.23 11.21 3.05 0.281 0.214
F590 4.61 6.93 12.61 4.64 0.0502 0.173
L60C 4.72 11.32 14.11 3.48 0.190 0.270
T61C 5.38 10.42 15.59 3.92 0.129 0.261
R62C 4.29 8.07 8.80 4.50 0.0842 0.100
1636 4.52 9.20 11.61 4.06 0.115 0.175
K64C 4.61 9.48 12.25 3.77 0.129 0.203
L656 4.78 9.34 12.55 3.92 0.117 0.199
F66C 4.09 8.82 7.15 3.48 0.136 0.0880
L67C 4.42 9.60 7.28 3.05 0.170 0.0940
Mutants P1,2(N2) P1,2(Air) P172150 mM NiEDDA) ABM, 11(Air) II(NiEDDA)

N726 3.73 7.45 6.67 4.29 0.0869 0.0686
T736 4.24 7.43 11.75 3.85 0.0832 0.196
V74C 5.77 7.95 13.51 4.62 0.0472 0.168
H756 4.58 8.03 10.12 2.86 0.121 0.194
Y76C 4.58 8.18 13.33 4.62 0.0781 0.190
I77C 5.38 8.54 12.62 4.07 0.0708 0.178
L786 6.02 8.85 13.14 3.63 0.0784 0.196
Mutants P,,2(N2) P1,2(Air) 13,,2 (50 mM NiEDDA) we“, 11(Air) II(NiEDDA)

G83C 3.99 6.00 10.76 4.62 0.0435 0.147
F84C 3.66 5.88 9.29 4.95 0.0448 0.114
W856 4.37 6.88 11.37 5.17 0.0485 0.136
N866 3.76 6.90 11.71 4.18 0.0754 0.191
V876 3.43 8.55 12.13 3.08 0.167 0.283
V886 5.64 8.17 10.02 4.4 0.0576 0.0997
N896 3.11 6.26 5.58 4.51 0.0480 0.0551
N906 4.85 7.57 11.28 4.07 0.0670 0.158

 

 

 

 

 

 

 

P’1/2 (DPPH) =9.96

115

 

 

 

The line shapes for proteoliposomes are significantly broader than proteins in
solution (compare Figures 3.13-15 and 3.16-18). This is due to increase of the rotational
correlation time (TR) for the proteoliposome because the liposomal structures are

significantly larger than proteins and rotate much slower in solutions than solubilized

proteins.

116

3.13. CW-EPR spectra of COX-2 helix A spin label mutants in 30% w/v sucrose

117

Figure 3.14. CW-EPR spectra of Cox-2 helix B spin label mutants in 30% w/v
sucrose.

119

Figure 3.15. CW-EPR spectra of Cox-2 helix C spin label mutants in 30% w/v
sucrose.

121

Figure 3.16. CW-EPR spectra of Cox-2 helix A spin label mutants
incorporated into liposome.

123

Figure 3.17. CW-EPR spectra of Cox-2 helix B spin label mutants
incorporated into liposomes.

125

 

Figure 3.18. CW-EPR spectra of Cox-2 helix C spin label mutants incorporated into

liposomes

127

 

i
a

 

 

a g a
E F ?

 

 

—l— Air
030- -o— 50 mM NiEDDA

 

 

 

Amino acid sequence

.0
.5
t.

O
_.s
N

.0
.1
o

0.08

Normalized Amplitude ml 0

58 59 60 61 62 63 64 65 66 67
Amino Acid Sequence

Figure 3.19. A) Plot of relative accessibility of Helix A side chains.
B) Plot of mobility of Helix A side chains.

129

—l— Air
We 7 50 mM NiEDDA

 

 

Amino acid sequence

0
.4
O

8

0.02

Normalized Amplitude ml:O

 

73 74 75 76 77
Amino Acid Sequence

Figure 3.20. A) Plot of relative accessibility of Helix B side chains.
B) Plot of mobility Helix B side chains.

130

Normalized Amplitude ml 0

—l— Air
407 60 mM NiEDDA

 

I I I I I

I T T
82 83 84 85 86 87 88 89 90 91
Amino acid sequence

 

83 e4 85 86 87 88 89 90
Amino Acid Sequence

Figure 3.21. A) Plot of relative accessibility of Helix C side chains.
B) Plot of mobility of Helix C side chains.

131

3.9. Discussion

Figures 3.19-21 show the sequence-specific variation in Hoxygen along helices A-C.
Consistent with an ct-helical configuration of the backbone, a periodicity of 3.6 is
observed as demonstrated from the superimposed sinusoid. The oscillatory behavior
continues uninterrupted along the sequence although a shift in phase is detected in the last
turn of helix A. I'Ioxygen and I'lNiEDDA both have the same period and phase. This reveals
that one side of the protein is more exposed to the aqueous environment, while the other
side is exposed to the interior part of the protein (See Figure 3.12).

Values of l'INiEDDA are very low throughout the Helices A, B and C being a
maximum 2: 0.3 .The values of Hoxygen are also very low throughout the three helices
being a maximum z 0.2. For reference, in 20 mM NiEDDA, R1 sites exposed to water or
lipid or buried in a protein interior have typical IINiEDDA values of z 6, <01, and 0.01,
respectively (32). In these experiments 50 mM NiEDDA was used so it is expected to get
HNiEDDA values which are even higher than the above mentioned values. Also for
reference, R1 at buried sites, water-exposed sites, and lipid-exposed sites have Home"
210.05, 0.45-0.55 and 1, respectively. These low values of accessibility for both I'lNiEDDA
and Hoxygen indicate that both polar and non-polar paramagnetic relaxers do not access the
aminoacids located in the membrane-binding domain. One possibility is that the
membrane-binding domain of hCOX-2 is sandwiched between the main body of the 144
kDa hCOX-2 dimer and the lipid bilayer. The spatial organization of the dimer around

the helices in the membrane-binding domain results in an excluded volume that lowers

132

the diffusion coefficient of oxygen. Our data suggests that the side of the helix, which is
accessible to water, is not completely immersed in water. If that was the case, we would

have expected to see a 111412131311 higher than 6 for 50 mM NiEDDA. Instead the maximum
IINiEDDA was around 0.3. This reveals that NiEDDA molecules can not diffuse freely. One

possibility is that the water-exposed site of the helix is in a close contact with the
phosphate head groups at the bilayer interface. The concentration of NiEDDA is much
lower in the bilayer interface than in the aqueous phase and this is the reason of such low
TIMEDDA values.

E58R1 which is the last amino acid in the loop connecting the EGF domain to
helix A has a higher accessibility to oxygen than to NiEDDA. This suggests that this
amino acid participates in anchoring the protein to the lipid bilayers. The last two amino
acids in helix A F66Rl and L67R1 have higher accessibility to oxygen that to NiEDDA
suggest that they may also play a role in membrane binding.

The crystal structure (33) shows that in helix A side chains for amino acids F59,
R62 and F66 would be exposed to the surface of the of the protein and T61 would be
exposed to the interior part of the protein although, our accessibility data and also the
data from line shapes reveal that F59, R62 are exposed to the interior part of the protein
and T61 is exposed to surface of the protein (See figures 3.22-23).

The crystal structure shows that in helix B the side chain for amino acid H75
would be exposed to the interior part of the protein although, our accessibility data and
also the data from the line shapes reveal that H75C is exposed to the surface of the
protein (see Figures 3.22-23). The failure to observe oscillation in the O2 or NiEDDA

accessibility data are indicative of the helix B environment that is uniformly accessible to

133

the NiEDDA but interior to the protein. Accessibility data for 02 are also uniform but
show that helix B is inaccessible to this relaxation agent.

The crystal structure shows that in helix C side chains for amino acids G83, W85
and N89 would be exposed to the interior part of the protein and V87 would be exposed
to the surface of protein. Our accessibility data and also the data from line shapes reveal
the same thing. (see Figures 3.22-23).

As it is shown in Figures 3.22 and 23, helices A and B are rotated anti clockwise
by several degrees in liposome compare with their structure in crystal. Helix C maintains

same structure in both liposome and crystal.

134

 

3.22. The crystal structure of the membrane-binding domain of hCOX-2.
This picture was made by Dr. Anne Mulichak.

Images in this dissertation are presented in color.

135

 

3.23. The structure of the membrane-binding domain of hCOX-2 that was proposed by
site directed spin labeling.

* Mobile and solvent exposed amino acids

# The lipid exposed amino acids

This picture was made by Dr. Anne Mulichak.

Images in this dissertation are presented in color.

136

3.10. References

10.

ll.

12.

13.

14.

Picot, D., L011, P. J ., and Garavito, R. M. (1994) Nature 367, 243-249.

Wendt, K. U., Lenhait, A., and Schulz, G. E. (1999) Journal of Molecular
Biology 286, 175-187.

Hubbell, W. L., and Altenbach, C. (1994) Current Opinion in Structural Biology
4, 566-573.

Hubbell, W. L., Cafiso, D. S., and Altenbach, C. (2000) Nature Structural Biology
7, 735-739.

Read, R. J., and Wemmer, D. E. (1999) Current Opinion in Structural Biology 9,
591-593.

Drenth, J. P. H., Boom, B. W., Toonstra, J ., and Vanderrneer, J. W. M. (1994)
Archives of Dermatology 130, 59-65.

Georgopoulos, C., Ang, D., Wall, D., Wu, B., vanderVies, S., Richardson, A.,
Deloche, O., Kelley, W., Keppel, F ., Drijard, L., Zuber, U., Liberek, K.,
Wawrzynow, A., Wojtkowiak, D., Zylicz, M., Szyperski, T., Pellacchia, M., and
Wutrich, K. (1996) Molecular Biology of the Cell 7, 1954-1954.

Michel, J. P. (2003) Wiener Klinische Wochenschrift 115, 273-274.

Ketchem, R. R., Hu, W., and Cross, T. A. (1993) Science 261, 1457-1460.

Grigorieff, N., Ceska, T. A., Downing, K. H., Baldwin, J. M., and Henderson, R.
(1996) Journal of Molecular Biology 25 9, 393-421.

Subramaniam, S., Gerstein, M., Oesterhelt, D., and Henderson, R. (1993) Embo
Journal 12, 1-8.

Lewis, V. A., Hynes, G. M., Dong, Z., Saibil, H., and Willison, K. (1992) Nature
358, 249-252.

Braig, K., Otwinowski, Z., Hegde, R., Boisvert, D. C., Joachimiak, A., Horwich,
A. L., and Sigler, P. B. (1994) Nature 3 71, 578-586.

Engel, A., Hayerhartl, M. K., Goldie, K. N., Pfeifer, G., Hegerl, R., Muller, S.,
Dasilva, A. C. R., Baumeister, W., and Hart], F. U. (1995) Science 269, 832-836.

137

 

15.

l6.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

Valencia, A., Hubbard, T. J., Muga, A., Banuelos, S., Llorca, O., Carrascosa, J.
L., and Valpuesta, J. M. (1995) Proteins-Structure Function and Genetics 22,
199-209.

Cantor, J. M., and Haskins, K. (2003) Faseb Journal 17, C271-C272.

Gross, A., and Hubbell, W. (1998) Biophysical Journal 74, A44-A44.

Postnikova, G. B. (1996) Biochemistry-Moscow 61, 679-693.

Langen, R., Oh, K. J., Cascio, D., and Hubbell, W. L. (2000) Biochemistry 39,
8396-8405.

Weil J.A., B. J. R., Wertz J.E. (1994) Electron Paramagnetic Resonance, John
Wiley & Sons Inc., New York.

Andrews, S. H. (1999) pp 170, Universityof Virginia, Charlottesville.
Froncisz, W., and Hyde, J. S. (1982) Journal of Magnetic Resonance 47, 515-521.

Hubbell, W. L., Gross, A., Langen, R., and Lietzow, M. A. (1998) Current
Opinion in Structural Biology 8, 649-656.

Berliner, L. J. (1998) Biological Magnetic Resonance, Vol. 14, Plenum Press,
New York.

Salwinski, L. (1998) pp 269, UCLA, Los Angles.

Berliner, L. J. (1998) Spin Labeling the next millennium, Vol. 14, Plenum Press,
New York.

Boitrud, P. D., and Berliner, L. J. (1996) Journal of Protein Chemistry 15, 231-
242.

Likhtenstein, G. I., Bogatyrenko, V. R., and Kulikov, A. V. (1993) Applied
Magnetic Resonance 4, 513-521.

Altenbach, C., Froncisz, W., Hyde, J. S., and Hubbell, W. L. (1989) Biophysical
Jouma156, 1183-1191.

Altenbach, C., Flitsch, S. L., Khorana, H. G., and Hubbell, W. L. (1989)
Biochemistry 28, 7806-7812.

Altenbach, C., Marti, T., Khorana, H. G., and Hubbell, W. L. (1990) Science 248,
1088-1092.

138

32.

33.

Gross, A., Columbus, L., Hideg, K., Altenbach, C., and Hubbell, W. L. (1999)
Biochemistry 38, 10324-10335.

Kurumbail, R. G., Stevens, A. M., Gierse, J. K., McDonald, J. J., Stegeman, R.

A., Pak, J. Y., Gildehaus, D., Miyashiro, J. M., Penning, T. D., Seibert, K.,
Isakson, P. C., and Stallings, W. C. (1996) Nature 384, 644-648.

139

Chapter 4

Double Site Directed Spin Labeling Studies

140

4.1. Abstract

The goal of the research described in this chapter is to study conformational
changes in the mouth of the cyclooxygenase channel. Formation of the NSAID-COX
complexes has been proposed to result in conformational changes in COX (1). This
hypothesis is supported by studies that show that time-dependent inhibitors protect from
proteolytic cleavage at Arg-277, an amino acid located on a peptide loop distant from the
inhibitor binding pocket (2-5). Heme, which binds at a site separate from the NSAID
binding pocket, also protected COX from proteolysis, indicating a global flexibility in the
COX isozymes. Despite this circumstantial evidence that the COXs undergo
conformational changes, comparison of the X-ray crystal structure of the COX-2
holoenzyme with the structure of COX-1 and —2 co-crystallized with various time-
dependent NSAIDs reveals only minor structural differences (6). To investigate this
controversy, I have examined arachidonic acid, NSAIDs, and heme induced
conformation changes of hCOX-2 using site-directed spin labeling and electron
paramagnetic resonance spectroscopy.

Ten mutants located in the helices that form the opening of the cyclooxygenase
active site and the membrane-binding domain of hCOX-2 were constructed. Each mutant
contained the I98C substitution in helix D, and one of five amino acids substitutions in
opposing helix B (T73C, V74C, H75C, Y76C and I77C) or one of the five amino acid
substitutions in adjacent helix C (F84C, W85C, N86C, V87C and V88C). The dipolar
broadenings of the spectra in the presence and absence of NSAIDs, in the presence and

absence of heme and in the presence and absence of arachidonic acid were compared in

141

this chapter a brief review about various interspin distance measurement techniques will
be given and then the Fourier convolution/deconvolution technique will be explained in

detail. Finally, the experiments and the results will be discussed.

4.2. Dipole-Dipole Interaction

The two nitroxide spin labels attached to a biomolecule e. g. a protein interact with
each other through the electron-electron dipolar interaction (Figure 4.1).

The biradical spin Hamiltonian is expressed as:

2R3

 

where h is the Planck’s constant, 7e is the electron gyromagnetic ratio, J is the exchange

integral, 0 is the angle between the interspin vector and magnetic field 2 direction, (21 and
02 represent the resonance offsets for electron spins 1 and 2, and S12 and S22 are the z
direction components of the electron spin angular momentum vectrors S, and S2 ,
respectively. The first two terms of equation 4.1 give rise to the EPR spectrum in the
absence of the electron-electron dipolar interaction. However, the existence of the
electron-electron dipolar interaction perturbs the non- interacting EPR spectrum to give

rise to an overall spectral broadening. The electron-electron dipolar interaction is given as

142

Dipolar EPR

 

Figure 4.1. Dipolar coupling between two spin labels. Taken from (9).

143

the third term of the equation 4.1. The electron-electron dipolar interaction has an R'3

dependence on the interspin distance R. Because spin exchange interaction decays
exponentially as a function of R, the electron-electron dipolar interaction becomes
dominant when R is larger than 7 A unless there is a through bond interaction. The short-
range spin exchange interaction is not considered in Fourier convolution deconvolution
(F CD) analysis.

When nitroxides are attached to a biomolecule such as a protein, the flexible
nitroxide side chain allows significant degrees of freedom for the magnetic tensors at a
given 9. This analysis must be included in the analysis of the spin Hamiltonian. To avoid
complications arising from incomplete time averaging of 0 the EPR spectra must be taken
under a frozen conditions.

Under these conditions the EPR absorption lines at all spectral positions in the

magnetic field for a given 0 are split by 2B (in Gauss),

 

2
2B=(3/2)gefi(3Cos 0-1)

(4.2)
R3

where ge is the free electron g factor, B is the electron Bohr magneton. Here it is assumed

that the effect of the g anisotropy in negligible at 9 GHz (9).

144

4.3. Determination of Electron-Electron Distance from Dipolar Interaction between

Two Spin Labels Measured by EPR in Immobilized State

The task of methodology development is to maximize sensitivity of the spectral
display to the dipolar interaction and to separate the dipolar interaction from other
features of the spectra and thereby extract the distance information from the spectrum. A
variety of techniques have been reported for various types of samples and magnitude of

dipolar interaction as outlined in the following paragraphs and summarized in Table 4.1

(9)-

4.3.1. Dipolar Splitting is Significant Compared with Linewidths of Spectra for

Corresponding Monoradical

When the dipolar splitting (D) is significant relative to the linewidths of the
signals in the CW spectra, analysis of the lineshape changes can be used to determine the
magnitude of the dipolar interaction. A variety of techniques have been used. Each of
these techniques will be briefly discussed in the following sections. The convolution of
the lineshapes will be discussed in detail. This technique was used to measure the

distances between the two spin labels in various hCOX-2 mutants.

145

Table 4.1. Summary of methods for determining electron-electron distances.

 

 

Method aRMaxCA)
Simulation of CW lineshape 20-25
Deconvolution of CW lineshape 25

Relative intensity of half-field transitions 12

Ratio of peak heights 22
2+1 pulse sequence pulsed ELDOR 60
4-pulse DEER 80
Double—quantum coherence 80

 

a RMax is the maximum values of r that has been proposed as measurable by a particular
technique .

146

4.3.1.1. Analysis of Spectral Lineshapes by Computer Simulation

In principle, computer modeling of the complete spectral lineshape can define
both the interspin distance and the orientation of the interspin vector relative to the
magnetic axes of both paramagnetic centers, as well as separating dipolar and exchange
interactions. The simulated spectral lineshapes depend upon the degree of ordering of the
relative orientations of the magnetic axes, therefore, information can be obtained

concerning the range of conformations present in the sample (10).

4.3.1.2. Lineshape Deconvolution

This technique is extensively explained in chapter 4.4.

4.3.1.3. Relative Intensity of the Half- Field Transition

Anisotropic interaction between two spins (dipolar interaction and anisotropic

exchange) shifts the triplet state m8 = i 1 energy levels relative to the rriS = 0 level, and

causes the normally forbidden transition probability between the mS = -1 and m5 = +1

levels to become non-zero. This transition occurs at half the magnetic field required for
the allowed transitions (at constant microwave frequency), and hence is called the “half-
field” transitions. If the anisotropic exchange is negligible, and if r > 4-5 A, the resonant
field for the half-field transition is independent of the values of J (exchange interaction)

and r (distance). The ratio of the integrated intensity of the half-field transition to the

147

integrated intensity of the allowed transition is a function of r and independent of J.
Therefore, isotropic and anisotropic contributions can be separated by measuring the

relative intensity of the half-field transition (11, 12).

4.3.1.4. Lineshapes Distortion Determined by the Ratio of Peak Heights

Broadening of the frozen CW EPR solution of a nitroxide spin label due to dipolar
interaction with a second nitroxide changes the relative intensities of the center and outer
lines of the spectrum. Kokorin et al (13, 14) proposed that a ratio of peak heights could

be used to estimate the distance between two nitroxide radicals.

4.3.2. Dipolar Interaction is Small Compared with Linewidths of Spectra for

Corresponding Monoradical

CW techniques for measuring dipolar interaction depend upon the observation of
significant broadening of the lineshape due to the interaction. Smaller dipolar interactions
can be measured by pulsed techniques than by CW techniques. Thus, the pulse
techniques tend to be more useful at longer distances than are accessible by CW

techniques.

148

4.3.2.1. 3-Pulse ELDOR (Electron-Electron Double Resonance)

A spin echo is created by a two-pulse sequence at one microwave frequency. The
timing of a pulse at a second microwave frequency is varied. Examples of this technique

have been described by Tsvetkov and co-worker (15) and are reviewed in Milov et a1

(16).

4.3.2.2. 2+1 Sequence

Raitsiming and co-workers proposed a sequence that is similar to 3-pulse
ELDOR, but with all three pulses at the same microwave frequency (17). This is a
denoted as the “2+1” sequence. The upper limit of distances between nitroxides

accessible by this technique is about 60 A.

4.3.2.3. 4-Pulse DEER (Double Electron-Electron Resonance)

In the 2+1 pulse sequence and in 3-pulse ELDOR experiments there is an inherent
experimental dead time that limits the magnitude of the dipolar interaction that can be
characterized. Martin and Pannier (18) designed a 4-pulse experiment that eliminated that

dead time. The interspin distances up to 80 A can be measured using this technique.

149

4.3.2.4. Double—Quantum Coherence

Allowed double-quantum coherence (DQC) can be generated in ordered or
disordered sample containing pairs of radicals and random distributions or radicals. Pake
doublets obtained from DQC can be used to determine distances within pairs of radicals
and the decay constants yield concentrations in a distribution. The upper limit of

distances between nitroxides accessible by this technique is about 80 A (19, 20).

4.4. Fourier Convolution Deconvolution of the Dipolar Coupling in the Spin-

Labeled EPR Spectra

4.4.1. Theoretical Background

In the following three sections, a brief mathematical background will be provided
which is necessary for understanding the Fourier convolution deconvolution of the

dipolar coupling in the EPR spectra.

4.4.1.1. Convolution

Mathematically, a convolution is defined as the integral over all space of one
function at x times another function at u-x. The integration is taken over the variable x
(which may be a 1D or 3D variable), typically from minus infinity to infinity over all the
dimensions. So the convolution is a fimction of a new variable 11, as shown in the

following equations. The cross in a circle is used to indicate the convolution operation.

150

C (u)=f (X)®g(X)= l f (X)g(u-X)dx (43)
Space

C (u)=g(X)®f (X)= l g(X)f(u-X)dx (4-4)
Space

It doesn't matter which function is taken first, i.e. the convolution operation is
commutative. The convolution of two functions is complex. To convolute two functions,
the first function must be superimposed on the second at every possible position, and
multiplied by the value of the second function at that point. The convolution is the sum of
all of these superpositions. Figure 4.2 shows how one can thinks about the convolution,
as giving a weighted sum of shifted copies of one function: the weights are given by the
function value of the second function at the shift vector. If one of the functions is
unimodal (has one peak), as in Figure 4.2, the other function will be shifted by a vector
equivalent to the position of the peak, and smeared out by an amount that depends on
how sharp the peak is. But if we switch the roles of the two functions, we see that the

bimodal fimction g has doubled the peaks of the unimodal function f.

4.4.1.2. Fourier Transform

The Fourier transform is a mathematical function that performs a conversion
between real space and reciprocal space (See Figure 4.3).

The Fourier transform of f ((0) is defined as:

151

 

 

 

 

91x1
J//\§b/\\L
-2 0 2 4 5"
gtu-2)
iAK
. U
-2 0 2 4 6
9(u-Z.51
2
—2 o 2 4 5“
gin-3)
-2 0 2 4 6“
fixigtu-xl

WI.)
-2 0 2 4 6

Figure 4.2. Convolution of f (x) <81 g (x)

fix)

OODO
mammo—

 

-2 O 2 4 6

ft2igtu-21

 

-20246

f(2.5)g(u-2.5)

 

-20245u

£13IQIU-3l

 

-20245"

Convolution
2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

,. . w 1.... .. ., r , 4 q
i ii i : so;- -:
05 ~ 5 I I
I ‘840é €
5,00- - -30} 1
a . O b ‘
. L .1
1 E20} -j
-0.5 - - I I
l- 1 u d
. , lO- ~
L 4 E L i
-1.0 ‘ i 0"“ ‘rm" ‘+
10 20 30 40 50 60 70 0.0 0.5 1.0 1.5 2.0

Time (5) Frequency (Hz)
2’... 14.4.7....,....,....r....,....1 60_ . . T r I . 4 . . , . . . 4 , . . . . ‘
1 50} -j
ll - Z I
l : 7340; :1
l . 5‘0 I I
a 1 -- : :
.500 ‘ 32,3307 1
U) ’ 1 E E :1
Z 1 20¢ .1
-l - 4 C i
I ‘ 10:- -j
Z 1
_2 Lu.i...A1411. .11. 11.1.11.1111. I 1 . I . 1 1.. I 1 1
0 10 20 30 40 50 60 70 0.0 0.5 1.0 1.5 2.0

Time (sec) Frequency (Hz)

Figure 4.3. Fourier transformation of a signal

The picture was kindly made by Professor Michael Romalis (Princeton University) using
GenPlot software.

153

F (to)= exp(27ric0t) f (t)dco (4.5)

427: 2
where 0)=1/t

To avoid confusion, it is customary to write the Fourier transform and its inverse so that

they exhibit reversibility:

F (t)= exp(27ria)t) f (co)dt (4.6)

\127r :
4.4.1.3 Convolution Theorem

The convolution theorem is one of the most important relationships in Fourier
theory, and in its application to x-ray crystallography and spectroscopy.
Convolution theorem: The Fourier transform of a convolution of two functions is the
product of their Fourier transforms (equations 4.7-9) where the superscript * indicates the

Fourier transform firnction.

[C(u)]*=[f(x)®g(x)l* (4.7)
[C(u)l*=[ l fixigru—xidxi'“ (48)
Space
at: 1 :1: 1 at:
[C(u)] =f(—) .g(—) (4.9)
x x

154

4.4.2. Fourier Convolution Deconvolution of the EPR Dipolar Coupling

In most biological systems, except macroscopically oriented samples, the
distribution of 0 is isotropic. It is possible then to treat the spectrum of the two interacting
spins as the convolution of the non-interacting powder pattern absorption spectrum with a
dipolar broadening function D(R,B), known as a Pake pattern. The average splitting

<ZB> over the distribution D(R,B) is then:

213>=(0.75)(3/2)gep

 

(4.10)
R3
Thus, the EPR spectrum II(B) for the two nitroxides is described by:
11(13):)300O S(B ')D(R,B'—B)dB' (4.11)

where S(B) is the non-interacting EPR spectrum. However, in practice there will be a
distribution of interspin distances due to the flexibility of the nitroxide side chain as well
as the conformational variation of the macromolecules. In this situation the EPR

spectrum is described by:

11(3):):20 S(B ')M(B'—B)dB' (4.12)

155

where M(B) is the weighted sum of the D(R,B) over the distribution of the distances

P(R):
M(B) = ZP(R)D(R,B) (4.13)

According to equation 4.8, equation 4.12 can be written as:
H(B)=S(B')®M(B') (4.14)

Since II is approximated by a convolution of S and M in real space, equation 4.14

is simplified in Fourier space using the convolution theorem (equations 4.7-9):
n’(a2) = s‘(w) oM'(w) (4.15)

where w=1/B in units of gauss’l.

So:

M (to) = Him”

S'(w)

 

(4.16)

M(B) is then obtained from the inverse Fourier transform of the M‘(0)) :

156

M(B): exp(27ritoB)M*(a))d(a)) (4.17)

1 I00
J27t '00
if equation 4.16 is substituted to equation 4.17 then:

n*(w)
S*<co)

 

M(B)=:/——%_7;-jj'_%o exp(27tia)B) d(co) (4.18)

The average splitting <ZB> for a given P(R) is obtained form the following formula:

__ (:20 IZB|.M(B)dB

 

<ZB>— (4.19)
130M(B)dB
if equation 4.19 in substituted in equation 4.10 then (9, 21):
0.75 3/ 2
<R>=( ( )gefl )“ 3 (4.20)

 

<ZB>

4.4.3. Spectral Analysis

EPR spectra can be analyzed according to the equations 4.14-4.20 using fast

Fourier transform routines (Figure 4.4), and the resulting spectra at each step of the

analysis are shown in Figure 4.5. Both real and imaginary components of the dipolar

157

Figure 4.4. A flow chart of the algorithm for distance analysis from dipolar spin-spin
interaction. Taken from (9). Copyright permission was obtained.

158

Two Absorbance Spectra:

1) The double labeled spectrum, 1'I(B).
2) The single labeled spectrum, S (B).

Step 1: Preparation of Spectra

Includes baseline correction, spectral
normalization, and alignment of the spectra.

Step 2: Calculating the Broadening Function, M'(c0) in
the Fourier space

1) FFT on the 11(B) and S(B), getting min) and s‘(m)
respectively.

2) Calculating M'((0) by division of 11700) by S'(0)).

Step 3: Signal Analysis on M‘(c0)

l) Suppressing noise in M‘(00) by fitting it with the sum of
gaussian functions and a constant offset

2) Decontaminating the non-interacting species by subtracting
the constant offset.

Step 4: Calculating the Inter-spin Distance, R

1) Inverse FFT on M‘(0)), obtaining back M(B), the broadening function
in real space.
2) Calculating the average separation of the two spins, <R> from M(B) .

 

159

Figure 4.5. The resulting spectra at each step of analysis outlines in Figure 4.4. Taken
from (9). Copyright permission was obtained.

160

 

 

Field ( cuss

 

 

 

 

 

 

12 1 . r . v r
B. 0
1 P
o
0.01 . C '
. I
€03 . ' . . 1
r t 2' 0
20.4). . w... .
. o .9, .o o ‘- ,
0-2’ ’ 0:. "0. o 0 ii
.0 0 0‘ .5 9
’ 0*. Q .0
0 ’ ‘ ” "d‘ g 0.0” .
0' . . . . ' ...‘
0.2 9
O . O
-0.¢ # ~ e . - - 1 ° . ’ .
0 10 20 30 40 50 60 70 80 W IN
a) (rt/1024 Gauss")
C.1.2 '

 

V v f v f V v a

 

 

 

o‘i
.0 A

. 4 4 4* 4‘ . . Ll . .
o 10 20 30 40 so 60 so ‘90 100
D 8‘10.3

to («11024 Gauss‘ )

 

 

7.

6.

 

r W v

 

 

 

 

9200 .150 -1oo $0

. 0 50 100 150
Field (Gauss)

200

161

Step 1
........ Mono labeled S (B)
_ DOUbIC labeled H (B)
Step 2.
0 Real part of M*(c0)
+ Imaginary part of M*(0))
Step 3.
Fitting with two gaussians
plus the y-axis offset, with a
— cut off (1) of

60(11/1024 Gauss “)

Separated gaussian
components

Step 4.

Two gaussians may reflect
two distinct spin-spin
distances.

 

“*(w)

S (to)

 

spectrum in Fourier space contain high amplitude noise at high values of to

(Figure 4.5 B). The noise can be suppressed by fitting the spectral decay in Fourier space
with a sum of gaussian functions. In most cases, we have found that it is sufficient to use
a sum of only two gaussians. The final dipolar spectrum is obtained by inverse Fourier
transformation of the best fitting function. The distribution P(R) of interspin distance
leads to broadening fimctions that are sum of many Pake patterns. This in combination
with the EPR line width leads to broadening functions that do not look like Pake patterns.

The average distance is calculated using equations 4.18-20 (9, 21).

4.4.4. Monoradical Impurities

Monoradical impurities due to incomplete spin labeling reactions are a potential
problem when analyzing spin-spin interactions. EPR spectra with monoradical
contamination can be incorrectly interpreted as showing larger nitroxide-nitroxide
distance. The Fourier deconvolution method can separate the dipolar spectrum for the
interacting nitroxides from the spectral contribution from the monoradical contaminants.

The composite EPR spectrum for a mixture of monoradical and biradicals is given by:

nan-=51?co S(B ')M(B'-B)dB'+b-S(B) (4.21)

where a is the fraction of the biradical, b is the fraction of the monoradicals with

a + b =1 . Thus after Fourier deconvolution we will have:

162

M '*(a))=aM*(a))+b (4.22)

in Fourier space and the non-interacting species contributes to the y-axis offset. This y-
axis offset b can be subtracted from the dipolar spectrum in the Fourier space (Figure 4.6)

(9, 21).

4.5. Materials and Methods

Materials. 1,2-Dioleoyl-sn—Glycero-3-Phosphocholine (DOPC) and 1,2- Dioleoyl -sn-
Glycero-3-[Phospho-L-Serine] (Sodium Salt) (DOPS) were obtained from Avanti Polar
Lipids Inc. (Alabaster, Alabama). Oleic acid 2 99% was purchased from Fluka. All
restriction enzymes and DNA modifying enzymes were purchased from New England
Biolabs (Beverly, Massachusetts). Ni-NTA was purchased from Qiagen Inc. (Valencia,
California). Specially purified aqueous detergent solution (10% solution) of Tween® 20
was obtained from Pierce (Rockford, Illinois). The nitroxide spin label (1-oxyl-2,2,5,5-
tetramethylpyrroline-3-methyl)- methanethiosulfonate (MTSSL) was purchased from
Reanal Fine Chemical (Budapest, Hungary). Arachidonic acid was obtained from
Cayman Chemicals (Ann Arbor, Michigan). Flurbiprofen, Ibuprofen and Naproxen were
purchased from Aldrich. The COX-2 selective inhibitor SC-58125 was obtained from

Karen Seibert at Pharmacia and other chemicals were purchased from Sigma.

163

 

 

    

 

 

 

 

A. 1.2 . 2 . . e e
1
0.8 1
0.5 . al'l(B) + bS(B)‘
0.41
0.2 i
0 _ S(B)
“5656.46.56 0 20 4o 60
Field (Gates)
B. 15
1
:3, 0 5 . A . '... 0 TL
2 - r. - .. .
y-axis offset: b=--0.45
0.
—O.5

 

 

 

20 40
a) («11024 Gauss")

C- 0.6 a

A

M(B

M(B)

0.1 . ~ - .
450-100 450 0 50100150

 

 

 

o 5_ 5 function from ‘
' y-axis offset in
0.4 i )‘(m) i
0.3 r .
0.2 t
0.1 r
0

 

 

 

 

 

 

Field (Guns)

1: 10"
a -
s .

M(B) alter

4’ subsuaction of ‘

, the offset b.
3
2 i l
1 i' 1
o .

 

 

 

.1 - 4 2
-1 50 -1 00 -50 0 50 100 150
Field (Gauss)

Figure 4.6. Analysis of monoradical impurities. Taken from (9).

164

Sample Preparation. Protein expressions, purification, spin labeling and liposome

incorporation are as described in chapter 3.

EPR measurements. All EPR measurements were recorded on a Bruker ESP 300E X-
band spectrometer equipped with a TE102 Bruker rectangular cavity. The samples
typically 200 pl at concentration of 14 pM labeled protein, were loaded into a 4-mm OD
clear fused quartz sample tubes (Wilmad-labglass, New Jersey). The first derivative
absorption spectra of the samples at 183 :2 2K were obtained by averaging 4 scans at a
scan width of 110 Gauss and each scan time of 4 minutes. The temperature was
controlled with Bruker BVT 2000 liquid nitrogen temperature controller. A modulation
amplitude of l Gauss was used. The microwave power was maintained at 1 milliWatt to

avoid the saturation of the EPR spectra.

4.6. Results

The site-directed spin labeling technique requires a template cDNA that does not
contain reactive cysteines. A “cysteine-less” template was therefore prepared by mutating
cysteines at positions 299 and 526 in human histidine-tagged COX-2 to serines. The
cysteine-less enzyme retains 70% of the cyclooxygenase activity of the native His-tagged
COX-2, indicating that cysteines 299 and 526 are not required for catalysis and these
substitutions only minimally distort the structure of the protein. Four helices (A-D) form
the membrane-binding domain that is also the mouth of the cyclooxygenase active site

(22). These helices were chosen as labeling sites because this is the region in which the

165

 

,7

most flexibility is observed in the crystal structure of COX (23, 24) . By using a cysteine-
less template, ten mutants were constructed, each containing the I98C substitution in
helix D, and one of five amino acid substitutions in opposing helix B (T73C, V74C,
H75C, Y76C and I77C) or one of the five amino acid substitutions in adjacent helix C
(F 84C, W85C, N86C, V87C and V88C). Nine of the ten double cysteine mutants retained
significant cyclooxygenase activities ( 10-55% of the native) (Table 4.2). One mutation,
W85C, reduced cyclooxygenase activity by 95-98%, indicating that this residue plays a
critical role in protein structure or catalysis and this mutation significantly alters the
COX-2 structure. Conjugation of the cysteines with the spin label MTSSL had no effect
on the activity. These results indicate that with one exception the cysteine substitutions
and subsequent conjugation with the spin label do not grossly disrupt the native structure
of the enzyme. Several studies indicate that the nitroxide spin labels do not perturb the
structure and other properties of the biomolecules (25, 26).

Each of the ten pair-wise mutant proteins were purified and then reconstituted
with 2 molar equivalent of cobalt-protoporphyrin IX. Binding of Co-PPD( to COX is
identical to hemin (27). Hemin was not used because even a small amount of oxidized Fe
(III) in hemin could cause dipole-dipole interaction with the spin labels. Cobalt-PPIX is
diarnagnetic and will not have any magnetic interaction with the spin labels. F lurbiprofen
(time-dependent COX inhibitor), naproxen and ibuprofen (time independent COX
inhibitors), SC58125 (COX-2 selective inhibitor) and arachidonic acid (the primary
substrate for COX-2) were used to examine conformational changes in COX-2. The
inhibitors were added with 20 fold molar excess with respect to protein for each sample.

EPR spectra of apoenzymes, holoenzymes, NSAID-treated holoenzyrnes and arachidonic

166

acid-bound holoenzymes were recorded at 183 K. The first derivative EPR spectra were
integrated (absorbance) and normalized to the same number of spins. The normalized
absorbance spectra of solubilized spin-labeled COX-2 holoenzymes were overlaid with
the normalized absorbance spectra of the spin-labeled COX-2 complexed with various
NSAIDs, with arachidonic acid, or with the spectra of the spin labeled apoenzymes
(Appendix C). All the spectra superimposed on each other, indicating that binding of
heme, arachidonic acid, and NSAIDs had no significant effect on the conformation of
COX-2.

The normalized absorbance spectra of solubilized spin-labeled COX-2
apoenzymes and COX-2 holoenzymes complexed with various NSAIDs or arachidonic
acid were overlaid with the normalized absorbance spectra of liposome incorporated
spin-labeled COX-2 apoenzymes and COX-2 holoenzymes complexed with various
NSAIDs or arachidonic acid (Appendix C). The slight differences in the line broadening
of the EPR spectra in the presence and absence of liposomes indicate that the spin-spin
interactions in mutants T73C/I98C, V74C/I98C, H75C/I98C, Y76C/I98C, I77C/I98C,
F84C/I98C and W85C/I98C were slightly stronger when the proteins were in the
detergent. This observation is consistent with the interpretation that there may be a small
conformational difference between the COX-2 in the detergent versus the phospholipid
bilayer.

' In helix C mutants (F84C, W85C, N86C, V87C and V88C), the dipolar line
broadening is due to the interaction of each spin labeled residue with the spin labeled
I98C residue. The interactions between the two subunits (inter-helix spin-spin coupling)

do not contribute to the line broadening because the residues are far apart from each other

167

(Figures 4.7-4.11). The Fourier convolution/deconvolution method can only be used
when the distances between the two spin labels are between 8-25 A (21 ). The distances
between each residue in helix C and I98C are in the range that can be measured by
Fourier convolution / deconvolution method. Distances more than 25 A do not contribute
to the line broadening. On the other hand for helix B mutants (T73C, V74C, H75C, Y76C
and I77C), the dipolar broadening is due to the interaction of more than two spin labeled
residues (Figures 4.12-4. 16). Because more than two spin labeled amino acids give rise to
the line broadening, it is not possible to measure the distances between the helix B
mutants and I98C.

All the mutants in helix B show a stronger spin-spin interaction in the detergent
than in the liposomes. This indicates that in liposomes the spin labels are further apart
from each other. There are two explanations for this observation.

1) There is a distance increase between the two spin label residues within each monomer
e.g. an increase in distance between spin labeled-H75C and spin labeled-I98C within
a monomer.

2) There is an increase in the distance between the two subunits i.e. an increase in the
distance between two spin labeled residues between the two monomers.

To determine which explanation is true the distance measurements were done on a
series of singly labeled residues in helix B (T73C, V74C, H75C, Y76C and I77C) in
detergent and in the liposomes. The normalized absorbance spectra of each mutant in
liposome and in detergent were superimposed on each other (Figure 4.17). The
superimposition did not show any changes in the line broadening which indicated that

there was no distance change between the two spin labeled residues between the two

168

monomers. The conclusion is that the slight change in the line broadening is due to the
increase in distance between the two spin label residues within each monomer.

Fourier convolution deconvolution analysis of the broadening functions resulting
from spin-spin interaction of the doubly labeled COX-2 (between helix C mutants and
198C) was used to calculate the molecular distance between the nitroxides by comparing
the broadening of the doubly labeled protein to the singly labeled reference spectrum
using a FORTRAN program written by Professor Yeon-Kyon Shin for Matlab ® (13)
(see Appendix B for the cutout of Matlab analysis). In this study the H75C single mutant
was used as the reference spectrum. The inter-nitroxide distances between the five
mutants in helix C and I98C in helix D in the detergent and in the liposome were

calculated (see Table 4.2).

169

5.; . ‘- f l . _.-.. ‘
"- a}. jf K., ’C ' a ' I
I” —-“ I ~ 1"?" [“7 ‘
t" - 3 . ‘1
(T ‘ . i ""6. 4 - 4 m
' ‘1’ 1 ' \ ‘11.,
c 5 , . ., ..- .
2 . If

Figure 4.7. The X-ray structure of murine holo-COX-2 (5COX) with two cysteine
substitution mutations at residues F84and I98.

I98C is shown in green. F84C is shown in red.

The distance between the two C01 s of the two I98C residues in dimer is 66.32 A.

The distance between the two Co. s of the two F84C residues in dimer is 48.43 A.

The distance between the C01 of I98C and C01 of F84C within a monomer is 22.01 A.

170

~ . x» 4r-
.am 1.4
e rigs. fleet-0. ..
a.» ' . ‘ \_ ' . (. v ,1
" 1:1 f VF‘FT‘P i r “A '2. 1
I '1" r, ' l" , ‘ i ‘ .
*ri/fi “' 33. "11311.: , t
'3" . .. f 7 :1'12.
. 1., 1 .154 F
F 1:?!
w—‘Q h" 1 -
..
1 "--’*.
_ u" .- i} 4. 1
f ,
" ‘11 . - t :'
‘. I 7'

Figlre 4.8. The X-ray structure of murine holo-COX-2 (5COX) with two cysteine
substitution mutations at residues W85 and 198.

I98C is shown in green. W85C is shown in red.

The distance between the two C01 s of the two I98C residues in dimer is 66.32 A.

The distance between the two C01 s ofthe two W85C residues in dimer is 52.89 A.
The distance between the C01 of I98C ended of W85C within a monomer is 22.72 A.

171

Figure 4.9. The X-ray structure of murine holo-COX-2 (5COX) with two, cysteine

substitution mutations at residues N86 and 198.

I98C is shown in green. N86C is shown in red.

The distance between the two Ca s of the two I98C residues in dimer is 66.32 A.

The distance between the two Ca s of the two N86C residues in dimer is 51.00 A.
The distance between the Co of 19sc and'Ca ofN86C within a monomer is 21.32 A.

172

a; "
x“ ' "
. 7 I . .5 \
J / t f ". ' - 5' “a" ‘
a 3.1““. . 1 » k l l .H t
. \ i (I .‘ . 7
" d” ; ;"\ \ “ “ - .-
J v ' ' ‘ i I t
, J., ’ k r' “S
‘» 6. . I kyu L ‘
V .. f T‘ 4’ V x ‘
§ ,‘ I’K‘.‘
, , y.
. ' ‘ '
I" V '\\
, .i‘
u , -_ .
~ g I ‘ ‘ \f“ )
1‘ '\ '_“
, l
. I I... '
. J/ _ ' "I O r.
g ‘ . .’ t '.- 1’
x / .. . ‘ ; ‘ '1. t
t' .f. « ~
to. x ‘ : ‘ . , g”. I, U
a a V .~ .‘.
i 7 A . . P Y
\ —v - l ’
‘~- ’ ’ \’ , . l
' . ' l I / z I-
‘ V .. ; i ,1 7 ‘1‘."
'-‘ f . ' I I ‘
L ‘ _ '
e -.» t
"1 -. g”
. 1 .
k H ’

Figure 4.10. The X—ray structure of murine holo-COX-2 (5COX) with two cysteine
substitution mutations at residues V87 and 198.

I98C is shown in green. V87C is shown in red.

The distance between the two Ca 3 of the two I98C residues in dimer is 66.32 A.

The distance between the two Ca 3 of the two V87C residues in dimer is 58.02 A.

The distance between the Ca of I98C and'Ca of V87C within a monomer is 19.97 A.

173

A A\\
I?‘ ..J

Figure 4.11. The X-ray structure of murine holo-COX-Z (5COX) with two cysteine
substitution mutations at residues V88 and I98.

I98C is shown in green. V88C is shown in red.

The distance between the two Co. 3 of the two I98C residues in dimer is 66.32 A.

The distance between the two Ca 8 of the two V88C residues in dimer is 57.05 A.

The distance between the C0. of I98C and Co of V88C within a monomer is 18.14 A.

174

 

Figure 4.12. The X—ray structure of murine holo-COX-Z (5COX) with two cysteine
substitution mutations at residues T73 and 198.

198C is shown in green. T73C is shown in red.

The distance between the two Ca 3 of the two I98C residues in dimer is 66.32 A.
The distance between the two Ca s of the two T73C residues in dimer is 44.96 A.
The distance between the Ca of I98C andTCa of T73C within a monomer is 39.67 A.

175

'J‘er‘ - _
‘ “. I

Figure 4.13. The X-ray structure of murine holo-COX-Z (5COX) with two cysteine
substitution mutations at residues V74 and 198.

I98C is shown in green. V74C is shown in red.

The distance between the two Ca 8 of the two I98C residues in dimer is 66.32 A.

The distance between the two Ca 8 of the two V74C residues in dimer is 44.66 A.

The distance between the Ca of I98C andCa of V74C within a monomer is 38.59 A.

176

5"" J:
a 4’ . ~ "
5v"‘-1C’L' rg’ ' ."v‘t
.. ”f <- 1‘ .11 . 4,
,1 4 - z‘y‘X , ’
t“ 7;" r * fl" ”‘K/
V.“ '1‘ 1" 1., ' TR ‘ A
r ‘J A ‘ t;$/
r
. \-
3.,
7. 1.:
V
. 1”". ‘
1 p '3‘"
. v 1’ 2', .r‘
. > \‘5 J; _ I h; A‘
y n ,
-. 14,7 ’5

Figure 4.14. The X-ray structure of murine holo-COX-2 (5COX) with two cysteine
substitution mutations at residues H7 5 and I98.

I98C is shown in green. H75C is shown in red.

The distance between the two Co. s of the two I98C residues in dimer is 66.32 A.
The distance between the two Ca s of the two H75C residues in dimer is 37.60 A.
The distance between the Ca of I98C andch of H75C within a monomer is 35.70 A.

177

I '9' \fit'
to? 1‘ ~ ( ' ‘ fix
We 5’: 1 v“ ’ 13‘7“?)
V _ .w. e . ..
’ 1n. . {fr-"W ‘ $3.;
r a 131‘: 1
3 i j;
.2 .
1 a ‘ 7’
. ‘3, '~
It . '
a.
‘,
' 1 C.
V) M t
1). k - x, L‘ ,
* ’ ‘, 1" I,
3% L'li . ,3
tr .. ‘f
xi 1 A»
‘25,“; '3 i” A., %

Figure 4.15. The X-ray structure of murine holo-COX-Z (5COX) with two cysteine
substitution mutations at residues Y76 and 198.
I98C is shown in green. Y76C is shown in red.

The distance between the two Ca 5 of the two I98C residues in dimer is 66.32 A.

The distance between the two Ca s of the two Y76C residues in dimer is 40.24 A.

The distance between the Ca of I98C and:Ca of Y76C within a monomer is 34.78 A.

178

 

 

. V . 1‘ v,‘ .
. . .\/ ’V 1,”: t : -‘. a
.'_ ’_ .‘ "1.. --?L I Pk I d L...
. _..< . - . _.,.
NP, \x : .\ . . 5'. - A v“ 4‘
.s— . \ r 1,.“ T '.
”1" I FV , \ h. i? ‘ l} r
4 . 1‘ _ v
,- . . 2 t .
.1”! V \ ’er '
'—\J‘ 4 ‘ v i ,1 r
. 1“,; \
\
:4“
"a J”? ,
‘1
.\ v
.. ' ' , h
i 2, fig 1‘
v r. \ I
' i
5. ‘af \ . )1 ‘_ ‘ N
J6 Ar , x I ' "
P .
‘ frk I ’r ,j-l ‘
.- ‘ 4 . , , “J f
f
)1 , %
~.‘ .

Figure 4.16. The X-ray structure of murine holo-COX-Z (5COX) with two cysteine

substitution mutations at residues I77 and 198.

I98C is shown in green. I77C is shown in red.

The distance between the two Ca 3 of the two I98C residues in dimer is 66.32 A.
The distance between the two Ca 8 of the two I77C residues in dimer is 45.73 A.
The distance between the Ca of I98C andCa of I77C within a monomer is 34.34 A.

179

 

 

1,___~-“7#.;,_-~_n_.;_-vmf i-e.”mm _

0.9 :— rh75c|a.dat ’ 4
1 H75C apoenzyrm in lbosorre ‘

0.8L “a” ~
1 rh75cda.dat

0-7 ” H75C apoenzyn'e in Tw een 20
5 (green)

0.6 ‘— ‘ -

0.5 §~ -

0.4;» I ~

0.3 e

(12— i"‘\ \ -

0.1— —

-60‘ -40 I20- 0 20 4o 60

 

Figure 4.17. Overlay of absorbance EPR spectra of H75C in liposome (red) and
H7 SC in Tween 20 (green). The two spectra completely superimpose on each other
and show no change in the line broadening.

180

I

 

 

4.7. Discussion

All NSAIDs function by blocking prostaglandin biosynthesis through inhibition
of COX, but different drugs achieve this via different mechanisms. Some NSAIDs such
as ibuprofen are classical time independent competitive inhibitors, which form reversible
enzyme-inhibitor complexes (EI). Other NSAIDs such as flurbiprofen are time
dependent, slow tight-binding inhibitors that initially form a reversible EI complex,

which is then gradually converted to an inactive form (EI*).

E + I e E1 (Time — independent)

E + I <-> E] H E! *(Time-dependent)

The difference between time-dependent and time-independent inhibition is
clinically relevant, as it underlies the isoform-specific inhibition seen with selective
COX-2 inhibitors. COX-2 selective inhibitors are time-dependent inhibitors of COX-2
forming tight-binding and slowly reversible complexes with this isozyme, but are simple
competitive reversible inhibitors of the COX-1 isozyme.

An understanding of the EI—)EI* transition associated with time-dependent
inhibition should therefore be useful in the rational design of novel therapeutics.

Different hypothesis have been advanced to explain this phenomenon, including
the invocation of a structural rearrangement between EI and EI“ and the suggestion that

time dependent and time-independent inhibitors occupy distinct binding sites (28-30).

181

 

To gain deeper insight into the mechanism of NSAID binding we have used
double site-directed spin labeling combined with EPR spectroscopy. Our results also
indicated no detectable conformational changes in the structure of COX.

Thus if the EI—>EI* transition corresponds to a conformational rearrangement, it
must be small or in the region of COX-2 that we did not observe.

The slight differences in the line broadening of T73C/I98C, V74C/I98C,
H75C/I98C, Y76C/I98C, I77C/I98C, F84C/I98C and W85C/I98C mutants in the
detergent versus liposome show a small conformational change between the structure of
the protein in the detergent and the structure of the protein in the lipid bilayers. This
observation is consistent with the interpretation that the mouth of the cyclooxygenase
channel in more open in the liposome environment. Previous studies on other proteins
have also suggested a slight conformational difference in the structure of membrane

proteins in the presence and absence of lipid bilayers (31).

182

Table 4.2. EPR determined inter-nitroxide for mutants in Helix C and I98C.

 

 

INTER NITROXIDE INTER NITROXIDE

SAMPLE DISTANCE (A) INDETERGENT DISTANCE (A)INLIPOSOME

 

F84C/I98C 15.358 i 0.420 15.536 i 0.976
W85C/I98C 16.086 i 0.231 16.734 i 0.180
N86C/I98C 14.581 i 0.481 15.188 i 0.277
V87C/I98C 16.194 i 0.352 16.489 i 0.244

V88C/I98C 14.384 1: 0.301 14.226 i 0.603

 

183

 

4.8. References

10.

11.

12.

13.

14.

Smith, W. L., and DeWitt, D.L. (1996) in Therapeutic Immunology (Austen, K.
F., Burakoff, S.J., Rosen, PS, and Storm, T.B., Ed.) pp 119-131, Blackwell
Scientific Publications, Inc., Cambridge, MA.

Kulmacz, R. J., and Lands, W. E. M. (1982) Biochemical and Biophysical
Research Communications 104, 758-764.

Chen, Y. N. P., Bienkowski, M. J ., and Mamett, L. J. (1987) Journal of Biological
Chemistry 262, 16892-16899.

Kalgutkar, A. S., Crews, B. C., and Mamett, L. J. (1996) Biochemistry 35, 9076-
9082.

Kulmacz, R. J. (1989) Journal of Biological Chemistry 264, 14136-14144.

Selinsky, B. S., Gupta, K., Sharkey, C. T., and Loll, P. J. (2001) Biochemistry 40,
5172-5180.

Koteiche, H. A., and McHaourab, H. S. (1999) Journal of Molecular Biology 294,
561-577.

Berengian, A. R., Parfenova, M., and McHaourab, H. S. (1999) Journal of
Biological Chemistry 274, 6305-6314.

Berliner L.J., E. S. S., Eaton GR. (2000) Distance Measurements in Biological
Systems by EPR, Vol. 19, Kluwar Academic/Plenum Publishers, New York, New
York.

Hustedt, E. J., Smimov, A. I., Laub, C. F., Cobb, C. E., and Beth, A. H. (1997)
Biophysical Journal 72, 1861-1877.

Eaton, S. S., and Eaton, G. R. (1982) Journal of the American Chemical Society
104, 5002-5003.

Eaton, S. S., More, K. M., Sawant, B. M., and Eaton, G. R. (1983) Journal of the
American Chemical Society 105, 6560-6567.

Kokorin, A. I., and Formazyuk, V. E. (1981) Molecular Biology 15, 722-728.

Kokorin, V. V., Osipenko, I. A., Polotnyuk, V. V., Zolkina, S. V., and Shirina, T.
V. (1982) F izika Metallov I Metallovedenie 54, 48-53.

184

 

 

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

Milov, A. D., and Tsvetkov, Y. D. (1997) Applied Magnetic Resonance 12, 495-
504.

Milov, A. D., Maryasov, A. G., and Tsvetkov, Y. D. (1998) Applied Magnetic
Resonance 15, 107-143.

Raitsimring, A., Peisach, J ., Lee, H. C., and Chen, X. (1992) Journal of Physical
Chemistry 96, 3526-3531.

Martin, R. E., Pannier, M., Diederich, F., Gramlich, V., Hubrich, M., and Spiess,
H. W. (1998) Angewandte Chemie-International Edition 37, 2834-2837.

Saxena, S., and Freed, J. H. (1997) Journal of Chemical Physics 107, 1317-1340.
Saxena, S., and Freed, J. H. (1996) Chemical Physics Letters 251, 102-110.

Rabenstein, M. D., and Shin, Y. K. (1995) Proceedings of the National Academy
of Sciences of the United States of America 92, 8239-8243.

Picot, D., 1.011, P. J ., and Garavito, R. M. (1994) Nature 36 7, 243-249.
Kurumbail, R. G., Stevens, A. M., Gierse, J. K., McDonald, J. J., Stegeman, R.
A., Pak, J. Y., Gildehaus, D., Miyashiro, J. M., Penning, T. D., Seibert, K.,
Isakson, P. C., and Stallings, W. C. (1996) Nature 384, 644-648.

Luong, C., Miller, A., Barnett, J., Chow, J., Ramesha, C., and Browner, M. F.
(1996) Nature Structural Biology 3, 927-933.

Langen, R., Oh, K. J ., Cascio, D., and Hubbell, W. L. (2000) Biochemistry 39,
8396-8405.

Postnikova, G. B. (1996) Biochemistry-Moscow 61, 679-693.

Malkowski, M. G., Ginell, S. L., Smith, W. L., and Garavito, R. M. (2000)
Science 289, 1933-1937.

Copeland, R. A., Williams, J. M., Giannaras, J., Numberg, S., Covington, M.,
Pinto, D., Pick, S., and Trzaskos, J. M. (1994) Proceedings of the National
Academy of Sciences of the United States of America 91, 11202-11206.

Llorens, O., Perez, J. J., Palomer, A., and Mauleon, D. (1999) Bioorganic &
Medicinal Chemistry Letters 9, 2779-2784.

So, 0. Y., Scarafia, L. E., Mak, A. Y., Callan, O. H., and Swinney, D. C. (1998)
Journal of Biological Chemistry 273, 5801-5807.

185

31. He, M. M., Voss, J ., Hubbell, W. L., and Kaback, H. R. (1997) Biochemistry 36,
13682-13687.

186

 

 

 

Chapter 5

Conclusion and Future Direction

187

The cyclooxygenase isozymes have been studied extensively because of their
essential and regulatory role in prostaglandin synthesis and also because they are the sites
of action of the Nonsteroidal anti-inflammatory drugs (NSAIDs) such as aspirin,
flurbiprofen, naproxen and new selective COX-2 inhibitors. However very little is known
about how the membrane binding domain of COX isozymes interact with the lipid
bilayers. It in not possible to orient the COX isozymes in a membrane, or to determine
lipid binding sites in the isozymes from a crystal structure, as few lipid or detergent
molecules are visible using X-ray analysis.

The goal of this dissertation is to study the membrane-binding domain of hCOX-2
using site-directed spin labeling technique.

Site-directed spin labeling (SDSL) is emerging as a promising tool for the
determination of protein topography. In the research described in chapter 3 of this
dissertation, 25 mutant enzymes that contained single reactive cysteines substituted for
amino acids within the hCOX-2 membrane-binding domain were constructed and spin
labeled. The accessibility of each protein-bound spin label with freely diffusing oxygen
as a non-polar paramagnetic reagent, and NiEDDA as a polar paramagnetic reagent, was
measured using power saturation EPR spectroscopy. Our results show the accessibility
parameter (H) for both oxygen and NiEDDA, has a periodicity of 3.6 for helices A and
C. These results are where the helix axis lies along an interface with protein.
Furthermore, I'loxygen and HNiEDDA both have the same period and phase. Our results
indicate that both polar and non-polar paramagnetic relaxers do not access the amino

acids located in the membrane-binding domain due to the very low values of I'INiEDDA and

I‘loxygen. This indicates that the membrane binding domain of hCOX-2 is located in the

188

 

interfacial region of the membrane and is sandwiched between the main body of the 144
kDa COX dimer and the lipid bilayers. We have identified 3 amino acids that participate
in anchoring the protein to the lipid bilayers, but most of the amino acids in the
membrane-binding domain are not in contact with the hydrocarbon core of the lipid
bilayers. hCOX-2 is the first peripheral membrane protein mapped using site-directed
spin labeling. COX isozymes and squalene hopene cyclase are the only members of the
new class of monotopically-associated peripheral membrane proteins. The results of the
SDSL experiments in this dissertation might have general applicability to this new class.

To fully investigate the MBD of hCOX-2, mutant enzymes that contain single
reactive cysteines substituted for amino acids in the loops of hCOX-2 membrane-binding
domain need to be constructed and investigated by site-directed spin labeling method.

Another aspect of my project was to study the conformational changes in the
mouth of the cyclooxygenase channel. The purpose of this research was to determine
whether the COX substrate and inhibitor binding sites are flexible, e.g., can be opened
and closed.

In the research described in chapter 4 of this dissertation, 10 cysteine double
mutants located in the helices that form the opening of the cyclooxygenase active site and
membrane-binding of hCOX-2 were constructed. The dipolar broadening in the presence
and absence NSAIDs, arachidonic acid and heme were compared in detergent. The same
experiments were also performed in liposomes. The inter-spin distance measurements

were done using the Fourier convolution/deconvolution technique.

189

 

Our results indicated that binding of heme, arachidonic acid, and NSAIDs had no
significant effect on the conformation of hCOX-2. However, a slight conformational
change was observed in the structure of the enzyme in the presence of the lipid bilayers.

In order to firrther investigate the possibility of the conformational change in the
mouth of the cyclooxygenase channel, time resolved EPR experiments on liquid samples
at room temperature need to be done in the presence and absence of various NSAIDs,

arachidonic acid and lipid bilayers.

190

 

Appendix A

The Data Analysis Procedures for the Experiments Described in Chapter 3.

191

 

 

are—noun name-thumb
[21L ‘lel’l EEBBEJ mama FTC F1

' ““m""t.tm§i- n More Mum-.. 111111.“: ‘ ‘ 1.13.; 11* nitrite}-
mM-meaoameboaaelamiblmmw, Eta—575333“

 

Figure 1A:

Step one: To open the spectrum with Bruker’s SimFonia software.

192

 

innaafirm [ring We r1

wmr M: warm; 1n vii-114.111

 

 

“digits.

‘_~] A .0 51C! I. ‘ TEn—Lhmmfiéjum— W ' ' A lm’fi! .36 m

Figure 2A:
Step two: To export the spectrum to WinEPR sofiware.

193

 

 

mun', .. , on.) ":
t ,, us. 1‘ 1

 

’ "no ‘ ”1 . 1.-..11.“"'-'- - :-
mes-meueanaeoaoauau £251.21me mar“ .111 1337'

Figure 3A:
Step three: To filter the noise in the spectrum.

 

 

 

 

LL

.,1W.“J r' V
mlumauelmsvtonael

:- .
‘1» Ar A

a“ ,’ 3!; 1:: m2" ” ' ,g‘“ 15;, .21 Ill—re- senescent“

Figure 4A:

Step four: To measure the peak to peak amplitude.

195

 

 

R M M M ‘ .
‘ ‘ anon - ' w

.1. . 1., 'u:..t. ”.1
.' v

, 1 ‘. 1 4. .1‘_;v‘ '1'. . ‘ -; .-~ _..1 .. 11“.” . .
£1,1lflBBOCBQOOBDIIat‘:mlfieflflflmt uni: u-

1
H

Figure 5A:

The stack plot of EPR spectra. Each spectrum was acquired in a different microwave
power.

196

 

Figure 6A:

Step five: Peak to peak amplitude mp0 of a first derivative EPR signal is a function of the
square root of the incident microwave power.

197

 

 

 

 

Figure 7A:

Step six: To plot the peak to peak amplitude of spectra as a function of the square root of
the incident microwave power.

198

 

 

 

 

Flglu'e 8A:

Step seven: Curve fitting the curve to the following equation:
Y=(P1*X)/((l+((2"(1/P2)-1)*((X"2)/P3)))"(PZ))
This equation is same as equation 3.22 in chapter 3.

 

 

 

 

Figure 9A:
Step eight: To obtain the unknown values.

Pl=I (scaling factor), P2=E (spectral homogeneity factor) and P3=Pm.

200

15: go- {the (gum loath

gig-gru- 5.1155! 3&- ”

 

 

 

 

 

 

 

[gums-run ‘AA - 'A 7" n.“ v“— ,-_ WJ— filth-

£21 Imaeoamae mun-.m- .‘_J9_“:J§_J__J!_"_Jl—H_J'au!!uernmm

Figure 10A:

A new software. was recently written in C programming language in collaboration with
Professor Michael Romalis in Princeton University.

The sofiware will automatically load the spectra, filter them and measure their peak to
peak amplitude mp0 as a fimction of the square root of the microwave power and save the

results in a *.dat file.

201

Appendix B
Application of MATLAB ® Scripts for Measuring the Distance between Two

Nitroxide Spin Labels

202

 

 

A FORTRAN program has been written for MATLAB® by Professor Y.K. Shin
that can be used for Fourier convolution deconvolution analysis of dipolar spectra.
Output of an example analysis that compares the singly labeled H75C with the doubly

labeled N86C/I98C mutant is presented below:

Computer Screen Explanation
EDU» load v8698dhl.dat Load the spectrum of the double-labeled

N86C/I98C mutant. EPR data are saved

from the instrument as Y-coordinates only.

EDU» load rh75cdh.dat Load the: singly labeled spectrum of H75

used as a reference (Y coordinate only).

EDU»v8698dhlf=eprfile(v8698dhl); Defines the double mutant data file as an
EPR spectrum. The f at the end of the file

name specifies that the file is used as a

function.

sweepwidth in Gauss?l 10 This is the sweep width of the EPR

spectrum.

203

 

absorbance( l) or differential(0)?0

Shift Field? Yeszinput number of points, No=00

Correct Field? Yes=l, No=01

EDU» rh750dhf=eprfile(rh75;cdh);

sweepwidth in Gauss?1 lO

absorbance( 1) or difi‘erential(0)?0

Shifi Field? Yeszinput number of points, No=00

Correct Field? Yes=1, No=01

204

EPR data is usually collected as a

first derivative spectrum.

Do not shift the field.

This corrects for minor variations in
the magnetic field that occur from

sample to sample.

Defines the single mutant data file as

an EPR spectrum.

The sweep width for both single and

double mutant should be same.

EPR data is usually collected as a

first derivative spectrum.
Do not shift the field.

This corrects for minor variations in
the magnetic field that occur from

sample to sample.

 

EDU» specintin=v8698dh1 f;

EDU» specnonin=rh7Scdhf;

EDU» fibroadf

what color would you like?'r'

Current plot held

what color would you like?'g._

Current plot released

EDU» sweep_width=220;

This defines v8698dh1f as the doubly

labeled spectrum.

This defines rh75cdhf as the singly labeled

reference spectrum.

This calculates the broadening fimction for
the doubly labeled spectrum over the
distribution P(r) and carries out a Fourier
transformation to isolate the electron-

electron dipolar interaction.

‘Red’ is the color of singly labeled

absorbance spectrum (See Figure Bl ).

‘Green’ is the color of doubly labeled

absorbance spectrum (See Figure Bl).

The Fourier transformation of the calculated

broadening function is plotted.

Input value as 2x the actual sweep width.

205

 

EDU»tau=[0.460.6 6O 0]; Use Figure B3 to determine the biphasic
decay components and the offset

representing the percentage of monoradical

within the sample.
EDU» fflowfit Command to solve for the distance between
the nitroxides.
What is the cutofi?58 Value obtained from the Fourier transform

where the decay function reverts to noise in

Figure B4.

Current plot held The program is plotting the various output

displays seen in Figure BS.

Current plot released

Current plot held

Current plot released

206

tau=

0.4962 6.9169 0.5031 50.1327 0.0007

sum_of_components =

1 .0000

what color would you like?'r'
what color would you like?'r'
Current plot held

what color would you like?'r'

B_dev =

7.8738

what color would you like?'r'
Current plot held

what color would you like?'r'
what color would you like?'r'

what color would you like?'r'

avg_2B = 6.4716

Calculated tau values. The sample was

monoradical free.

‘Red9

‘Red,

(Red,

‘Rw9

(Red,
(Red,

(Red,

This is the calculated value for
(23)

207

 

 

distance =14.7723 This is the calculated value of r.

finish(0)? 0 End of the calculation.

208

 

 

2.5 ~ —

1.5— —

0.5 -— —

0 L__ _._-..:. _. ..,_____ hhfifl-~_._ ~ _

-60 -40 -20 0 20 40 60

 

Figure Bl. Overlaid Absorbance spectra of singly and doubly labeled proteins. The
overlay spectra of the singly labeled H7 5C mutant (red) and the doubly labeled
N86C/I98C mutant (green) are shown above.

The spectrum of doubly labeled protein is defines as:

m3) = Tsw ')M(B'— B)dB-'

For detail discussion see chapter 4.

209

 

 

 

5 1 ‘ 4

 

of if if“ 10061 1500 2000 2500

Figure 82. Fourier transform of 1'I*(oa), the doubly labeled spectrum. Only the ends of
this curve contain usefiil information. A close-up of this is shown in the next figure.

210

 

 

 

4+ 3'4

 

Offset (b) 5

 

 

0'10. 20 do 40' 50 ab 7b so 90 100

Figure B3. Zoom of Fourier transform of doubly labeled spectrum. This figure is used to
obtain the five components referred to as the tau. Tau is made up of 5 components

[1, 2, 3, 4, 5]. The first two components are the rise and run of the initial decay, the third
and fourth components are the rise and run of the second decay and the final component
is the offset.

211

 

 

1.5~ —

_._.______ ._ __ ,_1___. .. __ _ __ ,H._ ___4. ___—__ _______

-1-- _ . _ -. . -. . _._:
0 20 40 30 80 100 120 140 160 18 200

 

Figure B4. Fourier transform using calculated tau components. This figure is used to
determine the cut-off for calculations, where the signal is totally overcome with noise.
For this measurement, the cutoff of 58 was used (black- line).

212

 

T—

 

1 eat—em _.. wee-— ~-- a, 0.03 mm“ x ewe-fl

 

 

 

o— 0.02r
A , B .
-1 * “ 0.01 f
2 ~* 0
-3 ~—~- -_ . --—~ -o.01 ? . ‘- . ___,
o 20 4o so -200 -100 o 100 200
4— -_._._._a__________-m_, 0.03. - . a
0.024
c n ;
1 0.0-1 "
. 0 _
__ ___ -0 o m n

 

 

- ' . 1* .
100 -200 -100 o 100 200

Figure B5. Spectra resulting form F CD analysis of dipolar coupling.

Panel A. shows the fit of the biphasic decay function, where the green dots are the
experimental data, the upper red line is the best fit of the equation and the lower line is
the discrepancy between the two.

Panel B. shows the inverse Fourier transfonn of M*(m) fit to 2 gaussians.

Panel C. shows the absorbance spectrum of the doubly labeled protein prior to (green
line) and after (red line) subtraction of the monoradical component b - S(B).

Panel D. shows the inverse Fourier tranform of only the biradical component aM*((o).

213

 

Appendix C

EPR Spectra of Ten hCOX—Z Double Mutants

214

 

 

Figure 1C. Absorbance EPR spectra of T73C/I98C double mutant in detergent (Tween
20) and liposome and in the presence of Co—PPIX, arachidonic acid and various COX
inhibitors.

Spectra were normalized to maximum absorbance and were uncorrected for any
contribution by monoradical species.

Images in this dissertation are presented in color.

215

 

 

 

ob .

ob

0.4 .

 

.°
7 W, # 7

 

p
(D
17 T‘ —T_“‘“’ ’W

P
M

. 4 .1 t.
a.
<§ § a.
«28 88.556 .5 at 3: no .
.33 .n
$883 a2 ..
«9.8 3983,28 Booms sea! ...
.5 a! co: mo. .
5983 \t.
\
L8 .8 c
4 r _ . l .4 ,.
388‘ am» .
«Q8 zoeoaxau 68m... :28. ..
+ 8583: 5 at 8: mo ._
A83 .n
3883 .8. _...
~88 30.03325 603.. 303$ ..
a. 7.. 8.3. No.
6.83 .._.
\\
.\
.8 .8 o

__L__ . _4L__fi__L.__L_.._.J

 

...._____~_~._ _._~..4

 

pm .
ob .
on a

Pm r

 

paw
oh .r
oh
9».

o... ..

 

O

l4lr ,yr.. l4. I1. .r w. r‘. i. u y u A:
sax. ..,.
$082.9. 5 ..
~28 30.8333 A903 :3! ,J
+ 3.3.68.3 5 .333 Mo. ._
=08 ‘ ..
,4
$88.: .8. \ ..
.350 3063356 A9693 33.! a ,
5 4! 8: no. .\ ...
633. .. ..
x a.“
- /.
‘\. p;
.\
\.
x
s H a I | u a
he -Mo 0 No ac
$883.8" ..
~28 30.8336 503.. 33.8 a
+ 3935 .5 4! co: no. \.
A38 \
<Vu§d hm" .\
~28 3.85.8 .98.. 333 \
5 at we: no. \
689.: \ ,.
\ s/
\ /,
R .
.... ,
~\ 4
/I!
Lo .8 o No .8

216

 

 

 

 

 

I‘ll .. i
l.
d 1 a 1 11 41 1 Na . 4 . 1 14] 1 d 1 l. I I
. . 11... >
. . , T < ,
ob <§E .. ob 3.836.333 .08.— .53...
.28....62535538. . .zsiasgsq .38
cm 65.831838. . cm. s .
. .33 \ a . .35
PM 38931.9” 4 9V T <§..QB
_. 3.8 35.53.. .08... 3.8. . 3.8 3.353. .98.. :8...
0.0 54.50038. . , L. 00 - 5410038.
. . 6.8...
Aggy \\ .
9m . _ pm - ..
\ _ .
o.A_r .. .x m ch 1 \.
. ,, . .
o... .. .. . 9w . ._
.. x. . ,. I m \.
9». . o.» . .
_ / _ ..
o; W \. \ .z... O.‘ r x...
O_V 11..-- . .r . .1 L. . 1 . . . F . .h 1,. L. 07 1 r 1
.8 .50 .3 o 8 8 8 .8 .2. .8
d . 1 4 . r 1 41: .1 . ., . . . A n. 11 1 14 1 _
a . .2 ._
.... .. M 582”?»
0.0 T <§g . .. fl 0.0 r g §<§ Ag 30:6.
f savanna—551.835. .. .. . +5185€888
oh .3... . ,. M ob T .3...
\ . .
o u - 5&8... . .. ,. oh 389......
$3.835 .88.. 3.8. \ . w 3.8 8.8.5.3. .083 3.3.
and. .. ._ s 188.3 .
ob - 638. \ .. r. o... T 6.8...
.. _
9m .. - .. _.. .. 3
o ‘ \\ , o b .m
. . .. .. _ . . .
o... ._ .... o... ..
_ . _ _ \
o.» __.. ... . o.» M. \.
_ . . . \
/ .
O . .. . M , . .r , 1.
8 Lo -8 o no .8 8 8 so -8

-l' '

, ,
__..._..;._. -..._

217

v 1 . .
fiWi _....- H... . a..- . .— ._..--._.

._..s--.-. -.

L._-.._L.

 

 

‘11!

 

11 ..Ifl , .. A. 1.1- .1} . . a a
o 0 389.9. . ..
. 8 gas .08..- 388. .
+ 883.8 5 683!
Pm - =8. .
°.V f. <V§f§
8 3.8330 60.8.. .633 .
5 880.3. ._ ,.
0.0 w Aggy \ ..
9m . x. \ .
. .\
a... _. . ..
_. ... .,//
ob” ..., , /
o.» _
. x
o.‘ .. /
_ /.
O #1. . . - .1 . . . . .
.8 to -8 o No 8
d a- 11. A . . _ . ..A., . . . a
o o W $88.9. .
. 8 330 A909: :33 ..
_ + 8.8 s 889.8. ..
o... w .3... .
.
o u .38.}. ..
. fl 8 3603!? A83. .83: ....
s 6836. ..
oh. 633. .
o... , \
p. .
.5 I.
.
_ .
o.» _. .,
.
o.. a ,.
O .. r . . .
8 .8 mo c we 8

 

__. ._ -¥_-.-...__ .4

 

9 .o
M 0)“

9
d

. m.
a:
_

91

-..I1- .4. . 1.1 . 11

<32!

.38 38333. .8 .63..
+ 393.... 5 538'

.85

<§. .g

8 3.8336 .93. :33

5 883'
.9003 \.
\\
...,
..
..
.\,
\\
.\.\
.8 no
i. .
$89.!
.38 3.83:5 A083 :23.
+ >30; and 5 883.0.
.39
$83... .9.
.38 3.83.23 .83.. :38.
5 8836.
.98...
\M
..
\ .
.8 mo

.

1
-_'_._ __J..---__. _...._.-._L__.'

 

218

 

 

‘fi. 111 Ill. -. .n .. \f... .I. 4| .3
385mg /__
ob - 38o 88.5.33 5 €836 \ W,
Ag _ ..
Pm . W. ,1
388.8. . W H
Pu . ~28 88.536 5 4.28.. no ..
6303 ._ .
ob . _
Pm . .
0.; .1 x.
o... 1 \W
o.» \
\
o. \\

 

go
3
8
o
23
s

 

.— . 1 1 1 .. . . . 1 .
o m , $899.
. 3.8 3.8.5.3. 68... :35
+ 688.3 5 680.3.
ob 1 .35
o N W4 <§d.&
. 350 3685.3 .99.- 3.3.
+ 683.3 5 428: no. \
o... T 6.83 .
Pm 1 1
0.5 _. .1
_ ..
W .
ob W m.
o n W W.
x.
s_
_
O W W h _ _
.0o .5 .3 o no .8

 

 

Po 1

ob 1

PV 1

9m 1

ob ..

9w ._

o.»

o; ..

 

3‘?

T.

3..

 

.0
w

9
to

a | a 1 .\J/1

. r
W ,_

$82.53 W

3.8 360333 6%.: :23. .

+ 3.36.2.5 s 5833. .

.38 .

5898. ..

\\
\ h 1 1 _ 1 H
.8 .8 o
. . A J. .
(385.3
3.8 3.5.3 .98- 33. ..
o 39.9.! 5 8893. .,
is
$808.8. L.
~28 3634.3 .93.. 3.3. W
+ 26.9.5 5 as. 8.. 8. W
8.8... x
\ \.
1 _ _
.3 .8 o

 

,W._,W.___ WWWW... _ . ~W_-_.W WW.._. 7 7.

mo

219

 

9o.
2..
o...
0.3
on:
of
a.
3..

0...

8°. 1-

a . 1 «1 .1 .x/ J I 1 q
._ .
$3090.00. .... .
swam 30635.3 .93.. :23. .
603.»... .5 €803 ..
.8... ..
$088.3.
Noam 30633.3 .93.. :93. .
$823 6 a... 00.. mo. ...
.93... ..
\.
\
\\
. . n
to no o we #o
u/.
<38... .8. . .
wag 3060.553 .08.... 30.3.
5 .6803 ..
.3...
38%. .8. ..
8.8 3.1833... .082. 33.... ._
5 4500.. no .
.98...
..
w.
\.
.8 mo 0 mo .8

.0
on

o...
9w.
9».

o. 1

2°-

.1u. 111—

$888.00.

3.8 30.83.50 .g 30.3.
+ 202.636 86 5 43.00.. No.
6.00...

$823.00.

3.8 3060.533 .98.. 30.3.
+ 30020036 and 5 60.3.3.
.3...

220

Figure 2C. Absorbance EPR spectra of V74C/I98C double mutant in detergent (Tween
20) and liposome and in the presence of Co-PPIX, arachidonic acid and various COX
inhibitors.

Spectra were normalized to maximum absorbance and were uncorrected for any
contribution by monoradical species.

Images in this dissertation are presented in color.

221

 

 

 

 

 

 

 

it .4 i I «
ob .. SEhB
3.8 88.5.3 5 418.. no.
o a - .3...
$389.03
oh 1 38 3.3 .88.. 3.3.
5413...
oh I .038.
Pm , . ,.\\
\
\
ob .. ..
ob r
o.» -
_.
O —- lk . F
.8 18 no
‘ «ft! . t 4. .
o @ $823.9.
. fi .28 85253.98. .63...
+ 853.3 5 .230: no
ob .. .33
. . 38 g3 .98.. 8.3.
s «to... .8.
Pa . .38....
o... _. x
p. ..
_
3.
_ ..
o.» _ \
_ .\
0.. . x
o w. . F
.8 .8 .8

Br

A..—-—A—_~—y —‘—r’ HEAR—4

_-n-—-A‘—_.—m-+‘ -

 

d—fii -l1.Y|.«|4¢ t I- ,1 ...
<32:
ob .- 28 3615.... .08! :8...
+ 50.0: 3 4‘00: 8.
ob - .3...
o a T 3309...!
. 2.8 3.833 .93- 3.3.
5.733 8.
oh 1 .98... \
ob _. . \
ob ..
.
93 .
_ x.
o.» _ \
. \.
0.. .
. .A.\
o - f ,. L L.
.8 .8 .8
d .. . . . .
_
ob $895.02
_, 3.8 20.8.5.3 .98- 3.3.
our +§oxssqtooa~9
. _ .38
oi- 3.89.3
_ .38 33.3.. .98.. .53.
o o .. 541.38.
. _ .98...
pm. -\
_ .
o... ...
_ .
9,... \
h .
O.” .. \.
p. _. .\
0 fl . t. . .
.8 so no

 

222

 

‘

’ "fl” '7

 

.0

 

 

 

 

. -.H... t i it 1 ..t\[ . t . i. . t . 1- 11.-. ‘1‘": dlllvr.( 3'. ~-II.
3.88.8. . , o 0 - 3683.0...
V200 30.00333 .900.— 393. .... .. . V0.8 §<3 .933 :03.
+§0.wms...1002~0. .. . +£82.03... 00.050.30.30
..00. .. .. oh I ..00.
.
E898. _ . o 3 - 3.89.8.
~50 30.00333 .9603 30.3. .... f, . V50 §N<nd .900.- 3026.
s 4100.. m0. .. 5 43.00.. m0
.98... .. .. - 0.0 - .98...
\\\ / .\\\
/ 9m ..\ \
_ \
.. / o... H.
. / . _
x .. 0.0 _. .
... _ _ ...
.. ,., o.» W.,
v 0.. _ .
.r- t t . t . . .-_ O W . AI... . _ l
.00 -mo 0 m0 .8 mo .00 #0 M0
.l .1 ..-.1 .1 . ..J t 1.. x A ... (1* in it- ._ ii ..la ...1
3.80.8. .. ‘ 0.0 . 3.00.0...
.3200 0000233 5 800020 . 3200 006023.43 .9000 30.3.
..00. .. . + 30.06.10... ... .0803
A o... .8...
.
9:83.00. .. _
2 .. . .. . 35003.00.
.38 502520 .080. 28. . .. o... 2.00 8.83.20 .08.... .020.
.... 380.20 .. ., A 5 68020,
- 900... \. . , ob r .Qooa.
\
M O.“ .---.\\
x .1 _ \
... .. . x
.. ... . o... .
.. x . . ..
/ .. 0.0 r .
_ _ .
m 0.» . .

_‘_,l"‘

 

223

.-.-J _

 

. ... . -
.
0.0. $000.00.
a ...—.00 206023.20 .0000.» 2020.
+ 08.0.02 .2 $8020
.0... 3.8.2....
3000 20.002320 .0000: 2020.
... €80.20
...... .08...
o... ..
o... _
_
0.0 __ ,_..
._
0.0 _ _.
_ ..
.... __. . ...
o o. l .. .
.00 .00 -mo
‘ 0+ 1.1! . . 4
fl
0 0. $5000.00.
. . 30.00 20.83.20 6000.. 2020.
. + 0000.00 ... 88020.
0.0.. ..8.
o... 0 3.82.8.
. 3...... 8003.20 .98.. 202...
fi 5 $8026. ..
pm .1 .,
_ 683V ...
w x.
0.0 . .
,.\
o... .
. ..
o... .
o.» .
_ ...
0.. _ .
_.
o . w . ‘ _
.00 #0 .M0

0.0 .

o...

0.0 ..
0....
0.0 ..

0.» .

 

 

.1'$i|1\i|fiv ulxl- .iI .. ..o. In. bifl‘xl. . L o. 1 11 11.1
..
3.82.0... .. /
V0.00 20.002530 .900- 20.3. . ..
+ 2006...: s 600020 .. .
.30. .. .
3.0020... .. /
30.00 20.00333 .0000: 2020. ..
.5 {80.3. . .
.068. \\ /
.
a \\ /
\ /
. /
\\ II!
x.\ .....n.
.00 -mo 0 M0 8
‘. .-li .‘. 2!, ...? .w 1 . .1.
3.00.8.8. ....
0000 206023.20 .9002 20:0. , .
+ 3002.002... 000 ... 68020. _ .
..00. ,.
3.82.00. .. ..
3.00 20.002320 .900: 2020. w
.5 {80.3. .. .
.3002. ... f
x .
Sx .
..
. ,
..
\ . /
\
#0 -mo 0 N0 8

x I
-. A ._..__._.__.__..

..8

224

92‘
9:.

oh .

 

Q 4. : a, El
Suwannhn. H‘
~28 38333 5 ...: 3: Mo H
683 H

‘
<§8§§ 5.

:3 803335 5 €893 k
=8.

.x‘ ,ILfi ‘ _ .I. ‘1

$82.3.

«Q8 3.8338 A003: 33.3 u

+ 858.3 5 .6823 a
:8.

<V§§.QN~ ..
$28 30.0333 603.. :35

+ 853.0: .5 at 3: no.

632:

 

r—

- ,
fl , figt..____1

 

ob

9m

PM

Pa

71.

 

 

“V‘llt #1.: 1

$309.33

SS6 30.3353 Booms :28.
+ 32290.8 5 €836

A33

<w§h§

Exam 30.8333 889: 33.3
+ 3:65:33 .5 4&8: no.

633

3585.8"

Nam 30.83.36 803: 33.3

+ 399:: .5 582.8
:8.

3383.9:

~28 30.83356 Boom: 3:3
+ 399:: 5 ...!oo: no.

633

‘ .

L.___Lfi_AA_;.—_g .r __.—V _gwnx.-- __4_ “4*

225

 

.0 9

.0
Va:

 

 

. II -III - I I
III , : I.I , . II ,J— I III IQI I ,1 I I I A I
..x ,, 4 o o - 5383.3.
7 5.88.3 N . 4 . 38 8.853. A98. 833
38 3.83<§ A983 :23. 4.. w + 28283 85 s a: 8: 8
+ 88.»... 5 «s8: 8. “ ,J _ Pa 688.
003 ... .,. ..
Ac v N . w O.N r (Nghmn
. $.83». \ I4 7“. 38 §<i 689.. :33
~28 30.8333 632. :33 \ w. n o m . + 233.836 won 5 680.8
. + @828 5 .883 é . Ass
53 _M o ..
1 .m I
.__ _
V, o.# F ,.
. ,* H.
_K _ .
w ob I ,
__H X
M o.»
I_.
_
O _ . III .III II _ _
.. i _ I- -I .. .. ...- _ . ,II L. [II ,u two .50 IMO
Lo .wo 0 mo 8 mo
_VII I I I w I I H I d , H ’ I a
_ 4 M
H $3839,
A 38 zoaoa<§ A98.” :33 _ _,.,
! 5 :28: no. _h h
_ 6.83 _
_. <§83a2 x.“ ;
; :80 30.003335 A003: :32 _
_ s 6823 _
_ Ase w
_ I”
, 5
9L .
i x A
O _ H I _ I I

 

1
_-4

 

226

Figure 3C. Absorbance EPR spectra of H75C/I98C double mutant in detergent (Tween
20) and liposome and in the presence of Co-PPIX, arachidonic acid and various COX
inhibitors.

Spectra were normalized to maximum absorbance and were uncorrected for any
contribution by monoradical species.

Images in this dissertation are presented in color.

227

:III I I J. I _.
fl

3..

ob

.0
a:

.0
A

p o
co 01

.. /w 4 a
Damage.
.380 980333 5 «i 3: no 1,.
A33 2‘ a,
c3838. 7 F...
~98 30.83225 3 at 8: Mo , u
683 , 2.
M
I 2,
.\ ,.
\ r
’-
x /
/
\ //
\ ,
c /
X x.
\ /
l H _ H i _ . i
be -mo o no .8
a I w . . » a 2 J
Qflmwgzufln ., ,2,
~28 3.8333 Boom: you! ,2
+ 8583: 5 :28: no .. .,
A69 4.
c3253.?» ..
wag :o_oo=~<:.o Boom: 3:3 .. 4.,
5 at can mo 1,
8.83 .t
\ r
\ ..
.u xx
its /
m
.\ ///
\c
\ 7
\\
. /.I
be .8 o no 3

 

 

We

, I I
,. V». 2,fi_.._~r-

mo

ob I

ob .

oh I

9m I

 

 

.0
(p d

.0
on

9m

o.m__

0.;

9w .

 

..IIu.... . In. I

c3839.

«28 30.33336 8085 :28.
+ 3:66.23 .3 «23: no

as...

93083.9.

umam 30.8333 803.. :93.
5 ...i 8: no

6.83

a I ~

05893.3.

«mam 30.83336 Anew»: 73$
+ 399.5 3 as. oo: no.

:8.

c.3899:

umBm 30.8333 A009: :33
5 A! we: no

A98...

#0 .No

; no

8

228

ob I

Pu I
91
Pm I
oh
oh .
pm.

on.

 

8‘.”

ob.

ob

oh

 

.-.l..|lf-Jl-.- 4

02.89.93

38 Eva; A90.- :33

+ 338 5 «to! no

.33

ugh-B

~28 SE: .98.. 333

5 ..f 8: Ma

8483 x

.4.
Is I 8
33352
ago 88353 5 8823
.35
c.3832
«Q8 §<So A082 333
5 g
. 63.3

ob I

Pm .

9% I

ob

 

 

 

I,

\

s

«
A
_..

-. II- I4II -- 4.-. I.
a?
«28 3.833! A9; :3!
+ 283538 on: 5 ...!!! no
is
389...... 4
was 3.8333 689- :3! .
5 «:3 no
633. x.
\
Mo .8
«
cumming
ES 353%? Ag :33
+ $28 5 €3.46
:8.
0389.3.
~28 §<=d 889.. :33
5 8888
683
.8 No

 

__4L_-¥,; -..u_,1.,

_'__-_-_L“!,_;v

229

 

 

.0
01

..O
‘5 _._..

.o .o .o .o .0
__-.._...‘" 9’ “ °° 5° ."

.o
:5

4 a . I
0.38.2:
. 348 30.8.5443 Boon: :33
+ 8533: 5 6804.8
. cos
- nun-.83.?"
348 30.8353 303.. 3234
5 E836
- 383
I. p I?
Lo .8
4 - 44
9.33.9»
348 30633418 Ag: 46:!
+ @084 Na 5 880.3
:84
cum-.8394
Noam 3.8336 8002. 3:3
5 .5838
6.83
4
.\
I8 -8

I'll-[h
I} 1 I
... 4
4 .4
.4
.4 4.
s .1
b. h -
a no
\4/
44 44,
4 .4
4 4.
4 4
a 4
. 4
N“ I
.4.
4 -_
0 mo

 

____..__4_ #4.__ - 44.44

8

Po 4-
o..~ I

Pm

 

 

-, III 4
0385.84
3.48 3.8338 .93.. 3:3
+ 288.3 5 .6836
438
0389.03
.348 3.8353 Agog :23.
5 .9836 .
4933 x.
\I.
.\
4x4
.44
4
I8 -8
I 4
03343.94
$48 30.83356 Ag :02!
+ 332838 >05 .5 {80:6
:84
c.3839:
3% 30.8333 689.. :33
5 680.3
8.83 ..
x
‘ \. I. \.
~\.‘
m
.48 .8

SI

 

 

4.4.4 ._ _._4_. 4 4

8 L

230

_u—I.

 

‘4-

Po .

ob

 

ob ..

ob.

o.» r

 

. .o .o .o .o
N (A) A 0|
.4 4- 44444444

0
cs

 

O

12: 4a. -1; J!
c.5883».
.548 383%..“ .5 4.4444 8: no
583
oummEhm.
~28 283336 .5 4680.8
:84
.44
4.. 4 w
..8 -mo
. ; 4 . 4‘
0389.8.
348 46400353 400344 7234
+ 84.6343 5 4.28: no
8803
Emma-hm.
~28 30.83336 4969.. you!
+ 888.3 5 4.6835
. :84
\x...
4K
4
\4.
.4.
. 1 4 w 4
.Ao -mo

4

.4

ho

 

 

ob T

Pm I

Pu r

PA ..

 

.0
to

p
o .5
f—T“—r -

é:
o

4 4.

4

9o 4
4

9m
o.

4
4
4
4
V 4

 

 

Illwti «4|:0 III 4. - .-N‘ \ ‘. 41 ‘ a 1‘ ‘04
3254.84 .. 4. 4
.548 30.83436 Bag.— 33: .4 4, 4
+ 323.643 5 at 8: no .4 _
6.834 44
cummmahg 4... .44
33¢ 30.8333 683 .63: .
+ 32290.8 4: 4433.0 4
Aqav 4

_
4
4.. 4
.4
\ 4
4 - 4
.4 4
.. .4
4... m. _
r/ 4.
/ 4
\ L ‘ 4 x .7 .4 Y ~11- . L
Lo .8 o no 48 mo
4..
c.3858. .4
.548 45.84.4335 Boom: 453$ .
+ 2229.5 4... 460330 .
489
c.3899:
~98 340354.43 40394.4 30304,
+ 2885: 5 44:83 no _
68034 4
._
.44
\4
4
4
\ 4
4 4
..4 4
4
_
» b{ r b! 4
Ac -mo o no 3 mo

231

PL.

004

o..\ 4.

car

99”
0.».

P:

.4.

0.?

ob.

91.

99..

‘1.4 ,|

0588.8.

.548 30.003350 40000.. 30.3.
+ @0843 .5 4.4.30: no

633.

c.3838.

2.048 30.00353 4003.. 30.8.
+ @998 5 {80.3

4.2..

. \

05083.00.

$48 30.003336 RUE... 30:0.
5 4t 00: mo

638.

0383.0».

.548 30.003358 40000.. 303.0.
5 {80:5

.48.

.K‘\

«8 -8

O
44.4.4.4_4.4

4 444*4—hi 4 4 k

.44-

pl

4

4
067

ch

(liix ‘ [[Lfi

05890.8.
348 30.00333 8000.. 323.
+ 2003.002... and .5 ...too: no
633.

@385QO.

348 30.003536 Ag: 303.0.
+ 233.0006 and .5 88036
:8.

Lo -8

mo

.5

232

Figure 4C. Absorbance EPR spectra of Y76C/I98C double mutant in detergent (Tween
20) and liposome and in the presence of Co-PPIX, arachidonic acid and various COX
inhibitors.

Spectra were normalized to maximum absorbance and were uncorrected for any
contribution by monoradical species.

Images in this dissertation are presented in color.

233

 

d IIIIII. I; Ill'JI .. . II oIII .III .. 4| . I‘ — “ .lliIT III II
5
$8898. 5 .5
ob I gmgzadezooamo 5 5
5 5
A33 5 5.
ob . .5 5
5389.85 .. 5
~28 358258 59.8.. 5.985 5 5
o.» s as 8: 8. .5 5 .
8803 ,5 ..
Pm 5.. 5.. M
5 5. .
0...5. x... 5, ._
. / 5
a... 5 / 5
. 5
9w 5. , .5
_ /.,/ 5
o.» _ . / 5
_ ....
o5 _ .... .,_,

_ x 5
OF ..-\II\W 5 5 I I L ..T. , .. I 5
.mo #9 mo 0 no .8 mo
d wI .2!!! 5 II. I q . . II. ...I\\4,.. . 5 I .|.| 5 II.. I A—

, 5.38985 5. 5

ob 3.8 362353 59855 3.85 .5 ... 5

+ 8.538: 5 ...: 8: mo. 5. ,. 5

9m _. 5.85 .5 5 5
.. 5

_ 5389.35 5 5 _

o.» 5. $8 363535 5885.. 8.85 .. .5

5 5 a! 8: mo. .. 5

9m 5 8385 ...

5

:5 5

5 _

o... 5 5_

5 5. /,.. 5

0.0 _ c, /. /. ._

5 ./

o.” w. \ / 5

N . 5

s5. . 5

5 .I\\\ r 5
O _ . \. . 5 5 I 5 . 5 I 5 5
.8 L8 -mo 0 no .8 mo

 

 

9o -
ob I
o.» .
ob I
9m ,
.ok. ..
o.» ,
o.» a

o;

I

 

pm I
3 .
o.» ,-
pa .5-
pm .
o...

ob I

 

o.»
S 5.

05
.8

II- :4- : I a - -
$388.25
38 8.833.. 889- .633
+ 9.6.63.3 5 4! 8: 8.
:85
3883.3.
V28 363355 5003.. :33
5 as oo: mo. \
8803
., \
\.x\.
s ~\.
I .I . _ I I
#0 .8
3883.3.
S 3.09523 5003.» :03:
+ 7385: 5 4i 8: mo.
:85
<§hm~
38 55.88258 598... 3.85
5 qt 8: no. . ..
6303 ..x.
\
\.
...\
\I. _ r
#0 .No

 

. I

 

234

' 5- 5 55555555J

 

 

I I I I
...__.__~_5_.5___ 5

 

 

dfil I. Il..|. .5. ..IJfiI I.I I . .Iw
<§§.&5 ....
ob .358 §<§ 593- 3.85 .
+ 55558 5: :38 B. .
pm . 5.85 ..
$827552
oh I 358 3683558 59855 8535 5.
5: 5.5: on: no. ..
9m I 8335 ..\
5 ..
9m \\\
55.55. 5.
5
3 5.
\5.
o.» 5- \
5 5
95 5. \\
5 I \
O .I \I \I 5 5 II I 5
.8 to .8 o
d I II 5 5 5.5 .
9m 5.385252 ....
358 8855558 5: 6898 ..5
pm . 5.85
5.385558. .5.
PV. ~28 30.823553 50825 56:55 ..
5: 56805.8 ..
5:5 - 555.82 \
9m I \
o; 55
5
ob 5r
PM 5.. ....
5 ....
55.5 5 ....
5
O . I \x 5
he be .3 0

ob r

ob

9V 5.

ob

 

9m .5.
55......
on
o.»

0.5

 

.0
V

.0
a)

.0
or

II I 4 I I J. #4.. II I d J
.. 55
Sgan. ... 5
3.58 358.536 593- .633 5
+ 28:50.8 8.5 5 at on: no. _ 5.
538 _. 5
$88.53 .5. 55
.358 30605356 508955 .532 5... 5
5 at 8: No. .... /
598.55 5/
s . I./l
.. /,
M
.5.
... ../
5
\\ _
\\ .x/
/’
L 5 I 5 . I
Lo .8 9 mo 58
5 In I 5 5 II
$895.85
358 358.5353 508955 3:!
+ 35586352.. 5: 5.60853. .. ...
533 5
538555.85 . ..
as 333.... 5.985.. 55......
5 88058. \. 5
6803 ... 5.5
.5.
.\ MI/
5. ..z
x ..z
N
.5,
. \. ,
/l
5 5 5 5 ul I
.58 .8 o no 8

I - I I
5.__ __.—__.“- (mg—.54

 

235

0.0

on.

 

.0
V

.9
a:

.0
01

382.8"

~28 3,835.8 5089: :3!
+ 858.3 5 8823.

539

$8278. 5.

38 8.8:, 3:8 598: .633
5 .6823. ..
6.83 \

\I

8%...
3

A II —

_. $8225.

~28 3.8.523 .98: 3:!
+ «02:8 5 .6823.
533

3895.8.

~98 363%? 5089: 8:2 .
5 .6823. .
6.83

 

ob.

Pm I

9V r

Om .

 

8.
o; 5..
3
on”.

35

3:
9m.
3 5.
of
o.._.
8.
S5

95..

I .«III . . a
$385.8
~28 3.83336 5089." :35
+ 508...: 5 E323.
583
383.3. .5
~28 36033.3 593: .533 .
.... $80.3. \\
6.8... ..
\ .
\
.\
5.
L8 mo
.5
<w§.§
~28 3.83:6 5003: 3:!

+ 283.828 you a .6893
589

<38.th

~28 3.08335 5985 .622 .
5 689.3 \
683 x

\
.\
..
L5 mo

1
p.

p—

u
c
.5.“

555.555 _55i._

236

.0
a:

.0
u:

o:

6395

I Ifi I 5 I I J
5.. .
..5
$8258. 5
~28 28853.0 .3 ...: was no. .,
6303 5,
$3358. ... /
~98 933335 5 5.5323 f
:2: 55
. I I . 5
Lo .8 0
~ 5 I ‘x. 4
\ .
$8838 5‘ ./
~§ 3633.3 Boom: 3033 5 ...
+ 838.3 5 .7: 3: no.
6803 ..
$88.3.
~28 72832.3 .083 :28. ~
+ 8583: 5 .5823. \
533 \
.\
5.
_ _ 5
L5 .8 o

no

 

 

 

ob -

o.~ .

oh I

 

9m 5.
Pa. 5.
3
o.»

o; I

 

2:?

. III. III... III IIIIAII

38.3 5
~28 388:...» .98.. 3.8. 5

<~m§.3~

~35 3060333 593.. :23.
+ 3:66.28 5 43.3: no.
6.83 \

$955.8. 5.
~28 30.83323 603: 32.3 x
+ 399:: 5 5823. 5..
538. _.

<~a§9am~ \
~28 3283338 5088. :28.
+ .6085: .5 .7... oo: mo.

AQooa

Lo -No

...

 

237

 

 

"'
fly“
'4 ,fi.

...." " ‘

d I. .- I I. IJYI IIII I4 . ,... I1; \J/II
383.3. .55
9o . ~28 3.83:8 598.. 83: 5
+ m9? mm 5 6823. .55
ob - 583 55.
o N I <~mmmnmn§ m
. ~28 8633.3 598.. 3:3 ...
. 8&5... 5 :38 8.
0.3 8.83 \
9m _.
o;
Pm .
:5
s5.
o 5. ... .. _ F I
.8 Lo .8 o
A «II I 55. a I \J
o o 5 $823.9" 5..
. _ ~28 3.8353 598.. 3033 _
5 s 6885. .5
Ob r :95 m
._
. ~28 3§<am 508mg .633 .
o m 5 45003 no. \
. 5 A9003 \
9m 5 \ -I \
o... .\
5 .5
Pa _ ....
S5 _
o.. _. x
.5. .

no

 

 

 

 

 

-.J5J

 

ob .

ob I

o.~ I

9m .

Pm

 

I .. I... I—I ..I I I.~..

$883.3.

~28 3.8335 .082. :23.
+ 28:58.0 >o§ 5 «too: mo.
6.83

3828.3.
~28 3.8.555 5080: 333 5
+ 28:62.3 >06 5 680.3.
=69 \
,\
\

~
~
.~
~

 

81...— 5-555-555

238

Figure 5C. Absorbance EPR spectra of I77C/I98C double mutant in detergent (Tween
20) and liposome and in the presence of Co-PPIX, arachidonic acid and various COX
inhibitors.

Spectra were normalized to maximum absorbance and were uncorrected for any
contribution by monoradical species.

Images in this dissertation are presented in color.

239

dfiIIII I. I JII . IJIIII . IIwJfia, I
$393.8. ... f.
ob - ~18 88333 5 «too: no 5. 5.,
3.83 5 5
ob . 5. ..
03895.3. \
o~. -§Z§<§5§§v 5.
. 5 :38 mo.
533
APO . .....w.»
9m 5. I ..
.\
Db. . \
.~
~
9w . ...
.7
o.» .. ...
o; .. ...
o I I » I II _ I 5
.8 .3 .3 o
d I IIJ I I. « IIII- III,
ob f 9333.3.
~28 3633.3 53% 333
+ 85x28 5 45!! mo. .
°.Q 5 Aug .5
o.~ 5 5.38998. .5
3.8 35%.: 59.3... .33 5
5: 4:8: 8.
ob T 5983 \\
9m . I x ,
.\
ch ......
pm
5 _.
o.»
\
o; ....
o .I F I
.8 Lo -8 o

 

 

I
5.555 ‘_ ___5555'555_L5_5.I

-15555 55..l_.__._'

 

 

9489...»:
ob . ~39 35.33. 593.. 8.85
+ 2653.! 5 ...ioo: no. .
oh 5 535 5.
$393.8. ..
c.~ . ~38 §<§ 68: 8:3 5
5 #38 no. ...
9a . 683 I ._
\.
9m 5. ... ..
5 \
ob I \
ob \
o.» 5 5
o; 5 ..\
O 5I I p . . u
.8 .8 no
‘55. II I IfiII- II1I
c.3891 ha"
90 - ~33 3.8.5.3 5902. 3235
+ 738...: 5 458: 8.
pm I 539
0389.3 ._._
c.~ . ~38 3.8333 593: 3:3 I
s #33 B.
3 I 683
.\ c.
9m I
o; 5 ..
5 _5
, 5.
9w _
5 ..
o.» 5 .5
5 \.
x
S 5.
5
O 5- 5 I
.8 .8 -mo

5.

 

240

 

C.
(D
:3“ ~ .5.555-I._ —1

35.

9+5

I I. . I .

3383.3.

~38 30.83538 5925 3:!

¢ $08.8 5 ~58: no.

583

2383 8.

~38 20.8338 590»: you! .

5 dz 8: no.

5983

x.
3
,\
\~
~._
5
.8 no
5. 5
: n-o®5aw~
~38 8833.3 5 .6893
589
©3833.
~38 3063355 5003.. 30:2
5 s83
5983
N
Am.
\..
. \
be no

~

5.

>—

 

90 ..

ob

o.~ -

 

9m 5,
Pm
.0.»

ob

 

o.»

o;

 

0.0.

99

o.~

 

J

93891.8.

~38 30.8338 593.. :33
5 qt 8: no.

5933

5582:!

~38 30633.3 508m: .633 \

+ 283.8% Sn 5 dz 8: 8 \
568 \

0.38.3.3»

~38 363:3 5003: 33.3
5 58026.

583

0382 .3.

~38 3063326 5080: you! + \

$3.353 5 339.8.
538

 

241

ob .
ob .

o.~ ..

 

0.8
0...,
o... __,.
o.»

o:

. ‘ I 41“: , 4. 4 . ~ ‘ \
C3893.
~38 30.00333 6030.. 30.8.
o 8:03.03 .5 880.3
.30.

0383.00. ..
~38 30.00338 .98.. 30:3 .
.3 .0893 3..

6.83 \

038.000.

~38 30.00333 8000.. 3030.
+ @898 .3 58033.

.88

0383.00. ..
~38 30.00338 .002... 3036. .
6.83 x

V . .
___-_ ,...._.._..____ . __. 3 .-..-.

 

. , ‘

A
_. .-.—.a.

Po 7

ob .

o.~w

 

0.3

Pa:

99.

‘1‘11‘4 \aT . 11 J
\ .
0385.0». \ ,
~38 30.00338 .0000. 30.3.
+ 306...: .5 {80:6
:8.
c3833.
~38 30.00333 6039.. 30:6.
5 88036
68... \
h\.\\
_\x
8 I .8 , o
x . ,,,
ugh». . .
~38 86833 .082 33a. ‘3
+ 283.0035 mod .3 .0803 3
.88 ..
. 0383.8. .
, ~38 30.00333 .0032. 30:8. \
.5 E8036
.0003 \
‘\
.3.
#0 .3 o

. . ‘ s
“fir _AUA-; 7,—— --__.___—_4_—._-

 

242

 

ob. 9.38.0.8.
~38 88:33 5 €803
pm .‘ .39
9383.3»
Puf 3‘8 8033365.?838
8303
9?
ob h.
ch .
ob r
o.» ._
o; 1.
o f 3‘ ..fi Lr H
.8 .8 -mo
. . I l‘ A. _
ob, aloe?"
~38 363226 Boom: 321
9m 1 + 8588: 5 .8823
.83
o.» 7 0383.8.

Pa

«38 3069.535 A008: :33
r + 0533: S 458: we

 

6303
oh I \2
oh , x
3 .
o.» .
9. . \\
_ .\\
O.‘ u r ~
ho Lo .8

 

. 1
Po 1
obfi
oh 7
ob ‘
Pa
0; 3
ob 7
o.»

o;

 

Pm“
ob a

93

.t‘lw IIJ‘ltl .u

 

In- ...»; ,I 4

c.3839. .... ..

38m 3.33:6 A9094 :3! .m. w
+ £18203 s 589.3 .m .fi
A33 , a

9383...... . w

38 §<§ A98: 3:! ..w

+ 38.23 .5 at 3... no _w
Aggy \

«i ‘ ‘ ¢ ‘ _ NJ

@3859. .. .
.38 3.83<§ 68s.. 33 \ f.

+ 22:93 5 .6823 \ A,
:8. _ ,

“38%. .8. . _,
38 30.83325 A93: 30:3 \
+ 75??! .5 4s. 03 no \
8.83 \

“l ll 1

 

 

243

ob r

ob .,

oh 7

ob.

Pm r

oh 1

ob .

o.» ..

 

:7

_
ob _
ob 4.

P?

v

o?

v

9%..
......
pm“
on...
......

0..
8

..-.Jf 1:11:13 3 .:I .
3.383.393 .. ,.
338 3060333 A930: 33! 3. 3..
$82 mm 5 .6888 .3 .,
can: ‘3 ..
a
n35»... .93
«.28 30.8333 393.. 30363 \ 3.
5083 mm .5 d... 83 mo 3. .3
8803 ....
N...
_kx
3
3,3
\ 3 _: wwh, ll —
Lo .8 o No 8
‘ ‘fi 3 . x ‘ 3.3 4 ‘ } a
030g.§ ... .
dam 30.83336 Agog 3233 .3 ..
5 880:5 ..
3s... .. ,_
238%. .83 ...
33.8 30603348 393.3 3033 .. .3
5 4s. 33 no \ ,
638. .\ ,.
.\ M,
‘.\)
\s
3
.3
3.“
.. /
.3.
\\

‘____. _L _. __.; .3..- __;_ 3. 7__l____ 3.1.—3...;

 

8

ob 3‘

9m 3,

9V 3.

 

9m 3.

lzfillt . .r.

3.38.8.8.

338 3063336 3080.3 333
+ 28353.0 won 5 £828

388

3.38%.».va

.338 30.83336 3939: 394.3
+ 9803.838 85 5 3:3: mo

6383

L“:

l

)

L . - , v .
__.—3 __.__._-3_ _3—.+_~_3_3..__3___._.-__—.4

 

244

Figure 6C. Absorbance EPR spectra of F84C/I98C double mutant in detergent (Tween
20) and liposome and in the presence of Co-PPIX, arachidonic acid and various COX
inhibitors.

Spectra were normalized to maximum absorbance and were uncorrected for any
contribution by monoradical species.

Images in this dissertation are presented in color.

245

 

3

ob

ob

9V .

3!.1; . _.

.3908». do.

mtg 38333.1 .5 .753 83 mo.
3383

<28?th

gm 3283.36 80g: 323!
.5 «133 no.

6383

 

.o .o
N 0)

.0
o _s

90 .

ob ‘

ch.

38qu hm.

gm 30.03218 Booms 303.3
+ 8:903: 5 3.! 83 mo.

:83

oh

. <28a3~§

. 5 3233 no.
633

 

.099
HMO.)

O

38 35333.3 398.. 3533 \

 

 

L3-

mo

 

8

 

 

 

3 - - .- .194 . ,i
o o 8.89%. \ 3.
. .. 2323;33398333 3
+ 3:33.333 5 5.13: no. .
9o . 3a... 3
. 2% 3063336 Boom: 33.3
5 4538: 8.
9a , 683
Pm. \ xx
oh ..
ob . ._
o.» .. 3.
x
O . . 3 F
he be .8 o
d _. I A t t .q M1!
0 o . <28aa§ 33
. 9.8 30525333 Boom: 323. 3
+ 399.5 5 as. 33 .8. .
o..~ $383.3. ..
38 36.8333 398.. 3.233 \
5 .3283 no. , ..
Pm ...
3
o; .3.
ob .
o.» .. ,
o 3 \\
. \\
,\_
o .. ... 3
.mo 5o .8 o

 

 

 

 

246

ob .

oh

oh 1

 

 

$588. an. \‘ w

38 §<§ .92.. 3.5
+ $898 3 .728: no.
:8. .1

<m$mavma$ x.
atom 30.8336 Soon: :3! x

5 4! co: no. \
6.83 \

a . 1 z a ‘a

$392 .9:
«#8 8833.3 .3 689.3
.83

3589: an" ‘.
«:8 §<§ .98.. 8.3 .
5 {80.3 x.

6.83 . .

>—

31-8 o1

‘ y l
_ --.—._. ._———_~—.—« _—

‘ . ‘ . , ‘
...—____4‘__..______—._—.. ‘ .. .1..-

mo

' l
__._.__..-_. -.

_h_,__.—_.——._ _,L ,__

ob .

ob

ob -

9m.

ob

 

I 1,4 . . In;-4

§§.&.

230 30.8333 Boon... 3:5
+ 28.356 86 5 at 3: no.
:8.

38quth

2am 30.8233 Soon: 33.3
5 A! 00: mo.

33:.

3393.8.

99.8 30633.8 283.. :33
+ 3:823 .5 6836

.33

<28! .8.
«£8 §<au A0025 :3!
.3 8835
6.83

L5 -mc

‘

.‘J. 4 \
.. ./
. J,
.
.
. ..
M., _,
__ a
‘ .
_, _
H
r
.,
/
/
/
l.
I
.x/ A
.i/
p b. } I l
o no 8
a | I _
..
_
__.
.H.
/ .
/l/ .
..//
_ _ w v
o no be

247

0.0

ob

o..~

 

99

o .o

.0
a:

I [[24 .4
$585.9»
0.28 30.8333 Bog.» :94!
+ 083.9: .5 6828
A83
<98! .9:
mtg 30.8333 803.. 3:3
.3 8893 1 . ‘
A983
\.
\L
\\\
\\
he no
fl
$58.3 .9.
gm 30603<§ A933 30.43
+ «99.3 5 5823
A33
<98! hm.
mtg 3.0.8333 A93: :23.
5 5893
8303
x.
\
be no

,,,J._vf_,a..‘ _,;,,

ob

ob Y

Pu .

ob .

 

1 6 1-, . 41
<®u§hfi
9:8 30.83336 88% :36.
+ 3‘93 .5 880.3
:8.
<28: an.
9:8 §<a A0009.» 3&3.
5 6893
6.33
‘\
N\
N
x
\
x
\
\\
3 -8
55869.8.
38 §<§ .900.— 30:!

+ 232.820 and .5 .6893
.88

<98... .9.
9:8 §<§ A98: 333
.3 {82.8
6.83
\
—\~
.\
\
\\

 

H

248

.-- _ ‘-k_.__.,_ M, Aifi—L.___,,L.__,~ ,,

<wb§h2 .... I.
«#8 0000353 .5 at 00: no .. I
6.83 \ .,

<28§ .8. ... .
«£8 000033.20 .5 .0828 .
=03 .. I

.8, .8 o no .8

$535.02

93 30.00333 I080: 30:!
+ 8008.03 .5 €823

I83

<§§4 8» u .
mtg 30.000330 I930: 303: . .
+ 066.3 s 238 no .

8303

mo

ob

9m .

PM 1

ob .

 

ob...

0.0.
09.

ob

_

3..

Ob.

9w ..

PM.

o; I.

362: .00.
«£8 30.03336 8030.. 33.3
+ «66.68.00 5 680.3 \

I83

augmha ...

$8 333...... 88». :33 I
+ 20.6.03: .5 «(03 mo ...... I
A983 _. .. .... ....

.5 -8 ,. o i no: 8

3893.02
«£8 30.00333 8030.. 30:3 ._ I

+ 508...: .. {80.8 .I .
:8. 0 /
55803.02 ..I I.
0:8 363%? A98: 3:! ._
+€0x534133w9 I
AQOOSV . p

.5 -8 - o mo .6

 

 

A

:8

. . 4

249

 

..-

ob .7

. .. 4. 1.

308.3 .00.
0&8 863:3.98... :93.
+ $08.3 .5 880:0
.30.
. 3.80.... .00.
9.8 30.83:: .930: 30.3.
608.8 .5 ...: 00: .8.
.9003.
...
.\
..
.
x.
#0 -mo
$58.3. hm.
9:8 30.00333 .93.. 30:0.
5 .6888
:8.
<3803N00.
wag 30.003330 .0038. 30.8.
.3 at 00: no.
680:.
_\
..
3.
L5 no

 

 

I
.m

 

‘ 1:]. t , ‘ . . 4 . ..
$5800.00.
ob . 38 3.8.53 .98.. 3.8.
+ 20035038 00.0 .5 32003 no.
ob .033.
o .x . 6.8.00.0».
. 9.8 30.83.30 .932. 30.3.
. + 2003.00.38 00.5 .5 {80.3.
ob .. .38
pm _ . K R...
9b w. ...\
. .
8 ..
PM _.
_
S _.
o .. . 1 .
- o Lo .8

_ __ _L___~J

 

.L__.._,_;.._ _ ...Li--..fiva_’._.-- -.

p

8 §

250

Figure 7C. Absorbance EPR spectra of W85C/I98C double mutant in detergent (Tween
20) and liposome and in the presence of Co—PPIX, arachidonic acid and various COX
inhibitors.

Spectra were normalized to maximum absorbance and were uncorrected for any
contribution by monoradical species.

Images in this dissertation are presented in color.

251

 

           

 

 

 

 

 

 

 

_ a q 0 4 .— - :01.-I:l.-.4.
0.0 v . 1 ob r gown—Nab. \ /,
$0880.80 0000 3.8303. 680.. 003. ‘0
58 800.530 .0 .3503 no + 30668.00 5 3:03 8 x
0.0 1 :80 0.0 a .33 \
0.0 , $895.00. 0.0 r <889..§ .
88.3.8003 880.. _ 0200 80253 .008. :33 0
o m _3 0.5.00: mo .5 4100: no
. . 6.80 L 0.0 _r as...
\
0.0 1 1 0.0. xx.
_
0.. , 0.0 . \
oh 1 4 9w . xx
0
0.0 T 0.0 . 0.0
.\
0.0 1 0.. - \
o k1 F p p P o .. ..L . a. . . . _
.mo to -mo 0 mo .8 mo 8 Lo .8
..w _ \‘0 xx d 0 t 0 ‘0- 0 1‘1'0
0 0: _
0.0 .A <8§§ _. 0 0 0.3 <88? .9. ...
_ 0200 3.853 5.8.. 0230 0 0, L, _ 0000 300.53 .0800 .532 “
om_t +£90.0054t003mo .... .0 u omw +§0x554i005~o .\
. _ .3000 fl ., . _ 080 \
owl. <wmooaghm. __ f ., oh - 3.0083002
“ 0000 00.83000 6080 :35 , 0 _ 0000 3.8300 882. 33,2
0.3 3588 0.. / 00”.. 3338
. 6.83 \ / _ 6.83 x.
0.0. I 0.2 \\
u x , --
0 0 0 / , 0 0W \
0.00 00— \
_ 0. _ ,0
_ _ . W n \
0.»: / _ 0.»; a
_ \. / _ D ..
0.. 0, ,.. . 0.. __.
a ,t /. _ ..
O _ .. 0 0 0 t. t. . .- _ O . 0 a
.mo to .8 o no go mo 8 #o .No

00+...-

252

‘._..7 A.__‘

 

 

 

 

 

 

 

‘fi Ltitsl.t,1‘.. ‘ . vii—w. ...... 4. t4.
<880000. .. w
ob . g 30.003~<:0 .98.. 30.3. . ,_
+ $08.»... 5 5.200... no. u
00- ..a. .
<8oma3..00. fl
0.0 .. 00.00.00.025. .30 .98.. 3030.
s «:80 00. .
Pm 6003 ....
0.... -.
0.. ..
m
0.0 __
z , _
. ._ ... /, __
0.. .- .
O ... \ . L ...r . . .u,
- o be no o Mo .5 mo
‘ q. q \1/ d d
O 0 v (OMGQDNCO— / ..
. 8.00 30005.30 .0 0000030
.89
ob v .
éwQth.
PM , gm 30.003350 .0000- 3030. .4
.3 0000030
ob f .9003 ..
Pm T r
0.. T .
ob r ff .
O.N r L
c. y 1
C L h _
.mo 0 mo #0 mo

 

 

90

Pm.

9N 7

 

9
OD
- -___-_._T_ _ _- __.—

2=> .0 .o
-.- AV on (0 d

.0
a:

1.... . ,- ,. - : . -
$5038.00.
ago 30.00.3330 .93.. 3020. +
20036036 00... 5 4! 00: no
.60.
$3003. .00.
8.00 30.00333 .90.... 30:0.
5 0.1003 no

.9003. .

\\
\\ .
\,
0‘
..
.
is.\‘l\\\.
i ‘ Li. L
.8 no
111‘ A J .

$009.00.

880 30.003320 .96.... 30.3.
9 2068.0: .5 .6830

.80.

$000.3. .00.

8.00 30.00333 .0000: 30:0.
.... 0080.3

.9003.

\\

 

 

L- -.__..l._.._.L ..___'.__.

‘__.__.J
253

 

..-

ob .-

oh 1

o...‘ ..

ob ..

oh .

 

ob
_
ob _.

0;

o:
.00

. .-
8.0
Pa -

on:

0.0.

 

9m -

0....

0.0.

0.0.
.
0.. u.

o...
.00

.-. . - . . 1 ..
8002.00.
0000 30.83330 .98.. 3000.
+ 8:08.00 .5 {00000
.80.
80003. .00.
0000 3060033.. .0002. 30:0.
5 0000000
.9000.
it. \
.\.,\.
\3
....
.8 no
.\4
8000.000.
800 30.000300 .0000: 3000.
+ 008. mu .0 ..0000:0
.30.
<88... .00.
0200 30.000300 .0008. 3000.
.5 180000 .
.0800.
\\
\x
8 no

 

___F__ _..-

ob

ob .

Pu

ob

 

0.0.
0...
0.0 .
0.0 .

0..

 

0.0 .

ob .

0.4.

Pa

9m

0...

 

0.0.

pm.

I . a .4
00003.00.
0000 30.008300 .98.. 3000.

+ 303.3 .0 {000:0 ..
.35 ..
30003. .00. ..
00.00 30.0003:0 .0002. 30:0.
.5 0.000000 .
.9000. ...
\
\
\
.
.
..
.
.
\
.\
8 -mo 0
t l . . .1}.
3000.00.00. . .

0000 30.000300 .0002. 30:0. +
0.0030000 00.0

.80.
<88! .00. .
00.8 30.00333. .93.. 30:0. 0
.5 ..000000 ..
.9000.
x ,.
....
. \\
w 0 -8 0

t y |
__. 4..—. __.—__..

b
__.—... —._ __L ...

-__ -...L-..._. J-,- ,7 .

 

r .
.... __.. -7

,00

254

 

 

 

.— vl I 4|. . .x
0.0 ..
3000.00.00.
o m 0000 00000300 .0 600000
. 0 .se
...... r 000000000.
00.00 00000300 .0 02000 00
Pm . 80000.
0.0 ..
_ \.
0.0+
0.0 . ...
. ..
0.0.
0.. r
. ...
o . . l ‘ .11.? .
-00 8 no
. _0 _ ‘ 0
_ 00000.00.
0.0 - 00.00 3.82000 .98.. 0000.
+ 0:08.00 .0 00000 00.
0.0 . .088.
o N. 50000.00.
. 000.00 30.000300 .9000 0000.
+ 0:03.00 .0 500000.
ob . .000.
0.0 .. x.
_ .x
0... My ....
0.0 _
.
0.0
0.. .
.
o m . .. . _
.00 8 mo

00 00 _-

ab 7
ob -
0.:
ob y

Pm -

 

o.» ...

0.. T

 

 

0’
O
80

. I. ‘.1

0.0 .

ob .-

oh .

 

 

-1-‘4\:1; .24:
300008.00.
00.00 00.000300 + 0800.63.00
5 .00 000 no.
.0300.
000000- .00.
00.00 30.000300 + 3000.90.00
5 600000 .
.80.
\
.
\.
w
u..
in F. .
.8 mo
.
3000.000.
00.00 30.000300 8000.. 3000.
+ 7008.00 .5 500000
.80.
<00om00. .00.
00.00 30.000300 .0000; 3000.
+ 7008.00 5 00000 no.
.0300.
.\
.\
..
.
.
.
Lo .8

.F- i...

- .0 ‘——~_.‘ g—Jw .__“

L.__. --.__..___._‘.-- ..-H..-_._.-

mo

. v
0. __.. _ .-....0

 

255

 

93

Pm ‘

 

.0
(n

O
\l
“"'—*x— A___~‘___+,_._ ___

<88auhm.

83m 30.03330 .93: :33
+ magma s as 83 no.

6.83

<88$h2

mmbm 30.0333 A902 32::
+ 59.8 5 €823

8.33

<88! .8.

88m 30.8353 803: 33.3
5 889.3

53

(88%: ha

mug 30.83.23 808: 32.3

.5 4! 8: No. ,

6.83

 

 

\
v

 

a \‘ fi ‘
..x ,_,
H ,.
. ,
. .
.
fl
,
.
,.
,,
h 1% L‘ c
o no .8
.3.. A 4 -
u /
__. ...
,,
.,
,,
.,.
J.
_ _ _ ‘
o mo 8 8

 

 

_ 4 { . \ ‘ ‘ \fw ‘ a w a M
‘ ,, _
ob T $88.8. \ ._, A
83 3635 A83.— Saa. \ I _
oh 7- s 6893 ‘ I .4
A33 \ , _
oh I Emgfi. ..V. / ,m
8% 8633. A93: :3! .. fl .
ob v- s «28: .8 \ M #
683 x m _,
3 V \ r, A
x J., _
o... x M
x m
S _P , A
. .. ,. , u
99 . , *
o; 4.
_ 7 . ‘ .. M
O_, 1.31:: h ‘. 11 _ F , _ ‘ ~ ;
-mo ..8 -mo 0 no 8 ac

256

 

 

 

Figure 8C. Absorbance EPR spectra of N86C/I98C double mutant in detergent (Tween
20) and liposome and in the presence of Co-PPIX, arachidonic acid and various COX
inhibitors.

Spectra were normalized to maximum absorbance and were uncorrected for any
contribution by monoradical species.

Images in this dissertation are presented in color.

257

 

 

T

 

 

 

 

 

258

 

 

 

— J\ I, tx \ \l/ 4 it ‘4 \rw d i llllntli I 4 l ‘4 t r x 4 ti 4
\ / /.
éogm‘hwn . \\ . . 0.0 . <§NQ3 a. ,
$8 8853 5 as 8: 8 s , $8 36823 68% :23. to .,
A89 \ y + 3.3.903: 5 4580: no .
\ , pm , is x 4,
<88§..§ \. , _a
8.6m 30.83548 .3 3 3: no .\ / i o.» x 3383 on. x
6.83 , 88 8683.3 A98- gv \ ..
f , o m . 5 4.28: no \ ..,
M . 8303 _\ J.,
/ t x
o m t
,f/ a. \\
x x w ob _r
t / m _ /
c / _ _ /,
m‘ r// 4 o.“ .T \ I
_ l, ./ m __ \
a. ,/ o.» ”t \
at /. 4_ , x 1
\. .,. _
.\. #_ o g _ \\ x
\ /.// . _ .\.\
.... \ P _ \t k . L _ . _ o v t . .1 .x _ II c r 1‘ h p
.8 .8 o no .8 8 .8 .8 .8 o no .8
a i i, J, a t 4 d t. t 1‘ ...t t ._ t it} > , H <<<<<
\\ if. _ a /
338... no. 5., . o o” <8oma3h2 /.
88 3.8325 808.. 83: \ c, _ . _. 83 8§<3 A983 33v s. /
+ 8:908: 5 ...: 3: mo. \ ,., m + 363.5 5 «too: no \ ..
$88.: .9» \ t. 5309. .3»
38 30.8338 A93: :93. .028 30.8323 A0000: 33.3
.5 2.30: no. \ 5 atom: .8
383 \ M 3803
\ , t
7
4
.k
, m
/ _ _
k z
/ _ 9. _5
Eu _ a ~ _ o .I L, \ c c F

L5 fine to no bot mo .8- .Ao -Mo .oj no 3:3

 

 

 

 

 

‘1' 1., ... s
o m imaging
. 3am 30.02535 A902. 30:!
+ 598 5.38: no
Pm 1 :08
o N r $6803.02
. §m§<=¢ .000»: 3033
.5 «$00: no
oh 7 AQooav
\
99., . .
m x,
o; 1. to
ca .
o.»
o; .
o .11 ‘ ‘Lfiv 2 Li
8 Lo -8
d —t t a | w 1.
Po Emma—3.02
88m 00003336 5 680:6
pm - .33
3821.0».
oh §m§<§ 8000: 30:3
3 €893
0.? 6303
out .\
0.?
9w ,.
o.»
S x
o - F a
be Lo -mo

9 i, it t t
.2 /
,2
,
f
0.,
f,
/
/
‘—
o no
~41 —
,/
..,
.,
V.,.
,.
2
,
I.
,
//
/
r.
//
./
/,
/.'
no

 

~ v
- _,-#¥,, A, A WA -.4

 

 

rldth.|iitliiti4v , , INN/Mt -.
<88¢3§ ..2 /
0.0 r 82 8.8.53 683 8.3. a.
+ 20020008 86 .5 «too: mo ... I,
3 T 5.5 2.. .,
o u I (engage 2.x ,
. 88 §<3 682. 333 .2 f
5 .5100: no .. f
3 i 683 \.. ,
f
.\
9m \ \
\\
\
ob ,- \
ob .. \
o.” . ..N
m ...
o; L,
on, t w 1. I -7 ‘LF ‘ 2
.8 Lo -8 o
o m _ <8§.8. 5,5
. 38 30.03355 A98: 7036. H
+ 306.533: .5 680.3 . ,2
o.“ M Aqav ..V ,.
o L 3582.8. 2 ,,
. m 38 §<§ 298.. :23. ..s ,
. 08 0..
ob .7 A9 v \ f,
\ /
o.m _7 \tt» \ /
__ \,
ob ”r \
o.“ "a a,
w to
o.» v ..2
H .\
o; \
w
o _. t r. n _
.mo Lb .8 o

 

 

 

 

mo

259

Po -

ob ‘

 

.= .= .a .c .D .= .c
(a) ‘5 0| 0) V a) (D d
"'fi ‘—‘-.'_"_ 'T '7‘ '_"'T1_" T-“'—‘_T'"'_' *7

.0
to

.°
-‘

29.

. «
<88: .03
8B0 30.00333 6000.. 30:!
+ 0:06.00 5 580.3
A33
6083 .00. .
«28 30.02536 6000.. 30:3 .
6803 ..

0.1.
\\
8 -8

A 1 a
$895.09.
88 00.83.30 608.. :33
+ 88.8
:09
<88! .00.

8g 30.000326 A93.» 30:!
6:33

,_

 

 

 

 

oh 7

ob _

oh 1

ob 7

 

.0
a:

o
‘5

9m.

. .0 .0
m on

O
—A

...I1. 4 ,‘\~‘

308...: .00.

028 30.00333 A900: 30:!
+ €93

:8.

(88.3.3.

800 583‘3 A08». :28.

.9003

53900.00.

gm 30.00033 8000: 30:!
+ 20020008 no.0

:8.

<8oma: an.

830 30.09536 A0000.» 30:3 V

A9003

,”
A ’

n
. ,
. m_‘__ _.,,k-, Wfiwifirivi-fi_-_t0siwi

1

 

/

J
..__V_

3
8.0
8

260

 

 

dfiiul -. q - .— I .I\ .J/ 4. \\1\.. 4 .
0.0 .. <88su§ \ J.
8.8 00003<3u 5 880.3 x ,,
on ...
ob .. a . ..., /
<888. .8. .. .
9i §m§<3654£33mo ., 5..
6.83 \ ,,
°.Q 1 .\. M.,
/
9m 5,
ch Fr .../
0.3 ,
pm.
o; 3 //
o _, . .L \ % _
.8 Lo no o no 3
d . a / a q
o m .- <80838. .
. $8 36833 598.5 :83 .. 5,
+ 80.28 s 3.8: 8 _. ...
0.0 . 6.83
o u T <80§0a
. 88 30.03336 6000.. 30.43
+ 80083: 5 {8030 _
O.“ Anav \
99 ,,,.
o; ...
5 /
ob . ...
_ \ ,
0.»: \ /
//
o.d W- \ //
o _ p _
.8 L8 .8 o 8 3

__..-..__ -- . ._

mo

‘ .

obi

ob 7

9V 1

ob T

 

9m.

.0

A

O
(a)

..N.

O

o.

d

_
_
_
_
. _T
_
v
n
_
i

o _
.mo

.0
a:

q «LIL:

508928.

83 30.03336 8000» 303:
+ 3.3.66.3 .5 41003 no
688.

33981.53 5.

83 3.833 .98. 33. .
+ 30368.0: 5 5803.0
.83

<8wm03. .00.

028 30.003330 A900: 3033
¢ $608.3 .5 32000 .8

8.83

508...: .02

82 30.03336 A900: 303.3
+ 23.35 5 {80.3

:8.

 

-8

261

ob 1

ob ..

 

.0 .o 9 o
0') ‘1 on (D
- ’3' ””“Tgk7 ’ —"-—T'—-" ‘7

.0
or

0.0...
0.”.

9;,

.mo

3.82098. .
8.8 30.00336 .93.. 3036. H.
+ $898 .5 {80.3 .

.33

88 30.00353 .93.. 30.3. x
+ @0098 .5 at no... no N
.983

. a ..( .4, H ‘ \j .

<88; an.
8.8 30.03235 .93.. 303.0. ..
.5 500036
.8...

$8803. .8.

8.8 30.003338 .0000: 303.0. _
5 «£003 8
.98...

mo

. .- - 1 -
ob _ 3383.3.
" 8.8 583.3..98. :83
“ +£33.83... 8554(838
94 50880.8. .
8.8 30.83.55 .93.. 30:5.
cam. émogafisasaaea
. H .80. \.
n .\
9... H.
.
o.a_.
_
our: ..
‘ .
o.» _
9. .
_
o W. .- . ‘ .
- o .8 .8

‘k4.

 

262

Figure 9C. Absorbance EPR spectra of V87C/I98C double mutant in detergent (Tween
20) and liposome and in the presence of Co-PPIX, arachidonic acid and various COX
inhibitors

Spectra were normalized to maximum absorbance and were uncorrected for any
contribution by monoradical species.

263

 

 

 

 

a. . . .. . . 1m I -. J . . . 1
. <mug. .8. ,,
ob 3.0m 0000353 5 «(003 no . ..
.v Aqav ... ..
Pm <3009.9. __.. .. -
Guam 335.6 .93.. 303.6.
0.1.. 53388 _H
. .9000. ._ _.
9m ., . ..
m .
. ..
a.
a
03 .3 .. /
ob .r
. ...
ob .. .... .
0.. .
o _ | _ . . _ .
.mo #0 .8 o no .5 8
d ... . . . ...J. . . .
_. «0300: .8. .. .
0.0 _- 0000 3603...... .98.. :23. _. , ..
u + 0:08.05 5 42003 no . .. A
o... _ ..a. . __
o 0. 8089.8. . ._
. . 3am 30.00358 .0030» 30:0. _ J
5 as. 005 no _. , m
ob .Qooa. ._ .
m x _
9m " ... L.
0... .. .. .
_ .
0.0 . ... / .
. x z .
. . ., .
J . / .
S. .. u
C .. t . _ . . 7
.mo Lo .8 o no 06 mo

 

 

- I . .. - . . . . 1
$0803.00.
0.0 r 800 8.8.5:... .98.. .6307
3.3.63.0: .5 4100.. no
0... . .se
Eugen». .
0.0 - 8.8 3.8353 .98.. 3:0. .
.5 4200: no ..
0.0 .033. \.
\\.
0.... - .
..\
0.. .
ob ..
PM .
0..
o . _ x. U - ..
.8 L5 -mo
...! - . . _
ob ,. gmaamhm.
0.58 30.00353 .9335. +
303...: .3 43.003 no
0.” Y Aqav
0.0 . 30803.8. ..
3.3 30.00358 .98.. 30.8. ._
5 4208 00 ..
o.m.. .9000. .
0.0.
_ \
0.... ..
.
0.0.
o.» _.
o.. .
o .. .. .. ..
.8 Lo -mo

..

 

 

264

 

 

‘ m. . .
c o . 300002.00.
. 0.50 30.8330 .003... 30:0.
+ 50.3 5 as. 003 we
ob . 3.00.
oh 1 50803.00.
00.00 30603.30 .0000: 30:0.
.5 4.2003 .8
pm ., x
_
o...
_
_
o0.
...».
o.
o . +. .
.2. Lo no
A ..- 1 4
m
o o .r $32.50».
. _. 03.00 00003.33 .5 {00030
n :00.
ob ..
<38... .00.
o N. 3.8 30.003<30 .0030: 302.0.
. . .5 600030
A 00:
oh . 9 .
9m
o.) ..
_ .\
9Q _. ...
...».
o. ..
O . h ..
-mo ..8 -mo

.. H

 

 

‘ .. - a _
o 0 6.3.0038.
. T 8.8 8603.21900- .53. .
+ 2003.00.10 00.0 5 32003 No ..
Pm . .30. ..
o w .- $08998. .\
. 00.00 30603.30 .0000: 303.0. ..
.5 4.0003 mo ..
0.0 - .038.
0....
o... _. W...
ob. ..
_ ...
...». ..
0.. m. \
_
o .. . _, _
.8 .8 -8
‘ Ho L .. _
o 0 .. 830.20...
. _.. 8.8 3.8330 .98.. 3:0.

. + 20.6.0.0: 5 600030 .
0.0. .3... .
o w 00003.9. ...

. H 00.00 30.00330 8000.. 30:0. ...

. .5 {0003.0 ..
...? .98...

. ‘\
9m .,

_

o...
.
0.0 w
on”
.
o. ..
O ..
-mo .3 no

. .
.r
y
..
.
.
I
..
.
..
444/
.,
. I
/
o No 3
.
/
.
.
.,.
H
I
I
/r
/
. _ _
o we so

265

 

 

 

. il 5....-.
3002.00.
0.0 T 800 3.83.20 .98.. :30.
+ €0.03 .3 580:0
o... - .3...
<0~om§hfi
Pu . 00.00 30603<308000=303£
.3 300030
0.0 - .060...
Pm .
o... . .
9w
oh
o. .- .
. \
o . . . -
0.. L5 -8
. _.. ... .
. <mߤfam~
ob . 0.50 30.8330 .003... 3030.
+ ”mafia ..3 €003
ob _ .3...
. 00.00 30.00330 .0000: 30:0.
.5 58030
0.0 _- .080...
o: .\
_
...:
_
o0.
_
PP.
.... ..
o. ...

0o , .5 -8

 

 

 

 

 

. .J‘- ‘ .1— d o IIIIIII.||..«!. ..I i ..ll. .|,|.. litlwamlll - I! \ \ .w.
_ 3.83.9. H
0.0 fi 800 860338 .08.... .53. _. ...
+ 50.9.5 .5 30003.0 . .
0.0 . .8... .. .
6803...... .. .
.. 0.0 - 8.8 3.0833 .98.. .63.. . .
_. 5 68030 \ ...
oh 7 .9003. \ .
_. . H
. o0
_ \.
,_ o... .
. ..
. o...
_ ,,
J. _
.. .
/ .. o.
. .. .. W O K x . . k .1
mo .5 mo .mo .8 -mo o no .8
w . a. ‘ J_ .— . \4 .
. o 0 .. 30800.8. . /
. 300 8.83.3 .98.. .63.. .. ..
_ + 23.685 8... s .0880 . ..
.. o... . .33 . ..
_ _ ~ .
.. o 0. 30003.8. .. ..
. . 0300 30603.30 .0000: 30:0. . _.
_ 5 {8030 .. ...
w 0.0 _ .98... ..
H _ .x. ..
. pm .. . . ..
M _\ ..
o... _
_ ..
. ob _. ..
. .
. o.» _.
x o. . ,.
. _ r _ C _ ll .. r .r. . .
mo 8 mo .8 Lo .mo 0 No .5

 

we

I..- - __.Q. .—

 

 

 

mo

266

 

..
. :8.

5.800. .00.
3.00 00003050 5 ...... 00: no
.0003

3.8.0.000.
00.8 00002350 5 {000:0

<88... .00.

3.8 00.00330 .0000: 00:0.
+ 0003.0: .5 680:0

..8.

$080: .8.

8.00 00.00330 .0000: 00:0. ..

+ 0:08.00 5 .....00: no
.0003

.
\‘0
xi

-.

 

x
H _-_____.-

 

8

 

:1 i-.. , . . -41-...
50803.08
0.0 1 0.00 §.3.9§ :30.
+ 20.06.00 .3 ...!000 no
ch. .9003
o N . 30003.00.
. 00.00 00.000350 .903 00:0.
+ 20.08.03 5 680:0.
ob .30.
00..
_
.
ob .. .
.
0.“. ...
o.» .0
o...
.
o n. ‘ ‘ . _ ‘
two ..8 no
. ..
0.0. <mv8a3mbm~
. 8.8 3.8.53 .08.... 3:0.
. +€0x554¢<0008
0.0“, .038.
0.3 3.00.0000.
_ 00.8 00.003330 5000.. 00:0.
0 m. + 600.55.09.30
. H. .38
0.0.
o...
obi.
...» _.,
o.. ..
O. .. .
.8 L5 .8

>—

267

o.
m
..-—... —1— __.... --_.

0.0..
0.0 r
0.0 _.

0.0..

.50800900.

- 00.00 30.003530 .900: 3030.

0 0000.00 .0 02000
.9000.

.0300. .00.

. 0.00 00.83.00 .98.. 0000.

+ 0000.00 .0 {080:0
..8.

<38... .00.

3.8 006003.40 .0000: 00:0.
.5 {000:0

.80.

0.3800. .00.

0...... 8.83.0.0 .0000... 00:0.
.5 4s. 00: mo

.93...

 

. ‘
__.___.—_--- .. .7 - k --- . A-,..--.._.__

 

.. . ,4 .
o o «0.80.0000.
. 3.00 00.000300 .900. 00:0.
+ 20006020 00.0 5 580:0
0.0 . .80.
9V. 308000.00.
00.00 00.000530 .900. 0000.
+ 30005030 00.0 .5 41000 no
°.m .y Aqgv
0.0 _. _
_ ...
o... _.
fl .
o w. ..
_ _
_ ..
PM W ..
.... ..
O H , . ‘ ‘ .
.00 be no

_
..

.

.

.8

268

Figure 10C. Absorbance EPR spectra of V88C/I98C double mutant in detergent (Tween
20) and liposome and in the presence of Co-PPIX, arachidonic acid and various COX
inhibitors

Spectra were normalized to maximum absorbance and were uncorrected for any
contribution by monoradical species.

269

oh -

 

.9
to

D
.—A
’ ’AT——'“T"_'_'~_.'"""_ "' _

.o .o
on 0 '-‘
' ""‘ ”'“fi— __.—“f--_.1

.0
V

0.0 ..
o...
0.0

0.»...

- 1 4- - .
008000.00.
8.00 00000~<=8 ... ...... 00: no
.8...
008000.00.
00.00 00.00.5200 .0000. 00:0.
.0 4! 00: me
.0800.
.\.x.
. \ \\
.3 .8
i . . :.
0880.00.
8.8 00.000550 .0000- 0030.
+ 8:03.00 5 .7: 00: mo
.30.
008000.00.
«mam 00.000530 .0000- 0030.
5 at 00: mo
.0300.
\\.
\.
\
.\
...
\\
-8 .mo

 

...m ..

ob

 

ob.

ob .

oh

ob -

 

.4 -

0880.00.

8.00 00.000330 .900- 0030.
+ 3.8.63.0: 5 .0: 000 8

.8...

08080.00. \.

0000 8.88.08 .080. .53. ._
5 ...!00... mo

830:. x.
\
.\
W.
.
..
x.
\
.00 -mo
ot.7. .
08800.00.
08m 00.000530 .0000. 0030.
.0 0:00: we + 308...:
.30.

000800.00.
8.00 00.00333 .0000- 0030.
.0 .0200: no .
.0300.
._
.k.
\.
_, ._
5.. -mo

270

 

 

 

 

.0999
Aroma»

0

ob

271

...m r

9m .

....»

 

I» — ‘1‘., 2.24. - .. l .4 1. . ‘ . - d ...i 4- ,1» .,
, 0 .
000008.00. _. . 000000800.
0000 3.8300 .98. 33». . . . 0.0 r 00.8 3.833» .98. 33».
+m80.~m»3?000~0 . m +>q000.000.000.054t000~o
.80. a . ob : ...00.
000000000. _, 000000000.
800 8.833» .080. .03». ... w 0.0 800 3.853» .080. 83».
5 0:8: .8 0 r 5 028: 8
3.8:. \ ... 0.0 6.8:.
\\\ _
\x\ 9m m
0 0.0 H
.. . ob _
\x. A., O.” k
\ 0.. U.
.-\\-\ m
. ‘ 0 0 O ‘- up
-8 .mo -00
t .0 0 0 -
08000.00. . . . 0.0 Y
00000 0000353 .0 60000.0 0. 0., W 8.8 00.000330 .0000: 0000.
.80. .. .0 o m T .0 6000306800.
00000000. ..., .. 00000.. .00.
g §<26 .0060: 30:6. ... fl . GUN . 8% $3516 .003: 3026.0
WNW...» ... . fl . 0:80.20: 5 .0803» .80..
. o.» m
o...»
0.0 n
_
ob _
o.»
.... _
m
_ r ‘ C _
so -8 0o

 

 

, v
--_ --.. _ ...-___.

 

 

 

 

. . . . . .. . .
. _
.. carom: 2: W W “was?
ob” 00.00 00.033.00.900: 00:0. . ob. 00.00% 00.00000 0000.
m . 8:03.00 .0 8000.3 ., ., m 0 308...: .0 “000030
c...“ ..a. , .. ...». .3... .
0880 0». . _ . . ._
0.3 _ o.» - 000003.00. .
_ wehwhfisza» .08.... 33.. .. , . .. 000.. 3.8.53» .93.. 33». ..
Pm. .0300. \..... ., oh. .5 5000.3 .
_ .. . . .9000. .,\
oh. . Pm .
o... 0.0
_ . _
0.0 ... 0.0.
...». ...»
_ _
_ _
o . . _ o . . .
.00 L8 mo 0 no 00 mo 0.. .3 .mo
4 J 4 d a, _ .
.
0880.00. _ 0030.80».
0.0 7 00.8 3.833» .08.... \A 0.0. 80.. 3.8308 .93.. 33».
+ 8.8.00.0 . _ +£80.03... 0905:0020
0.0 y .80. 0.0. .60.
00080 00. M 0803.0...
0.0 I 00.00 3.80003 .08.... 0.0. 00.8 3.8330 .93.. 330. .
.0 500030 5 3000.3 .
o... 1 .0800. A 0.0.. .9000.
4 0.0..
ob
...... ..
_
L o.» ,
.... ..
b h P O u. . .
o no 00 mo .wo .00 -mo

 

 

no

272

...};

ob r

Pa 1

T

9
V

 

.0
____cn

P
01

.o .o
0) b ____

ob
9m

9V.

 

20890.0». .

0000 00000330 .0 008000
.2. .

000000000. .
00.00 00000330 .0 .00 000 no
.0300.

nag-hm. ..
00.00 00.805000 .900- 00.3.
+ 6:08.00 .0 6800.0 .
.60. ..

000000.00. ..
0000 8.82030 .98.. .530. ...
+ 353.0: .0 as. 000 no
.0300.

 

 

 

. EE-‘ [TIL 4.0
000000.00.
0.0 I 0000 00.8303 .080. 023.
+ 30668.00 .0 6800.0
ob , .30.
0 u , 000000.00.
. 00.00 00.80300 .0000- 0030.
+ 3.6.08.0: .0 41000 no _
ob 0 .0800. x.
0.0 W \.
ob ..
W
0.0 W.
W
...m ..
W.
.0... W
W
C 7 L ‘ \ \Li 1 ,..
.8 ..8 .mo
., .. .
000000.00.
0.0 .0 0000 8.83030 .98. 0030.
+ 363.00 ..0 6000.3
0.0 :8.
. 000000000.
0.0 W 0000 8.83.08 .98.. 00.8.
W +§0x554£000mo \
92 8.8... \
W .
an
_
0.:
W
9.:
00W
0.. W
O W ‘ I . .
do .00 -Mo

\

.

0
x
.

.

..

x

x
.

W

H

273

 

 

 

 

 

 

 

.— 3 ‘\ 1 \1 - _ ‘ ‘1? \ _
gash», \
Po 1 83m 5833 88% 83 M
+ magma 3 8835 ‘N fl
\ ,_
c 3888.8. ‘. ,
.4 I 8.8 8.853 808.. 333 ..\ ,.
+ m89~m s at 8: 8 \. ,.
9m 8.83 \ x,
3 _f . d.
_ \\
ob
our
_
o.» . /
n;
O M. L. 1‘ \\ NHL F‘- \ fl ‘ h L
-8 .3 .8 o no 8
d 7 - .g ‘ ; \4/
r c8858. ..‘ ,,..
ob 88m 30.8353 68% :oBm. “ ..,
s 8886 x ,...
9Q r :03 “. .fl
o u . 358:8. \ fl
. 82 33.28 68“., 33.2 “x M
5 4‘8: 8. \x p
oh r 632: \x w
9m ‘ x .
ob.
9w
oh.
8
O H H H
g #0 -mo 0 no 3

 

 

6.0 y
obj

PM .

 

> it I «II _
c2389.
3 3060353 280% gag
+ 28368.0 and 5 6836 \
=8. \
..
38889.3 \

8% 8.853 882 33 \
+ 28:53.0 won 5 at no: no N...

6.33 \
\w
.
\
Ax.
\ H r 1
.8 -mo

 

 

274