>55.

 

{LE3

6...! .1:

7
3+
‘3 4

1.

rant 4%“ 4
c2. ‘ 1 3!...

. 2
Erase it... 2 sinus;
3.99... E... a 9.
{napapnhcgt _
39.5,... ;
s. s. s... . :
$37!. 3.
it i.’ u

.
, s
I; :3 a

nvmomfluum hawk.

win, mmnfiw

z ., n
x 1 tn. 5”.
.fifismi
. a.
6.. mun.“
m... the.
I

5
h

This is to certify that the

dissertation entitled

PULSED ELECTRON PARAMAGNETIC RESONANCE STUDIES OF THE
HISTIDINE AND WATER LIGATION OF TYPE 1 AND TYPE 2 COPPER
SITES IN FET3 PROTEIN, A MULTICOPPER OXIDASE FROM YEAST

presented by

Constantino Ponsaran Aznar, 111

~ has been accepted towards fulfillment
of the requirements for

Phi)- degree in _flb_em_i_51£y_

 

Major professor

Date 5? /ly/OO

MS U is an Affirmative Action/Equal Opportunity Institution 0— 12771

 

 

 

' LIBRARY

Michigan State

University

 

 

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

 

002
ONT-$106.2 2o 3

 

r6963 {1:3

 

 

 

 

 

 

 

 

 

 

 

 

6/01 chlRC/DateDue.p65-p.15

 

PULSED ELECTRON PARAMAGNETIC RESONANCE STUDIES OF THE
HISTIDINE AND WATER LIGATION OF TYPE 1 AND TYPE 2 COPPER SITES IN
FET3 PROTEIN, A MULTICOPPER OXIDASE FROM YEAST

By

Constantino Ponsaran Aznar, III

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

2000

ABSTRACT

PULSED ELECTRON PARAMAGNETIC RESONANCE STUDIES OF THE
HISTIDINE AND WATER LIGATION OF TYPE 1 AND TYPE 2 COPPER SITES IN
FET3 PROTEIN, A MULTICOPPER OXIDASE FROM YEAST

By
Constantino Ponsaran Aznar, III

Electron spin echo envelope modulation (ESEEM) spectroscopy has been used to
study two variants of Fet3p generated by site-directed mutagenesis. A homolog of
ceruloplasmin, Fet3p is a multicopper oxidase from the yeast Saccharomyces cerevisiae.
Like human ceruloplasmin (th), Fet3p catalyzes the reduction of molecular oxygen to
water with ferrous ion as the reducing agent. This ferroxidase activity is essential to the
physiologic role the two enzymes play in organismal iron homeostasis. The two mutant
Fet3p studied were proteins in which either the type 1 or type 2 Cu(II)-binding sites had
been depleted so that EPR/ESEEM studies could be accomplished without the
complication of spectral overlap. l“N ESEEM was used to determine possible histidine
ligation to type 1 and type 2 Cu(II) sites. Computer simulations of three-pulse spectra
taken at magnetic field positions corresponding to g" and g i suggest that the type 2 Cu(II)
shows one equatorially-bound histidine imidazole ligand contributing to the echo-
envelope modulation. For the type 1 Cum), two magnetically distinct histidine imidazole
ligands were found. Water ligation to these copper sites was also studied by the
deuterium-exchange technique. The results suggest that both the type 1 and type 2 sites

are readily accessible to water. However, water is an inner-sphere ligand to only the type

2 copper site with one axial and one equatorial D20. The 2H modulation function of the
type 1 copper site is due entirely to ambient deuterons with the nearest metal-to-deuteron
distance being 3.7A. Comparison of the data obtained in this work to the x-ray crystal
structures of the multicopper enzyme homologs ascorbate oxidase, laccase and human

ceruloplasmin was made.

To Mama and Papa

iv

ACKNOWLEDGEMENTS

MARAMING SALAMAT! It is a Filipino way of expressing their gratitude. There
are several people I need to thank for their contribution to the completion of this

dissertation.

Thanks to Professor John McCracken, my advisor, for patiently guiding and
supporting me throughout my five- (and more ?)year stay at MSU. He is not only a
very good researcher, but an excellent lecturer as well. His lecture and talk during
group and one-on-one meetings were very helpful and enlightening. Not to forget to

mention were his amusing stories, mostly about his experiences.

Thanks to our collaborators, Richard Hassett; Daniel Yuan; and specially Prof. Daniel
J. Kosman. Without the Fet3sample from Kosman's Lab, this dissertation would not

be possible.

Thanks to the guys in the electronic shop for gladly providing me the electronic

devices that I needed to finish up the pulse-generating boxes.

Thanks also to Vada O'Donell.

Thanks to the past and present group members of the McCracken Lab, specially Dr.
V. Bouchev, Dr. V. Singh, Dr. E. Saari, Dr. R. Muthukumaran, and Dr. S. Smith. I
thank Pearl (R. Muthukumaran) for his help in the collection of CW—EPR and

ENDOR data.

Thanks to my Filipino and non-Filipino friends, students and non-students, here in
East Lansing, specially to Aileen, Cecilia, Vivienne, Charisse (for the sotanghon and
bilo-bilo, ©), Beni, Chris, Ian and Oggie. Activities such as tennis, volleyball, get
together for the purposes of eating and playing card games or charades, out of town
and state trips, the ever exciting practices and performances for Michigan and

Global Festival, etc. made my stay here less lonely, fun and exciting.

Thanks to my family - mama and papa, brothers and sisters, 1010 and lola, aunts and

uncles, for their love and prayers. I want to thank Auntie Eva and her family, they are

the nearest family I have here in the US.

And To God, a very sincere MARAMING MARAMING SALAMAT!

vi

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES
LIST OF ABBREVATIONS
INTRODUCTION ........................................................................... 1-2
Chapter 1
Introduction to Continuous Wave and Pulse-Electron
Paramagnetic Resonance ................................................................ 3-54
1.1 Interaction of Free Electron with a Magnetic Field ........................ 4-6
1.2 A Simple CW-Experiment .................................................... 6-8
1.3 Spin-Hamiltonian ................................................................ 8-9
1.3.1 Electron Zeeman Interaction .................................... 9-12
1.3.2 Nuclear Zeeman Interaction .................................... 12-13
1.3.3 Hyperfine Interaction ............................................ 13-17
1.3.4 Nuclear Quadrupole Interaction (NQI) ........................ 18-19
1.4 ESEEM ......................................................................... 19-20
1.4.1 Effects of Microwave Pulses ................................... 19-51
1.4.2 Formation of Two-Pulse Echo (The Hanh's Echo) .......... 20-23
1.4.3 Formation of Three-Pulse Echo
Echo (Stimulated Echo) ......................................... 23-27
1.4.4 Nuclear Modulation ............................................. 27-30
1.4.5 Density Fonnalism of ESEEM ................................. 37-48
1.4.5.1 Density Operator ....................................... 37—39
1.4.5.2 Application of Density Matrix to ESEEM .......... 39-45
1.4.5.3 Modulation Formulae for S=1/2, I=1/2
System ................................................... 45-48
1.5 Two-Pulse ESEEM vs. Three-Pulse ESEEM ............................. 49-51
List of Refences .................................................................... 52-54
Chapter 2
Introduction to Fet3 Protein, A Multicopper Oxidase From Yeast ............. 55-96
2.1 Coordination of Cu(I) and Cu(II) ........................................... 56-58
2.2 Classification of Copper-Binding Sites .................................... 58-65
2.3 Redox Behavior of Copper Complexes .................................... 65-69
2.4 Mutlicopper Oxidases ......................................................... 69-71
2.5 Fet3 as a Multicopper Oxidase .............................................. 71-72
2.6 Trinuclear Cluster ............................................................. 72-77
2.7 Mechanism of 02 Reduction ................................................. 77-79

vii

2.8 Regulation of Iron Transport in Yeast ...................................... 79-81

2.9 High-Affinity Transport of Iron in Yeast .................................. 81-83
2.10 The Iron-Copper Connection ............................................... 83-85
2.11 Ferroxidase Acivity of Ceruplasmin ....................................... 85-87
2.12 Fet3 vs. th .................................................................. 87-88
List of References .................................................................. 89-96
Chapter 3
Instrumentation .......................................................................... 97-138
3. 1. CW-EPR ....................................................................... 98-99
3.2 Pulse-EPR Machine at MSU ............................................. 100-100
3.2.1 Microwave Bridge .............................................. 100-100
3.2.1.1 Transmitter ............................................ 100-103
3.2.1.2 Receiver ............................................... 103-105
3.2.2 Cavity and Cryogenic .......................................... 105-105
3.2.3 Magnet ............................................................ 105-106
3.2.4 Computer Interface, Experiment Timing
and Data Collection ............................................ 106-106
3.2.5 Modification of the Pulse Logic Module ..................... 106-109
Appendices ...................................................................... 1 10-138
Appendix 3A ..................................................................... 110-129
Appendix 3B ..................................................................... 130-136
List of References ............................................................... 137-138
Chapter 4
“N ESEEM: Histidine Ligation to Type-1 and Type-2 Copper Centers in
Fet3p ...................................................................................... 139-184
4.1Experimental ................................................................. 141-142
4.2 ”Nmodulation and Computer Simulation .............................. 142-148
4.3 Comparison of Nuclear Transition Spectra ............................. 148-152
4.4 Results and Discussion .................................................... 152-180
4.4.1 CW-EPR ......................................................... 152-155
4.4.2 l“N ESEEM ...................................................... 155-180
4.4.2.1 TlD Fet3p ............................................. 155-170
4.4.2.2 T2D Fet3p ............................................. 170-179
4.4.2.3 Comparison of MN Hyperfine Interactions
at the Type 1 and Type 2 Cu(H) in Fet3p ................. 179-180
4.5 Summary .................................................................... 180-180
List of References ............................................................... 181-184
Chapter 5
2H ESEEM: Water Ligation to and Accessibility of the Type-1
and Type-2 Copper Sites in Fet3p ................................................... 185-224

viii

5.1 Experimental Procedures ................................................. 187-187

5.2 2H Interaction and Computer Simulation ............................... 187-189
5.3 Results ...................................................................... 190-194
5.4 Discussion ................................................................... 195-217
5.4.1 Theoretical 2H-ESEEM Traces for Ambient, Axial and
Equatorial Water ....................................................... 195-205
5.4.2 Frequency Domain Spectra of Ambient, Axial and
Equatorial D20 ......................................................... 206-206
5.4.3 Analysis of T1D Fet3p Data .................................. 206-213
5.4.4 Analysis of T2D Fet3p ........................................ 213-215
5.5 Effect of Distal N-D on the Modulation Function ..................... 215-217
5.6 Summary and Conclusion ................................................. 217-222
List of References ............................................................... 223-224

ix

LIST OF TABLES

Table 2.1. Coordination and stereochemistries of Cu+1 and Cu+2 ........................ 57
Table 2.2. Distribution of Copper in Some Multicopper Oxidases ..................... 70

Table 4.1. Ligand hyperfine parameters for the stimulated echo ESEEM simulations
aimed at the determination of number of histidine ligands covalently bonded to the
type 2 center of Fet3 (Fig. 3). (1,6, and y are the Euler angles that relate the principal
axis system of the nuclear-quadrupole interaction to the Cu(H) g-tensor axes. ON and
(1),, accomplished this transformation for the hyperfine tensor .......................... 166

Table 4.2 Ligand hyperfine parameters for stimulated ESEEM simulations for the
histidine ligands of the type 1 site in Fet3. The simulations are shown in Figure
4.16 .............................................................................................. 176

Table 5.1. Tabulated ratios of the amplitudes of the peaks at the deuterium Larmor
frequency (VD) with the corresponding sum combinations peaks at twice the Larmor
frequency. Peak amplitudes were measured directly from ESEEM frequency. The
distribution of ambient water was modelled assuming the nearest non-coordinating
water to be 4.0A from the Cu center and ni = (41t/3)(pN)(13,,,-r3,) ....................... 212

Table 5.2. Experimental and stimulated peak ratios of the amplitude of the deuterium
fundamental frequency to its sum-combination frequency for the type 1 center of
Fet3 ............................................................................................. 215

LIST OF FIGURES

Figure 1.1. Dependence of the energy of interaction of a classical magnetic dipole
and external magnetic field on 0. 0 is the angle between the axis of the magnetic
dipole and the direction of the applied field. ................................................ 4

Figure 1.2. A typical CW-EPR experiment. A variable external magnetic field splits
the degeneracy of or and B electron spin states, while a fixed microwave radiation
induces transition between the two states to trace an absorption or first derivative
EPR spectrum. ................................................................................... 7

Figure 1.3. Schematic diagram for the vector addition of electron spin angular
momentum, S and orbital angular momentum, L to yield an effective angular
momentum J. ................................................................................. 10

Figure 1.4. Dipolar interaction between the electron magnetic moment, It, and
nearby nuclear magnetic moment, M. The pictorial description assumes that u, and
LIN are aligned along the direction of the magnetic field. The vector r represents the
separation between the two dipoles and 0 is the angle between r and the direction of
the magnetic field. The dipole-dipole energy can be approximated to be

Ed”, =[(1— 3cos2 6)/r3];t,v2u¢z = Bump“. Depending on the value of 0 , the Blocal may
enhance or oppose the applied external magnetic field. ................................. 15

Figure 1.5. Isotropic hyperfine interaction. Energy diagram showing splitting of
energy states for the case on S=l/2, I=1/2 system. The splitting due to nuclear
Zeeman interaction is not shown. ........................................................... 17

Figure 1.6. Nuclei with quadruple moment, (a) oblate and (b) prolate. ............... 18

Figure 1.7. (a) An inhomogeneously broadened EPR spectrum comprising of
several homogenous spin packets with linewidth AB,. (b) "Burning of a hole" of
some width in the a inhomogeneously broadened EPR spectrum as a result of
applying a microwave pulse of some duration. ........................................... 22

Figure 1.8. Vector model description of the formation of two-pulse electron spin
echo. A). Equilibrium condition, the paramagnetic system has a characteristic bulk
magnetization prior to the application of the first pulse. B) A 90° pulse is applied so
that a transverse bulk magnetization is created in the Y-axis of the rotating frame. C).
Decay of the magnetization to individual spin packets' magnetization during waiting
period, I. D) Reflection of the dephase pattern of magnetization about the X2 plane
upon the application of a 180° pulse. E) Formation of an 2-pulse echo when the spin
packets refocused at Zr after the first pulse. .......................................... 25-26

Figure 1.9. Classical description of the formation of three-pulse ESE.

xi

A) Equilibrium condition. B) Application of the first 90° pulse creating a transverse
magnetization along Y. C) The transverse magnetization M' is allowed to dephase
during time t. D) Application of the second 90° pulse, this time the dephasing
magnetization pattern is on the XZ plane. E). Formation of precessional
cones/storage of magnetization along the Z axis during time T. F) Cone formation
right after the third 90° pulse is applied. G) A stimulated echo is formed. ...... 28-29

Figure 1.10. Classical explanation of the origin echo modulation. An isotropic g-
factor is assumed implying that the S2 is aligned along the direction of the applied
field. (A) An electron in motion produces a local magnetic field, B“, at the nearby
magnetic nuclei. B1c adds to the external magnetic field B0 to yield an effective
magnetic field, BM. The precession of the nuclei I about Blcff produces a time
varying local field AB”(t) at the electron and consequently modulates the electron
Larmor frequency. (B) Application of the resonance pulse reverses S2 to -S2 resulting
in a new effective magnetic field, B2,,fo about which the nucleus precesses. If the
duration of the pulse is shorter than the nuclear Larmor frequency, not all the nucleus
will able to precess about BM. Thus the electron Larmor frequency is modulated by
AB,,(t) and AB2,(t) at nuclear frequencies proportional to Bleff and B2,“, respectively.
Interference of these nuclear frequencies produces beats in the electron prescession
signal at the nuclear frequencies. The modulation of S2 can be detected in a form of
an echo if a second resonant pulse is applied. ............................................. 32

Figure 1.11. Combination of quantum mechanics and vector model of echo
formation to explain the origin of echo modulation. A) Four-energy level diagram
showing branching of transitions. B) Vector model showing the interference of spin
packets as result of branching of transition in a two-pulse echo sequence. ....... 34-35

Figure 1.12. Nutation and free-precession periods in two- and three-pulse ESEEM.
................................................................................................... 40

Figure 1.13. A) Set of energy levels that interact with a resonant microwave pulse.
For spin echo modulation to be observed, it is essential that branching occurs. That
is, there must be transitions from on one given state in or electron spin manifold to
more than one final states in the B manifold. B) Distribution of values of B0 in an
inhomogenously broadened line. ............................................................ 42

Figure 2.1. Splitting of a d-orbitals in an octahedral, tetragonal or square-planar
ligand field. .................................................................................... 59

Figure 2.2. Theoretical CW-EPR spectra of Cu(H)-aquo complex. A) Absorption
spectum, B) first derivative spectrum of A , and C) first derivative spectra showing
hyperfine interaction in the g" region ........................................................ 61

Figure 2.3. Absorption Spectroscopy. d-d and ligand to metal charge transfer
(LMCT) transitions observed for a d9 copper. ............................................. 63

xii

Figure 2.4. Comparison of spatial arrangement of the different copper sites between
ascorbate oxidase(AO) and human ceruloplasmin (th). ............................... 76

Figure 2.5. A proposed pathway of transfer of electron from type-1 copper center to
type 2-type3 trinuclear center in multicopper oxidases based on the X—crystal
structure of ascorbate oxidase from zucchini. The X-ray structure showed the S(Cys)
ligand is flanked by two histidines ligands from the two copper ions of the type-3
copper binunclear site. Theoretical calculations predicted the large overlap between
the Cu dx2-y2 and S(Cys) p1t orbitals. These two findings provide an efficient
pathway for the transfer of electron from the type 1 to the trinuclear copper site. ...78

Figure 2.6. Model for the regulation of metal transport in yeast (adopted from
Radisky et al., J. Biol. Chem, 1999, fl, 4481-4484). High affinity metal
transporters (TH) are selective for the target metal and are transcribed according to
metal need by transcriptional activators (TA) which release DNA upon metal binding.
Low affinity transporters (TL) are less responsive to metal need and are less metal
selective. Although Fet4 was identified as low-affinity transporters, it can transport
other metals. ORF, open reading frame. ................................................... 80

Figure 2.7. Hi gh-affinity uptake of iron across the cell of yeast. The essential
components of the transport system include Fet3, integral membrane exhibiting a
ferroxidase catalytic activity, Ftrl the iron transporter and Frel and Fre2, which
generate the substrate, Fe(II) for Fet3. The ferroxidase activity of Fet3 is dependent
on the three copper centers (one type 1, one type 2 and one type 3) that are localized
outside the plasma membrane protein. ...................................................... 82

Figure 2.8. Scheme for the delivery and loading of copper to apo Fet3. Two
pathways are proposed. The primary route is Ctrlp-Atxlp-Ccc2p-Fet3p. The
secondary route involves Ctrlp—Ccc2-Fet3p. ............................................. 84

Figure 2.9. Iron mobilization in mammalian plasma. Acting as ferroxidase,
ceruloplasmin converts Fe(H), released from the cell via an unknown transporter, to
Fe(III). Two moles of Fe(III) subsequently binds with transferrin (Ti) and
transported to other cells. ..................................................................... 86
Figure 3.1. Diagram of the Bruker EPR spectrometer. ................................. 99

Figure 3.2. Design of the pulse-EPR machine at Michigan State University (MSU).
................................................................................................... 100

Figure 3.3. Transmitter section of the microwave bridge of the pulse-EPR machine.
................................................................................................... 102

Figure 3.4. Receiver section of the microwave bridge of the pulse-EPR instrument.
................................................................................................... 104

xiii

Figure 3.5. A schematic diagram of the pulse- EPR instrument built at Michigan

State University. Schemes in Figure 3.3 and 3.4 are incorporated. ............. 107-108
Figure A3.Al. Timing diagram for a 2-pulse ESEEM measurement. ............... 111
Figure A3.A2. Module Interconnections. ............................................... 112

Figure A3.A3a. Pulse Generating Subunit (A input, page 116 and B input, page
117)

The IC's used are numbered. They are l (74F00, Quadruple 2-input positive-
NAND gate); 2 (74FO4, Hex-Inverter); 3, 3*, 5, 5*, 6 and 6* (74L8123, Dual
retriggerable monostable multivibrator); 4 (PPG 33F-10, 3-bit programmable pulse
generator); 4A (PPG33F-50, 3-Bit programmable pulse generator); and 7 (74F02,
Quadruple 2-input positive-NOR gate).

IC #4 (4A*) and #4A (4A*) create the narrow and wide pulses, respectively.
The width of the pulses coming from 4-2 (4*-2) and 4A-2 (4A*-2) is determined by
the logic input in pins 8, 9, 10 of IC#4 and #4A. Logic states of these pins are
depended on the states of the binary switch (BIN-SW). PPG's 4 (4*) and 4A (4A*)
are enabled by the double-pole-double-throw (DPDT) on/off/on toggle switch. A
logic 0 enables 4, while a logic 1 enables 4A. The width of the TWT and 4) pulses are
depended on the resistance of the rotary switches. ............................... 115-117

Figure A3.A3b. Circuit design of the pulse generating subunit for C (119) and D
(120) inputs. The numbers and identification of IC's used are the same as in Figure
A3.A3a. ................................................................................... 118-120

Figure A3.A3c. Circuit design of the pulse generating subunit for E (page 222) and
F ( page 223) inputs. IC numbers and identification are the same as in Figure
A3.A3a. .................................................................................. 121-123

Figure A3.A4. Pulse-combiner unit. The IC's used are 8(74F30, 8-input positive-
NAND gate); 9 (74F00, Quadruple 2 -input positive-NAND gate); 10 and 12 (74140,
Dual 4-input positive-N AND 50-ohm line driver); and 11 (7420, Dual 4-input
positive-NAND gate). .................................................................. 124-125

Figure A3.A5. Receiver Control Subunit. IC's used are 13 and 14 (7493, 4-bit
binary counter); 15 and 23 (74LS123, Dual retriggerable monostable multivibrator);
l6 and 17 (7404, Hex-Inverter); 18, 19, and 22 (7400, Quadruple 2-input positive-
NAND gate); 20 (7420, Dual 4-input positive-N AND gate); and 21 (74140, Dual 4-
input positive-NAND gate). .......................................................... 126-127

Figure A3.A6. Phase Control Subunit. The IC's are 24,25,26,27, 28, 29, 30, and 31
(74LS373, In-phase module Octal-D type latches); 32 (74128, 5052 NOR line driver);
33 (74LSS4, 4-wide AND-OR-Invert gate); 34 (7454, 4-wide AND-OR-Invert gate);
35 (74L8123, Dual retriggerable monostable multivibrator); 36 (74138, 3-Line to 8-
Line decoder); 37 (7494, 4-bit shift register), 38 (7400, Quadruple 2-input positive-

xiv

NAND gate); and 39 (7474, Dual-D-Type positive-edge-triggered flip-flops with
preset and clear). ...................................................................... 128-129

Figure A3.Bl. Davies-ENDOR pulse sequence. ....................................... 130

Figure A3.B2. Transfer of spin populations, represented by small circles, in
Davies-ENDOR measurement for a S=1/2 and I=l/2 spin system during the
preparation, mixing and detection periods. ............................................... 131

Figure A3.B3. Davies-ENDOR spectra of Cu-dien-imid recorded at 3200 G using
the inversion and detection microwave pulse widths of (a) 220, 116 and 220 ns and
(b) 40, 20 and 40 ns. Other experimental conditions: v = 9.241 GHz, 1 = 900 ns and
tIf = 7000 ns for (a) and t = 450 ns and trf = 4400 ns for (b). ........................... 134

Figure A3.B4. (a) Mims ENDOR spectrum of Cu-dien-imid complex at 3200
Gauss, obtained using preparation pulse of 20 ns, t = 250 ns and RF pulse of 36800 ns
. (b) Davies ENDOR collected at similar conditions as in Figure A3.B3b. ......... 135

Figure A3.B5. Computer simulation (solid lines) of Davies ENDOR data shown in
figure A3.B3b. The parameters used for the simulation are A,, = Ayy = A,, = 34.5
MHz, equ = 3.3 MHz, 1] = 0.13 and NQI Euler angles with respect to B0 of (0,0,0).
................................................................................................... 136

Figure 4.1. Reference coordinate system for ESEEM calculations. The rotation angle
0 and the azimuthal angle 4) describe the direction of the static magnetic field with
respect to the g PAS. The nuclear position of the remote nitrogen of imidazole is
defined by 0N and 6N. ........................................................................ 143

Figure 4.2. Plot of constant g factors (rhombic symmetry case, gl < g2 < g,) in a unit
sphere showing correspondence to a magnetic field position in the CW-EPR
spectrum. ...................................................................................... 144

Figure 4.3. Discrete orientations defined by 0 as a function of external magnetic field
for copper complex with axial g-tensor. The EPR spectrum is generated with g, = gy=
2.05, g, = 2.25 , A, = Ay = 35 MHz and A, = 586 MHz Hamiltonian parameters.

................................................................................................... 146

Figure 4.4. Comparison of different nuclear transtition spectra. A) ENDOR
spectrum calculated by a Matlab program Spint1.mat, B) and ESEEM spectrum
calculated by orientation averaging using the Fortran program SESPNl, C) ESEEM
spectrum at a discrete orientation, 0N = 90°, D) ESEEM spectrum at a discrete
orientation, 0N = 0°, and E) ESEEM spectrum at a discrete orientation, 0N = 45.° C-D
spectra are calculated using the Fortran program ANGSESPNI. All spectra are
generated using the same magnetic parameter for an assumed S = 1/2, I = 1 system: gn
= 2.26, A,so = 4.0, re, = 2.6A, B, = 3200 Gauss, v = 9.2 GHz, 1 = 200ns an T = 40ns.

XV

The g principal values used to calculate spectra C, D and E are g, = 2.25, g), = 2.25
and g, = 2.05. ........................................................................... 150-151

Figure 4.5. Continuous-wave EPR spectra of (A)wild-type Fet3, (B) type 2 depleted
(T2D) Fet3 and, (C) type 1 depleted (TlD) Fet3. For (A), measurement conditions
were: microwave frequency, 9.478GHz; microwave power, 2mW; modulation
frequency, 100kHz; modulation amplitude, 10G; temperature, 16K. For (B),
measurement conditions were: microwave frequency, 9.481GHz; microwave power,
0.998mW; modulation frequency, 100kHz; modulation amplitude, 5.054G;
temperature, 16K. The simulated spectrum (dotted line) of (B) was obtained using g,
= gy = 2.050, g,= 2.198,A, = Ay = 25 MHz and A,= 271 MHz. For (C), measurement
conditions were: microwave frequency, 9.481GHz; microwave power, 0.998mW;
modulation frequency, 100kHz; modulation amplitude, 5.054G; temperature, 16K.
The simulated spectrum (dotted line) of (C) was obtained using g,= gy= 2.054, g ,=
2.250, A,= Ay= 35MHz and A,= 586MHz. ....................................... 153-154

Figure 4.6. Representative experimental MN-ESEEM spectra for the type-2 copper
(T 1D Fet3). Spectrum (A) was collected at B0: 2690 Gauss and for t = 260ns.
Spectrum (B) is the cosine Fourier transform of (A). Other measurement condition
are: microwave frequency, 8.78 GHz; microwave power (43 dB); sample
temperature, 4.2 K. ........................................................................... 156

Figure 4.7. Representative experimental 14N-ESEEM spectra for type-2 copper (TlD
Fet3). Spectrum (A) was collected at B0: 3050 Gauss and for 1: = 230ns. Spectrum
(B) is the cosine Fourier transform of (A). Other measurement condition are:
microwave frequency, 8.78 GHz; microwave power (43 dB); sample temperature,
4.2 K. .......................................................................................... 157

Figure 4.8. (A) Energy diagram describing 'exact cancellation condition' for I = l.
(B) Typical ESEEM spectrum of 1“N following exact cancellation condition.
............................................................................................. 159-160

Figure 4.9. Determination of number of histidyl imidazole ligands for the type-2
center of Fet3. The simulation (dotted lines) models and parameters are listed in
Table 4.1. The stimulated-echo ESEEM (solid lines) and simulations (dotted lines)
were collected at g = 2.33 using a ‘r-value of 260 ns. Other common experimental
conditions common to the ESEEM measurements are given in Figure 4.6. The
simulation given for (a) considers only one coupled nitrogen, those shown for
simulations of (b)-(e) are for 2 coupled nitrogens with Hamiltonian parameters being
given in Table 4.1. . ......................................................................... 162

Figure 4.10. Determination of number of histidyl imidazole ligands for the type-2
center of Fet3. The simulation (dotted lines) models and parameters are listed in
Table 4.1. The stimulated-echo ESEEM (solid lines) and simulations (dotted lines)
were collected at g = 2.06 using a “It-value of 230 ns. Other common experimental

xvi

conditions common to the ESEEM measurements are given in Table 4.1.
................................................................................................... 163

Figure 4.11. Time-domain computer simulations (dashed lines) of the normalized
l“N ESEEM data collected for TlD Fet3 protein at g = 2.33. Trace A is for one
equatorially coordinated histidyl imidazole ligand (Am: 1.75 MHz and reff = 3.0 A).
Simulation B assumes two magnetically equivalent histidyl imidazole ligand. Trace
C assumes one equatorial and one axial (Aim: 0.52 MHz and r,fr = 3.2 A) histidyl
imidazole ligands. Other parameters are listed in Table 4.1. .......................... 164

Figure 4.12. Time-domain computer simulations (dashed lines) of the normalized
l4N ESEEM data collected for TlD Fet3 protein at g = 2.06. Trace A is for one
equatorially coordinated histidyl imidazole ligand (Aim: 1.75 MHz and ref, = 3.0 A).
Simulation B assumes two magnetically equivalent histidyl imidazole ligand. Trace
C assumes one equatorial and one axial (Aim: 0.52 MHz and reff = 3.2 A) histidyl
imidazole ligands. Other parameters are listed in Table 4.1. .......................... 165

Figure 4.13A. Tl D F et3 ESEEM data and computer simulations at B0 = 2690G for
different r values: A) 260ns, B) 350ns, C) 435ns. Hamiltonian parameters are the
same as in Figure 4.9a in which one magnetic l4N nuclues couples with the electron
spin. ............................................................................................ 167

Figure 4.13B. TlD Fet3 ESEEM data and computer simulations at B0 = 3050G for
different 1: values: A) 230ns, B) 310ns, C) 385ns. Hamiltonian parameters are the
same as in Figure 4.10a in which one magnetic l4N nuclues couples with the electron
spin. ............................................................................................ 168

Figure 4.13C. TlD F et3 ESEEM data and computer simulations at B0 = 31406 for
different 1: values: A) 225ns, B) 300ns, C) 375ns. Other Hamiltonian parameters are
the same as in Figure 4.13 A and B in which one magnetic l4N nuclues couples with
the electron spin. ............................................................................ 169

Figure 4.14. Representative experimental l“N-ESEEM spectra for type-1 copper site
(T2D Fet3). Spectrum (A) was collected at B0: 2825 Gauss and for 1: = 250ns.
Spectrum (B) is the cosine Fourier transform of (A). Other measurement conditions:
microwave frequency, 8.836 GHz; microwave power, 43 dB; sample temperature,
4.2K. ........................................................................................... 171

Figure 4.15. Representative experimental l“N-ESEEM spectra for the type-1 copper
site (T2D Fet3). Spectrum (A) was collected at B0: 3090 Gauss and for 1: = 230ns.
Spectrum (B) is the cosine Fourier transform of (A). Other measurement conditions:
microwave frequency, 8.836 GHz; microwave power, 43 dB; sample temperature,
4.2K. ............................................................................................ 172

Figure 4.16. Determination of number of histidyl imidazole ligands for the type-1
center of Fet3. The stimulated-echo ESEEM (solid lines) (A) were collected at g =

xvii

2.23 using a ‘r-value of 250 ns, while that of (B) were collected at g = 2.04 using a I-
value of 227 ns. Other common experimental conditions common to the ESEEM
measurements are given in Figures 4.14 and 4.15. The Hamiltonian parameters used
to calculate the simulation (dotted lines) curves are given in Table 5.2. ............. 174

Figure 4.17. Computer simulations for single nucleus, a) N1 and b) N2 for static

magnetic fields, 2825G(A) and 309OG(B). Spectra c in A and B are ESEEM product
of ESEEM functions of N1 and N2. Hamiltonian parameters used to generate a and b
are found in Table 5.2. ...................................................................... 175

Figure 4.18 14N ESEEM data for T2D F et3p (a) and computer simulations (b) at
various 1 values. .............................................................................. 178

Figure 5.1. Continuous-wave EPR spectra of (A) wild-type Fet3, (B) type 2 depleted
(T2D) Fet3 and, (C) type 1 depleted (TlD) Fet3 protein in D20 buffer. For (A),
measurement conditions were: microwave frequency, 9.48GHz; microwave power,
2mW; modulation frequency, 100kHz; modulation amplitude, 4.014G; temperature,
15.6K. For (B), measurement condition were: microwave frequency, 9.48GHz;
microwave power, 2mW; modulation frequency, 100kHz; modulation amplitude,
4.014G; temperature, 21K. For (C), measurement conditions were: microwave
frequency, 9.48GHz; microwave power, 2mW; modulation frequency, 100kHz;
modulation amplitude, 4.014G; temperature, 15.6K. .............................. 191-192

Figure 5.2. Two-pulse 2H-ESEEM experimental traces for type-2 copper site in
Fet3 protein (TlD). (A) Two-pulse ESEEM data in D20 buffer, measurement
conditions: v = 8.79GHz; microwave power, 44dB; B0: 3055 G and 1: = 250ns. (B)
Two-pulse ESEEM for TlD in H20 buffer, measurement conditions are the same as
A. (C) Quotient (A/B) modulation envelope of T1D Fet3p. ........................... 193

Figure 5.3. Two-pulse 2H-ESEEM experimental traces for type-1 copper site in
Fet3 (T 2D Fet3p). (A) Two-pulse ESEEM data in D20 buffer, measurement
conditions: v = 8.79GHz; microwave power, 44dB; B0 = 3090 G and t = 250ns. (B)
Two-pulse ESEEM for T2D in H20 buffer, measurement conditions are the same as
A. (C) Quotient (A/B) modulation envelope of T 2D Fet3p. ........................... 194

Figure 5.4. Theoretical two-pulse 2H ESEEM traces for (A) equatorial, (B) axial and
(C) ambient water. Hamiltonian parameters for ESEEM simulations for equatorial
D20: A,, = -0.6, Ayy = -1.2, A,, = 1.4 MHz, gn = 0.85741, magnetic field = 3055 G, t
= 250ns, equ = 0.22 and 1] =0.5. Parameters for axially bound D20 are the same
except for the hyperfine tensor, A“ = -0.5, Ayy = -0.6, A,,= 1.1 MHz. For ambient
D20, A,, = -02, Ayy = -02, A,, = 0.4 MHz which corresponds to an r,,, of 4A. 196

Figure 5.5. Comparison of 2H-ESEEM traces calculated by angle-selection
(ANGTPSPNI, dotted curves) and spatial orientation averaging (TPSPNl, solid
curves) for D20 axially coordinated to copper. All the traces are calculated with A“,
= 0.15 MHz, reff = 2.8A and no NQI contribution. Additional parameters needed to

xviii

generate the dashed curves are diagonal elements of the g-tensor and A-tensor of
copper: gm: gyy = 2.05, g,, = 2.25, A,, = Ayy = 35 MHz, and A,, = 586 MHz. Dotted
trace 5.4A has 0,), 9n = 0°,0°. Dashed trace 5.4B has 0D, 4),, = 35°,0°. .......... 198-199

Figure 5.6. Comparison of 2H-ESEEM traces calculated angle-selection
(ANGTPSPNI, dotted curves) and spatial orientation averaging (TPSPNI, solid
curves) for D20 axially coordinated to copper. 2H NQI is included (equ=0.22 MHz,
11:0.50). Dashed curve 5.5A assumes 00, on = 35°,0° and NQI Euler angle coincident
with the g PAS (or = B = y = 0). Dashed curve 5.5B assumes GD, (1),, = 35°,0° and NQI
Euler angle coaxial with HFI PAS (or = y = 0, B = 225°). All other parameters are the
same as in Figure 5.5. .................................................................. 200-201

Figure 5.7. Comparison of 2H-ESEEM traces calculated by angle-selection
(ANGTPSPNI, dotted curves) and spatial orientation averaging (TPSPNl, solid
curves) for D20 equatorially ligated to copper. All the traces are calculated with A,”
= -1.0 MHz, reff = 2.5 and no NQI contribution. Additional parameters needed to
generate dotted curves are diagonal elements of the g-tensor and A-tensor of copper:
g,, = gyy = 2.05, g,, = 2.25, A,, = Ayy = 35 MHz, and A,, = 586 MHz. Dotted trace
5.6A has GD, (1),, = 90°,0°. Dotted trace 5.4B has DD, (1),, = 125°,0°. .............. 202-203

Figure 5.8. Comparison of 2H-ESEEM traces calculated angle-selection
(ANGTPSPNI, dotted curves) and spatial orientation averaging (TPSPNl, solid
curves) for D20 equatorially coordinated to copper. 2H NQI is included (equ = 0.22
MHz, 11:0.50). Dotted curve 5.7A assumes OD, (to = 90°,0° and NQI Euler angles, (or
= y = 0, B = 270°). Dotted curve 5.7B assumes DD, 4),, = 125°,0° and NQI Euler angles
(or = y: 0, B = 325°). All other parameters are the same as in Figure 5.7. .....204-205

Figure 5.9. Splitting of Ms electron spin level by the nuclear Zeeman and hyperfine
interaction of nearby nucleus with 1:1. Quadrupole interaction is neglected.
............................................................................................. 207 -208

Figure 5.10. Broadening of the fundamental frequency as function of distance of
water deuterons. Traces, (A) equatorial, B) axial and, C) ambient D20, are Fourier
transform of the simulations shown in Figure 5.3. ...................................... 209

Figure 5.1]. Computer simulation (dotted) of (A)time and (B)frequency domain 2H-
ESEEM data (solid) of T1D Fet3p. Calculated ESEEM traces assume a product of
the modulation function of one axial D20, one equatorial D20 and a distribution
(n,=(41t/3)(pN)(r3,,,-r3,)) of ambient D20. .................................................. 211

Figure 5.12. Computer simulation (dotted) of (A)time and (B)frequency domain 2H-
ESEEM data (solid) of T1D Fet3p. The ESEEM traces were generated using the
modulation function of a distribution (n,=(21t/3)(pN)(r°,,i-r3,)) of ambient D20. No
covalent waters were included. ............................................................ 214

xix

Figure 5.13. Frequency ESEEM data of T1D and T2D Fet3p in H20 and D20 buffer
system showing only the modulation frequencies due to proton nucleus. ............ 216

XX

INTRODUCTION

Multicopper oxidases (e.g. laccase, ceruloplasmin, ascorbate oxidase) are
enzymes that couple four, one-electron oxidations of a substrate to the four—electron
reduction of molecular oxygen to water. A recent addition to this group of enzymes
is Fet3 (Fe = iron, t = transport) protein from the budding yeast Saccharomyces
cerevisiae. Among the multicopper oxidases, only Fet3p and ceruloplasmin possess
ferroxidase activity, that is, specificity towards Fe(II) as the substrate. This unique
enzymatic activity underscores the specific role that these two multicopper oxidases
play in the iron homeostasis of their respective organisms. Yeast is an excellent
model for the study of metal homeostasis because the genetics of the organism is
easily manipulated.

Two of the copper sites in Fet3 protein, in their divalent oxidation states, are
paramagnetic and therefore, can be probed by the Electron paramagnetic resonance
(EPR) techniques. This thesis will present the biological application of electron spin
echo envelope modulation (ESEEM) with Fet3 protein the system studied. Chapter 1
describes spectroscopic techniques used in the study. The fundamentals of
continuous wave (CW) and pulse-EPR will be discussed. In Chapter 2, the
chemistry of copper; properties of multicopper oxidase; and the biochemistry of Fet3
protein will be introduced. Chapter 3 is a description of the pulsed-EPR instrument at
Michigan State University. Recent changes and modifications on the instrument are
discussed. The last two chapters give details on the experiment, results and analysis.

Chapter 4 discusses the ”N ESEEM studies of the histidine ligation to the type 1 and

type 2 copper sites in Fet3 protein. The coordination of these copper sites to water

and their accessibility to water solvent, determined using 2H-ESEEM technique, are
the main the focus of Chapter 5. Chapter 5 also integrates the results obtained from
the 1“N ESEEM and 2H-ESEEM experiments to give the proposed structures of the

type 1 and type 2 copper sites in Fet3 protein.

Chapter 1

INTRODUCTION TO CONTINUOUS WAVE- AND PULSE-ELECTRON
PARAMAGNETIC RESONANCE
Magnetic resonance spectroscopy has proven to be an essential spectroscopic
tool in the investigation of many biochemical and biological systems. Electron
paramagnetic resonance (EPR) spectroscopy, for example, has found wide range of
application in the studies of photosynthesis, bioenergetics, membranes and electron
transport.

EPR Spectroscopy was experimentally observed in 1945 by a Russian

scientist, Zavoisky.1 Gradually, it developed to be a powerful tool for studying
atoms, ions and molecules that have unpaired electrons. Systems with unpaired spin
are said to be paramagnetic. The paramagnetism could be inherent to the system or
could be produced by irradiation or other physical or chemical treatment of
nonmagnetic substances. Paramagnetism implies a nonzero angular momentum and
therefore the existence of magnetic dipole moment. EPR deals with the interaction
of this magnetic moment with an external magnetic field. The interaction is termed
the electronic Zeeman Interaction and is probed by applying an oscillating
microwave magnetic field perpendicular to the applied external magnetic field.

The first objective of this chapter is to present a brief description of the

phenomenon of EPR. Magnetic interactions , other than the Zeeman interaction, that

contribute to the spectra will be discussed}8 The second goal is to describe the

pulse version of EPR, in particular, Electron Spin Echo Envelope Modulation

(ESEEM).9'16 ESEEM is an EPR technique used to resolve hyperfine interactions

that are masked in the continuous wave-EPR (CW-EPR) spectrum. Combination of
the CW-EPR, and ESEEM can provide a better picture of the geometric and

electronic structure of paramagnetic substance.

1.1 Interaction of Free Electron with a Magnetic Field
Classically, the energy as a result of the interaction of the magnetic moment,

vector, It, and applied magnetic field vector, B, is mathematically described as

E: 1.10 B (1.1a)
or equivalently,
E: -uBcosO, (1.1b)

where 0 is the angle between the magnetic moment and direction of the magnetic
field (Figure 1.1). For macroscopic magnetic moment, the energy is a continuum as 0

can assume any value from 0 to 21:.

\S/

 

   

 

 

/N\

Figure 1.1. Dependence of the energy of interaction of a classical magnetic dipole
and external magnetic field on 0. 0 is the angle between the axis of the magnetic
dipole and the direction of the applied field.

Equation 1.1-1 has a quantum-mechanical counterpart, or operator, for the
microscopic system of the electron. The magnetic moment of an electron is

expressed as an operator in equation (1.2).

A

it = 19.138 “-2)
S is the operator for the electron's intrinsic angular momentum, while gc is the

electron g-value (2.0023) and B is the Bohr magneton and is equal to

eh/2m,c=9.27401 * 102‘] T". Using the value of u, we can write the quantum-
mechanical counterpart of equation (1.1-1a) as

36 = gfié- B (1.3)
If the z-direction is chosen to be the direction of the magnetic field (B=B,k=Bok,
B,=By=0), then equation (1.3) reduces to

at = gfié. -B. (1.4)
Borrowing from the results of the rigid rotor model problem for describing the

quantum mechanics of angular momentum:
SISmS) = ,/S(s +1)h2|Sm,)

s,|5ms) = msh where m,= -s, -S+1, ...+3
and (8,) = <Sy> = 0,

where S is the total angular momentum operator and m, represents the z-component
of the angular momentum. Experimentally, it is known that a single unpaired
electron spin gives rise to two Zeeman energy states. Following the results of the
rigid rotor problem, S=1/2 so that the unpaired spin can exist as two energy states

described by m5 = i1/2. These two energy states are typically labelled IO> =

|1/2,1/2> and |B> = |1/2,-l/2>. Operating with the Hamiltonian operator of equation

(1.4) on the |0t> and |B> states yields

13, = $16330 (1.5a)

1
and E5 = —5Bg,Bo (1.5b)

1.2 A Simple CW-Experiment

Figure 1.2 shows the diagram describing the splitting of the two degenerate
electron spin states in the presence of an external magnetic field. The low-energy
spin state corresponds to the states with spin projection m, = —1/2 and the high-energy
spin states corresponds to spin projection m,=1/2. The population of the spin states

follows Boltzmann distribution so that the relative population of |0t> to |B> state is

E = exp[—AE/kT] (1-6)

m

where ns. is the population of the state, k the Boltzmann constant, T the absolute
temperature and AE (equal to gBBo) the energy difference between the two states. It
is essential that the ratio should be less than one (n5 > n,,) for a net absorption of the
microwave field energy to occur. Decreasing the temperature is one way of
increasing the population difference.

In a typical EPR experiment, the energy difference is the measurable
property. It corresponds to the energy required to the change the orientation of the
electron's magnetic moment. In general the energy difference is probed by irradiating
the paramagnetic species with a fixed microwave frequency with magnetic field

strength, B,, while the varying the external magnetic field, 8,, to yield an absorption

~<GW1TJZFJ

CW-EPR signal as shown in Figure 1.2. For a transition to occur, Bl should be
directed perpendicular to the B0 field, following the selection rule, Am, = i1 and Am,
= 0. Depending on the detection system used in the instrument, a first-derivative EPR
signal can also be recorded. Resonance is achieved when the energy of the radiation
matches the energy difference between the two states,

hv = g,,BB0 or [1.7a]

at: geBBo/h. [1.7b]

ms=1/2
E=(l/2) gBB

 

ms=-ll2
E=(-1/2) gBB

 

 

Figure 1.2. A typical CW-EPR experiment. A variable external magnetic field splits
the degeneracy of or and B electron spin states, while a fixed microwave radiation
induces transition between the two states to trace an absorption or first derivative
EPR spectrum.

Classically, (no, the electron Larmor frequency, corresponds to the precession of the
magnetic moment of the electron about the external magnetic field. In quantum
mechanics, it corresponds to the frequency required to flip spin states of the electron

from m5 = -1/2 to m, = 1/2.

1.3 Spin Hamiltonian
The Hamiltonian operator that describes the energy of an atom or radical with

unpaired electron spin in the presence of nuclei that also have magnetic moments

consists of several energy terms,6"7

36136,,+3£ch+3615+9£55+3£&+3£”+3£& +36H+J£Q, where
at

C

, = electronic Hamiltonian operator consisting of the sum of the kinetic energy of
each electron and the potential energy of the electron relative to the nuclei and the
coulombic interaction ( 104-10° cm“),

3%. = Hamiltonian describing the interaction of the electron orbital with the electric
fields of the crystal or the chemical bond of the molecules (103-10‘ cm"),

36,, = spin-orbit interaction Hamiltonian , AL'S (102 cm"),

Jigs = electron spin-spin interaction (0-10 cm"),

£7, = electron Zeeman interaction operator (0-100m"),

M=hyperfine interaction operator describing the coupling of electron-nuclear spin
(0-102cm"),

fl,” = nuclear Zeeman (0-10'3 cm"),

.96" = nuclear spin-spin interaction,

EfQ = nuclear quadrupolar interaction Hamiltonian (0-10'2cm").

.76, and .760. operate only on the spatial term of the total electronic
wavefunctions . Both, plus the .9615, are often much greater than magnetic
interactions that give rise to an EPR signal so that their specific forms need not be
known. The relevant terms that contribute to the spin Hamiltonian, therefore, are

36=9£h+3£35+36HF+3£b+3£Q+3£W
The effect of spin-orbit coupling on the g-value is included in the effective electron
Zeeman , 39,. For the copper system studied in this work, however, the energy
operators that give rise to the observed EPR are

fl=3£h+3£fl+ign+flo. (1-8)

These four Hamiltonian operators will be discussed in the succeeding sections.

1.3.1 Electron Zeeman Interaction

The resonance condition, equation (1.7) was derived for the case of an
electron coupled to an external magnetic field. In addition to its intrinsic spin angular
momentum, the electron in an atom has orbital angular momentum, L. This orbital
angular momentum couples with S to yield a resultant angular momentum J. For

light elements such as Cu+2 and other transition metals, the coupling of L and S is

adequately described by the Russell-Saunders coupling scheme,3

J=L+S.
The vectorial addition of the two angular momenta (Figure 1.3) results in an effective
magnetic moment vector which can be derived to be equal to

11: 'gr bJ,

3J(J+l) + S(S+l) - L(L+1)
2J(J+l)

 

where gJ = assuming g,= 2.0. Hence, an appropriate

electron Zeeman Hamiltonian would be ,

file = gJBJ'B (19)

HI. 9

L1

 

”L 111.

Figure 1.3. Schematic diagram for the vector addition of electron spin angular
momentum, S and orbital angular momentum, L to yield an effective angular
momentum J.

and the corresponding resonant condition would be

AB = hv == gJBBo. (1.10)
The g-factor, g,, is characteristic of the sample and should be experimentally
determined.

When atoms are part of the molecular framework, the degeneracy of the
orbitals are lifted and orbital angular momentum becomes zero. We say then, that the

orbital angular momentum is "quenched" and therefore expect that the g-factor will

10

be equal to the electron g-value. However, spin-orbit coupling restores some orbital

angular momentum by admixing orbitals.

Another way of writing the equation is .761, = ( BL + geS) 0 B. A more
succinct form of the equation is

flu=$og0B (1.21)
In the equation, 8 is the effective spin and g is a 3x3 tensor that takes into account the
dependence of effective moment u,- on the orientation of external magnetic field.
The elements of the symmetric g-tensor,

gxx gxy gxz
g: gxy gyy gyz
gxz gyz gzz

can be derived to be equal to3

gal3 : gcaufl ‘1‘ 21.2 (nlLa:"Z-<g'.l.llfl|n> . (122)

 

In a coordinate system that diagonalizes g, the electron Zeeman becomes

A A A gxx O 0 11130
are: fi[s. 5y 8.]. o gyy o . hie. , (1.23)
O 0 gzz h3Bo

where the h,., are the direction cosines in polar coordinate so that the electron Zeeman

Hamiltonian becomes,
39,. = BBO[(g,, sin 6cos ms, + (ng sin 63in ns, +(g,, cos as] (1.24)
In an axial g-tensor, meaning g,, = gyy = g 1 and g,,= g”, the electron Zeeman

Hamiltonian reduces to

.761, = BB,,[gl sin 6(cos $3. + sin oSy) Hg" cos 0)S:]. (1.25)

11

The corresponding energy eigenvalue for the above Hamiltonian is

El, = B m, (g ,2 sin20+ (gn2 c0320)"2 B0 = B ms g B0 (1.26)

g = (g ,2 sin20 + g"2 c0320)”.

The energy equation assumes the form that we are more familiar with where gc is
replaced by the effective g-factor. Clearly, equation (1.26) implies (28 + 1) energy
levels that equally space even in the presence of an anisotropic g.

Anisotropy in g is lost in a system in which molecules tumble and rotate
rapidly, so that an average g-value equal to the average of the trace of the g-tensor is

observed. This averaging will allow us to simplify the electron Zeeman

Hamiltonian,
at. = rests . B
lfillBo
‘xlc zgisodsx Sy 8: hZBo (1.273)
h3Bo

or simply , in the case of the external applied in the z-direction,

96.; g...fl§.(h.B.> = gsfiéfi. (1.27!»

1.3.2 Nuclear Zeeman Interaction

Several nuclei have nuclear spin and an associated magnetic moment, 11,.
Similar to the electron, the nuclear magnetic moment of the magnetic nucleus is
proportional to its nuclear spin angular momentum,

“N = gNBNIN, (1.29)

where gN is the nuclear g-factor which is usually assumed to be isotropic and whose

magnitude depends on the identity of the nucleus, BN (5.05078 * 10'27 JT') is the

12

nuclear magneton which is 1838 times smaller than the electron Bohr magneton and,
IN, is the spin angular momentum for the nucleus which can assume an integral or
half-integral value. The absence of the negative sign in the right hand of the equation
implies that the nuclear magnetic moments and nuclear spin are parallel, although that
is not always the case. The magnitude of BN relative to B is very small because of the
large rest mass of the proton with respect to the electron. Consequently, the electron
Zeeman is about 103 stronger than the nuclear Zeeman.

The Hamiltonian for describing the nuclear Zeeman is given by,

fin. = -gnfinin -B. (1.30)

This Hamiltonian adds to the electron Zeeman to yield the total spin Hamiltonian,

yer: gBSOB—gNBNINOB. (1.31)
The above equation tell us that or and B electron spin states are split according to the

projection of 1, ml, in space.

1.3.3 Hyperfine interaction

The nuclear magnetic moment creates a local magnetic field (Bum) which may
couple to the moment of the unpaired electron spin to enhance or oppose the external
magnetic field (B,,,) yielding an effective magnetic field (B,,,). The resonance
condition is still valid if B is replaced by Be“. Because there exists a 21+1 multiplicity
nuclear spin states for a nucleus with spin 1, there would be 21+1 different values of
the resonance field strength.

The interaction of the electron spin and nuclear spin magnetic moments is

termed the hyperfine interaction. It is responsible for the additional (21+l) splittings

13

that we observed in the CW-EPR spectrum. The energy of the hyperfine interaction

is defined by the Hamiltonian,

3c: sag-B or (1.32a)
A... A... A... i.

H: [3. 5y 5.]. A... Ayy Ayz . Iy , (1.32b)
sz Ayz Azz i

N

where S and I are the electron and nuclear spin operators and the A the hyperfine

tensor. A, a 3X 3 matrix, is a sum of the anisotropic interaction and the Fermi

contact interaction,
A: T + Aiso 1. (1.33)
Aiso , a scalar quantity, is the isotropic hyperfine coupling and is proportional to the

probability of finding an electron in a nucleus,
87:
A.» = -3- gfigNIivl M0) |2 (1.34)

It vanishes for orbitals with l > 0 because |\1!(0)|2 is zero. The anisotropic part of the
hyperfine interaction is most simply modelled according to the classical interaction of
two point dipoles. The energy of interaction of two point dipoles is described

classically as

 

301.» r)(u~ or) _ ()1. HM} (1.35)

E . = —{
d1 lar
P0 I.5 1.3

In the above equation, r represents the vector joining the unpaired electron and

nucleus. Vectors It, and 11,, are the classical electron and nuclear magnetic moments.

For a quantum mechanical system, the magnetic moments in equation (1.35) must be

14

N

 

j)
J)

7)

C6

K?

 

 

Figure 1.4. Dipolar interaction between the electron magnetic moment, It, and
nearby nuclear magnetic moment, “N. The pictorial description assumes that it, and
It], are aligned along the direction of the magnetic field. The vector r represents the
separation between the two dipoles and 0 is the angle between r and the direction of
the magnetic field. The dipole-dipole energy can be approximated to be

Ed, = [(1 — 3cos2 H)/r°]y~zu,, = Ewan“. Depending on the value of 0 , the Blocal may

enhance or oppose the applied external magnetic field.

15

replaced by their corresponding spin operators to give,

(1.36)

 

r3

3S0 Io S-I

Equation (1.36) can be expanded in Cartesian coordinates to yield,
I 2-2... 2-2.... 2-2....
3" 5 r s.I.+3y 5 r 5.1.4.3Z 5 r 3.1.
at: gNflvgefl< 3" r 3 r 3 >. (1.37)
x .. ~- 2” .. x2” ..
+——y-(s.1.+s.1,)+-L5(s.1,.+s.1.)4—5—(s.1.+5.1.)J

 

 

 

t r5 1' r

 

 

Integrating the Hamiltonian over the spatial variable yields a compact Hamiltonian

that depends only on spin operators,

)

35",: -I-i. (1.38)

T is a traceless tensor whose diagonal elements are given by,

 

.2__ 2
T. = gNBNg.13<3‘ 5 r >, i: x,y,z (1.39)
r
and the off-diagonal elements are given by
3ij
Tij = gNBNgeB<—g‘>. (1.40)
r

In liquid phase, the molecules are in constant motion. If the rate of rotation and
tumbling of the paramagnets is larger than the microwave frequency, the anisotropic
hyperfine averages out leaving only the Fermi contact, Am,

A... = (A... + Ayy +A..)/3. (1.41)
as consequence of T being traceless. Figure 1.5 depicts the splitting of energy levels

that will be observed for an S=l/2, I=1/2 system characterized by isotropic hyperfine

interaction. Shown also is the expected CW-EPR spectrum.

16

 

 

 

 

 

 

 

 

 

 

 

 

     

Figure 1.5. Isotropic hyperfine interaction. Energy diagram showing splitting of

energy states for the case of S=1/2, I=1/2 system. The splitting due to nuclear
Zeeman interaction is not shown.

17

1.3.4 Nuclear Quadrupole Interaction (NQI)

Nuclei with spin I 2 1/2 have a quadrupolar moment, eQ, that is collinear with

the nuclear magnetic moment. Quadrupolar nuclei may have either oblate or prolate
charge distributions (Figure 1.6) compared to nuclei with I = 1/2 which have a

spherical charge distribution. The magnitude of eQ is a measure of the deviation
from a spherical charge distribution in the nucleus. Mathematically, eQ is 3
eQ = 27:] (pm, 0)(3 cos2 0—1)r4drsin 6d6 , (1.42)
0 0
where the dependence of r on 0 is removed and pN is the charge distribution function

of nuclear charge.
Nuclear quadrupole interaction is the interaction of the quadrupolar moment
with electric field gradients at the nucleus. These electric field gradients arise from

the non-uniforrn electronic charge distribution caused by the atoms in the vicinity of
1, pt

1...

(a) (b)

   

Figure 1.6 Nuclei with quadruple moment, (a) oblate and (b) prolate.

18

the nucleus. The magnitude and orientation of these gradients are determined by the
electronegativity and the geometry of the atoms surrounding the nucleus.

Since the axis of the of the quadrupole moment and the nuclear dipolar
moment are coincident, we can describe the nuclear quadrupolar interaction as the

nuclear spin self-coupling,

A A A Qxx Qxy sz Ix
.g‘ = (Ix 1y Iz Qxy ny Qyz Iy (1.43)
On Qyz Q12 12

where Q is the traceless quadrupole tensor that describes the anisotropy of the
quadrupole interaction. In the coordinate system that diagonalizes the Q-tensor,

the principal values of the tensor are given by

1
xx = -—A 1—
Q 2 ( 77)
l
ny=--2-A(1+n) (1.44)
sz = A ,
2 _
where A = flfindn = M. NQI will be discussed further in chapter 4
21(21 — 1) sz

where MN coupling with the electron spin is considered.

1.4 ESEEM

Knowledge of the hyperfine and NQI tensors would give us information
about electronic and geometric environment of our paramagnetic system. These
magnetic parameters are often difficult to extract from the CW-EPR. Biological

samples are usually prepared as frozen solutions so that the CW-EPR spectra are

19

broad, powder-like spectra. These spectra are comprised of EPR spectra of
individual single crystal-like paramagnetic species whose molecular axes are
randomly oriented with respect to the direction of the external magnetic field.
Consequently, the hyperfine and NQI tensors are obscured because they contribute to
the broadening of the spectra. To circumvent the poor resolution of the hyperfine
tensors, one may resort to studying single-crystal samples . However, preparation of
single crystals of biological samples is a difficult task. A more reasonable solution is
to make use of the sensitivity of the Electron Spin Echo Envelope Modulation

(ESEEM) technique in measuring weak hyperfine couplings.

ESEEM is pulsed- EPR method that requires the application of a sequence of
intense microwave pulses to induce a paramagnetic signal in the form of a spin echo.
The intensity of the spin echo is measured as function of time separating two
microwave pulses. Spin echoes monotonically decay with time. In addition, they
are modulated by neighboring magnetic nuclei. An envelope of modulated echoes
contains information about the electron-nuclear interaction as well the nuclear
quadrupole interaction. The two most commonly used pulse schemes in ESEEM are
the two- and three-pulse sequence. The physics and mathematics of spin echo

formation will be discussed in the next sections.

1.4.1 Effects of Microwave Pulses

The formation of spin echo is based on the concept of an inhomogeneously

broadened EPR spectrum as shown in Figure 1.7.10 The EPR spectrum centered at

B0 has a full width half maximum (FWHM) of AB. It is an envelope of many

20

homogeneously broadened lines of width ABi and center B,. These individual lines
are known as spin packets. A spin packet represents electron spins that experience
the same average effective magnetic field.

The effect of microwave (mw) pulses of duration tp is to burn a "hole" in the

inhomogeneously broadened EPR resonance line (Figure 1.6b).10 A typical EPR
spectrum covers a frequency range of several megahertz (10-102 MHz for radicals
and 103 MHz for metals) so that even the shortest microwave pulse (10ns) that can be
generated cannot cover the entire EPR spectrum. In contrast, NMR spectrum spans
only a few kilohertz so that radiofrequency (RF) pulses can adequately excite almost
all the nuclear spins in the sample to the same extent. Narrowing the microwave
pulses can widen the burnt hole as the narrower the pulse, the wider the frequency
distribution (1/tp 5 Am). However, a narrow microwave pulse needed to excite the
entire EPR spectrum may result in low intensity of the microwave field so that
electrons are weakly excited. Thus, unless the height of the pulse is simultaneously
increased, the depth of the hole becomes shallower and consequently signals are too
weak to be observed. This can be viewed better by considering the mw pulse as a
oscillating magnetic field that torques the magnetic moment of the electrons in the

paramagnetic sample by an angle 00,9, where 00, is the frequency of oscillation about

the microwave B, field and tp is the width of the microwave pulse.3r10 In this view
the purpose of the microwave pulses is to provide a magnetic field B, oscillating in

resonance with as many of the spin packets as possible. If B, is very weak and tp is

21

(a)

 

 

 

 

(b)

 

 

Figure 1.7. (a) An inhomogeneously broadened EPR spectrum comprising of
several homogenous spin packets with linewidth AB,. (b) "Burning of a hole" of
some width in the a inhomogeneously broadened EPR spectrum as a result of
applying a microwave pulse of some duration.

22

long, only a very few spin packets will be affected by the microwaves. A large B,
field with short tp, however, can excite spin packets over a broad range of resonance
frequencies about the microwave frequency (30-50 MHz bandwidths are common).
Therefore, the microwave pulse, of frequency 0),, , can be completely described as an
envelope of a time(t)-varying field of intensity B,cos((0,,t), where (130 = g,BBo and can

burn a hole of some width in the EPR spectrum.

1.4.2 Formation of an Echo in Two-pulse (The Hahn's Echo)

In 1950, Erwin Hahn discovered the first nuclear spin echo using the 1t/2-r-1t

RF pulse sequence.17 The first electron spin echo (ESE) was observed by R. Blume

in 1958 in his studies of electron-spin relaxation time of sodium-ammonia

solution.18 In the experiment, he used the Carr-Purcell rt-t-rt/Z pulse sequence.

Any sequence of two microwave pulses can give rise to an echo, but the
maximum echo intensity occurs in a 1:!2-t-n pulse sequence. The formation of a 2-
pulse electron spin echo can be described pictorially using the vector model (Figure
1.8) and quantitatively by the Bloch equation. In this description, a rotating
coordinate system (X,Y,Z) is chosen. The coordinate system rotates at angular
frequency equal to the frequency of the microwave pulse whose magnetic field, B,, is
directed in the X-direction and is perpendicular to the external magnetic field, B0
(BOIIZ). Prior to the application of the first 1t/2 pulse, the spin packets assume an
equilibrium value, the bulk magnetization Mo, in the presence of the external

magnetic field. M0 is the sum of the magnetization of the homogeneous spin packets.

23

The time-dependent evolution of M0 upon the application of the microwave can be

evaluated using the Bloch equation,

dMo
dt

 

= VB... x M. (1.45)

where B,ff is the effective field (B,,, + B,) in the rotating frame. Solution to this
equation can be viewed as the precession of the bulk magnetization about 8,... At
resonance (i.e. the B, field oscillate at (00 ), the magnetization will precess about B, at
0), = yB,. Following this constraint, the exact solutions to the Bloch equation are,

M2( to) = Mocoswi In and v('rp) = Mosinwit'p , (1.46)
where t, is the time duration of the microwave pulse and ‘D is the transverse
component of the magnetization defined along the Y-axis of the rotating frame. If the
microwave pulse is applied such that (0,1,, = N2, the M, component is zero while 1) =
M0 = M' and is lying on the Y-axis of the rotating coordinate system. When the first
1t/2 microwave pulse is turned off, the magnetization starts to dephase because the
individual spin packets comprising M' have different precession frequencies about
B,,. This process is called a free induction decay (FID) and it is what is most often
measured in solution NMR. Those spins packets that have precession frequency (00,)
greater than (00 will rotate away from Y counterclockwise at a rate Au) = (1), - 000,
while those spin packets with precessional frequency less than (1)(, will rotate
clockwise away from Y -axis by -A(1)= (no-(1),. After some 1, spin packets a and b
have rotated through an angle Au); and Awb‘t, respectively, and spin packets c and d -
Amer and -A00d1:, respectively. The dephasing or decay of the magnetization is

determined by the phase memory time T M which is in the range of the spin-spin

24

Figure 1.8. Vector model description of the formation of two-pulse electron spin echo.
A). Equilibrium condition, the paramagnetic system has a characteristic bulk
magnetization prior to the application of the first pulse. B) A 90° pulse is applied so that
a transverse bulk magnetization is created in the Y-axis of the rotating frame. C). Decay
of the magnetization to individual spin packets' magnetization during waiting period, 1:.
D) Reflection of the dephased pattern of magnetization about the X2 plane upon the
application of a 180° pulse. E) Formation of an 2-pulse echo when the spin packets
refocused at 21 after the first pulse.

25

Two-pulse echo formation

 

 

 

 

 

 

 

 

 

 

 

I . _L ._.
rt ‘ echo
\. I (0,,
pulse 1 Mo: M
tp = W2
A) t = 0 B) application of 1t/2 pulse
1:
c),
(”A
’—
(”Br pulse 2
(0 k T = n
c p 1,.
(”D
D) application of 1: pulse C) dephasing of spin packets

21:

a) w.

 

 

E) refocusing of net magnetization

26

relaxation T2. At time r after the application of the first pulse, the second 1: pulse is
applied. The effect of the second pulse is to reflect the dephased pattern of the
magnetization about the X-Z plane. This results in the change of the accumulated
phase angle for the spin packets a and b from Awwr to (it - Adi),b 1:). Similarly, the
phases of the spin packets c and d are changed to (n + Awmr). The spin packets still
precess at their former rate and direction so that at time 2r after the application of the
first pulse, spin packets a and b have accumulated phase equal to [(rt - Am,“ + Awa'br
= 1r) ] and spin packets c and d have the same phase, [(1t + Au),_dr - Awwr = n)]. Thus,
the spin packets converge to form transverse magnetization, M'(2r), in the -Y axis.
The refocussing of spin packets results in the emission or radiation detected in a form

of spin-echo.

1.4.3 Three-Pulse Echo formation (Stimulated Echo)

The classical description of he evolution of magnetization in the (1t/2-r-rt/2-

T-rt/2-echo(r)) three-pulse sequence16 is shown in Figure 1.9. The same coordinate
system will be used in the description. As expected, the 90° pulse will rotate the bulk
magnetization to the Y-axis. After some time r, the transverse magnetization M' will
dephase into its components. Spin packet A will form an phase angle, AwAr (where
Ari)A = (DA-(1)0) while spin packet B, -A(0,,r ( where -A(1),, = (no—(1)3) with respect to the
Y-axis. A second 90° pulse is applied at time r, rotating the dephased pattern of the
M' magnetization into the XZ plane where they are allowed to precess for time T.

During the long waiting period T, further dephasing occurs and a series of cones are

27

Figure 1.9. Classical description of the formation of three-pulse ESE.

A) Equilibrium condition. B) Application of the first 90° pulse creating a transverse
magnetization along Y. C) The transverse magnetization M' is allowed to dephase during
time 1:. D) Application of the second 90° pulse, this time the dephasing magnetization
pattern is on the XZ plane. E). Formation of precessional cones/storage of magnetization
along Z axis during time T. F) Cone formation right after the third 90° pulse is applied.
G) A stimulated echo is formed.

28

Three-pulse echo formation

 

 

 

 

  

 

 

 

 

 

 

 

l— r I T I T—I
1V2 1112 NZ echo
Z
\ I (00
X
Y
B) right after the application of
A) at t = 0 1V2 pulse :
z
Z V
0.) A J
\ a)
’5, °
Y «1 ......................
2nd 1r/2 pulse Y
D) right after application C) 31:33:? magnetrzatron dephases
of second 1r/2 pulse
z
V
00A
3rd 1112 pulse
X ............................
E) right after application
‘ of third «I2 pulse
E) magnetization stored at .
-Z at time T I".
Y

 

G) formation of stimulated echo
at time (Zr + T)

29

formed as projections of spin magnetization along the direction of B,,. These cones
allow the phase information contained in AmAr and -A0)Br of each spin packet to be
stored along the Z axis. Application of the third 90° pulse after time T, will bring the
stored Z-magnetization back to the XYplane where the spin packets will refocus at
time r after the third pulse. The three-pulse echo is called a stimulated echo because
the second 90° pulse puts the spins into a "holding" pattern in the Z direction until the
third pulse generates the rephasing in the XY plane to stimulate echo production.
The actual shape and width of the electron spin echo are dependent on the
sample and spectrometer parameters. In atypical ESE experiment, the integrated
echo intensity is recorded as function of r, in the case of two-pulse echo, and T in the
case of three-pulse echo. In a typical three-pulse experiment r is held fixed and T is
varied to generate an echo decay curve. The three-pulse echo as a function of T
decays with the spin-lattice time, T, since during time T the electron spins are
oriented in the z direction. Thus, the three pulse echo decays much more slowly than

the two-pulse echo.

1.4.4 Nuclear Modulation

In an Electron Spin Echo Envelope Modulation experiment, the intensity of
the electron spin echoes are studied as function of increasing one of the microwave
pulse spacings. For two-pulse ESEEM experiment, the time r is incremented, and in
a three-pulse ESEEM measurement, T, the spacing between the second and third
pulse is varied. In both experiments, an overall decrease in echo intensity is observed

This decrease is due to random and irreversible loss of magnetization by each

30

individual packet which is dependent on T2, spin-spin relaxation, for the case of the
two-pulse experiment and T,, spin-lattice relaxation, for the case of three-pulse
measurement. Interference effects that originate from electron-nuclear hyperfine
coupling lead to modulation of these intensities and provide a means for recovering
this information that is lost due to inhomogeneous broadening. The physics of these
modulations will be discussed in this section. Both classical and quantum

mechanical descriptions will be discussed.

The classical explanation”,13 is as follows (Figure 1.10). We assume that
S, is quantized along the direction of the external magnetic field. Nearby nuclei with
spin 1 of some distance r from the electron will experience an effective field B,eff
equal to the sum of the external magnetic field and the dipolar magnetic field, B,, due
to the electron. The magnetic nuclei will precess about this effective field. Similarly,
the magnetic nucleus will generate a dipolar field, BN whose magnitude is almost
negligible compared to the B0 so that the electron will precess about ~Bo. However,
the precession of the magnetic nuclei about B,eff will cause variation in B,,. This
time-dependent variation of the field, ABN,(t), causes modulation of the electron
Larmor frequency by the nuclear Larmor frequency (1),“. When a microwave pulse is
applied, the projection of S, will change and consequently will alter the effective
magnetic field experienced by the magnetic nucleus from B,eff to B,,“. If the
microwave pulse width is shorter than the nuclear Larmor frequency, not all nuclei
can adiabatically precess around the new B,,“ direction. Some nuclei will precess

about B,eff at WM and some nuclei will precess about B2cff at wm. Thus, microwave

31

 

 

 

 

82

Figure 1.10. Classical explanation of the origin echo modulation. An isotropic g-
factor is assumed implying that the S, is aligned along the direction of the applied
field. (A) An electron in motion produces a local magnetic field, B,, at the nearby
magnetic nuclei. H,e adds to the external magnetic field B0 to yield an effective
magnetic field, B,,“. The precession of the nuclei I about B,eff produces a time
varying local field A H,,(t) at the electron and consequently modulates the electron
Larmor frequency. (B) Application of the resonance pulse reverses S2 to -S, resulting
in a new effective magnetic field, B,,,“ about which the nucleus precesses. If the
duration of the pulse is shorter than the nuclear Larmor frequency, not all the nucleus
will able to precess about BM. Thus the electron Larmor frequency is modulated by
AB,,(t) and AB2,(t) at nuclear frequencies proportional to B,eff and B,,“, respectively.
Interference of these nuclear frequencies produces beats in the electron prescession
signal at the nuclear frequencies. The modulation of S, can be detected in a form of
an echo if a second resonant pulse is applied.

32

pulse causes a branching of the precessing nuclei into two sets that produce two
different nuclear-modulation frequencies at the electron. Interference between these
two nuclear frequencies produces beats in the electron precesssion signal at the
nuclear frequencies. The modulation of S, can be detected by applying a second
microwave pulse to generate an echo. The amplitude of the echo depends on the
orientation of the B,,,” and hence on the electron-nuclear distance r and on the
number of equivalent nuclei at r. Thus, an analysis of an echo-modulation pattern
gives information on the number and distance of magnetically coupled nuclei as well
as their identity from the modulation frequency.

A quantum-mechanical view of echo modulation can be described using the

energy level diagram for an S=1/2 and I=l/2 spin system shown in Figure 1.11.11
Anisotropy in hyperfine interaction contributes to the inhomogeneous broadening of
the EPR linewidth as electron spins experience different local magnetic field due to
the magnetic nucleus randomly oriented with respect to the direction of the external
magnetic field. This anisotropy also results in the spin states whose wavefunctions
are mixtures of nuclear spin states. As a consequence of the mixing of states, the two
forbidden transitions can be observed in addition to the two allowed transitions (AmS
= i1 and Am, 2 0). Clearly from the energy level diagram, it is possible to induce
transition from one initial spin states to different final spin states upon the application
of the microwave pulse. This branching of transitions is responsible for the
modulation of electron spin echoes.

The origin of nuclear echo modulation can be explained by combining the

vector model explanation of two-pulse ESE formation and the quantum mechanical

33

Figure 1.11. Combination of quantum mechanics and vector model of echo formation
to explain the origin of echo modulation. A) Four-energy level diagram showing branching
of transitions. B) Vector model showing the interference of spin packets as result of
branching of transition in a two-pulse echo sequence.

34

 

|1> = cos(0a/2) |1/2 1/2> + sin(0/2) |1/2 -/2>

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I
(1),, A : .1)—
? , A ; |2>= -sin(0a/2) |1/2 1/2> + cos(0/2) |1/2 -/2>
mg'ozs Diner.
A)(D13 C)0~)24 (us
i 1' T |3> -..- cos(0,,/2) |-1/2 1/2> + sin(0/2) |-1/2 -/2>
(DB i
Y i i |4> = -sin(0,,/2) |-1/2 1/2> + sin(0/2) |-1/2 42>
0=quantization axis
Z
X
Y Z
(013
it ulse
p X
..A/‘Dn
0’23
Z
2 r
X
4%”; Y

35

description of branching of transitions. For simplicity, the spin packet that undergoes
transition from E, to E3 at frequency (1),3 will be considered. After the first
precession, the spin packet will a develop a phase angle with respect to the Y axis.
Application of second pulse at time r after the first pulse will rotate the spin packet to
180°. In addition to this rotation, splitting of the magnetization vector will occur as a
result of branching of transition. The new spin packet that was generated will precess
at (0,3. Because of the difference in the precessional frequency, the new spin packet
will precess at a different frequency while the original spin packet will maintain its
original frequency. At time t after the second pulse, the original spin packet will
align at the Y axis and contribute to the echo. The echo, however, was modulated
because the second spin packet does not refocus and therefore interferes with the
formation of the echo. The extent of interference that results from the branching of
the EPR transitions depends on the projection of the branched spin packet on to the
Y-axis, cos{ (00,,-00,,)}r, weighted by the probability for the branching. The echo,
therefore, is modulated as one varies the time between the 90° and 180° pulses with a
frequency |(0),3-0)23)| = 00,.

The origin of modulation of the three-pulse electron spin echo is the same as
for the two-pulse echo. It essentially depends on transient excitation of branched
EPR transitions involving two nuclear spin orientations. One difference is that three-
pulse ESEEM is modulated by the nuclear frequencies 00,, and 0),, whereas the two-
pulse ESEEM is modulated by the sum and difference of (0,, and (1),, in addition to

modulation by the nuclear frequencies at each electron spin manifold. This difference

36

will be understood better by discussing formulation of the modulation of function of

the two and three-pulse ESEEM using the density matrix formalism.

1.4.5 Density Formalism of ESEEM
1.4.5.1 Density Operator

In ESEEM, we are concerned with measuring the bulk magnetization in the
form of a spin echo. Calculation of the time-dependent evolution of an ensemble

average can be done without knowing the exact state of the system by invoking the

concept of the density matrix. 19'”
Consider a system described by a set of orthonormal states and that the
probabilities that the system is in each of these states are w,. The expectation value of

an observable A is
(A) = EMA/1.0m), (1.47)
which is a quantum and statistical average. A density operator, p, can be defined as

f; = Zwimxm. (1.48)

We can express the set of orthonormal states as linear combinations of basis state so

that,
lVr>=ZC§”|¢.>- (1.49)
I

The statistical average, therefore, can be rewritten as,

(A)wu .—. 2w,2cj”‘(¢j lAz c,‘;"|¢,) (1.50)
1' j k
= z2[2.,.,.,,,.;.](,|n.,)
j k i

37

= 22011.41. = Tr (pA)
j k

in which the density-matrix p,,, the matrix representation of the density operator in

this basis, is defined by
pk}, = chmwiczil = (@1445!) (1.51)

If the physical observable evolves with time, then we should write the orthonormal

states as |w,(t)), the basis states as |¢,(t)), the coefficient as c,(t), and the observable

as (A(t)). We should also described the time-dependence of the density operator as,
[30) = 2w) uz,(r>)< 14ml. (1.52)
whose matrix-elements on, is

pkja) = 203;). (”W-CPU) = <¢,(r)[a(r)|¢j (1)). (1.53)

In a case in which the state of the system is known, all probabilities are zero except
for one (i.e. pure state), so that density operator becomes,
{30) = I W>>< W)! (1.54)

and the matrix elements as
p.,-(r)=c;(t)c.(t>=(¢.(r)|fi(t>|¢,(t)). (1.55)

Important properties of density matrix operator that need to be mentioned are that a)
the Tr(p)=1, b) it is a Hernritian operator, and c) in all its representation the diagonal
elements are always real and lie in the interval [0,1].

The density operator equation of motion is given by ,

fizz _ =1
(1, h{p(t)H Hp(t)} firearm. (1.56)

38

where H may time-dependent or time-independent Hamiltonian. If H is time-
independent, p(t), the density operator at time, t, can be detennined from p(0) at time

zero by ,

90) = emf?”

 

}p(0)exp{-IZ—t} . (1.57)

Therefore, if p(O) is known we will be able to calculate the average eigenvalue of the

some observable (A) at time t.

1.4.5.2 Application of Density Matrix to ESEEM

The detailed discussion for the derivation of the modulation function for two-

pulse and three-pulse ESEEM is presented in the manuscripts of W. Mims.9r23 In
this section, a brief explanation will be presented.

In ESEEM, we let our spin system (i.e magnetization) evolves in time during
the application of microwave pulses (nutation period) and when the microwave pulses
are absent (free precession period) as shown in Figure 1.12. We can track the time-
dependent motion of the magnetization of our perturbed system using the density
operator equation of motion. At some time, teem, we observe an echo which is due to
the refocused y-component, (My), of the transverse magnetization. The echo signal,
Emod, can be quantitatively known, if p(techo), density operator at the time echo
appears, is known. This echo signal is proportional to the Tr{p(t,c,,o)S,}.

The general form of echo signal is given by,

Emod = 7777mm... )fl}

39

where r] is constant that depends on the experimental arrangements and .76, is the
time-dependent Hamiltonian portion of the total Hamiltonian,

360,: 3t, + 3%,. (1.58)
360 is the static interactions in the sample that include the Zeeman interaction,

hyperfine interaction and nuclear quadrupole interaction. 36,, on the other hand,

a) TWO-PULSE ESEEM

III“ 1:1‘“/\_

b) THREE-PULSE ESEEM

11ml] . Lin/L

N. : nutation period

P,: precession period

Figure 1.12 Nutation and free-precession periods in two- and three-pulse ESEEM.

describes the interaction of our system with the field of the microwave pulse. The

explicit form of these Harrriltonians are

A

36.: BS-g-B-BNgNI-B + S-A- +i-13-i... (1.59)

36 = g,BB,S, cos(a),t). (1.60)

40

The trace should be summed over (a) the |or,> and |Bi> states (Figure 1.13a)
and the sum over the systems which make up the inhomogeneous line (Figure 1.13b).
These sums can be calculated separately assuming that a) the spread in the
inhomogeneous shifts AB,k (B,,, = B,avg +AB,,,) is large compared to the intervals in
the or and B manifolds and (b) AB,k only varies the spacing between the or and B
manifolds and does not lead to an inhomogeneous distribution of intervals within
each manifold. With these assumptions, the sum over or and B manifolds gives the
envelope modulation function and the sum over the inhomogeneous line yields the
form of the echo signal.

It is necessary to express 36,,“ .76, and .96, in a rotating-frame coordinate
system. The transformation to this coordinate means that we fix the direction of the
microwave field B,, say along X, thereby removing its dependence on time, t. This

transformation can be done by applying the similarity rotation operator,

R(6=wt)= Exp[%w2,], (1.61)
where X, =[é 31]

has an analogous meaning as the Pauli matrix 0,. I is identity submatrix of
dimension equal to the number of nuclear spin states (21 + 1). In the rotating frame,
.76, and .76, are expressed as 36', and 30,, respectively. After the appropriate
transformation to the rotating frame, the density operator during the nutation period

can be expressed as,
150,): R.b<t.)R~. (1.62)

where RN = exp{i(3c; + at; )(n — n)/ h} = exp{i(.7£, + at; )rp/h (1.63)

41

a)

 

 

 

 

 

 

 

 

 

 

 

 

 

|or> ‘ A A
i L A a— I
(1)
V V
r r
|B> __Y_
‘ v i
b)
Bmve
+AB0J:
Bo

 

Figure 1.13. A) Set of energy levels that interact with a resonant microwave pulse.
For spin echo modulation to be observed, it is essential that branching occurs. That
is, there must be transitions from on one given state in or electron spin manifold to
more than one final states in the B manifold. B) Distribution of values of B, in an
inhomogeneously broadened line.

42

and the free precession as
90,) = Iii/30.)): (1.64)
where R, = exp{i(36,)(tf —tr}/ h}. (1.65)

Using these exponential operators, the density operator when echo appears, [3

echo ’

can be determined after the sequence of nutation and precession with respect to the
initial density, f), , by the transformation

(3“,, = R*[)R (1.66)
where for the two-pulse sequence ,

R = RN, R,R,,,2 Rm (1.67)
and for three pulse

R = R,,,2R,RN2 &RN3R,. (1.68)
The initial density, [2,, is a time-independent equilibrium density matrix prior to the
application of the first pulse. It is diagonalized in the representation that diagonalizes
36', and the diagonal elements represents the thermal equilibrium initial population of

the alpha and beta electron spin states . Assuming all the sublevels in alpha and beta

electron spin manifolds are equally populated, the initial density in matrix form is

_ p; 0 _ a1 0
parting-'[O p;:|-[O bl] (1.69)

We can think of a and b as the average Boltzmann population of the several sublevels
in the alpha and beta electron spin manifolds.

36', can be diagonalized (i.e. transformed into Imsm,> basis), by using the
unitary transformation operators, M, and M,. M, and M, are (21 + l) x (21 +1)

matrices describing the mixing of nuclear spin states caused by the hyperfine

43

interaction in the alpha and beta manifolds. Diagonalization of 36', is essentially
diagonalizing its two block diagonal (21 + 1) ( (21 +1) submatrices for each alpha and
beta spin state. The exponential operator, R, and 36', should also be expressed in this
representation.

After all the transformations, we can figure out the density matrix at the time
of the appearance of the echo in the diagonalized representation for the two— and
three-pulse echo sequence. For the two pulse experiment, the echo is generated at
time Zr and therefore we are interested at knowing the p,,,,, at time 2r

9....(21)= R*p(0)R

The corresponding normalized modulation function for E(2r) is,
1
E r = —— Tr +M+P+M MIPM+H.C. . 1.70

In the equation, H.C. stands for Hennitian conjugate. I has the usual meaning, it is
the spin of the nucleus that gives rise to the observed superhyperfine couplings. P,
and Q, are (21 +1) x (21+1) diagonal matrices describing the evolution of the electron
spin manifolds during the preceesional period. They consist of elements (P,),, = P,i =
exp(iw,,,r) and (Q,),,k = Q,,,,= exp(i(1),,,,r). M is the unitary matrix M,t M3. If more
than one nucleus is coupled to the unpaired electron spin, the echo modulation

function assumes the form,
Em,,(r)=I]E;,,,(r). (1.71)
i=1

For the case of the three-pulse echo, the echo appears at time (2r +T). The

normalized B,,, is then,

(1.72)

 

1 HQ:M*(P;P:>MQ.M*(P.P.)M+ ]

E ,r =
“d” ) [4(21+1) (Q;Q;)M+P;M(Q,Q,)M*P,M+H.C.

When more than one nucleus is coupled, the product rule applies. However, the

product rule is obeyed within the same electron spin state so that
1 " . .-
Em,,(r, T) = .2_[1'[ E,(r,T) + E,(r, 7)]. (1.73)

It should be noted that the generalized modulation function for S =1/2 (equations
1.70-1.73) are derived with the assumption that 36, >> 36,. With this assumption the

exponential operator for the nutation period is simplified because the 36, is neglected.

1.4..5.3 Modulation Formula for an S=1/2, I =1/2 System

In the rotating frame, the Hamiltonian describing the interaction of the
nucleus of spin I = 1/2 and the electron spin S = 1/2 is given by,9’24r25

36, = h[(w. — ans. — (Uri: + sj + 5.31)]. (1.74)
The above equation assumes g to be isotropic. Likewise, 36, in the rotating frame is
expressed as

54 = hwléx . (hw1= geBBI). (1.75)
The eigenvalue of 36',, can be determined in a representation that diagonalizes the S,

and 1,. This can be carried out using the unitary transformation operator, exp(i(pIy),
36,“, = exp(i(ofy)36, exp(-irpiy) = Mar. — (0)32 - hwofz (1.76)
The transformation is equivalent to rotation about the y axis through an angle (p. The

angle (p is determined by

45

B'ms

_, 1.78
A'ms—wl ( )

mnwz

and assumes the values (p, and (p, when m, = or and m, = B, respectively. Physically,
these angles determine the quantization axes of the nuclear spin, the Larmor
frequency of which assumes two values corresponding to two electron spin states,

the alpha and beta states,

, 1/2
A ' B2
a) = wi— +— . 1.79

We can view exp(i(pIy) as an operator that nrixes nuclear states at the same electron

spin projections and can write it in matrix form using the |m, > representation.

,. cosga— -sin&
exp(-i(paly) = (p2 (p2 = M, (1.80)
sin—l cosJL
2 2
. cosfl —sinfl
exp(-irps1.) = 2 = M, (1.81)
srn— cos——
2 2
Therefore, the unitary matrix M assumes the form,
cos L; W sin L. :p, v u
M = M M + = = . 1.82
a ’3 . (Pa-(Pp (Pa-(Pp [—u* v*] ( )
—srn— COS—T

The column of the matrix M, are actually the eigenvectors that corresponds to the
two eigenvalues in the alpha manifold. This also applies to MB. In mathematical

form, the energy eigenvalues are given by,

46

_ _ 2 2 1/2
E(—1/2,l/2)=E,‘4=W-§[(ar+§) +§J

2 2 1/2
E(—1/2,—1/2)=Ep.3=M+h (a) +A) +5
2 2 2 4

2 2 1/2
E(1/2,=l/2) E,, =M- h(wr—é) +42
2 2 2 4

2 21/2
E(1/2,—l/2)= E,l =M+ +h(wI—é) +E .
2 2 2 4

Corresponding to the above energy eigenvalues are the eigenvectors,

91(-1/2,1/2)=\11, =—sin—|— —,1/2 1/2)+cos—|— —1,/2— —1/2)
‘1’(—1/2,—1/2)=‘1’, =cos—|— —,1/2 1/2)+sin— (”Z-”l 1/2,— —1/2)
‘1’(1/2,1/2)=‘1’,,=—sin—|1/2, 1/2)+cos—|1/2, —1/2)

‘1’(1/2,—1/2)=‘1’,,=cos-‘Z—“ll/2, 1/2)+sin%9‘-|1/2,—1/2)

(1.83a)

(1.83b)

(1.83c)

(1.83d)

(1.84a)

(1.84b)

(1.84c)

(1.84d)

The notations E(ms, m,) and ‘I’(ms,m,) are incorrect for the case described above.

They are correct for the case in which the electron and nuclear spins are quantized

along the same direction. They are shown for the purpose of showing the

transformation of the eigenvalues and eigenvectors when hyperfine anisotropy is

present.

Having known the elements of M, and M, allows us to express the

exponential operator for both the precession and nutation period in the eigenvalue

representation,

47

Ma -th M+
R.= ex —- a + , 1.85
I i MB] p[ 7' A Mr] ( )

where H=36', when pulse is off and H=36', when pulse is on. Using the explicit form
of M for I=1/2, the analytic solution for two-pulse echo,
1- cos(w, r) — cos(wflr) +

k
Emod(T)=1—_ l l (186)
2 Ecoscosqw, + 0),]1') — ~2—coscos([w, - (0,317)

and for three pulse,

k sin2(&z){l — cos{w,(r+ T)] +
Em,,(r,T)=l-— . (1.87)
sin2(w7°1-){1— cos[wfl(r+ T)]

In both equations, k is the modulation depth parameter and is equal to

 

2
k = 4|vl2|u|2 = shim. w.) =[ ”'3 ] , (1.88)
w a),

where M2 andlvl2 are the transition probabilities for the forbidden and allowed spin

transition. |u| and|v| are also known as branching parameters. They describe the
mixing, which may be due anisotropy in hfi interaction, of each zeroth-order nuclear
spin function into each spin function of the total electron-nuclear system and are
essential for the appearance of the ESEEM. When either M or |u| are zero (i.e. no

mixing), only the spin-allowed transitions are excited and no ESEEM will be

observed.9’26

1.5 Two-Pulse ESEEM vs. Three-pulse ESEEM

48

As discussed earlier, two factors determine the intensity of the echo. The
experimental echo decay of is a product of a monotonically decreasing relaxation
decay function and the modulation function,

E =E E (1.89)

echo decay mod '

The exact form of the decay function is not known. Approximately, it is modelled to

an exponential function, exp(-r/r,)° where n=1,1.5,2 or 0.5.27

The analytic solution to the two-pulse echo suggests that the modulation of the
two-pulse echo occurs at hyperfine frequencies and their sum and difference
combination frequencies. These frequencies will manifest in the Fourier transform
of the time domain data. Because of the product rule, frequencies due to the
combinations of frequencies arising from the different nuclei will also manifest in
addition to fundamental and combination frequencies of one nucleus. Thus, adding
complexities. These additional frequencies compounded by the short phase memory
time, TM (approximately in the order of the spin-spin relaxation, T2) that describes
the decay of echo with time will contribute to poor resolution of two-pulse ESEEM
data. The B,,, of the three-pulse echo, on the other hand, is modulated only by the
fundamental frequencies. When several nuclei are coupled, the product rule is
applied to alpha and beta electron spin manifolds separately. Also the decay of the
echo is governed by longer relaxation processes (electron spin-lattice relaxation,
electron cross relaxation and nuclear spin-spin relaxation) at liquid helium
temperature leading to a better frequency resolution than typically found for the two-

pulse experiment.

49

One important property of the three-pulse echo modulation function is the
suppression effect which is helpful in the assignment of superhyperfine frequencies
belonging to alpha or beta electron spin manifold. The choice of r has an effect on
the intensity of the modulation amplitude. By setting r over the range of values
determined by the period of the complementary hyperfine frequencies, we may be
able to enhance or suppress the shf frequencies from an electron spin manifold. This
suppression effect is useful in experiment when one want to suppress the modulation
frequencies of protons to find out the modulation due to another nucleus, say 1“N.
This can be done by setting 1: proportional to the period of oscillation of a proton
nucleus at a given external magnetic field intensity.

One drawback from three-pulse ESEEM is that the sequence can generate
several two-pulse echoes which may overlap with the three-pulse echo resulting in
distortion in the ESEEM pattern. To eliminate the effect of overlapping echoes, a
phase cycling technique was introduced. The contribution of these unwanted echoes

is removed by cycling the phase of the mw pulses to produce alternating negative and

positive two-pulse echoes while the three-pulse echo remains always positive.28
Although it may appear that three-pulse is better than two-pulse experiment,
the use of a particular echo sequence will depend on what one wants to obtain. For
example, water ligation in a metal center will be determined better with two-pulse
ESEEM because information about the coupling of water proton to an electron is
contained in the sum combination peak. Also, two-pulse is more reliable in D20-
exchange experiment where division of modulation functions consisting of product of

modulations functions of individual nuclei coupled to the electron, is involved. For

50

the accurate determination of the 1“N hyperfine frequencies, three-pulse ESEEM has
the advantage over two-pulse because the three-pulse ESEEM gives better resolution
of peaks than the two-pulse ESEEM. In the work that I am going to present, two-
pulse ESEEM was used to characterize water ligation to copper center and three-pulse
was used to determine histidyl-imidazole ligation. The copper system studied will be

discussed in the next chapter.

51

List of References

1)Orton, J. W. Electron Paramagnetic Resonance An Introduction To Transition
Group Ions in Crystals; Iliffe Books Ltd: London, 1968.

2)Abragram, A.; Bleaney, B. Electron Paramagnetic Resonance of Transition Ions;
Dover Publications, Inc.: Mineola, New York, 1970.

3)Atherton, N. M. Principles of Electron Spin Resonance; Ellis Horwood & PTR
Prentice Hall: New York, 1993.

4)Wertz, J. E.; Bolton, J. R. Electron Spin Resonance: Elementary Theory and
Practical Applications; 2 ed.; Chapman and Hall: New York, 1986.

5)Gordy, W. Theory and Applications of Electron Spin Resonance; John Wiley &
Sons, Inc.: New York, 1980; Vol. 15.

6)Poole, C. P. Electron Spin Resonance: A Comprehensive Treatise on Experimental
Techniques; 2 ed.; John Wiley & Sons, Inc.: New York, 1983.

7)Poole, C. P.; Farach, H. A. Theory of Magnetic Resonance; 2 ed.; John Wiley &
Sons, Inc.: New York, 1987.

8)McMillan, J. A. Electron Paramagnetism; Reinhold Book Corporation: New York,
1968.

9)Mims, W. B. Phys. Rev. B 1972, 5, 2409-2419.

10)Norris, J. R.; Thumauer, M. C.; Bowman, M. K. Electron Spin Echo Spectroscopy
and the Study of Biological Structure and Function; in Advances in Biological and
Medical Physics, Lawrence, J. H., Gofman, J. W. and Hayes, T. L. ed.; Academic

Press, Inc.: New York, 1980; Vol. 17, pp 365-416.

52

11)McCracken, J. Electron Spin Echo Envelope Modulation; in Handbook of Electron
Spin Resonance, Poole, C. P. and Farach, H. A. ed.; Springer-Verlag: New York,
1999; Vol. 2, pp 69-84.

12)Kevan, L. Modulation of Electron Spin-Echo Decay in Solids; in Time Domain
Electron Spin Resonance, Kevan, L. and Schwartz, R. N. ed.; John Wiley & Sons,
Inc.: New York, 1979, pp 280-341.

13)Kevan, L. Development in Electron Spin-Echo Modulation Analysis; in Modem
Pulsed and Continuous Electron Spin Resonance, Kevan, L. and Bowman, M. K. ed.;
John Wiley & Sons, Inc.: New York, 1990, pp 231-266.

14)Schweiger, A. New Trends In Pulsed Electron Spin Resonance; in Modern Pulsed
and Continous Electron Spin Resonance, Kevan, L. and Bowman, M. ed.; John Wiley
& Sons, Inc.: New York, 1990, pp 43-118.

15)Keijers, C. P.; Reijerse, E. J .; Schmidt, J. Pulsed EPR: A New Field of
Applications; North-Holland: Amsterdam, 1989.

16)Liao, P. F.; Hartman, S. R. Phys. Rev. B 1973, 8, 69-80.

l7)Hahn, E. L. Phys. Rev. 1950, 80, 580-594.

18)Blume, R. J. Phys. Rev. 1958, 109, 1867-1873.

19)Blum, K. Density Matrix Theory and Application; Plenum Press: New York, 1996.
20)Cohen-Tannoudji, C.; Diu, B.; Laloe, F. Quantum Mechanics; Herman: Paris,
1977; Vol. 1.

21)Feynman, R. P. Statistical Mechanics: A Set of Lecture Notes; W.A. Benjamin:
Reading, Massachusetts, 1972.

22)Slichter, C. P. Principle of Magnetic Resonance; Springer-Verlag: Berlin, 1978.

53

23)Mims, W. B. Phys. Rev. B 1972, 6, 3543-3545.
24)Dikanov, S. A.; Shubin, A. A.; Parrnon, V. N. J. Magn. Reson. 1981, 42, 474-487.
25)Rowan, L. G.; Hahn, E. L.; Mims, W. B. Phys. Rev. A 1965, I37, A61-A71.
26)De Groot, A.; Evelo, R.; Hoff, A. J. J. Magn. Reson. 1986, 66, 331-343.
27)Blumberg, W. E.; Mims, W. B.; Zuckerman, D. Rev. Sci. Instrum. 1973, 44, 546-
555.

28)Fauth, J. M.; Schweiger, A.; Braunschweiler, L.; Forrer; Ernst, R. R. J. Magn.

Reson. 1986, 66, 74-85.

54

Chapter 2

INTRODUCTION TO FET3 PROTEIN , A MULTICOPPER OXIDASE
FROM YEAST SACCHAROMYCES CEREVISIAE

Copper-containing enzymes and proteins are ubiquitously distributed among
animals and plants. The copper ions of these proteins maybe involved in electron
transfer, participate directly in the catalytic process or may act only as internal
cofactors. Copper enzymes involved in the activation of molecular oxygen are
classified as oxygenases or oxidases. Oxygenases activate molecular oxygen for
incorporation into organic substrates. If one atom of the dioxygen is incorporated it is
classified as a monooxygenase, while dioxygenases catalyzes incorporation of both
the two atoms of molecular oxygen into the substrate. Oxidases, on the other hand,

couple one-, two- or four- electron oxidation of substrates to one-, two- or four-
electron reduction of dioxygen to hydrogen peroxide or water.1

The Fet3 protein,2’3 an extracellular membrane-bound enzyme found in yeast,
is an oxidase. It is a multicopper oxidase meaning it has multiple copper active sites
that are structurally and spectroscopically different from one another. It plays role in
iron homeostasis in yeast and has been studied to gain insight into the iron regulation
and metabolism of higher eukaryotes.

This chapter presents an introduction to the Fet3 enzyme. The coordination
chemistry of copper of copper, its EPR and absorption properties, and its redox

properties will be discussed. The second part of this chapter will focus on

55

multicopper oxidases, with emphasis to Fet3 and ceruloplasmin, both of which

exhibit ferroxidase activity.

2.1 Coordination of Cu(I) and Cu(II)

Regardless of whether copper ions are directly involved in the catalytic
process, they change redox states during reaction with a substrate as electrons are
transferred between the copper ion and the substrate or between copper ion and other
redox partner. The predominant oxidation states of copper in biological systems are
the +2 and +1 oxidation states. A neutral copper has a valence-shell configuration of
3d'°4s'. Removal of one electron will result in Cu"1 with 3d10 electron configuration.
Because the electrons are paired in the d orbitals, Cu“1 is diamagnetic. Removal of
another electron from Cu+1 will yield the paramagnetic Cu+2 (3d9) species. Although
further oxidation to Cu+3 is possible, Cu+3 compounds are short lived.

The oxidation state of copper determines the stereocherrristry of the copper
complex. Table 2.1 lists the common coordinations and stereochemistries of Cu"l and

Cu+2 complexes. Normally, Cu”-ligands favor tetrahedral coordination, while those

of Cu+2 favor a square planar bonding geometry.45

In biological systems, the ligands of Cu+1 and Cu+2 can be classified as
endogenous and exogenous ligands. The arrrino acids that bind to the copper ions are
called endogenous ligands. Copper binds to these ligands through the nitrogen,
oxygen and sulfur atoms which may be located at the N- or C-terrrrini of the

polypeptide chain, within the chain or in the side chains. Ligands not derived from

56

proteins are considered exogenous ligands. Water, hydroxide, oxide and sulfide are

examples of these ligands. 1’5

Table 2.1. Coordination and stereocherrristries of Cu"1 and Cu”.

 

 

 

 

 

 

 

 

 

 

 

common cln d shell common effective spin ionic
oxidation Z,ff coord. no. geometry State radii
(s) (A)
2 linear 0 0.46
+1 (110 7.85 3 trigonal 0
4 T,* 0 0.60
6 0h 0 0.77
4 Td 1/2 0.57
+2 d9 8.20 D,* 1/2
5 C, 1/2 0.65
D,,, 1/2
6 O,* 1/2 0.73
*favored conformation

The concept of "hard and soft acids and bases" (HSAB)5 is useful in assigning
ligands that may coordinate to a metal center to form stable complex. In this concept,
small metal ions and acids that are not easily polarized are classified as "hard" while
large and easily polarized metal ions or acids are classified as "soft". Ligands with
highly electronegative donor atom are hard bases while polarizable ligands are soft.
Generally, stable complexes are either formed by hard acids (metal ions) and hard
bases (ligands) or soft acids(metal ions) and soft bases (ligands). HSAB can be

applied to copper ligands as well. Thus, we expect that in a copper-protein complex,

57

 

the soft ion Cu+1 will preferably coordinate ligands having soft donor atoms such as
Cys, Cys’ and Met. In the same logic, the harder Cu“2 will complex with water Tyr,
Tyr‘, Ser, Thr or His which have 'hard' donor atoms, 0 and N.

Cu+3 is the hardest among copper oxidation states. It will only form
complexes with hard ligands such as Tyr, Tyr’ and OH‘ , because soft ligands will
reduce Cu”. Although arrrino acid ligands that form stable complexes with Cu+3 are
available, Cu“3 has yet to be found in a protein or enzyme. One reason is that, redox
reactions involving Cu+3 active site require extremely high reduction potential. Such
high magnitude of potential has not yet been needed biologically. The simple reason,
however, could be that while Tyr, OH’, and Tyr‘ are hard enough to stabilize Cu”,

such complexes may not be stable enough to provide part of the active site of an

enzyme.4

2.2 Classification of Copper-Binding Sites
Based on spectroscopic data obtained from copper in its (19 (i.e. Cu”)

electronic configuration, copper binding sites are traditionally classified as type -1,

type-2 and type-3 sites.6'9 These spectroscopic properties reflect differences in their
electronic and geometric structures.

According to ligand-field theory, when copper and other transition metals are
placed in an octahedral ligand field, the five degenerate 3d-orbitals will be lifted into

two sets of degenerate orbitals - (a) t,, (triply generate) and (b) eg (double generate)

sets.10 The manner of splitting is consequence of the way the lobes of the

different d-orbitals are directed to the ligands. The d“, d,,, (1,, orbitals are directed

58

 

—
o
o
o
o
o
o
o
o
o
o
a ""
o
l .g‘
a
s
‘ e
a s.
e
. s
a
a
e
a
e
o
c
e
o
I
a
a
c
I
o
o
l
o
a
l
o
o
e
l
a
-4
e
e
e
e
e
e
s
e
e
e
e
e
e
9 o
0 '0
9 o
e .0
a _— .o
e
‘ —..‘
e
—— e
e
e
e
e
e
s
e
e

+714 714

tetragonal distortion

free ion octahedral field

 

 

O
'
O
O
O
O
O
O
O
O
0
O
O
'
0
0
.' A
. l
--.*
‘
‘
‘
.
D
D
‘
D
‘
‘
O
O
‘ d
. —
O 0 xy
0‘ g
0
‘ o
‘. ‘
‘ 0
‘ O
. '
‘ C
. l
D
‘C
.. A
o e 2
O “
I. ‘
O ‘
0 Q
0 D
' ‘
. ‘—— 2
' z
0
C
U
0
U
0
O
0
0
0
0
l
'
C
0
U
0
0
...I‘.
.
Q
Q
Q
~
.
.
..
.

 

 

 

 

square-planar field

Figure 2.1. Splitting of a d-orbitals in an octahedral, tetragonal or square-planar

ligand field.

59

between the ligands whereas the lobes of the d,2 and d},2 are directed toward the

ligands. Consequently, the d,2 and d,,2-y2 orbitals are repelled more than the d,,, d,,,

d,,. Following J ahn-Teller theorem,10’11 further splittings will occur as a result of
tetragonal distortion in which distance of the ligand at the z-axis is lengthen and the
ligand at x and y are contracted. In the limit of tetragonal distortion, the ligands at
z-axis are completely withdrawn leaving a square planar arrangement. Distribution
of the valence electrons in Cu"2 among the 3d-orbitals will result in the unpaired
spin residing in the d,2-y2 orbital (Figure 2.1).

The unpaired electron makes Cu+2 amenable to EPR studies. Shown in Figure
2.2 is a typical powder EPR signal of Cu+2 in frozen solution. It is a characterized by
an axial g-tensor with gll > g , > g, = 2.0023. As explained in the first chapter, the
deviation of the g-values form the g, is attributed to spin-orbit coupling. For an axial

Cu+2 system in a square planar ligand field, the g values at the canonical orientations

are,IZ'15
g,=gy=g,+2?t/A,. (2.1b)

where the A's are energy differences as defined in Figure 2.1. The above equations
were derived assurrring an electrostatic interaction, that is the unpaired electron spend
most of its time in the 3d,},2 orbital. Usually, the calculated g values using these

equations are greater than the experimental g-values, indicating that there is too

60

gxa gy

gz

 

 

 

 

 

An

I I I I I I I I I I I I I I I I I I I I I L I I I I I I I I I I I

2400 2600 2800 3000 3200 3400 3600 3800 4000
Field (Gauss)

Figure 2.2. Theoretical CW-EPR spectra of Cu(H)-aquo complex. A) Absorption
spectum, B) first derivative spectrum of A, and C) first derivative spectrum showing
hyperfine interaction in the g" region.

61

much angular momentum added to the ligand field ground state wavefunction. This
deviation arises from the fact that ligand field calculation uses pure (1 orbitals. One
way to correct this discrepancy is to write the ground state function in terms of
molecular orbital of the d orbital and ligand orbital. Covalency or delocalization of
the electron into the ligand reduces the angular momentum associated with a
molecular orbital and therefore calculation should include its effect. A ligand

molecular orbital (MO) treatment for calculating Cu+2 g values are discussed in

several papers.13v16‘18

In the gll region of the EPR spectrum, additional peaks are observed because
of the Cu+2 hyperfine interaction. Cu+2 possesses a nuclear spin I = 3/2, which will
couple to the unpaired electron spin to yield (21 + l) resonances. Often times, the
principal axes of the metal hfi and the g-tensor are coincident so that assignment of
the hyperfine coupling constant is not complicated. Generally, the hyperfine parallel
splittings, A", for normal tetragonal Cu+2 complex are large ranging from 150 to 250
x 10“ cm". The perpendicular hyperfine coupling constants, A ,, are small (<35x10-
4 cm-l) and most often will not be resolved but will appear as broadening in the

powder spectrum. Mathematically, the hyperfine coupling constants are given

 

 

by14,l9
40:2 3A
All ——Pd( -—7—— 7g, —Ag"), (2.2a)
20:2 llAg
A =—Pd x— 1 2.2b
1 ( 7 14 ) ( )

where P = 2Bg,B,<r'3> = 396 x 10" cm", K is the Fermi contact and (12 is the percent

metal character in the d,2-y2 orbital.

62

Cu+2 is also accessible to UVN is absorption studies. Absorption bands
(Figure 2.3) can be assigned as d-d transitions and ligand-to-metal charge transfer

(LMCT) transitions. Because the orbitals are split by the ligand field, the d-d

 

 

 

L 41...
ll

 

Lo

Figure 2.3. Absorption Spectroscopy. d-d and ligand to metal charge transfer
(LMCT) transitions observed for a d9 copper.

transitions are a sensitive probe of geometry of the copper-complex. For example,
tetragonal complexes exhibit d-d transitions in the region from region ~500nm to 1

um. Copper-complexes that are constrained to assume tetrahedral geometry have d-d

63

transition in the range of 1-2 mm. LMCT are high energy transitions. They provide

information on the degree of overlap between the metal and ligand orbital.8’19

Type-1, type-2 and type-3 copper sites have characteristic EPR and

absorption properties.4r6‘9 The deep-blue colored type-1 copper centers show
characteristic absorption bands at approximately 600nm with molar extinction
coefficients, 8, of about 5000M'1 cm". This band is assigned to a cysteine-sulfur to
copper charge transfer transition. The cysteine-copper bond is covalent and is one of
the reasons why the A” of copper (43-95 X 104cm") is smaller compared to those of
tetragonal complexes. Because of the covalent character of this bond, the unpaired
spin is delocalized into cysteine ligand and thus reduces the interaction of the Cu
nuclear spin with that of the unpaired electron spin. Plastocyanin is an example of a
well-studied type-1 copper site. Like other type-1 coppers, it functions as an electron
transfer agent. The metal is characterized by a high positive reduction potential and
ligands which are diposed in space to assume a structure that is a compromise of Cu“1
and Cu*2 stereochemistries. Type-2 copper sites feature tetragonally-distorted
octahedral coordination geometries similar to those typically found for Cu+2 in
solution. They exhibit EPR spectrum that have All in the range from 158 to 201x10’
4cm". Although they exhibit d-d transitions, they are very weak so that they are often
not detected in the optical spectrum These copper sites are found in oxidases,
oxygenases, nitrite reductases and Cu/Zn superoxide dismutase where they are
directly involved in oxygen metabolism. With regards to ligation, type-2 copper
sites are not as uniform as type-1 copper sites which have generally a conserved

2Histidine and 1Cysteine equatorial ligation. One common feature, however, is that

at least one histidine ligand is coordinated and there is one coordination site that is
always free for oxygen binding. Type-3 copper sites are binuclear centers that are
antiferromagnitacally coupled. The two copper ions are covalently linked by
bridging ligand OR‘, allowing the unpaired spins to couple and thus making the
binuclear sites EPR-silent. Type-3 copper sites show a strong absorption band at
about 330nm with e~3000-5000M"cm'l and is believed to be due to ligand(OR')-
metal charge transfer transition. These copper sites can be found in hemocyanin and
tyrosinase.

There are other copper-binding centers that do not belong to the above

classifications. Among these are the trinuclear sites comprised of a single type-2 and
single type-3 copper centenzo‘26 the binuclear CuA found in cytochrome oxidase27'
30 and nitrous oxide reductase,30a31 CuB-heme in cytochrome C32 and recently the

CuZ in nitrous oxide reductase33r34. Except for the trinuclear copper sites, these
copper centers will not be discussed in this dissertation. T rinuclear copper centers

will be discussed in the multicopper oxidase section of this chapter.

2.3 Redox Behavior of Copper Complexes

As mentioned previously, Cu+1 ligand complexes will favor a tetrahedral
conformation, while those of Cu+2 will favor a square planar conformation.
However, in the copper-binding sites of proteins, ligands and conformations may
differ greatly from those normally preferred. To be a relevant component in
biological catalysis, the copper-binding centers must be able to change oxidation

states with relative ease thermodynamically. A compromise between ligands and

65

conformations of Cu+1 and Cu+2 must occur to optimize the redox potentials for
specific processes. This is particularly true for type 1 copper whose primary function
is long-range electron transfer.

In aqueous solution, the redox potential of the redox couple Cu"2/Cu“l is E°,.,

= +153 mV35. This is in contrast to the relatively high positive reduction potentials,

E°,,d ,that are found in copper proteins complexes specially in type 1 copper which

varies from 183 mV for stellacyanin to 765 mV in type 1 copper site laccase.4 The
relatively high reduction potentials of copper sites means that Cu+1 is stabilized more
than the Cu+2 species. Ligand-binding interactions, hydrophobicity of environment

and solvent accessibility of the metal site are variables that affect the reduction

potential of the Cu“2/Cu+1 protein complex.36'48
Studies on model copper complexes were made to determine the effect of
replacing water ligands with ligands having N and S donors on the E°,,, of the

Can/Cu“1 half-redox reaction. Yandell et 81. were able to correlate the potentials

observed for the copper complex models to the equation”,

13°... = EOCu+2lCu+l+ z(nr AEL) (2.3)
where E°C,+,,C,,, is the standard reduction potential of the aqueous Cu“2/Cu+l couple,
AEL is the variation in redox potential introduced by a donor atom in ligand L which
displaces the water ligand and nL is the number of occurrences of ligand L.
Replacement of water by thioether sulfur substantially increases E°,,,_ The effect of
nitrogen donor is dependent whether they are aromatic or saturated.

The disposition of the ligands which define the geometry of the complex also

affects the reduction potential of the copper complex. Tripodal ligands that force the

66

copper(II) complex to assume a distorted tetrahedral geometry increase the reduction

potential of the Cu*2/Cu+1 couple appreciably.39 We may expect therefore that for
the same set of ligands, higher reduction potentials for complexes with almost
tetrahedral structure than for those complexes whose ligands are distributed in square-

planar arrangement will be observed. A study on the effect of the geometry on redox

potential was done by Addison and Yokoi.38

The observed variation of E°,,d in model complexes can be translated to copper
proteins. The E°,,, of a protein Cu*2/Cu+1 couple can be increased by introducing rt-
acceptor or "soft" ligands such as the thioether (Met) and thiolate(Cys) which favor
Cu+1 to the copper binding sites. Therefore some copper bindings sites are ligated
with soft ligands that favor Cu+1 and hard ligands that favor Cu”. A positive E°,,, can
also be achieved by destabilization of the Cu*2 ion and this can be done by
constraining the geometry of the metal center to those of low LFSE which usually
favor Cu“. A typical example of this is the small type 1 copper plastocyanin in
which the structure of the oxidized form is a distorted tetrahedron. Plastocyanin has
the typical conserved ligation HiSZCys comprising the equatorial ligands. An axial
Met ligand which is ~3.0 A from the copper center is also present. From electronic
structure studies of both Cu‘“l and Cu+2 species by Solomon et a1, it was suggested
that the elongated axial Cu-Met bond contribute to the high-reduction potential of the
plastocyanin. This elongated bond allows for a small change in the geometry as

copper changes oxidation states and consequently a low Frank-Condon barrier or

reorganizational energy suitable for fast electron transfer.45r46

67

In addition to choice of ligands and the geometry imposed by the ligands on
the metal center, the reduction potential of the copper protein couple is also
dependent on the dielectric constant of the protein matrix and/or solvent accessibility
of the copper site. Because of low charge of Cu+1 complex, a hydrophobic
environment will favor Cu“l with respect to the Cu+2 species. This will result in the
more positive reduction potential as compared to those observed in aqueous solution.
Access of solvent water to the metal center can be limited by the choice of ligands.
Amino acids with bulky hydrophobic side chains will inhibit entry of water to the
metal site and will result in the stabilization of Cu“1 with respect to Cu”.

Experimentally, these ideas have been proven using systems with type 1 copper

proteins.44

A theory to account the effect of solvation of metal-protein complex was

developed by Churg and Warshel.41 Their explanation is based the on the
thermochemical concept that the standard free-energy change, AG°red when adding
one electron to an oxidized molecule A°", to give the reduced form, A”, is related
to E°,,, through the well-known equation

AG°,,,= -JfE°,,, + constant. ( 2.4)
f in the above equation is the Faraday constant. The above equation can be used to
determine the environmental effects of different solvents upon the reduction of the of
some molecule A°". The relation of solvation energy and change in redox potential
is given by,

-JfE°,,, = A[AG°,,,,(A°") -AG°,,,,(A’°°) = AAG,,,, = AAG°,,,. (2.5)

AAG,,,v is approximated to be equal to the sum of (i) the interaction of complex

68

charge and protein permanent dipoles, (ii) the interaction of complex charge and the
induced dipoles of associated with polarizable atom, (iii) solvation energy of the
complex by the surrounding water molecules and the (iv) interaction of complex with
ionizable groups. This mathematical approach has been applied to explain the redox
variation of redox potential of cytochrome c and the octapeptide-methionine/histidine
complex formed by the hydrolysis of cytochrome.

To sum up the points discussed in this section, metalloproteins (i.e.
copper protein) tune their redox potentials for specific reactions by the appropriate
choice of ligand, a compromise between the structure of oxidized and reduced forms

of the metalloproteins and control of solvent access to the metal center.

2.4 Multicopper Oxidases

Multicopper oxidases are class of enzymes that contain type-1, type-2 and

type-3 copper centers.8’9 These blue oxidases couple the four-electron reduction of
dioxygen to water to the four-electron oxidation of a substrate. Among the well-

characterized multicopper enzymes are laccase (Lc), ascorbate oxidase(AO) and

ceruloplasmin (Cp). Recently, an enzyme from yeast, Fet32»3, has been identified as
belonging to this group. All of these enzymes oxidize their substrate by one electron
processes.

Laccases are widely distributed in plants and fungi where they catalyze the
oxidation of numerous aromatic phenols, amino phenols and diarnines via complete

reduction of dioxygen to water. By far the most studied plant laccase is the Rhus

vernicifera. It was discovered by Yoshida in 18839. Plant laccases are believed to be

69

involved in lignin formation and wound healing. Fungal laccases have been isolated
and purified from various sources such as basidiomycetes, ascomycetes and
deueromycetes. They are postulated to be involved in sporolation, pigment
formation, lignin degradation, and pathogenesis.

Ascorbate oxidase occurs exclusively in higher plants. The two most common
sources of these enzymes are cucumber and green zucchini squash. It is thought the
enzyme play a role in ripening, growth control and disease control.

Ceruloplasmin is an enzyme that is found in the plasma of mammals and

birds. At least four functions have been attributed to ceruloplasmin, those of copper
transport and, ferroxidase activity49, amine oxidase activity and as an

antioxidants),50 A wide range of organic and inorganic compounds may be utilized

as substrates. However, biological and kinetic data indicate that the physiologically

Table 2.2. Distribution of Copper in Some Multicopper Oxidases.

 

 

 

 

 

Enzyme Type 1 Type 2 Type 3 Reference
Laccase l 1 1 8,71
Ascorbate 2 2 2 24,25

Oxidase
Human 3 l 1 26
Ceruloplasmin
Fet3 1 1 1 57,59

 

 

 

 

 

 

 

70

relevant substrate of ceruloplasmin is Fe (H). Plasma ferroxidase activity is essential

to iron homeostasis in human.51»52

1.5 Fet3 as a multicopper oxidase

As a multicopper oxidase, Fet3 oxidizes four moles of iron and
concomitantly reduces molecular oxygen to water. The role of Fet3 as a multicopper
oxidase is supported by several studies. Sequence analysis revealed an open reading
frame with homology to other multicopper oxidases such as laccase, ascorbate
oxidase and ceruloplasmin. Homology is particularly strong in the regions that are
involved in the binding of the catalytic Cu“2 prosthetic groups. The copper catalytic
domain is exposed on the extracellular side of the plasma membrane and is attached
to the plasma membrane by a single transmembrane unit. This configuration implies
that another molecule is needed to transport iron across the cell membrane. Ftrl with

multiple transmembrane domains and a potential ferric binding motif is the apparent

transporter which is believed to form complex with Fet3.53'55 Enzymatic activity

studies also supported Fet3 as a multicopper oxidase and confirmed its ferroxidase
activity.3r56’57 Experiments have shown that the best substrate for Fet3 is the

ferrous ion.55 Yeast cells that were grown in copper-deficient media, had profound

deficiency in iron transport which was directly related to the absence of ferrous

activity3’58. Addition of copper restored iron transport activity. This implies that

copper loading of apo-Fet3 is essential for multicopper activity. Studies on Fe(II)-

dependent O2 consumption3 by yeast cells provided evidence of Fet3’s ferroxidase

activity in that an approximate 4:1 ratio of the rate of Fe(II) oxidation to O2

71

consumption was revealed. Its relevance to high-affinity transport is supported by the
absence or reduction of O2 consumption in cells with mutated Fet3, cells grown in
high iron-media and cells grown in copper-free media. 02 consumption using cells
stripped of cell walls also indicated that the ferroxidase domain of Fet3 is located on
the extracellular surface because removal of the yeast cell wall did not hinder iron-
induced increase in 02 consumption but addition of trypsin did.

In addition to sequence analysis and catalytic activity studies, support for the

multicopper sites of Fet3 comes from the characteristic UV/VIS absorption bands

and CW-EPR spectra of wild-type Fet3.57’59 Four copper cofactors have been
determined based on the atomic absorption spectroscopy analysis. Spectroscopically,
these copper sites can be differentiated as type-l, type-2 and a binuclear cluster, type-
3. Type-1 and type-2 are paramagnetic whereas type-3 site is EPR-silent. This is
evident in the overlap of EPR signals coming from type-1 and type -2 copper sites.
The type-3 copper site is EPR inactive because a hydroxide (OH’) bridges the copper.
Absorption spectroscopy was used to detect the presence of a type-3 copper in Fet3.
The absorption band at 300 nm is assigned to OH’ to metal charge-transfer transition.
Another intense absorption band is observed at 600 nm and assigned to type -1 copper
site. As expected, no detectable UV-Vis absorption due to type-2 copper site was

observed.
2.6 Tri-nuclear cluster

Type-2 and type-3 copper sites in multicopper oxidases are believed to form a

trinuclear site. This trinuclear copper site is considered to be the binding and

72

reduction site for molecular oxygen. Support for the existence of a trinuclear site
includes the azide binding and oxygen reactivity studies on laccase, the crystal
structure of AC from Zucchini and the kinetic and crystallography studies on human
ceruloplasnrin (th).20'26

The copper sites in plant laccase from Rhus Vemicifera have been extensively
studied by Prof. Solomon's group at Stanford University. With the help of laccase

variants, they were able to suggest that type-2 and type-3 Cu sites form a trinuclear

cluster necessary for oxygen binding.22’64’65 The type-2 copper can be selectively
chelated to produce a type-2 depleted (T2D) laccase variant containing only type 1
and type 3 copper centers. Type-1 copper can be selectively replaced by Hg*2(T1Hg)
leaving the type-2 and type-3 copper sites intact. The use of these variants avoids
difficulty in spectral analysis as the contribution of either one of the paramagnetic
copper sites is eliminated. As we will find later, similar technique was used in the
spectral analysis of Fet3 by EPR.

Azide binding studies in both native laccase and Tng laccase utilized low-
temperature magnetic circular dichroism (LTMCD), absorption and EPR
spectroscopies. Results of these studies indicated that azide bridges the type 2 and
the type 3 copper suggesting that type 2 copper is near to type 3, (<5.2 A) and
therefore capable of forming a trinuclear cluster.20’21r23 A combination of these
spectroscopic techniques allowed ligand field (d-d) and charge-transfer (CT ) spectral
features associated with the paramagnetic type-2 center to be differentiated from

those of the antiferromagnetically coupled type-3 center.

73

Further proof of a trinuclear site came from the study dioxygen reactivity
using X-ray absorption edge spectroscopy.22 Qualitatively, a preedge ls-4p
transition, assigned empirically from polarized single-crystal studies by Smith et
al.66, near 8984 eV is present in the X-ray edge absorption spectra of Cu(I) but is
absent in Cu(II) spectra. Thus, the 8984 eV feature is useful in probing the reduced
copper site and its coordination environment. Exposure of T2D with oxidized type 1
and reduced type 3 copper sites to air, did not result into the substantial change in
the appearance of the 8984 eV feature. Similar results were obtained for the T2D
laccase where both the type-l and type-3 copper sites were reduced. This means that
none of the reduced copper sites were oxidized and therefore no reaction with 02
occurred. However, exposure of fully reduced Tng laccase to air resulted in the
total lost of the 8984 eV transition. The results implies that the trinuclear cluster
represents the minimum structural unit for the reduction of molecular oxygen.

The existence of a type2-type3 trinuclear cluster is also supported by the x-ray

crystallographic studies on ascorbate oxidase by Albrecht Messerschimdt et al.24,25
The crystal structure resolved at 1.9 A of AC from Zucchini revealed that the
multicopper oxidase is a homodimeric enzyme with subunits weighing 70 kD and a
chain length of 552 arrrino acids. Each of the subunits or each monomer, like laccase,
consists of three domains and contains four copper atoms. These domains are
characterized by B-strand. Domain 1 is made up of two four-stranded B-sheets
forming a B-sandwich structure. Six-stranded and a five-stranded B—sheet comprise
domain 2. The final domain 3 is shaped by two five-stranded sheets forming a B-

barrel and a four-stranded B -sheet. The type 1 copper is located in domain 3 and the

74

trinuclear copper center is situated between domains 1 and 3. The binding site for the
reductant is on the surface of the enzyme, near the type 1 copper center. The oxygen
binding site, i.e. the trinuclear copper center, is embedded inside the enzyme and
accessible by two channels of different sizes.

The mononuclear type- 1 copper in domain 3 is coordinated by His445(N5),

Hi8512(N5), Cy8507(S), and Met517(S). These ligands are the canonical ligands that

are found the mononuclear type-1 copper sites of plastocyanin“:68 and azurin69.
Like these two type-1 copper centers, the methionine forms a distant
noncoordinating axial ligand. The trinuclear copper site, consisting formally of type
2 and type 3 copper centers, requires eight histidines, one OH‘ or (02') and a second
OH‘ (or H20) coordination. Six of the eight histidine ligands coordinate to both
copper ions of the type 3 copper center to form a trigonal prismatic structure. The
two copper centers of the type-3 site are inequivalent with respect to histidine
nitrogen ligation. Histidine ligands of the first copper ion include I-Ii8102(N8),
His450(N8) and HisSO6(Nt-:) and those of the second copper ion are His62(N8),
Hile4(Ne) and Hi8508(Ne). Additionally, a OH‘ (or 02‘) bridge links these two
copper atoms resulting in a strong antiferromagnetic coupling. The type 2 copper ion
is coordinated by two histidine residues, His60(N2) and His448(N2) and possibly by
hydroxide or water. These coordinations are shown in Figure 2.4.

In 1996, a 3.1 A resolution X-ray crystal structure of human serum

ceruloplasmin was published and confirmed the presence of a trinuclear cluster in the

enzyme.26 The trinuclear cluster of type-2 and type3- copper sites may therefore be

75

A) Ascorbate Oxidase

<\
/1

\/

B) Human Ceruloplasmin

 

 

  

 

 
   
 

‘- ‘ Type 1
.v- " His445, Cy3507
HisSlZ, Met517
Type 3 .
His106, His450
HI8506, OH ’00 12.7A
3.9.1 "
Type 3
, . 3,7,; His62,Hile4
H3508, OH
Type 2
His60,His448, OH

I

 

   

Type 1
17.9,, His637, Cys680
/ [ H3685, Met690
Type 1 ‘ i
No Methionine
His276, Cys319,
His324 17.8A
17.5.1
hfi- Type 1
His975, Cys1021
12.5.1 . HileZ6, Met1031
Type 3
Hisl63, 1115890
Hisrozo, 13'“
4.0A
we 3
_, 3.4;, His103,His161
Type 2 Hrsrozz

Hilel,His978

Figure 2.4. Comparison of spatial arrangement of the different copper sites in
ascorbate oxidase(AO) and human ceruloplasmin (th).

76

present to all multicopper oxidases. The crystal structure reveals that the molecule is
comprised of six plastocyanin-type beta barrel domains arranged in a triangular array
(Figure 2.4). It has about 1046 arrrino acids. There are six copper atoms; three form a
trinuclear cluster situated at the interface of domains 1 and 6 and there are three
mononuclear sites in domains 2,4 and 6. Each of the mononuclear coppers is
coordinated to a cysteine and two histidine residues and those in domains 4 and 6
also coordinate to a methionine; in domain 2 the methionine is replaced by a leucine
residue which may form a van der Waals type contact with the copper. The latter

type-1 copper was found to be permanently reduced and therefore not relevant to

catalysis.7O

2.7 Mechanism of 02 Reduction

Figure 2.4 provides the schematic representations of the structure of th and
A0. The spatial arrangement of the coppers sites domain 1 and 6 of the th bears
resemblance to the arrangement of the copper sites in the monomeric unit of AO.
This leads to the idea that the three copper sites in laccase and Fet3 are geometrically

arranged in a similar fashion to AD. In fact, the X-ray crystal structure of type-2

copper depleted laccase from Coprinus cinerus71 confirmed that its type-1 site is
approximately 13 A from the type-3 center.

With resolution of the structure of A0, a mechanism for the catalytic activity
of the multicopper oxidase has been proposed. The widely accepted mechanism
involves the type-1 site as the primary electron acceptor, abstracting an electron from

the reducing substrate and transferring it to the trinuclear site. One relevant structural

77

feature is that type-1 copper cystiene ligand is placed in between the two histidines
that are ligand to type-3 site. This together with the covalent character of the Cys-

(type-l copper) bond, as determined experimentally and theoretically from the studies

on plastocyanin8’46’72, provide an efficient pathway for the delivery of the electron

from type-1 to the trinuclear center (Figure 2.5).

(His)N

T 3 -
N(His) ______________ if: (H's)

(His)
Type 1 / / Type 2
L—@—§(Cys) 11 @ on2
\ (His)
N(His)

N(His)

 

(His)N

Figure 2.5. A proposed pathway for the transfer of electron from type-l copper
center to type 2-type3 trinuclear center in multicopper oxidases based on the X-
crystal structure of ascorbate oxidase from zucchini. The X-ray structure showed the
S(Cys) ligand is flanked by two histidines ligands from the two copper ions of the
type-3 copper binunclear site. Theoretical calculations predicted the large overlap
between the Cu (1,2,y2 and S(Cys) p1: orbitals. These two findings provide an efficient
pathway for the transfer of electron from the type 1 to the trinuclear copper site.

78

Binding to oxygen requires all of the copper sites in the enzyme to be fully reduced.
This can be achieved by four single-electron reductions of type 1 copper with
substrate and transfer of three electrons to the trinuclear copper site. The formation
of water upon reaction of dioxygen with the fully reduced substrate may involve

formation of an oxygen intermediate, hydro-peroxide or oxygen radical.

2.8 Regulation of Iron Transport in Yeast

Transition metals such iron, copper and zinc play a vital role in biochemical
processes in eukaryotes. They stabilize protein structures and act as cofactors in
catalysis. Because of their hi gh-redox potentials, they are important components of
biochemical redox reactions. Their redox properties, however, also make them
potentially harmful because they are capable of generating toxic free radicals if the
concentration of these metals in the cell is not tightly regulated. Therefore, a balance
between the amount of metal essential for biochemical processes and the amount that

is potentially toxic must be maintained. To meet the required concentration, cells

utilize a specialized mechanism, controlled transmembrane metal transport.73 74'76
The yeast Saccharomyces cerevisiae provides an excellent model for the
study of metal homeostasis in eukaryotes because its genetics can be easily

manipulated. This simple eukaryote employs two transmembrane transport systems,

one of low- and the other of high-affinity (Figure 2.6).73‘76 The high-affinity
transport system is very specific to its metal target and is regulated according to metal

needs, whereas the low-affinity system is less specific for the metal it transports and

79

 

    

Plasma membrane

\

 

 

 

 

 

TH:
Fe: FET3/FIR]
Cu: CTRl/CTR3
Zn: ZRTl

Mm?

   

 

 

 

 

 

 

 

TA (Transcriptional Activators)

Fe: AFI’l
Cu: MAC]
Zn: ZAP!
Mm?

 

 

 

Figure 2.6. Model for the regulation of metal transport in yeast (adopted from
Radisky et al., J. Biol. Chem, 1999, 2_'7_4, 4481-4484). High affinity metal
transporters (T,,) are selective for the target metal and are transcribed according to
metal need by transcriptional activators (TA) which release DNA upon metal binding.
Low affinity transporters (TL) are less responsive to metal need and are less metal
selective. Although Fet4 was identified as low-affinity transporters, it can transport
other metals. ORF, open reading frame.

80

and less responsive to metal requirement. These two transport systems provide for
metal homeostasis.

In an iron-replete condition, the low-affinity transport system, (Michaelis-

Menten constant, Km = 30 11M) mediated by the product of Fet4 gene, 77,78 is
activated for the transport of iron across the cellular membrane. Fet4, an integral
membrane protein localized in the plasma membrane, is a Fe(II) transporter. The low
affinity-transport system for iron can also acquire other metals such as manganese
and cadmium. Under conditions where iron is limiting, high-affinity iron transport

(K, = 0.15 11M) takes place. An essential component of this transport system is Fet3.

2.9 High-affinity Transport of Iron in Yeast
Two major proteins are involved in the high-affinity transport of iron (Figure
2.7). These are the Fet3 enzyme, an integral membrane protein with an extracellular

multicopper domain and Ftrl (Fe t_ransport) protein, a transmembrane iron perrnease.

Genetics studies50v53’54 have suggested that these proteins work in concert and that
simultaneous expression of these proteins is necessary and sufficient for high-affinity
iron transport. Fet3 enzyme acts as a ferroxidase by sequentially converting four
moles of Fe+2 into Fe”. The four moles of electrons that are released in the process
are transferred to molecular oxygen, thereby reducing it to water. The oxidized iron,
then, enters into the cell, where it may then be utilized for cellular processes, via the

transmembrane permease, Ftrl. Ftrl has six potential transmembrane units and an

iron-binding motif similar to that found in mammalian ferritin.53 Similar to low-

affinity iron transport, the activity of the cell surface ferrireductases, Frel and

81

4Fe(II)

v Fe(III)

‘

4H+ + 02 @‘/ i
> \ @ 4e@ y ie

\
3,0 @ mean)

 

 

 

 

k.

 

    

 

 

 

Plasma membrane

\

Out ""

 

 

Fre1IFr02

In

 

l 4FeIII)

Figure 2.7. High-affinity uptake of iron across the cell of yeast. The essential
components of the transport system include Fet3, integral membrane exhibiting a
ferroxidase catalytic activity, Ftrl the iron transporter and Frel and Fre2, which
generate the substrate, Fe(H) for Fet3. The ferroxidase activity of Fet3 is dependent
on the three copper centers (one type 1, one type 2 and one type 3) that are localized
outside the plasma membrane protein.

82

Fr62,79'82 is also required in the hi gh-affinity iron uptake because they convert iron
into a soluble form, Fe+2 and hence make iron available for transport. The

components of the high-affinity transport system are transcriptionally regulated by the
product of AFT 1 gene76, an iron responsive DNA-binding protein. Transport of iron

via the high-affinity system occurs only under aerobic condition.83 In an anaerobic

environment, the low-affinity transport is Operative.

2.10 The Iron-Copper Connection

The ferroxidase activity of the Fet3 protein depends on the presence of the
four copper ions. Any defects in the genes and enzymes that are involved in the
delivery and loading of copper to apo-Fet3 result in defects in iron uptake. These
copper delivery and loading genes/enzymes were identified in the process of
searching for the genes involved in iron transport in yeast .

The CT R1 gene encodes the enzyme required for high-affinity transport of

copper in yeast. A plasma-membrane protein, Ctrl is responsible for the trafficking

and delivery of copper to various cellular targets.58984 One of the cellular
destinations is the mitochondria where it is needed to activate cytochrome oxidase.
Loading of cytochrome oxidase with copper is mediated by intracellular transporter
protein Cox1785. C0pper can also be loaded via Lys786 to the cystolic copper
protein, superoxide dismutase. Furthermore, copper ions are also delivered to

secretory pathways and one of the targets for this mode of delivery is Fet3. Loading

of copper to Fet3 is mediated by a cystolic enzyme Cc0287. The CCC2 gene is

83

plasma mebrane

 

 

 

 

 

Figure 2.8. Scheme for the delivery and loading of copper to apo Fet3. Two
pathways are proposed. The primary route is Ctrlp-Atxlp-Ccc2p—Fet3p. The
secondary route involves Ctrlp-Ccc2-Fet3p.

homologous to the human Wilson and Menkes disease genes. Both the yeast and
human genes are believed to function in the translocation of copper across

intracellular membranes into the secretory pathways. In another pathway, copper can

be loaded to apo-Fet3 via the Ctrl-Atxl-Ccc2 route.88 The ATXl gene is believed
to the cystolic carrier of copper specific for the Ccc2 protein. Therefore, copper can
be delivered to Fet3 in two pathways as shown in Figure 2.8. Any defects in the

genes involved in the delivery and loading of copper to apo-Fet3 will abrogate the

ferroxidase activity of Fet3 and consequently the iron uptake will cease.58,83’88

84

2.11 Ferroxidase Activity Ceruloplasmin
The other multicopper oxidase that exhibits a ferroxidase activity analogous
to Fet3 is the serum enzyme from animals, ceruloplasmin. This enzyme is

synthesized in the liver and secreted into the serum with six copper atoms, required

for the ferroxidase activity, during biosynthesis.50’51’89 Figure 2.9 shows the
involvement of ceruloplasrrrin in iron mobilization by acting as a ferroxidase,
converting ferrous ion to ferric ion. The substrate ferrous ion is released from the cell

through a yet to be known enzyme. Two moles of the oxidized iron binds with the

transfenin (Tf), an 80-kDa glycoprotein that circulates in vertebrate blood.89 To
transport iron into the cell, the diferric Tf then binds with membrane glycoprotein ,
transferrin receptor (TfR) . The Tf—TfR complex enters the cell via endocytosis. At
the right pH, the ferric ion is released from the complex in the acidified endosome,
and transferred to the cell cytoplasm. The iron in the form of Fe+3 is then stored in

fenitins (Fts). From Ft, iron can be transported to other cellular components. The

mechanism of intracellular transport is still ambiguous.89'92 In the acidified
endosome, the apon is still complexed with TfR. Ultimately, the apon-TfR returns
to the surface and in the basic environment of the blood, apon is released to harbor
more iron.

The ferroxidase activity of ceruloplasmin is supported by several
studies.8’50’89 In 1968, a hematology study, with pigs as the subjects, showed that
copper-deficient pigs developed an iron-deficiency anemia. Copper-deficient

animals were capable of absorbing iron but were incapable of releasing iron to

plasma. Oral and intravenous administration of iron failed to correct the anemia but

85

Plasma membrane

 

| Fe(II)

Figure 2.9. Iron mobilization in mammalian plasma. Acting as ferroxidase,
ceruloplasmin converts Fe(II), released from the cell via an unknown transporter, to
Fe(III). Two moles of Fe(III) subsequently binds with transferrin (Tf) and
transported to different cells.

86

results in the accumulation of iron in duodenal enterocytes and reticulo-endothelial
cells. Examination of the animal's blood revealed a deficiency in ceruloplasmin.
Administration of ceruloplasmin resulted in the release of iron to the blood and
corrected the observed anemia. It is apparent that copper deficiency caused the
blockage of iron release from the cell. Implicated in this deficiency is
ceruloplasnrin, a copper-protein with ferroxidase activity. Copper-deficient pigs have
impaired ceruloplasmin activity and cannot oxidize ferrous iron to the ferric state.

Therefore, the transfer of iron to the plasma did not occur. The human analogue of

this disorder is aceruloplasminerrriag3 and is characterized by diabetes, retinal
degeneration and neurologic symptoms. Affected patients have parenchymal iron
accumulation in conjunction with the absence of serum ceruloplasmin. Molecular
genetics analysis reveals inherited mutations in the ceruloplasmin. All of these
findings point to the essential role played by ceruloplasmin, ferroxidase activity in

iron metabolism.

2.12 Fet3 vs. th
Genes that are involved in the loading of copper to Fet3 have human

homologues. A putative human plasma-membrane copper transporter with homology

to Ctrl has been identified and is called as hCTR194. As stated before, Menkes and

Wilson disease proteins, ATP7A and ATP7B respectively, are homologues of

Cc02.75v87 Menkes disease is characterized by copper deficiency and Wilson disease
by copper overload. ATP7b is expressed in the liver. Defects in and loss of Wilson

disease protein, Atp7 result in defective transport into the intercellular vesicle where

87

ceruloplasmin normally becomes copper loaded. Wilson disease patients show low
levels of active ceruloplasrrrin and may exhibit anemia similar to that observed in

copper deficient pigs. Hahl, human homologue of the yeast Atxl has also been

identified.95 Thus, the genes involved in the transport of copper to the multicopper
oxidases appear to be completely conserved between yeast and human. These human
genes may be able to substitute for their yeast counterparts allowing the study of the
structure and functions of these genes in the host yeast.

The multicopper proteins Fet3 and ceruloplasmin therefore catalyze similar
reactions. Both of them exhibit ferroxidase activity. This ferroxidase activity is
necessary for the mobilization or transport of iron. They differ in that Fet3 mediates
iron uptake whereas ceruloplasmin mediates iron egress.

The main goal of this dissertation is to characterize the structure of two COpper
sites, the type 1 and type 2 sites of Fet3 using the ESEEM technique in conjunction
with CW-EPR. In particular the ligation of these two paramagnetic sites are

determined.

88

List of References:

1)Holm, R. H.; Kennepohl, P.; Solomon, E. 1. Chem. Rev. 1996, 96, 2239-2314.
2)Askwith, C.; Eide, D.; Van Ho, A.; Bernard, P. S.; Li, L.; Davis-Kaplan, S.; Sipe,
D. M.; Kaplan, J. Cell 1994, 76, 403-410.

3)de Silva, D. M.; Askwith, C. C.; Eide, D.; Kaplan, J. J. Biol. Chem. 1995, 270,
1098-1 101.

4)Abolmaali, B.; Taylor, H. V.; Weser, U. Evolutionary Aspects of Copper Binding
Centers in Copper Proteins; in Structure and Bonding, Williams, R. J. P. ed.;
Springer: Berlin, 1998; Vol. 91, pp 91-190.

5)Hughes, M. N. The Inorganic Chemistry of Biological Processes; 2nd ed.; John
Wiley & Sons: New York, 1981.

6)Reinhammar, B.; Malmstrom, B. G. "Blue" Copper-Containing Oxidases; in
Copper Protein, Spiro, T. G. ed.; John Wiley & Sons, Inc.: New York, 1981; Vol. 3,
pp 110-149.

7)Jameson, R. F. Coordination Chemistry of Copper with Regard to Biological
System; in Properties of Copper, Sigel, H. ed.; Marcel Dekker: New York, 1981; Vol.
12.

8)Solomon, E. 1.; Sundaram, U. M.; Machonkin, T. E. Chem. Rev. 1996, 96, 2563-
2605.

9)Messerschmidt, A. Blue Copper Oxidases; in Advances in Inorganic Cherrristry,
Skykes, G. ed.; Academic Press, Inc.: San Diego, California, 1994; Vol. 40, pp 121-

185.

89

10)Huheey, J. E. Inorganic Chemistry Principles of Structure and Reactivity; 3rd ed.;
Harper & Row, Publishers, Inc.: New York, 1983.

11)Jahn, H. A.; Teller, E. Proc. Roy. Soc. 1937, A161, 220 - 235.

12)Atherton, N. M. Principles of Electron Spin Resonance; Ellis Horwood & PTR
Prentice Hall: New York, 1993.

13)McGarvey, B. R. Electron Spin Resonance of Transition-Metal Complexes; in
Transition Metal Chemistry, Carlin, R. L. ed.; Marcel Dekker, Inc.: New York, 1966;
Vol. 3, pp 89-201.

14)Abragram, A.; Pryce, M. H. L. Proc. R. Soc. London 1951, 205A, 135-153.
15)Abragram, A.; Pryce, M. H. L. Proc. R. Soc. London 1951, 205A, 164-172.
16)Kivelson, D.; Neiman, R. J. Chem. Phys. 1961, 35, 149-135.

17)Maki, A. H.; McGarvey, B. R. J. Chem. Phys. 1958, 29, 31-34.

18)Maki, A. H.; McGarvey, B. R. J. Chem. Phys. 1958, 29, 35-38.

19)Solomon, E. I. Commun. Inorg. Chem. 1984, 3, 225-319.

20)Allendroff, M. D.; Spira, D. J .; Solomon, E. 1. Proc. Natl. Acad. Sci. USA 1985,
82, 3063-3067.

21)Spira-Solomon, D. J .; Allendroff, M. D.; Solomon, E. I. J. Am. Chem. Soc. 1986,
108, 5318-5328.

22)Cole, J. L.; Tan, G. 0.; Yang, E. K.; Hodgson, K.; Solomon, E. I. J. Am. Chem.
Soc. 1990, 112, 2243-2249.

23)Cole, J. L.; Clark, P. A.; Solomon, E. I. J. Am. Chem. Soc. 1990, 112, 9534-9548.
24)Messerschmidt, A.; Rossi, A.; Ladenstein, R.; Huber, R.; Bolognesi, M.; Gatti, G.;

Marchesini, A.; Petruzelli, R.; Finazzi-Agro, A. J. Mol. Biol. 1989, 206, 513-529.

90

25)Messerschmidt, A.; Ladenstein, R.; Huber, R.; Bolognesi, M.; Avigliano, L.;
Petruzzelli, R.; Rossi, A.; Finazzi-Agro, A. J. Mol. Biol. 1992, 224, 179-205.
26)Zaitzeva, I.; Zaitzev, V.; Card, G.; Moshkov, K.; Box, B.; Ralph, A.; Lindley, P. J.
Biol. Inorg. Chem. 1996, I, 15-23.

27)Andrew, C. R.; Fraczkiewicz, R.; Czemuszewics, R. S.; Lappalainen, P.; Saraste,
M.; Sanders-Loehr, J. J. Am. Chem. Soc. 1996, 118, 10436-10445.

28)Blackbum, N. J .; de Vries, S.; Barr, M. E.; Houser, R. P.; Tolman, W. B.; Sanders,
D.; Fee, J. A. J. Am. Chem. Soc. 1997, 119, 6135-6143.

29)Gamelin, D. R.; Randall, D. W.; Hay, M. T.; Houser, R. P.; Mulder, T. C.;
Canters, G. W.; de Vries, S.; Tolman, W.; Lu, Y.; Solomon, E. I. J. Am. Chem. Soc.
1998, 120, 5246-5263.

30)Houser, R. P.; Youbg, V. G.; Tolman, W. B. J. Am. Chem. Soc. 1996, 118, 2101-
2102.

31)Neese, F.; Zumft, W. G.; Antholine, W. E.; Kroneck, P. M. H. J. Am. Chem. Soc.
1996, 118, 8692-8699.

32)Malmstrom, B. G. Chem. Rev. 1990, 90, 1247-1260.

33)Rosenzweig, A. C. Nat. Struct. Biol. 2000, 7, 169-171.

34)Brown, K.; Tegoni, M.; Prudencio, M.; Pereira, A. S.; Besson, S.; Moura, J. J .;
Moura, I.; Cambillau, C. Nat. Struct. Biol. 2000, 7, 191-195.

35)Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry; 5th ed.; John Wiley
& Sons, Inc.: New York, 1988.

36)Kassner, R. Proc. Nat. Acad. Sci. 1972, 69, 2263-2267.

91

37)Yandell, M. A. A.; Addison, A. W.; Karlin, K. D. Inorg. Chim. Acta 1981, 55,
L35-L37.

38)Yokoi, H.; Addison, A. W. Inorg. Chem. 1977, 16, 1341-1348.

39)Karlin, K. D.; Yandell, J. K. Inorg. Chem. 1984, 23, 1184-1188.

40)Addison, A. W.; Rao, T. N .; Sinn, E. Inorg. Chem. 1984, 23, 1957-1966.
41)Churg, A. K.; Warshel, A. Biochemistry 1986, 25, 1675-1681.

42)Addis0n, A. W. Inorg. Chim. Acta 1989, 162, 217-220.

43)Canters, G. W.; Gilardi, G. FEBS Letters 1993, 325, 39-48.

44)Malmstrom, B. G. Eur. J. Biochem. 1994, 223, 711-718.

45)Solomon, E. I.; Penfield, K. W.; Gewirth, A. A.; Lowery, M. D.; Shadle, S. E.;
Guckert, J. A.; LaCroix, L. B. Inorg. Chim. Acta 1996, 243, 67-78.

46)Guckert, J. A.; Lowery, M. D.; Solomon, E. I. 1994.

47)Battistuzzi, G.; Borsari, M.; Loschi, L.; Righi, F.; Sola, M. J, Am. Chem. Soc.
1999, 121, 501-506.

48)Messerschmidt, A. Metal Sites in Small Blue Copper Proteins, Blue Copper
Oxidases and Vanadium-Containinng Enzymes; in Structure and Bonding, Hill, H. A.
O., Sadler, P. J. and Thomson, A. J. ed.; Springer: Berlin, 1998; Vol. 90, pp 37-68.
49)Osaki, S.; Johnson, D. A.; Frieden, E. J. Biol. Chem. 1966, 241, 2746.
50)Kaplan, J .; O'Halloran, T. Science 1996, 271, 1510-1512.

51)Gitlin, J. D. Pediatric Research 1998, 44, 271-276.

52)Mukhopadhyay, C. K.; Attieh, Z. K.; Fox, P. L. Science 1998, 279, 714-717.
53)Stearman, R.; Yuan, D. S.; Yamaguchi-Iwai, Y.; Klausner, R. D.; Dancis, A.

Science 1996, 271, 1552-1557.

92

54)Askwith, C.; Kaplan, J. J. Biol. Chem. 1997, 271, 401-405.

55)Askwith, C. C.; Kaplan, J. J. Biol. Chem. 1998, 273, 22415-22419.

56)de Silva, D.; Sandra, D.-K.; Fergestad, J .; Kaplan, J. J. Biol. Chem. 1997, 272,
14208-14213.

57)Hasset, R. F.; Yuan, D. S.; Kosman, D. J. J. Biol. Chem. 1998, 273, 23274-23282.
58)Dancis, A.; Yuan, D. S.; Haile, C.; Askwith, C.; Eide, D.; Moehle, C.; Kaplan, J .;
Klausner, R. D. Cell 1994, 76, 393-402.

59)Kosman, D. J .; Hassett, R.; Yuan, D. S.; McCracken, J. J. Am. Chem. Soc. 1998,
120, 4037-4038.

60)Graziani, M. T.; Morpugo, L.; Giuseppe, R.; Mondovi, B. FEBS Letter 1976, 70,
87-90.

61)Malkin, R.; Malmstrom, B. G.; Vanngard, T. Eur. J. Biochem. 1969, 7.
62)Reinhammar, B.; Oda, Y. J. Inorg. Biochem. 1979, 11 , 115-127.
63)Spira-Solomon, D. J .; Solomon, E. I. J. Am. Chem. Soc. 1987, 109, 6421-6432.
64)Morie-Bebel, M. M.; Morris, M. C.; Menzie, J. L.; McMillin, D. R. J. Am. Chem.
Soc. 1984, 106, 3677-3678.

65)Morie-Bebel, M. M.; Menzie, J. L.; McMillin, D. R. The Preparation and
Characterization of a Mixed Metal Derivative of Laccase; in Biological and Inorganic
Copper Chemistry, Proceedings of the Conference on Copper Coordination
Chemistry, Karlin, K. D. and Zubieta, J. ed.; Adenine Press, Inc.: Guilderland, NY,
1986; Vol. 1, pp 89-95.

66)Smith, T. A.; Penner—Hahn, J. E.; Hodgson, K. O.; Berding, M. A.; Doniach, S.

Springer Proc. Phys. 1984, 2, 58-60.

93

67)Guss, J. M.; Freeman, H. C. J. Mol. Biol. 1993, 169, 521-563.
68)Penfield, K. W.; Gay, R. R.; Himmelwright, R. S.; Eickman, N. C.; Solomon, E. I.
J. Am. Chem. Soc. 1981, 103, 4382-4388.

69)Norris, G. B; Anderson, B. R; Baker, E. J. Am. Chem. Soc. 1986, 108, 2784-
2785.

70)Machonkin, T. B; Zhang, H. H.; Hedman, B.; Hodgson, K.; Solomon, E. 1.
Biochemistry 1998, 37, 9570-9578.

71)Ducros, V.; Brozowski, A. M.; Wilson, K. S.; Brown, S. H.; Ostergaard, P.;
Schneider, P.; Yaver, D. S.; Pedersen, A. H.; Davies, G. J. Nat. Struct. Biol. 1998, 5,
310-316.

72)Pierloot, K.; De Kerpel, J. O. A.; Ryde, U.; Roos, B. O. J. Chem. Phys. B 1998,
102, 4638-4647.

73)Askwith, C. C.; de Silva, D.; Kaplan, J. Mol. Microbiol. 1996, 20, 27-34.
74)Eide, D. Curr. Opin. Cell Biol. 1997, 9, 573-577.

75)Askwith, C.; Kaplan, J. Trends Biol. Sci. 1998, 23, 135-138.

76)Radisky, D.; Kaplan, J. J. Biol. Chem. 1999, 274, 4481-4484.

77)Dix, D.; Bridgham, J.; Broderius, M.; Eide, D. J. Biol. Chem. 1997, 272, 11770-
11777.

78)Dix, D. R.; Bridgham, J. T.; Broderius, M. A.; Byersdorfer, C. A.; Eide, D. J. J.
Biol. Chem. 1994, 269, 26092-26099.

79)Dancis, A.; Klausner, R. D.; Hinnebusch, A. G.; Barriocanal, J. G. Mol. Cell Biol.

1990, 10, 2294 -2301.

94

80)Dancis, A.; Roman, D. G.; Anderson, G. J.; Hinnebusch, A. G.; Klausner, R. D.
Proc. Natl. Acad . Sci. USA 1992, 89, 3869-3873.

81)Eide, D.; Davis-Kaplan, S.; Jordan, 1.; Sipe, D.; Kaplan, J. J. Biol. Chem. 1992,
267, 20774-20781.

82)Shatwell, K. P.; Dancis, A.; Cross, A. R.; Klausner, R. D. J. Biol. Chem. 1996,
271, 14240-14244.

83)Hasset, R. F.; Romeo, A. M.; Kosman, D. J. J. Biol. Chem. 1998, 273, 7628-7636.
84)Dancis, A.; Haile, D.; Yuan, D. S.; Klausner, R. D. J. Biol. Chem. 1994, 269,
25660-25667.

85)Glerum, D. M.; Shtanko, A.; Tzagoloff, A. J. Biol. Chem. 1996, 271, 14504-
14059.

86)Cullota, V. C.; Klomp, L. W. J .; Strain, J .; Casareno, R. L. B.; Krems, B.; Gitlin,
J. D. J. Biol. Chem. 1997, 272, 23469-23472.

87)Yuan, D.; Stearman, R.; Dancis, A.; Dunn, T.; Beller, T. Proc. Natl. Acad. Sci.
1995, 92, 2632-2636.

88)Lin, S.-J.; Pufahl, R. A.; Dancis, A.; O'Halloran, T. V.; Culotta, V. C. J. Biol.
Chem. 1997, 272, 9215-9220.

89)Winzerling, J. J.; Law, J. H. Annu. Rev. Nutr. 1997, 17, 501-526.

90)Watkins, J. A.; Altazan, J. D.; Elder, R; Li, C.-Y.; Nunez, M.-T.; Xue-Xiang, C.-
T. ; Glass, J. Biochemistry 1992, 31, 5820-5830.

91)Thorstensen, K.; Romslo, I. J. Biol. Chem. 1988, 263, 8844-8850.

92)Nunez, M. T.; Escobar, A.; Ahumada, A.; Gonzales-Sepulveda, M. J. Biol. Chem.

1992, 267, 11490-11494.

95

93)Harris, Z. L.; Takahashi, Y.; Miyajima, H.; Serizawa, M.; MacGillivray, R. T. A.;
Gitlin, J. Proc. Natl. Acad. Sci. 1995, 92, 2539-2543.

94)Zhou, B.; Gitschier, J. Proc. Natl. Acad. Sci. 1997, 94, 7481-7486.

95)Klomp, L. W. J.; Lin, S.-J.; Yuan, D. 8.; Klausner, R. D.; Culotta, V. C.; Gitlin, J.

D. J. Biol. Chem. 1997, 272, 9221-9226.

96

Chapter 3

INSTRUMENTATION

In both optical and magnetic resonance spectroscopy, a source of
electromagnetic radiation is used to irradiate a sample placed in specialized sample
vessel. Changes in the properties of the incident radiation (frequency, energy,
intensity etc.) detected by an appropriate detector will reveal some information about
the sample. One obvious difference between optical and magnetic resonance
spectroscopies is the need for an external magnetic field. A static magnetic field is
required to align the magnetic moments of the sample and to split the energy levels
to observe resonance. In EPR spectroscopy, other differences could be pointed out.
One is that in EPR, monochromator such as prism or grating is not needed to emit
monochromatic radiation. The microwave source (e. g. klystron or synthesizer) in
EPR is capable of emitting monochromatic microwave radiation. Also, in a typical
optical spectroscopy, a spectrum is obtained by varying the frequency of the
electromagnetic radiation and observing the change in the radiation as it passes
through the sample. An EPR spectrometer measures a spectrum by fixing the
radiation frequency and scanning the applied magnetic field.

As mentioned in chapter 1, the first EPR experiment was done in 1945
by a Russian scientist, Zavoisky who detected a radio frequency absorbance from a
copper dichloride sample. The availability of microwave components brought about
by World War H led to the widespread use of paramagnetic resonance. Advances in

design of microwave components and in electronics paved the way to modem

97

instruments and the construction of pulse-EPR instruments. The first part of this
chapter will present a brief description of the CW-EPR spectrometer used in our
study. Then, it will be followed by a description of the home-built pulse-EPR

machine at Michigan State University. A comprehensive discussion on EPR

instrumentation is presented by C. Poole1 in his book and would serve as an excellent

reference for those who wish to fully understand EPR instrumentation. A chapter on

CW-EPR instrumentation is also explored in the book by Wertz and Boltonz.

Discussion about pulse-EPR instrumentation can be found in several papers and

books.3‘7

3.1. CW-EPR

Data collection was done using the commercial Bruker ESP-300E

instrument.8 A simple schematic of the important components comprising the CW
instrument is depicted in Figure 3.1.

The microwave (MW) radiation source is a klystron (l), which is housed in
the microwave bridge. The microwave power is divided into a "power" arm and
reference arm. MW coming from the power arm are used to irradiate the sample,
placed in a resonant cavity, through a nonreciprocal device, the circulator (3). An
attenuator (2) is used to vary the power used to excite the sample. The reflected MW
from the cavity (7) are directed into a Schottky barrier diode detector (8), via port 3 of
the circulator, where it is combined with MW coming from the reference arm. The
reference arm has an attenuator (4) to bias the diode current making sure that the

detector diode operates in a region of linear response (i.e. diode current is about 200

98

11A). It also has a phase shifter (5) that matches the phase of the reference MW with
that of the reflected MW. The current from the detector diode then is fed to the signal
channel (9) for phase-sensitive detection, amplification and processing. In our
experiment, a TE102 cavity (7) was used and the temperature is maintained using an
Oxford ESR-900 continuous flow cryostat. The current needed for the generation of
the magnetic field of the external magnet is provided by an ER032 magnet power
supply. Using a Hall probe, the magnitude of the magnetic field is monitored and

regulated by ER032 magnetic field controller.

 

9) to signal channel

 

 

 

 

 

 

 

1) power arm 3) 3
2) 2.

 

 

a) u

 

 

 

 

 

 

Figure 3.1. Diagram of the Bruker EPR spectrometer.

99

3.2 Pulse-EPR Machine at MSU

A layout of the pulse-EPR machine is shown in Figure 3.2. The machine was

originally designed by Prof. John McCracken in 1990.9 Essentially, the pulse-EPR

machine has similar basic components as the CW-EPR machine.

 

 

 

 

 

reference arm “0'3”"
I 9 l l—I'l
Ell “—J
pulse-generating arms ampnfler circulator mixer
. . [ variable att - uator s
‘ amplifier
synthesizer to digital oscilloscope A
resonator I;

echo

video amplifier
[<— magnet

Figure 3.2. Design of the pulse-EPR machine at Michigan State University (MSU).

3.2.1 Microwave Bridge

The microwave bridge is comprised of a transmitter used to generate and
control microwave radiation, and a receiver, which has a detector for detecting MW
radiation. It also has rotary vane attenuator for varying the MW power that is fed to

the sample through a circulator.

3.2.1.1 Transmitter
Figure 3.3 shows the schematic diagram of the transmitter part of the

microwave bridge. A Microwave synthesizer (Giga-Tronics, Model 610) provides

100

the microwave radiation and is capable of operation over a frequency range of 6-18
GHz. A directional coupler (Omni Spectra, Model MN2025-6018-10) splits the
microwave power with 90% of it being directed to the reference arm and the
remainder to the pulse-forming channels. Acting as a current-bias and a phase-
reference, the reference arm has an adjustable phase shifter (ARRA, Model 9828A).
It also has an isolator (Innowave, Model 111R) which makes sure that there is no
microwave radiation reflected back to the source. The remaining 10% of the
microwave power from the source is fed to a power divider (Omni Spectra, Model
PN2089-6209-00) further splitting the microwave radiation into two independent
pulse-generating channels. Both of these channels have low-power, high-speed PIN
diode switches (General Microwave, Model FM864-BH) and high-speed 0°/180° wide
band phase modulators (General Microwave, Model F1938) whose width and phases,
respectively, are controlled by a home-made pulse logic circuit. One of the channels
has an attenuator (ARRA Model P9804-20), while the other one has a phase shifter.
The low-power MW pulses (~100uW) generated are combined, passed through an
isolator, and then amplified by a medium-power GaAs FET amplifier (Avantek,
Model SWL-89-0437). Through an attenuator, the GaAS FET-amplified MW pulses
~(100mW) are fed into the travelling wave tube (T WT) amplifier (Applied Systems
Engineering, Model 117). In order to vary the amount of MW power irradiating the
sample, the TWT amplified pulse (lkW) is allowed to pass through a rotary vane
attenuator (Hewlett Packard, Model X382A). In between the rotary vane attenuator

and the TWT amplifier is an isolator, designed especially for high-power MW pulses.

101

 

2.88598 3

 

.05;an mam—-023 2: no omens 03380:: 2: mo cocoa 8:25:88 .Qm ennui

 

 

.3535:
.5: 3:5
I l
53.55:."
E39.
933...;

 

 

:3

.3me S e 5.3 do.—

 

102

The MW pulses are then fed to a circulator (Microwave Associates, Model MA-
8K269), a non-reciprocal device like the isolator, which directs the microwave power
in one terminal into the cavity containing the sample and then directs the reflected

MW from the cavity to the detector through the next terminal in the sequence.

3.2.1.2 Receiver

The schematic diagram of the detection system is shown in Figure 3.4. The
reflected MW signal from the sample resonator is first fed to a GaAs FET amplifier
(Avantek, Model AWT-18635) through an isolator, a high-speed PIN diode limiter
(Innowave, Model VPL-6018) and a fast PIN diode switch (General Microwave,
Model F9114) which is triggered by the pulse-logic circuit. The diode limiter and
switch serve to protect the GaAs FET amplifier from hi gh-power microwave pulses.
For tuning purposes, 1% of the microwave power is extracted via a direction coupler.
Most of the microwave power is allowed to pass through a bandpass filter (K& L
Microwave, Model 3H10-2000/18000-0/0) to remove unwanted radio frequency
noises introduced by PIN switches, prior to its being directed to a double balance
mixer (RHG, MODEL DM2-l8 AB). In the mixer, the reflected MW signal from the
resonator is mixed with the MW power coming from the reference arm. The
homodyne arrangement of the double balance mixed generates signals at sum and
difference of the frequencies of the signal coming from the receiver and the reference
arm. The intensity of the echo signal is proportional to the difference between the

amplitude of the two input signals as well as their relative phases. It is then amplified

103

.EoEHEmE Mamba—am 2: mo ownwfi 0225828 of «o 5:08 Stoned in PEME

8::— _e.=:8 .5339.

..lAIJlA a. i m an

.83.. 3 353:... 82>

 

 

9:55 ..8
2.88536 3

Eu eflvhouvh BAY—u

188:9.

104

and adjusted through series of video amplifier (Comlinear, Model E220)-attenuator
(JFW, Model 50DR-OO3)-video amplifier and then finally fed to a digital

oscilloscope(Tektronix 620 B) to convert the analog signals to digital signals.

3.2.2 Cavity and Cryogenic

For ESEEM experiments, a reflection-type cavity with folded stripline half-
wave resonator10 is used. The use of the folded stripline resonator lowers the quality

factor, QL,1,2 of the cavity. A low QL is desirable for ESE experiment to cut down
the ring down time of the resonant pulse in the cavity and also to the decrease the
dead-time. QL is related to amount of energy stored in the cavity and thus to the
magnitude of B1. A low QL implies small Bl field and consequently low sensitivity.
However, the use of this resonant element allows the increase of filling factor, n, by

concentrating the Bl field seen by the sample. Coupling of the microwave pulse with

resonator is accomplished by a Gordon Coupler.11v12 By adjusting the position of
the this coupler, the QL of the cavity can be varied.

The cavity is immersed in a homemade liquid helium immersion dewar
allowing us to work at 4.2 Kelvin. At this temperature, the relaxation processes, T1

and T2 are increased allowing us to detect echoes at longer times.

3.2.3 Magnet
The external magnetic field is supplied by an electromagnet (Walker

Scientific, Model I-IF-12H) capable of generating magnetic field from 50 to 16000

105

gauss. The field values are monitored and regulated by a Hall-effect field controller

(Bruker, Model B-H15).

3.2.4 Computer Interface, Experiment Timing and Data Collection

Figure 3.5 shows a detailed schematic diagram of the set-up of the pulse
machine. The scope, field controller and the delay generator (Research Stanford
Systems, Model DG535) are connected to the PowerComputing Mac computer via
IEEE-488 interface (National Instruments). The delay and gate generator controls the
timing of the echo experiment. It is used to trigger the pulse logic unit and therefore
determines the pulse sequence. It is also used to trigger the oscilloscope for
measuring the intensity of the electron spin echo. The data acquisition is

implemented using LabVIEW 5 .0. 1(National Instruments).

3.2.5 Modification of the Pulse Logic Module

As mentioned earlier, the pulse logic module has three functions. It (1)
determines the width and phases of the pulses coming out from the two microwave
pulse-generating arms, (2) turns on the TWT amplifier, and (3) turns on and off the
receiver diode switch to protect the detection system

The original design of the circuitry for this pulse logic unit was made by Prof.

John McCracken and Dr. Hong-In Lee.13 It was modified for the present work to
allow for the generation of longer microwave pulses and reduce timing jitter
introduced by poor circuit layout. Non-availability of some IC's used in the old pulse

logic module at present time is also a reason for modifying the circuitry.

106

Figure 3.5. A schematic diagram of the pulse-EPR instrument built at Michigan State
University. Schemes in Figure 3.3 and 3.4 are incorporated.

107

TWT Switch

Receiver Gate

ABCD

Delay
Generators

for ESE-ENDOR

Mac computer

 

108

Using TTL textbooks and logic tables,15’16 the design for the present circuit was

made and their schematic diagrams are shown in the appendix section of this chapter.

109

APPENDICES

Appendix 3A
Pulse Logic Module

Timing Diagram for a Two-pulse ESEEM Measurement

Figure 3A.A1 shows a timing diagram for the creation of microwave pulses
in a 2-pulse ESEEM experiment. The A and B inputs are generated by a delay and
gate generator (DG535) which sets the separation between the two microwave pulses.
Shown also in the figure are the output pulses used to turn on and off the PIN diode
switches (General Microwave, Model FM864-BH), phase modulators (General
Microwaves, Model f 1938), TWT control, and the receiver PIN diode switch
(General Microwave, Model F9114). The PIN diode switches are closed by a
positive logic 0 input. A positive logic 1 input turns on the phase modulators and the

TWT control.

Module Interconnections

A general scheme on how these trigger inputs and pulse outputs are integrated
for the formation and amplification of microwave pulses is shown in Figure A3.A2.
Basically, inputs from DG535 are fed into a pulse forming unit. This unit creates "0"
logic pulses to drive the PIN diode switches (pulse P) and the TWT control circuit
(pulse TWT), and logic "1" to control the phase modulators (pulse (1)). Pulses P and

TW'I' are directed to the pulse combiner unit which generates output pulses shown in

110

A) INPUTS

 

A trigger I

 

B trigger

 

 

B) OUTPUTS

 

 

 

PIN

switch

TWT l
Phase

modulator

 

 

 

 

 

 

 

 

 

Receiver
switch

 

C) 2-Pulse ESEEM Microwave Pulse Sequence

 

I<—t———->|< '5

Figure A3.A1. Timing diagram for a 2-pulse ESEEM measurement.

111

Figure A3.A1, while pulse 4) is fed to the phase sorter unit which generates the output

pulse train for the phase modulators (Figure A3.A1). A pulse from the combiner unit

is tapped and directed to the receiver control unit, which operates the PIN diode

switch that gates the receiver.

trigger
from
DG535

 

 

Pulse
Forming
Unit(PFU)

 

 

 

 

 

 

 

 

j I'— P pulse __ T to PIN 1
> >switch
| I TWT pulse L Pulse. to PIN2
W p Combmer > switch
ls J—L_
M" e + to TWT
control

 

 

 

 

 

 

Figure A3.A.2. Module Interconnections.

112

\J

 

 

 

> too-1 arm

 

 

 

 

 

Phase
Sorter
> to (ti-2 arm
geceivelr ’
ontro to receiver
microwave
switch

 

 

Descriptions and Circuit Diagrams of PFU, Pulse Combiner, Phase Sorter and
Receiver Control Switch

The pulse forming, pulse combiner, the receiver control, and the phase sorter
units constitute the pulse logic module. Their circuitry is shown in Figure A3.A3,
A3.A4, A3.A5 and A3.A6, respectively.
W. Currently, the pulse forming unit (Figure A3.A3) can create six
P (PA-PF), TWT (TWT A-F) and (l) (¢A-F) pulses. The width of the six P pulses
(output 7-1 and 7-4, Figure A3.B3a-c) can be varied according to the logic inputs of
pins 8, 9 and 10 of PPG33F-50, a 3-Bit programmable pulse generator (IC#4A and
4A). The duration of the TWT (output 5-12 and 5*-12, Figure A3.A3a-c) and (I)
pulses (output of 6-5 and 6*-5, Figure A3.A3a-c), on the other hand, are determined
by the resistance between pins 7 and 16 of monostable multivibrator, 74LSl23 (IC#S
and 6, respectively).
Wt. The TWT pulses (TWT A-F) that corresponds to each
microwave pulse generated, are combined by an 8-input positive NAND (IC#8) gate
of the pulse combiner (Figure 3A.A4) to form train of pulses for turning on the TWT
amplifier. Three P pulses, designated as P-A, P-B and P-D, are combined in a dual 4-
input positive-NAND (IC#l 1) gate (Figure A3.A4) to create train of pulses for
triggering the PINl diode switch of the pulse-generating arm of the transmitter
(Figure 3.5). Similarly, the pulses P-C, P-E and P-F are combined in the other half of
the dual 4-input positive-NAND gate for triggering the PIN 2 diode switch.

W. The train of pulses from IC# 11-6 and 11-8 are tapped and

directed to the receiver control unit (Figure A3.A5). This unit counts the number

113

pulses created and turns on the receiver switch (General Microwave, Model F9114,
Figure 3.5) only when an electron spin echo appears.

Wit, The six 4) pulses from the pulse forming unit are fed to a phase
sorter unit (Figure A3.A6) whose function is to phase cycle microwave pulses for the

elimination of unwanted echoes found in a 3-pulse ESEEM experiment.

114

Figure A3.A3a. Pulse Generating Subunit (A input, page 116 and B input, page 117).

The IC's used are numbered. They are 1, 1* (74F00, Quadruple 2-input positive-
NAND gate); 2, (74FO4, Hex-Inverter); 3, 3*, 5, 5*, 6 and 6* (74LS123, Dual
retriggerable monostable multivibrator); 4, 4* (PPG 33F-10, 3-bit programmable pulse
generator); 4A, 4A* (PPG33F-50, 3-Bit programmable pulse generator); and 7 (74F02,
Quadruple 2-input positive-NOR gate).

IC #4 and #4A (4* ans 4A*) create the narrow and wide pulses, respectively.
The width of the pulses coming from 4-2 (A*-2) and 4A-2 (4A*-2) is determined by the
logic input in pins 8, 9, 10 of IC#4 (4*) and #4A (4A*). Logic states of these pins are
depended on the states of the binary switch (BIN-SW). PPG's 4 (4*) and 4A (4A*) are
enabled by the double-pole-double-throw (DPDT) on/off/on toggle switch. A logic 0
enables 4 (4*), while a logic 1 enables 4A (4A*). The width of the TWT and q» pulses are
depended on the resistance of the rotary switches.

115

.59: < dm<.m< 0.5»:—

2-326

 

 

 

 

 

 

 

 

 

 

 

00
Q

 

 

 

 

L.-.-.-.--

 

 

("l

 

 

 

 

 

 

 

L...-....... .

Jae < ..8
2 we 0

[\
N
i

—t

 

v

_.

116

.39: m .343 2:5

chB—é

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

117

Figure A3.A3b. Circuit design of the pulse generating subunit for C (page 119) and D
(page 220) inputs. The numbers and identification of IC's used are the same as in Figure
A3.A3a.

118

.59: o 522 gauz—

    

onBTe
mem om

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

119

4&5 a 52.2 2:5

         

2-32-6

 

 

 

 

 

120

 

 

U

3w-2_m
am _N om

 

 

 

 

 

 

 

 

 

 

 

 

Figure A3.A3c. Circuit design of the pulse generating subunit for E ( page 222) and F
(223) inputs. IC numbers and identification are the same as in Figure A3.A3a

121

.59: m .93.? ea»:—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

122

.39: m .8413: 95»...—

©7m\c_-w

 

 

 

 

 

 

 

 

 

 

 

 

- .3 o g:
S S o: I 6
mn— v

 

 

do. 5%
m

 

 

 

:55 m
o

 

 

 

 

 

 

 

 

 

 

>m >o

123

Figure A3.A4. Pulse combiner unit. The IC's used are 8(74F30, 8-input positive-NAND
gate); 9 (74F00, Quadruple 2 -input positive-NAND gate); 10 and 12 (74140, Dual 4-
input positive-NAND 50-ohm line driver); and 11 (7420, Dual 4-input positive-NAND
gate).

124

 

 
 
 
  
  

(4) IRS!

TWT-A
TWT-B
TWT-C
TWT-D
TWT-E
TWT-F
no input
no input

(4) 1m

14
6 TWTA trigger

 
 

10

 
 
 

'7

 
 
 
 

REC. CON 1 (13-14)

 

PC
‘ PINZ trigger
PE

PF
no input

   
 

(2) IRS!
REC. CON 2 (14-14)

  

 

 

125

 

Figure A3.A5. Receiver Control Subunit. IC's used are 13 and 14 (7493, 4-bit binary
counter); 15 and 23 (74LS123, Dual retriggerable monostable multivibrator); 16 and 17
(7404, Hex-Inverter); 18, 19, and 22 (7400, Quadruple 2-input positive-NAND gate); 20
(7420, Dual 4-input positive-NAND gate); and 21 (74140, Dual 4-input positive-NAND
gate).

126

to receiver
6 diode switch

 

A input
trigger

 

 

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

from 1 1-6

 

 

127

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

S c °> :Nm
a: 0 MM
< o
o<° ‘7 o
:4 E m
U M
2 E
._l “J.-
°P _
fl
_
E
h

 

Figure A3.A6. Phase Control Subunit. The IC's are 24,25,26,27, 28, 29, 30, and 31
(74LS373, In-phase module Octal-D type latches); 32 (74128, 5052 NOR line driver); 33
(74LS54, 4-wide AND-OR-Invert gate); 34 (7454, 4-wide AND-OR-Invert gate); 35
(74L8123, Dual retriggerable monostable multivibrator); 36 (74138, 3—Line to 8-Line
decoder); 37 (7494, 4-bit shift register), 38 (7400, Quadruple 2-input positive-NAND
gate); and 39 (7474, Dual-D-Type positive-edge-triggered flip-flops with preset and
clear).

128

 

129

Appendix 38
Davies -ENDOR on Cu-dien-imidazole Complex

The capability of the pulse-EPR machine to perform Davies-ENDOR was
tested using SmM sample of frozen solution of Cu(dien-imid) complex. The
experiment described here was done with the help of Dr. Rajendabose
Muthukumaran.

The pulse sequence for Davies-ENDOR is shown Figure A3.B1. It can be
divided into the (a) microwave preparation pulse (rt-pulse), (b) the mixing
radiofrequency pulse and, (c) the pulse-echo detection pulses. Davies-ENDOR is an
ESE-ENDOR which measures the change in the intensity of the echo as a function of
radio frequency. The change in the echo intensity is a consequence of the creation of

the spin population in the nuclear sublevels.

 

 

 

 

 

 

 

 

 

 

 

 

 

1! n/2 1:
*—— T :lrer —b <—1:—>
RFpulse

Figure A3.B1. Davies-ENDOR pulse sequence.

130

The effect of different pulses can be explained by what they do to the electron
spin populations in a four-energy level system (S=1/2, I=1/2, Figure A3.B2).
Initially, the spin system assumes an equilibrium population that follows the

Boltzmann Distribution Law,

1‘1 = exp[-AFJkT].
no

Because g,,l3,,Bo >> ganBo, spin population in the El and E2 levels, and E3 and E4
levels are essentially equal. The preparation microwave pulse is a it pulse of a

duration chosen to make it selective for one of the EPR transitions. For example if

E4 000 000 000
E3 000 0000 0000

 

__K __ —__f
___i_

 

E2 0000 000 0000
I.
E1 0000 0000; 000
Preparation Mixing Detection

Figure A3.B2. Transfer of spin populations, represented by small circles, in Davies-
ENDOR measurement for an S = 1/2 and I = 1/2 spin system during the preparation,
mixing and detection periods.

131

the transition between E2 and E3 is irradiated, the population between these two
levels is inverted. Consequently, a population difference is created between nuclear
sublevels El and E2 and E3 and B,,. This difference can be detected by applying a
long radio frequency pulse during the mixing period. If the RF pulse is resonant with
one of the NMR transitions, E,-> 1:?2 for example, a nuclear spin flip will occur
inverting the populations between these sublevels and equalizing the population
between El and E4 and E2 and E. If a n/2-r-1t-r-echo sequence is applied right after
the RF pulse, no primary echo will be observed. However, if the RF is not resonant,
an echo will be detected as long as a population difference between EPR transition E2
and E3 (or E1 and E4) exists.

Figure A3.B3 shows a Davies-ENDOR of SmM frozen solution of Cu-
diethylene-imidazole complex at B0 = 3200 Gauss. Figure A3.B3a is collected using
the pulse sequence: mw 1: pulse, 220ns; rf pulse, 7000ns; and detection pulse 1,
900ns. Two peaks are observed. The narrow peak at ~13.5 MHz is due to weakly
coupled proton which can be from proton of ambient water. The broad peak of low
intensity at around 18 MHz can be assigned to the nuclear modulation frequency of
the directly coordinated 1“N of the imidazole ligand. Using a narrower preparation
pulse, the contribution of the weakly coupled proton can be suppressed as shown in
Figure A3.B3b. This Davies-ENDOR data was collected using the pulse sequence:
mw 1t pulse, 450ns; rf pulse, 4400ns; and detection pulse 1:, 450ns. The experimental
data is simulated (solid line) spintl.mat ( Matlab program) with the following
Hamiltonian parameters: A“: Ayy = An =34.5, equ = 3.3 MHz, 11 = 0.13 and NQI

Euler angles with respect to B0 of (0,0,0). A spectrum obtained using Mims-ENDOR

132

sequence is shown in Figure A3.B4 for comparison. In Mims-ENDOR spectrum, the
ENDOR signals from weakly coupled protons dominate. These signal, however, are
suppressed in Davies-ENDOR spectrum. Application of these ENDOR methods to

the Fet3 protein were not successful, probably because of low sensitivity.

133

(a)

(b)

l l l L 1
0 5 1 0 1 5 2 0 2 5
RF frequency (MHz)

Figure A3.B3. Davies-ENDOR spectra of Cu-dien-imid complex recorded at 3200 G
using the inversion and detection microwave pulse widths of (a) 220, 116 and 220 ns
and (b) 40, 20 and 40 ns. Other experimental conditions: v = 9.241 GHz, 1 = 900 ns
and trf = 7000 us for (a) and 1: = 450 ns and trf = 4400 ns for (b).

134

(a)

(b)

l l 1 l l l
0 5 1 0 1 5 2 0 2 5
RF frequency (MHz)

Figure A3.B4. (a) Mims ENDOR spectrum of Cu-dien-imid complex at 3200
Gauss, obtained using preparation pulse of 20 ns, t = 250 ns and RF pulse of 36800
ns. (b) Davies ENDOR collected at similar conditions as in figure A3.B3b.

135

i ._
I |-
i ."‘
. .3
:i .. ..
n .
C
=v 5
9,:‘2 ,"
IF . .
at”: '0;
{1' ' e
a. I
*1 0‘ t,
. fa ‘el
0 t I ' I
I Q I
a I :
I: O” : I
r I
I :- : .I
I . . up. .l:' a
: ' ~ - .3". ~::
' ' | ' VI l on" 3' Q a
' ‘ . '. V I I a” 'e.‘ o
u'I'i ' - H‘ .1 ,\ o .
iron! “HI-1v, ,i o N,
Is ‘I...u I "\’| I' (- . . f ‘. ‘
: w ' w .. 3:... ..
.‘J' " A" " 0.5.!
1 §:-\:.$.‘ = ‘w' :6»: 1M ...
. «a. w.» are . s :4
My ‘0" "'0 OJ ’ ' a 1
‘~ ".m'x '04 a 17
'33118' "3:155 “in .Gn. ': z.
I.'\‘ I.I II 0| I. ”
'rg'l ' I 0' i I
0 n I I
I : | a
r . g
I
I

F-
—
—
n-
_
_
P
b

 

O
h

8 12 16 20 24 28
RF frequency (MHz)

Figure A3.B5. Computer simulation (solid lines) of Davies ENDOR data shown in
figure A3.B3b. The parameters used for the simulation are: A,,( = Ayy = An =

345MHz, equ = 3.3 MHz, 1] = 0.13 and NQI Euler angles with respect to B0 of
(0,0,0).

136

LIST OF REFERENCES

l)Poole, C. Electron Spin Resonance: A Comprehensive Treatise on Experimental
Techniques; 2 ed.; John Wiley & Sons, Inc.: New York, 1983.

2)Wer’tz, J .; Bolton, J. Electron Spin Resonance: Elementary Theory and Practical
Applications; 2 ed.; Chapman and HallzNew York, 1986.

3)Mims, W. B. Rev. Sci. Instrum. 1965, 36, 1472-1479.

4)Brown, I. M.; Sloop, D. J. Rev. Sci. Instrum. 1970, 41 , 1774-1777.

5)Blumberg, W. E.; Mims, W. B.; Zuckerman, D. Rev. Sci. Instrum. 1973, 44, 546-
555.

6)Norris, J. R.; Thumauer, M. C.; Bowman, M. K. Electron Spin Echo Spectroscopy
and the Study of Biological Structure and Function; in Advances in Biological and
Medical Physics, Lawrence, J. H., Gofman, J. W. and Hayes, T. L. ed.; Academic
Press, Inc.: New York, 1980; Vol. 17, pp 365-416.

7)Keijers, C. P.; Reijerse, E. J.; Schmidt, J. Pulsed EPR: A New Field of Applications;
North-Holland: Amsterdam, 1989.

8)Weber, R. T.; Heiss, A. A. ESP 300E EPR Spectrometer User's Manual; Bruker
Analytische Messtechnik, 1992.

9)McCracken, J .; Shin, D. H.; Dye, D. 1. Appl. Magn. Reson. 1992, 3, 305-316.
10)Lin, C. P.; Bowman, M. K.; Norris, J. R. J. Magn. Reson. 1985, 65, 369-374.
11)Britt, D.; Klein, M. J. Magn. Reson. 1987, 74, 535-540.

12)Gordon, J. P. Rev. Sci. Instrum. 1961, 32, 658-661.

137

13)Lee, H. 1. Electron Spin Echo Envelope Modulation Studies of Some Transition
Metal Model Complexes; Ph.D. Thesis; Michigan State University: East Lansing,
1994.

14)Davies, E. R. Phys. Lett. 1974, 47A, 1-2.

15) The TTL Logic Data Book; Custom Printing Company: Owensville, Missouri
1998.

l6)Lancaster, D. TTL Cookbook; Howard W. Sams & Co., Inc.: Indianapolis, 1974.

138

Chapter 4

1“N ESEEM: Histidine Ligation to Type-1 and Type-2 Copper Centers in Fet3p

The R side chain unit in the amino acid histidine is imidazole. Imidazole can
coordinate to various metal centers through its donor nitrogen atom. Nitrogen has
two isotopic forms, 14N and IN, both of which are magnetic. 1“N has nuclear spin,
1:1 and 15N has I=l/2. Therefore, histidyl imidazole ligation to a metal center can
be identified by EPR techniques. CW-EPR spectroscopy, however, may not be
appropriate because ligand hyperfine interactions are typically small compared to
metal hyperfine and g-factor anisotropy. As a result, these important features
contribute only to the broadening of absorption peaks in the CW-EPR spectrum. In
this situation, Electron Spin Echo Envelope Modulation (ESEEM) becomes handy.
ESEEM can measure weak hyperfine interactions that are otherwise masked by
inhomogeneous broadening of the CW-EPR spectrum.

Mims and Peisach pioneered the application of X-band ESEEM for the study
of copper proteins. Their studies on stellacyanin proved that copper with

histidine/imidazole ligands exhibited a strong modulation due the remote nitrogen of

the bound histidyl imidazole.1 At X-band (~3000G), the directly coordinated N is
strongly coupled so that the electron-nuclear coupling exceeds those of the 1“N
nuclear Zeeman and the quadrupole term. Consequently, no modulation will be
observed from the directly coordinated nitrogen as only the allowed transitions are
excited by the microwave pulse. The couplings are also too strong to excite the

"branching" of transitions for the weak forbidden lines that do exist. The first

139

assignments were made based on qualitative comparison of the two-pulse modulation
pattern of stellacyanin with those of Cu(H)-aquo complex, Cu(H)- glycylglycine,
Cu(H)-(imidazole),, Cu(H)-(guanidine)4 complexes, and for Cu(H) bovine serum
albumin which was previously found to have imidazole coordinated to copper(II).
The quantum mechanics and mathematical description of this l“N-ESEEM was also

done by Mims and Peisach as they successfully simulated the two pulse and three-

pulse spectra of Cu(H)-diethylene-imidazole and Cu(H)-(imidazole)4 complexes},3

Since these initial studies, several copper proteins have been studied using

ESEEM.4-8

Here, we report the study of histidyl imidazole ligation to the type-1 and type-
2 copper centers of the multicopper oxidase Fet3p using l“N ESEEM. Because both
type 1 and type 2 centers are paramagnetic, the determination of histidine ligation
requires that the contribution of one be eliminated. This was accomplished through
site-directed mutagenesis. The two mutants studied were proteins where either the
type-1 or type-2 Cu-binding sites had been depleted so that EPR/ESEEM studies can
be accomplished without the complication of spectral overlap. To quantify the
number of histidine ligands coordinated to the copper sites and determine how they
are coordinated, computer simulations of the three-pulse spectra taken at magnetic
field positions corresponding to the g" and g 1 were done. The type-2 copper site was
found to have one equatorially bound histidyl imidazole ligand contributing to the
echo-envelope modulation. For the type-1 Cu(II), two magnetically distinct histidyl

imidazole ligands were found.

140

4.1 Experimental

Construction of F et3p mutants and production of soluble protein. Mutant F ET3
alleles were constructed by site-directed mutagenesis by Prof. D.J. Kosman at SUNY-
Buffalo. Type-l depleted (TlD) Fet3p was prepared by Cys484Ser mutation at the
type 1 copper site, and type-2 depletion (T2D) was achieved by mutation at the
HisSlGln mutation at the type 2 copper site. These mutations were selected on the
basis of sequence alignment with the other multicopper oxidases. Following
amplification of the mutant-encoding vectors in bacteria, the FET 3 sequences were
confirmed by dideoxysequencing (Sequenase 2.0) and the vectors transformed into

yeast host strain M2*. Culture growth and mutant protein purification were

performed as has been described in detail for the wild type protein.9

C W-EPR. The continuous-wave EPR spectra of the Fet3 and its variants were
obtained at X-Band on a Bruker ESP300E EPR spectrometer equipped with an
Oxford ESR-900 liquid helium cryostat and utilizing a TE102 EPR. An EIP
Microwave 25B frequency counter was used to monitor the microwave frequency and

a Bruker ER 035 NMR gaussmeter was used to measure the external magnetic field.

The QPOWA10-12 (acquired from R.L. Belford at Illinois EPR Research Center)
simulation program was used to extract the Cu(H) magnetic parameters for ESEEM

simulation.

141

ESEEM. Pulsed EPR experiments were performed on a home-built spectrometer

described previously. 13 Three-pulse (rt/2-r-1t/2-T-1t/2-t-echo) ESEEM data were
collected at X-band using a reflection cavity that employed a folded stripline resonant
element. To eliminate unwanted echoes, a (0,0,0), (1t,1'c,0) two- step phase cycling of
the microwave pulses was used. The values of I were set at period in which the
nuclear modulations due to weakly coupled the proton were minimized. All
experiments were performed at liquid helium temperature, 4.2 K. Frequency

domain spectra were obtained by a Fourier transformation procedure that included

dead-time reconstruction14 as described elsewhere.

4.2 1“N Modulation and Computer Simulation

Computer simulations of the 1“N ESEEM data were done using the density
matrix formalism developed by MimslSv16 together with the angle-selection protocol

described by Hurst, Henderson and Kreilick17’18 for Cu+2 ENDOR spectroscopy.
The CW-EPR spectrum of a frozen solution of Cu+2 complexes consists of
contributions from all possible orientations of the metal's magnetic axis with respect
to the external magnetic field. In ESEEM, the external magnetic field is fixed and
hence the simulation starts with determination of the various molecular orientations

that satisfy the resonance field condition

B :hv - MIA(0, ¢)
Be g(0r¢)

 

(4.1)

where h is the Plank's constant, v is excitation frequency, MI is the metal's nuclear

spin quantum number, 0 and (l) are angles that describe the orientation of the external

142

gz

 

3y

 

 

 

Figure 4.1. Reference coordinate system for ESEEM calculations. The rotation
angle 0 and the azimuthal angle 4) describe the direction of the static magnetic field
with respect to the g PAS. The nuclear position of the remote nitrogen of imidazole
is defined by 0N and (DN.

magnetic field with respect to the g-tensor principal axis system (Figure 4.1), and

g(0,¢) and A(0,¢) satisfy the following condition,

 

3 1/2
g(0.¢)=[‘ (grin-)2]
[2 2 (Ajigl'hi )2]
A 0, = : i=1 j=l
( ¢) g(B.¢)

where the h,.s are the direction cosines. Explicitly, the direction cosines in the
spherical coordinate system are, h,= cos¢sin0, h2 = sin¢cosG, and h3 = c080.
Equation 4.1 tells us that for each set of (0,4)) at a given M, , there is a field at which
resonance can occur. A pictorial explanation of equation (4.1) is shown in Figure 4.2.

At fixed B, a curve S is traced on the surface of the sphere. S represents all the

143

82 (B)

 

(B)

(C)

(D) (A)

 

Figure 4.2. Plot of constant g factors (rhombic symmetry case, g, < g2 < g,) in a unit
sphere showing correspondence to a magnetic field position in the CW-EPR
spectrum.

144

molecular orientations having the same value of g but different values 0 and (1119,20
For an axial g and A and assuming their PAS are coincident, these molecular
orientations can easily be calculated because the dependence of Br on 4) vanishes. A
graphical solution, for the case of frozen solution of Cu(H) complex is shown in
Figure 4.3. It should be noted that an ESEEM spectrum, characteristic of a single
crystal, can be obtained if data are obtained at lower field corresponding to transitions
involving only the M,=3/2 sublevel.

The molecular orientations that satisfy equation 4.1 are then used to calculate
the ESEEM modulation function. From chapter one, the general stimulated echo
envelope modulation function is given by,

Emod(r,T) = [2(21 + 1)]'1 Tr{GCl + Gs} (4.2)

where G, = Q1+M+PT+PJ MQM+PTPM

and Gfl = Q:QT*M*P.*MQ.QTM*P.M.

In equation 4.2, Q,P and M are matrices of dimension 21 + 1, with P and Q being
submatrices of the rotation operators that describe the time evolution of the density
matrix during the free precession periods for the two electron spin manifolds. The
modulation depths of the frequencies contained in P and Q are given by the elements
of M where, M = Ma’MB. Ma and M, are matrices that diagonalize the spin
Hamiltonian describing the superhyperfine splittings in the M,=1/2 and -1/2 electron
spin manifolds, respectively.

For 1“N modulation the spin Hamiltonian that describes the superhyperfine
splittings consists of the nuclear Zeeman, electron-nuclear superhyperfine interaction

and the nuclear quadruple interaction.

145

100

20’

C
80
60-

40*

 

 

A)

B)

J l

2400 2600 2800 3000 3200 3400
Magnetic Field (Gauss)

Figure 4.3. Discrete orientations defined by 0 as a function of external magnetic
field for copper complex with axial g-tensor. The EPR spectrum is generated with g"
= gy= 2.05, gz = 2.25 , A, = Ay = 35 MHz and A2 = 586 MHz Hamiltonian

parameters.

146

.74,=—gnfinB-I+SO_A_OI+IO(_2-I (4.3)
In the calculation, the hyperfine tensor is restricted to be axial and the elements are

given by

A.. : fliqugiGninj _ 60)] + A (3.21

u hr} iso l)

 

(4.4)

where n, = costl)N sin0N , n2 = sintersinGN and n3 = cosBN. The angles 0,, and (I),J
specify the orientation of the l“N-hyperfine tensor principal axis with respect to the g-
tensor while r, is the effective dipolar distance. Assuming only the hyperfine
interaction accounts for the echo modulation, the fundamental modulation
frequencies can be derived to be equal to,

1/2
3 3
w(B,mS)=[2[g—(r;fi3[2gjthfl)—h,wo]] , (4.5)

i=1 j=i

gNfiNBo '

where too =
h

In its principal axis system, the 1“N nuclear quadrupole interaction can be

written as,
2
ya] 4%2—[315 -2 +170: 43)] (4.6)

To describe this interaction and relate the orientation of its PAS to that of the g-tensor
five parameters are needed. These are the quadrupole coupling constant(e2qQ);
n (asymmetry parameter) and the three Euler angles (0t,B,y).

After the appropriate transformation of the hyperfine and quadrupole tensors
to the g-axis system, the Hamiltonians in the alpha and beta electron spin manifolds

are separately diagonalized to give the eigenvalues and eigenvectors. The

147

eigenvalues are used to compute the elements of P and Q while the eigenvectors
which determine the intensities are used to calculate the elements of M. With P,Q,
and M known, the modulation function at a certain M,, Em, M,, can be calculated. It
is given by the equation,

Em, M. (er) = JISEmod(r,T,S)dS (4.7a).

S can be expressed as a function of g and (I) , S(g,¢), so that
21:
dS
B,,,M,(r.r)= j B,,,(mowdcr. (Mb)
0

(dS/d¢)d¢ gives the probability the paramagnetic assumes an orientation at a given g.

Consideration of distribution in field, (4.7b) expands to,
BO+ABZn (18
EM, M, (er) = j Immune), R(B)—dodB (4.8)
B,,-AB 0 CW}
R(B) in equation is a gaussian lineshape function. At a static magnetic field
corresponding to g ,, transitions involving all M,'s of copper maybe possible so that
the resultant Em, consists of four line integrals of the form described by (4.7b). The
resulting modulation function is then multiplied by the decay function , exp[-(r/t)“]

where values of r, t and n were chosen such that the decay of modulation of our

simulation is similar to the experimental time-domain data.

4.3 Comparison of Nuclear Transition Spectra
Prior to using the ESEEM angle-selection program, its output was compared
to simulations of spectra obtained by other methods. The purpose of this exercise

was to test the output of our orientation selection ESEEM simulations program by

148

comparing it to the predictions made by the more established routines developed for
ENDOR spectroscopy. An ENDOR powder spectrum for an assumed S = 1/2, I =1
spin system described by the Hamiltonian parameters: g, = 2.261, Aiso = 4.0, r,,, =
2.6A, B0 = 3200 Gauss, v = 9.2 GHz, is shown in Figure 4.4A. Two powder
lineshapes centered about the nuclear Larmor frequency of 5.52MHz are predicted.
The lineshapes are of the familiar "Pake" type that result from the sine weighting of
the electron magnetic moments. The most intense portion of each Pake lineshape
occurs when 0 = 90°, or when the electron magnetic moments are perpendicular to the
lab field. The frequency difference between those perpendicular features yields the A ,
principal element of the hyperfine tensor. If the electron magnetic moments are
aligned with the lab field, 0 = 0°, and one observes the largest splitting in the ENDOR
spectrum of All = 7.1MHz.

A corresponding powder ESEEM spectrum is shown in Figure 4.43. Just as
with the ENDOR spectrum, one finds two broad peaks, from the two electron spin
manifold, centered at the nuclear Larmor frequency. The ESEEM lineshapes differ

from the ENDOR powder spectrum pattern in that they are weighted by EPR

transition probability products.22 Because the branching of transition is most favored
for 0 = 45° and tails of to zero at 0 = 0° and 90°, a more complex response is
predicted. Now if g-anisotropy is added to this problem so that powder averaging of
the spin system is no longer appropriate, the ESEEM simulation of Figure 4.4C-E
are obtained. For these simulations an axial g-tensor with principal elements of g,“ =
gyy = 2.25, and gzz = 2.05 were used. Figure 4.4C shows the output obtained near g =

2.05, where only the 0 = 90° orientation is sampled. At g = 2.25, the 0 = 0°

149

Figure 4.4. Comparison of different nuclear transition spectra. A) ENDOR spectrum
calculated by a Matlab program Spint1.mat, B) and ESEEM spectrum calculated by
orientation averaging using the Fortran program SESPN l, C) ESEEM spectrum at a
discrete orientation, 0 = 90°, D) ESEEM spectrum at a discrete orientation, 0 = 0°, and E)
ESEEM spectrum at a discrete orientation, 0 = 45.° C-D spectra are calculated using the
Fortran program ANGSESPNI. All spectra are generated using the same magnetic
parameter for an assumed S = 1/2, I = 1 system: gn = 2.26, Aiso = 4.0, reff = 2.6A, B0 =
3200 Gauss, v = 9.2 GHz, 1 = 200ns an T = 40ns. The g principal values used to
calculate spectra C, D and E are gx = 2.25, gy = 2.25 and g2 = 2.05.

150

 

 

 

 

 

 

 

 

 

 

 

,JpLL

 

 

 

 

 

 

 

 

 

  

 

 

 

 

4‘ i 1 ‘
frequency (MHz) 0

151

orientation is studied and these results are shown in Figure 4.4D. The 0 = 45°
orientation is resolved at g = 2.15 and the predicted ESEEM is provided in Figure
4.4E. Part of the inequivalence in amplitudes for the peaks predicted in the ESEEM

spectrum stems from the r-suppression behavior of three-pulse ESEEM.

4.4 Results and Discussion
4.4.1 CW-EPR

X-band CW-EPR spectra of Fet3p and its variants are shown in Figure 4.5.
Figure 4.5A shows the spectrum of wild type protein with both type 1 and type 2
centers contributing. Figure 4.5B and 4.5C are of T2D Fet3p [spectrum of the type 1
Cu(II) only] and TlD Fet3p [spectrum of the type 2 Cu(II) only], respectively. The
spectra are characterized by axial g tensors with no hyperfine structure being
observed at g ,. Both mutants lacked enzymatic activity. These and other data
demonstrate that the sequence homology upon which the mutagenesis was based
correctly predicts the arrrino acid ligands to the type 1 and type 2 copper atoms in
Fet3p. The results of Figure 4.5C are supported by the visible absorbance data that
have also demonstrated that Cys484Ser mutation resulted in the loss of the type 1
Cu(n).23

These CW-EPR data were analyzed using the QPOWA program. In Fig.

4.5B, the spectrum of the type 1 Cu(II) site is simulated (dotted line) using the

parameters, g” = 2.050; g i = 2.198; A” = 271 MHz and A 1= 25 MHz, while the type

2 Cu(II) spectrum is simulated in Fig. 4.5C (dotted line) using the values g” = 2.054;

g 1 = 2.250; A” = 586 MHz and A ,= 35 MHz. The observed principal g values (gll >

152

Figure 4.5. Continuous-wave EPR spectra of (A)wild-type Fet3, (B) type 2 depleted
(T2D) Fet3 and, (C) type 1 depleted (TlD) Fet3. For (A), measurement conditions were:
microwave frequency, 9.478GHz; microwave power, 2mW; modulation frequency,
100kHz; modulation amplitude, 10G; temperature, 16K. For (B), measurement
conditions were: microwave frequency, 9.481GHz; microwave power, 0.998mW;
modulation frequency, 100kHz; modulation amplitude, 5.054G; temperature, 16K. The
simulated spectrum (dotted line) of (B) was obtained using gx = gy= 2.050, g2: 2.198,Ax
Ay = 25 MHz and AZ: 271 MHz. For (C), measurement conditions were: microwave
frequency, 9.481GHz; microwave power, 0.998mW; modulation frequency, 100kHz;
modulation amplitude, 5.054G; temperature, 16K. The simulated spectrum (dotted line)
of (C) was obtained using g,( = gy = 2.054, g z= 2.250, Ax: Ayz 35MHz and AZ:
586MHz.

153

 

A)

 

 

 

 

2000 2500 3000 3500 4000
field (gauss)

 

B ) I l W... l

....... ...LM

 

 

 

L

2000 2500 3000 3500 4000
field (gauss)

 

 

 

 

 

 

2000 2500 3000 3500 4000
field (gauss)

154

g i > 2.0023) of both type 1 and type 2 copper sites indicate that at each copper the

2

unpaired electron primarily occupies the 3d,, -y2 atomic orbital on the metal.24’25 The

smaller g values of the type 1 Cu(H) in relation to the values for the type 2 Cu(II),
particularly g“, is a result of the strong covalent bonding between Cu+2 and the

cysteine-sulfur ligand at the type 1 site. This covalent bonding leads to more electron
delocalization onto the ligand and consequently the metal hyperfine coupling constant

decreases.

4.4.2 l“N ESEEM
4.4.2.1 TlD Fet3p

Figure 4.6 and 4.7 shows the three-pulse l4N-ESEEM data collected for TlD
Fet3p at magnetic field positions in the g” (B,,: 2690 Gauss) and g i (B,,: 3050 Gauss)
regions, respectively. At both field strengths, the frequency-domain spectra show
sharp low frequency peaks at ~0.6, ~0.9 and ~1.5 MHz. A broad peak that shifts
from ~36 to ~42 MHz as the magnetic field is varied from 2690G to 3050G is also

observed. The spectra are similar to those obtained for ascorbate oxidase (A0) at low

field where only the type 2 Cu(H) contributes to the envelope modulation function.26
The overall appearance of the experimental modulation pattern are similar to

those obtained by ESEEM studies on model copper complexes and copper proteins

exhibiting 14N modulation arising from the interaction of the unpaired electron spin of

Cu(H) and the remote (non-coordinated) nitrogen of an equatorially coordinated

imidazole.2’4’8 Furthermore, the appearance of these spectra is typical of 1“N nuclei

characterized by an isotropic hyperfine coupling that is approximately equal to twice

155

400

 

A)

N (a)
O O
O O
l l
l 1

lntergated echo
3
O
l
l

 

y

 

 

-100

 

0 2000 4000 6000 3000 1 10‘
T + 1: (ns)

 

B)

 

Intensity

 

 

 

L l 1

2 4 6 8 10
frequency (MHz)

I-

 

Figure 4.6. Representative experimental "N-ESEEM spectra for the type-2 copper
(TlD Fet3). Spectrum (A) was collected at B,,: 2690 Gauss and for 1: = 260ns.
Spectrum (B) is the cosine Fourier transform of (A). Other measurement condition
are: microwave frequency, 8.78 GHz; microwave power (43 dB); sample
temperature, 4.2 K.

156

 

2000 , , , ,
A)

.5
0‘
O
O
l

l

..a
O
O
O
F
1

500_ _

lntergated echo intensity
O
l
1

 

-soo _. L’ _

0 2000 4000 6000 8000 1 10‘
T + 1(ns)

 

 

 

 

B)

Intensity

 

 

 

l J l 1

2 4 6 8 1 0
frequency(MHz)

 

Figure 4.7. Representative experimental l“N-ESEEM spectra for the type-2 copper
(TlD Fet3). Spectrum (A) was collected at B0 = 3050 Gauss and for r = 230ns.
Spectrum (B) is the cosine Fourier transform of (A). Other measurement condition
are: microwave frequency, 8.78 GHz; microwave power (43 dB); sample
temperature, 4.2 K.

157

The nuclear Larmor frequency. The consequence of this condition is depicted in
Figure 4.8. Because the nuclear Zeeman and hyperfine interactions cancel one
another in one of the electron spin manifolds, the electron spin energy levels are split
by the 14N nuclear quadrupole interaction (NQI). Consequently, three sharp lines are
observed in the low frequency region of the spectrum. In the other manifold where
the nuclear Zeeman and the electron-nuclear hyperfine interaction are additive, a
single, broad double quantum transition is expected at about four times the nuclear

Zeeman frequency.

This condition of "exact cancellation";27 enables us to estimate the
quadrupole coupling constant, eZqQ, and the NQI asymmetry parameter, 1], by

assigning the three low-frequency peaks to the zero-field N QI resonance frequencies:

v = equUr)
° 2
v_ = equ(3 - n)
4

v = equ(3 + n)
+ 4 '

From the positions of the experimental peaks and using the above equations we
obtained our first estimate of r] = 0.75 and equ = 1.60 MHz. The position and shape
of the broad peak near 4 MHz are determined by the symmetry and strength of the
hyperfine interaction, and by the relative orientation of the hyperfine and NQI

tensors.

158

Figure 4.8. (A) Energy diagram describing 'exact cancellation condition' for I = 1. (B)
Typical ESEEM spectrum of 4N following exact cancellation condition.

159

 

A) Ms M

 

 

 

 

' |1>
." ’N~~‘ ."-‘V‘ I
O. —| o
|-1/2> l0> x I
s - - - fi~ o. v
v_ +
..o '1> vo“’ ...m 7 ’

 

 

 

Nuclear
Hyperfine Quadrupole

 

Electron Zeeman Nuclear Zeeman

 

 

B) V0

 

 

 

 

160

 

Figure 4.9 and 4.10 illustrate various simulations of the l4N-ESEEM data
obtained for the type 2 copper site in Fet3p. Figure 4.9 shows comparisons between
experiment (solid lines) and theory (dashed lines) for data collected at g", while a
parallel set of comparisons is made for g i on Figure 4.10. In spectra 4.9a and 4.10a,
our calculations assumed one equatorially bound histidine imidazole ligand having an
isotropic hyperfine coupling constant (Am) of 1.75 MHz and effective dipole-dipole
radius, reff of 3.0A. In contrast, simulations 4.9(4. 10)b-d were generated with the
assumption that there are two equatorial imidazoles bound to the type 2 Cu(H).
Simulations 4.9b and 4.10b assume two magnetically equivalent imidazoles while c
and (1 consider two magnetically inequivalent equatorial imidazole ligands. All of
these simulations show minor peaks that range from 2.2 - 3.2 and that arise from
linear combinations of the narrow peaks resolved at 0.6, 0.9 and 1.5 MHz. These
combination lines are a consequence of the product rule and were not resolved in our
ESEEM study of the type 2 site of Fet3p. The simulations of Figure 4.9/4.10b-d were

generated for two reasons. Recent X-ray crystallographic studies of two other
multicopper oxidases, ascorbate oxidase (AO)28:29 and human ceruloplasmin

(th)30, identified two histidines close to the type 2 copper sites in these proteins.
Furthermore, sequence analysis indicated the presence of a homologous type 2 copper
site in Fet3p.31’32 Indeed, this analysis was used to design the mutagenesis strategy
that did, in fact, result in the successful production of the TlD and T2D mutants used
here and elsewhere.23 Our simulations, however, indicate that only one histidine
imidazole is coordinated equatorially to the type 2 copper in Fet3p despite the

likelihood that both histidines are present at that site.

161

     
     
     
     
     

lN/ Ab=1.75 MHz

2N/Ah=l.75Il.75MHz

2N/A.m=l.75,1.22 MHz

2N/Ah=l.75,1.05 MHz

2N] Ah=1.75 MHz(equatorial)
A.” = 0.52 MHz (axial)

I i I I I I I I I J I I I l I I I I L J

2 4 6 8 10
frequency (MHz)

Figure 4.9. Determination of number of histidyl imidazole ligands for the type-2
center of Fet3. The simulation (dotted lines) models and parameters are listed in
Table 4.1. The stimulated-echo ESEEM (solid lines) and simulations (dotted lines)
were collected at g = 2.33 using a r-value of 260 ns. Other common experimental
conditions common to the ESEEM measurements are given in Figure 4.6. The
simulation given for (a) considers only one coupled nitrogen, those shown for
simulations of (b)-(e) are for 2 coupled nitrogens with Hamiltonian parameters being
given in Table 4.1.

162

g = 2.06

IN] Ah=1.75 MHz

 

'

2N/Ah=l.75/l.75MHz

 

2N/Ah=1.75,l.22 MHz

 

'

2N/Ah=1.75,1.05 MHz

 

v

2N] Ah=l.75 MHz(equatorial)
A,” = 0.52 MHz (axial)

 

 

2 4 6 8 1 0
frequency (MHz)

Figure 4.10. Determination of number of histidyl imidazole ligands for the type-2
center of Fet3. The simulation (dotted lines) models and parameters are listed in
Table 4.1. The stimulated-echo ESEEM (solid lines) and simulations (dotted lines)
were collected at g = 2.06 using a 'c-value of 230 ns. Other common experimental
conditions common to the ESEEM measurements are given in Figure 4.7. The
simulation given for (a) considers only one coupled nitrogen, those shown for
simulations of (b)-(e) are for 2 coupled nitrogens with Hamiltonian parameters being
given in Table 4.1.

163

0.8

lntergated echo

0.8

0.4

lntergated echo

0.2

0.4

lntergated echo

0.2

 

 

 

 

 

 

 

 

 

 

 

 

(A) 1N, A; = 1.75 MHz 1
”_ "1
o 2000 4000 6000 3000 1 10‘
T + 1: (ns)
(B) 2N, A; = 1.75 a 1.75 rim:
.; -
o 2000 4000 6000 8000 1 10‘
T + r (ns)
(C) 2N, A__‘= 1.75 MHz (equatorial)
. ' A_ - 052 Mil: (axial)
o 2000 4000 0000 8000 1 10‘
'l‘ + 1 (ns)

Figure 4.11. Time-domain computer simulations (dashed lines) of the normalized
l“N ESEEM data collected for TlD Fet3 protein at g = 2.33. Trace A is for one
equatorially coordinated histidyl imidazole ligand (Am: 1.75 MHz and reff = 3.0 A).
Simulation B assumes two magnetically equivalent histidyl imidazole ligand. Trace
C assumes one equatorial and one axial (Am: 0.52 MHz and reff = 3.2 A) histidyl q
imidazole ligands. Other parameters are listed in Table 4.1.

164

lntergated echo lntergated echo

lntergated echo

 

(A) 1N, A.;: 1.75 MHz

 

0.2.. :

 

 

 

o 2000 4000 6000 6000 1 10‘
T +1 (ns)

 

(B) 2N, A_'= 1.75 a 1.75 6'le

 

 

 

 

0 2000 4000 6000 6000 1 10‘
T +1 (ns)

 

(C) 2N, A.;: 1.75 MHz (eqilalorial)
A_ = 052 M11: (axial)

    

0.6

0.4 _ _

0.2: _

 

 

 

o 2000 40T00 6000 6000 1 10‘
T +1 (ns)

Figure 4.12. Time-domain computer simulations (dashed lines) of the normalized
l4N ESEEM data collected for TlD Fet3 protein at g = 2.06. Trace A is for one
equatorially coordinated histidyl imidazole ligand (Am: 1.75 MHz and r,ff = 3.0 A).
Simulation B assumes two magnetically equivalent histidyl imidazole ligand. Trace
C assumes one equatorial and one axial (Am: 0.52 MHz and r,_.ff = 3.2 A) histidyl
imidazole ligands. Other parameters are listed in Table 4.1.

165

 

 

 

 

 

 

 

.386 5.8.8 8.8.5 5.8.8 5.8.8 5.8.8 5.8.8 5.8.8 - 5.8.8 23.5
85 8.0 m8 m3 3o 86 m3 m8 - 86 =
m3 ”3 no: 8.: m2 8: 8: 8.: - 8.: Oman
.c..o .98 .o..o 5.8 8.8 .o.8 8.8 .98 - .o.8 29.6
8.... an an 3 an ca 8m 3 - QM 968:7
moo n2 8.: m: 8: m5 m5 m2 - n5 Agiél
NZ ~Z NZ _Z NZ _ Z NZ ~Z NZ _Z «and:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

88:8 25893 05 com 588.8853

166

25 vosmzmfioncoa 29 98 2¢ .838 588% 33:0 of 9 8888:: 2823:8823: of mo E89? $8 RES—ca 05
822 85 8&5 83m 2: Ba > 98 d8 .33. was ad .wwb Gum Co 528 N 093 2: 2 cameos 35.830 85%: 36sz
do 898:: .6 cocaine—Bot 05 8 BER macaw—:86 Emmanu— ocoo Conn—:85 05 88 882523 25893 2wa 3. 035.

 

 

 

 

8
s _ _
53‘
'53
5 4 1:: 260ns _
*5 B =269OG
"‘ 2 l _
O ...
’ 2 i 3 6 é 1 0

frequency (MHz)

 

 

 

 

 

 

 

>1

H
'63

s:

3

:
~

é ‘4 s 6 1 0
frequency (MHz)
1 2 _ q

53'

’8

s:

O

H

:-

~

 

 

 

 

 

 

frequency (MHz)

Figure 4.13A. TlD Fet3 ESEEM data and computer simulations at B0 = 26906 for
different 1.” values: A) 260ns, B) 350ns, C) 43 5ns. Hamiltonian parameters are the
same as in Figure 4.9a in which one magnetic l4N nucleus couples with the electron

spin.

167

 

s
3‘
'5
54
E
"'2

 

 

 

 

 

 

' 2 i :1 is 6 1 0
frequency (MHz)

 

 

 

 

 

 

 

14 T
1 2 — _
>. 10 7. _
.5 8 4'3: 1:310ns .
5 6 5:3; Bo=3050G
:5 _. ' ‘
4
2
0
' 2 i 3 is a 1 o
frequency(MHz)
1 4
1 2 r— ..
i’.’ _
.5 “B38508 j
5 B =3050G
‘5 ° ‘*

 

 

 

 

 

A L

 

 

 

4 6
frequency(MHz)

Figure 4.13B TlD Fet3 ESEEM data and computer simulations at B0 = 3050G for
different 1: values: A) 230ns, B) 310ns, C) 385ns. Hamiltonian parameters are the

same as in Figure 4.10a in which one magnetic 14N nucleus couples with the
electron spin.

168

Intensity

Intensity

Intensity

 

 

 

 

 

 

 

 

A )
1.’ = 225ns _
B0 = 3140C
' ‘1
”' T
4 13 s 10
frequency(MHz)
B )
_ 1: = 300ns ‘
B0 = 3140C
4 is 5 1 o
frequency(MHz)

 

 

 

  

 

 

”l-

4 6
frequency(MHz)

169

10

Figure 4.13C. TlD Fet3 ESEEM data and computer simulations at B0 = 31406 for
different “C values: A) 225ns, B) 300ns, C) 375ns. Other Hamiltonian parameters are
the same as in Figure 4.13 A and B in which one magnetic 14N nucleus couples with
the electron spin.

Axial histidine imidazoles have a very weak hyperfine interaction (less than 1

MHz for the directly coordinated nitrogen) so they yield only a shallow modulation

that is easily obscured by 1“N at exact cancellation.33 Therefore, the contribution of
an axial histidine ligand to the modulation of the electron spin magnetization at the
type 2 Cu(II) would be obscured if an equatorially bound histidine was present as
well. To test this inference, a simulation was performed for one axial and one
equatorial histidine ligand (Fig. 4.9/4.10e). As expected, simulations a and e show
only modest differences. These simulations indicate that if a second histidine is
present as an inner sphere ligand at the type 2 Cu(II) in Fet3p, it is most likely axially
coordinated. One of these ligands would be His81, since its substitution in the T2D
mutant does result in loss of the copper atom. The other is most likely His4l6, based
on the sequence alignments with the other multicopper oxidases. A test of this
prediction by mutagenesis has not been performed.

Spectra obtained at different t's for each field are shown in Figures 4.13A-C.
These spectra show the effect of the choice of “c on the intensity of the low frequency
peaks. The 1.6 MHz peak grew in intensity with respect to other peaks as t was
increased. This trend is true for three Bo setting. Simulation curves (dotted lines)

replicate the this trend, thus adding confidence to our simulation parameters.

4.4.2.2 T2D Fet3p
l4N-ESEEM data for the type 1 Cu(II) site were collected from T2D Fet3p
samples. These are shown in Figure 4.14 and 4.15. Spectra taken at g" (28256,

Figure 4.14) show peaks at 0.4, 0.7, 0.8, 1.4, 1.6, 3.2 and 3.5 MHz. In the gl region

170

1500

 

 

 

 

 

 

A) l T H I l
O
£1000 _W~ _.
O
Q
E
w 500 _ ..
l-
0
H
I:
F1
0 _ _
o 2000 4000 6000 soon 1 10‘
T + 1: (ns)
5 I I I I
B)
4 _ _
3

intensity

 

 

 

I
N
.—
p-
r-
.—

 

frequency (MHz)

Figure 4.14. Representative experimental l“N-ESEEM spectra for the type-1 copper
site (T2D Fet3). Spectrum (A) was collected at B0: 2825 Gauss and for 1: = 250ns.
Spectrum (B) is the cosine Fourier transform of (A). Other measurement conditions:
microwave frequency, 8.836 GHz; microwave power, 43 dB; sample temperature,
4.2K.

171

 

6000_.B) -

5000r .

 

4000 __ _

3000 _ _

2000 _ _

lntergated echo

1000_ _

 

0_ L-J

 

 

 

o 2000 4000 6000 8000 1 10‘
T + 1.’ (ns)

 

B)

intensity

 

 

 

h

.2 1 1 1
2 4 6 8 10

 

frequency (MHz)

Figure 4.15. Representative experimental l“N-ESEEM spectra for the type-1 copper
site (T 2D Fet3). Spectrum (A) was collected at B0: 3090 Gauss and for 1: = 230ns.
Spectrum (B) is the cosine Fourier transform of (A). Other measurement conditions:
microwave frequency, 8.836 GHz; microwave power, 43 dB; sample temperature,
4.2K.

172

(3090G, Figure 4.15), l4N-ESEEM the peaks are found at 0.508, 1.4, 1.6, 3.2, 3.6
and 4.1 MHz.

The number of ESEEM frequencies was greater than what one would expect
for either a single equatorially-bound histidine imidazole, or for two magnetically
equivalent ones. Therefore, simulations were based on the assumption that there are
two magnetically inequivalent histidine ligands at the type 1 site in Fet3p. Generally,
type 1 or blue Cu(II) sites, feature two histidine and one cystiene ligands that conform
to a trigonally distorted tetrahedral array, and differ only in their “axial” ligation.
Assuming that the observed modulation intensities require a coupling near exact
cancellation, we were able to estimate two sets of quadrupole coupling constants,
eZqQ and r]. The sharp peaks at 0.4, 0.8 and 1.4 MHz observed at 2825 G were
considered as a first set of quadrupole peaks, while the frequency components at 0.7,
0.8 and 1.6 MHz comprised the second set. The position and the width of the double
quantum peak were estimated by adjusting Aiso and the reg. Both the relative
amplitudes of the lower peaks and the appearance of the additional features in the
double quantum peak are very dependent on orientation of the lab field with respect
to HFI tensor and the N QI tensor principal axes.

Figure 4.16 shows the best simulations (dotted lines) of l“N—ESEEM data for

the type 1 Cu(II) site in T2D Fet3p at both the g“ and g 1 regions. The proposed

histidine imidazole ligands (designated Hisl and HisZ) have the same hyperfine
tensor principal values (Aiso (1.47 MHz) and rcff (2.50 A)) for their remote nitrogen.

They differ only in their tensor orientation with respect to the Cu g-tensor and NQI

parameters. For Hisl, equ = 1.35 MHz and n = 0.55, while for H182, equ = 1.60

173

 

 

 

l l l l

2 4 6 8 1 0
frequency (MHz)

 

 

 

 

 

l l 1 l

2 4 6 8 1 0
frequency (MHz)

 

Figure 4.16. Determination of number of histidyl imidazole ligands for the type-1
center of Fet3. The stimulated-echo ESEEM (solid lines) (A) were collected at g =
2.23 using a 't-value of 250 ns, while that of (B) were collected at g = 2.04 using a I-
value of 227 ns. Other common experimental conditions common to the ESEEM
measurements are given in Figures 4.14 and 4.15. The Hamiltonian parameters used
to calculate the simulation (dotted lines) curves are given in Table 5.2.

174

 

 

 

 

.-
L
l
..
p.

frequency (MHz)

Am a,
1.... 1’

A J L A A A l I L L l l J I I

i A 4 6 s 1 0
frequency (MHz)

B)

 

 

 

 

Figure 4.17. Computer simulations for single nucleus, 3) N1 and b) N2 for static
magnetic fields, 2825G(A) and 309OG(B). Spectra c in A and B are ESEEM product
of ESEEM fiinctions of N1 and N2. Hamiltonian parameters used to generate a and b
are found in Table 4.2.

175

Table 4.2. Ligand hyperfine parameters for stimulated ESEEM simulations for the
histidine ligands of the type 1 site in Fet3. The simulations are shown in Figure 4.16.

 

 

 

 

 

 

 

ligand N 1 N2

A,,0(MHz) 1.47 1.47

reg/3.) 2.50 2.50
ON, 0,. i71°,180° :97°,77°

equ 1.35 l .60

n 0.55 0.80
0143,). 70°,90°,0° 0°,90°,60°

 

 

 

 

 

MHz and 11 = 0.80. These NQI parameters are assigned to the uncoordinated l4N of
the histidyl imidazole ligands. X-ray crystal structures of multicopper oxidases,
AO,28,29 th30 and laccase34 and mononuclear blue copper oxidases,
plastocyanin35, and azurin36'38 indicate that the type 1 copper is coordinated by the
”N5 of the histidyl imidazole. We assume this to be the case for type 1 copper site in
Fet3p as well. Previous studies showed a small variation in the NQI parameters of
type 1 copper centers (equ=1.40-1.49 MHz, 1] = 0.90-.94) as compared to those
observed in type 2 copper centers (equ = 1.4 -l.8, n = 0.48-0.98).39’4O It has been

suggested that these 1“N difference may arise in part from the inequivalence of

histidyl imidazole ligands when the remote nitrogen is of the 5 versus 8 type.39

However, numerical replication of the ESEEM data on T2D Fet3p was unsuccessful

176

using the modest range of NQI coupling parameters that typify 8 versus 8 nitrogens.
More likely, the difference in NQI parameters measured for the two histidyl
imidazole ligands of T2D Fet3p stems from the extent and/or strength of the

hydrogen bond involving the distal N-H of the ligands. Based on the study by Jiang

et al.40, we would expect the NH of HisZ (e2qQ = 1.60 MHz, 11 = 0.80) to be
hydrogen bonded more strongly than the equivalent N-H of Hisl (e2qQ = 1.35 MHz,
11 = 0.55).

Recent X-band ESEEM data on single crystals of azurin from Pseudomonas

aeroginosa revealed two sets of NQI parameters, e2qQ = 1.43 MHz and n=0.95 for

Hisl 17 and e2qQ=1.37, n=0.86 MHz for His46.41 The X-ray-crystal structure of

this protein revealed that His46 is hydrogen bonded to the carbonyl oxygen of Asp10,

whereas Hisl 17 is hydrogen-bound to a water molecule.371~°’8’42 Hisl 17 is protuding
through a surface in a shallow depression, called the hydrophobic patch, and is
experimentally proven by site-directed mutagenesis to be involved in electron
transfer.

The imidazole rings of the histidine ligands at the type 1 Cu(II), predicted to
be His4l3 and His489, also differ in their orientation with respect to the Cu+2 g-tensor
principal axes. The data indicate that one imidazole has the principal axis of its
hyperfine coupling i20° from the xy plane (values of 0,: 70° or 109°, 0,, =180°), while
the other has this principal axis i7° with respect to the xy plane (values 0n=83° or 97°,
(1),, =77°). These orientation angles are consistent with the distortion from square-

planar geometry that is characteristic of the trigonally distorted tetrahedral geometry

177

B)
A)

 

 

 

 

 

frequency (MHz) frequency (MHz)

I B's-309°C
t-305n5

M
M b)
J

frequency (MHz) frequency (Ml-12)

 

 

 

 

 

 

Figure 4.18. 14N ESEEM data for T2D Fet3p (a) and computer simulations (b) at
various 1 values.

178

found at type 1 sites. In AO, the homologous His ligands (residues 445 and 512),
together with Cys507, are arrayed on one side of the trigonal plane. Met517 is the

apical ligand with its sulfur atom at the long apex of a distorted trigonal pyramid.

Fet3p, like laccase34, has a Leu residue at the homologous point in its primary
sequence and therefore is most likely to have a trigonal rather than trigonal pyramidal

type 1 copper site geometry.

4.4.2.3 Comparison of 1‘N Hyperfine Interactions at the Type 1 and Type 2
Cu(II) in Fet3p

The observed Aiso (1.47 MHz) for the type 1 Cu(II) is smaller than that for the
equatorially-bound histidine at the type 2 Cu(H) (Am: 1.75 MHz). Previous ENDOR
studies have developed an inverse correlation between the magnitude of the isotropic

hyperfine coupling due to the directly bound nitrogen and the amount of tetragonal

distortion of the complex towards a tetrahedral geometry.43 These studies showed
that a deviation from square planar geometry results in the reduction of the overlap of

the Cu+2 dxz-y2 orbital occupied by the unpaired spin with the coordinating nitrogen 0‘

orbital leading to a decrease of the hyperfine coupling constant. Colaneri et al.44
provided an explanation for this observed decrease in Am. Tetrahedral distortion
would result in a decrease of the spin density of the directly coordinated nitrogen o
orbital but at the same time an increase in the delocalization of spin into the 1:

imidazole ring or the C-N bond orbitals. Combined with the decrease in the overlap

of the nitrogen o and copper dXZ-y2 orbitals, these changes in the electronic structure

would result in the observed decrease in nitrogen isotropic hyperfine coupling.

179

Irrespective of the precise mechanism for this correlation, the difference in Aiso values
for the type 1 and type 2 sites in Fet3p conforms to the expected geometry of each
site. If Aiso for the directly coordinated nitrogen decreases with tetrahedral distortion,

we would also expect Aiso for the remote nitrogen to decrease.

4.5 Summary

Simulations of the MN ESEEM spectra of T1D and T2D Fet3p showed that the
observed modulations are due to remote nitrogen of histidyl ligands. In the TlD
enzyme, one equatorial histidine is coordinated to the type 2 copper site. An axial
histidine might also be present. Two magnetically inequivalent histidine ligands that
are disposed ~10° below the equatorial plane, could be present in the type -1 copper
site of Fet3p. These two histidine are different in their NQI values. The difference is

explained by the degree of hydrogen bonding involving the remote nitrogen.

180

List of References

1)Mims, W. B.; Peisach, J. Biochemistry 1976, 15, 3863-3869.

2)Mims, W. B.; Peisach, J. J. ChemPhys. 1978, 69, 4921-4930.

3)Mims, W. B.; Peisachs, J. J. Biol. Chem. 1979, 254, 4321-4323.

4)Mondovi, B.; Graziani, M. T.; Mims, W. B.; Oltzik, R.; Peisach, J. Biochemistry
1977, 16, 4198-4202.

5)Kosman, D. J.; Peisach, J .; Mims, W. B. Biochemistry 1980, 19, 1304-1308.
6)Fee, J. A.; Peisach, J.; Mims, W. B. J. Biol. Chem. 1981, 256, 1910-1914.
7)Zweier, J. L.; Peisach, J.; Mims, W. B. J. Biol. Chem. 1982, 257, 10314-10316.
8)McCracken, J .; Pember, 8.; Benkovic, S. J .; Villafranca, J. J .; Miller, R. J .; Peisach,
J. J. Am. Chem. Soc. 1988, 110, 1069-1074.

9)Hasset, R. F.; Yuan, D. S.; Kosman, D. J. J. Biol. Chem. 1998, 273, 23274-23282.
10)Nilges, M. J. ; Ph.D. Thesis; University of Illinois: Urbana, Illnois, 1979.
11)Belford, R. L.; Nilges, M. J. Computer Simulation of Powder Spectra; Belford, R.
L.; Nilges, M. J ., Ed.: Denver, Colorado, 1979. .

12)Maurice, A. M. ; Ph.D. Thesis; University of Illinois: Urbana, Illinois, 1980.
13)McCracken, J .; Sin, D.; Dye, L. Appl. Magn. Reson. 1992, 3, 305-316.

l4)Mims, W. B. J. Magn. Reson. 1984, 59, 291-306.

15)Mims, W. B. Phys. Rev. B 1972, 5, 2409-2419.

16)Mims, W. B. Phys. Rev. B 1972, 6, 353-3545.

l7)Hurst, G. C.; Henderson, T. A.; Kreilick, R. W. J. Am. Chem. Soc. 1985, 107,

7294-7299.

181

18)Hurst, G. C.; Henderson, T. A.; Kreilick, R. W. J. Am. Chem. Soc. 1985, 107,
7299-7303.

19)Poole, C. P.; Farach, H. A. The Theory of Magnetic Resonance; Wiley-
Interscience: New York, 1972.

20)Hoffman, B. M.; Martinsen, J .; Venters, R. A. J. Magn. Reson. 1984, 59, 110-123.
21)Hutchison, C.; B, M. D. J. Chem. Phys. 1977, 66, 3311-3330.

22)De Groot, A.; Evelo, R.; Hoff, A. J. Magn. Reson. 1986, 66, 331-343.
23)Blackbum, N. J .; Ralle, M.; Hasset, R.; Kosman, D. J. Biochemistry 2000, 39,

23 16-2324.

24)Atherton, N. M. Electron Spin Resonance; Ellis Horwood Limited: New York,
1973.

25)McGarvey, B. R. Electron Spin Resonance of Transition-Metal Complexes; in
Transition Metal Chemistry, Carlin, R. L. ed.; Marcel Dekker, Inc.: New York, 1966;
Vol. 3, pp 89-201.

26)Avigliano, L.; Davis, J. L.; Graziani, M. T.; Marchesini, A.; Mims, W. B.;
Mondovi, B.; Peisach, J. FEBS Letters 1981, 136, 80-84.

27)Flanagan, H. L.; Singel, D. J. J. Chem. Phys. 1987, 87, 5606-5616.
28)Messerschmidt, A.; Rossi, A.; Ladenstein, R.; Huber, R.; Bolognesi, M.; Gatti, G.;
Marchesini, A.; Petruzelli, R.; Finazzi-Agro, A. J. Mol. Biol. 1989, 206, 513-529.
29)Messerschmidt, A.; Ladenstein, R.; Huber, R.; Bolognesi, M.; Avigliano, L.;
Petruzzelli, R.; Rossi, A.; Finazzi-Agro, A. J. Mol. Biol. 1992, 224, 179-205.
30)Zaitzeva, I.; Zaitzev, V.; Card, G.; Moshkov, K.; Box, B.; Ralph, A.; Lindley, P. J.

Biol. Inorg. Chem. 1996, 1, 15-23.

182

31)de Silva, D. M.; Askwith, C. C.; Eide, D.; Kaplan, J. J. Biol. Chem. 1995, 270,
1098-1 101.

32)Askwith, C.; Eide, D.; Van Ho, A.; Bernard, P. S.; Li, L.; Davis-Kaplan, S.; Sipe,
D. M.; Kaplan, J. Cell 1994, 76, 403-410.

33)Comelius, J. B.; McCracken, J .; Clarkson, R. B.; Belford, R. L. J. Phys. Chem.
1990, 94, 6977-6982.

34)Ducros, V.; Brozowski, A. M.; Wilson, K. S.; Brown, S. H.; Ostergaard, P.;
Schneider, P.; Yaver, D. S.; Pedersen, A. H.; Davies, G. J. Nat. Struct. Biol. 1998, 5,
310-316.

35)Guss, J. M.; Freeman, H. C. J. Mol. Biol. 1983, 169, 521-563.

36)Nor1is, G. B; Anderson, B. E; Baker, E. J. Am. Chem. Soc. 1986, 2784-2785.
37)Baker, E. N. J. Mol. Biol. 1988, 203, 1071-1095.

38)Nar, H.; Messerschmidt, A.; Huber, R. J. Mol. Biol. 1991, 218, 427-447.
39)Goldfarb, D.; Fauth, J. M.; Farver, O.; Pecht, I. Appl. Magn. Reson. 1992, 3, 333-
351.

40)Jiang, F.; McCracken, J .; Peisach, J. J. Am. Chem. Soc. 1990, 112, 9035-9044.
41)van Gastel, M.; Coremans, W. A.; Jeuken, L. J. C.; Canters, G. W.; Groenen, E. J.
J. J. Phys. Chem. A 1998, 102, 4462-4470.

42)Colaneri, M. J .; Vitali, J.; Peisach, J. Biochemistry 2000, 39, 584-591. Based on
the ESEEM studies and X-ray crystal structures of some biological and model type 1
and type 2 copper complexes, Colaneri et al . suggested that for 11.2.0.9 the distal
nitrogen is hydrogen bonded to water molecule, while 11<0.9, the remote N is likely

hydrogen bonded to the carbonyl oxygen of the backbone or side chain of protein.

183

Further simulations was done with one of the histidines of the Fet3p assigned to have
n = 0.90 for its remote nitrogen. At one 1: value, experimental data were fairly
replicated for both magnetic field settings. However, at higher 1' the experimental
data were poorly matched.

43)Iwaizumi, M.; Kudo, T.; Kita, S. Inorg. Chem. 1986, 25, 1546-1550.

44)Colaneri, M. J.; Potenza, J. A.; Schugar, H. J .; Peisach, J. J. Am. Chem. Soc.

1990, 112, 9451-9458.

184

Chapter 5

2H ESEEM: Water Ligation to and Accessibility of the Type-1 and Type-2
Copper Sites in Fet3p

The nature ligands and their arrangement at the metal center in
metalloproteins determine the reactivity and function of the metal center. In chapter
4, the histidine ligation to type 1 and type 2 centers in the Fet3 protein was
determined. In these chapter, water ligation to these copper centers is considered.

The coordination of metal by an exogenous ligand, water and its accessibility
to water solvent can be determined by ESEEM. One way of determining this is by
making use of the attributes of the sum-combination (t)Cl +115) peak. Sum combination
peaks are resolved in two and four-pulse ESEEM because the electron spin echoes are

modulated not only by the fundamental frequency at the alpha and beta electron spin

manifolds but by the sum and difference of these fundamental frequencies. 1‘3 Non-
coordinating water has a 1H-ESEEM frequency spectrum characterized by
overlapping fundamental nuclear modulation frequencies, 1)., and 115, centered at
proton nuclear Zeeman frequency, DH and a narrow sum-combination peak centered at
21)”. Previous ENDOR/ESEEM studies on some metal-aquo complexes revealed that

the protons of covalently bonded water have hyperfine interactions that are highly

anisotropic.4'7 As a result of this anisotropy, 1),, and “B will give rise to powder
pattern lineshapes that span a range of ~20MHz about 1),, and the sum-combination

peak, (1)., + DB) will shift from 21)”. This anisotropic peak is shifted from the twice

the Larmor frequency of proton by an amount equal tog,

185

9A1
161),,

 

As(va+vfl)—2v,,= (5.1)

Equation 5.1 was derived for an isotropic g and axial hyperfine interaction where A 1
is the perpendicular component of the hyperfine tensor (A 1 = A,“ = A”). The absence
of the shifted peak, therefore, signifies absence of axially and/or equatorially
coordinated water to Cu(II).

Water coordination and accessibility can also be established by DZO-exchange

experiments?)10 Here, ESEEM data are collected for systems in D20 and H20 buffer
under similar conditions. Ratio of the ESEEM function of a system in D20 buffer
over a system in H20 is obtained to yield ESEEM function that is dominated by
exchangeable deuterium and protons. Because the nuclear g factor, go, of protons is
6.5 time larger than deuterium, the modulation frequencies from each nucleus-type
occur in regions that do not interfere with one another. This technique was used in
this study to determine water ligation and solvent accessibility of the type-1 and type-
2 copper sites in Fet3p. DzO-exchange experiment does not only allow us to identify
coordinated water but also assess the contribution of ambient/matrix water to the
ESEEM function. We were able to determine the nearest solvent water to metal
center and describe the distribution of ambient water that accounts for the observed
two-pulse quotient modulation function. The ratio envelopes suggest that both the
type 1 and type 2 sites are readily accessible to water. However, the type 2 is more
open than the type 1 site. The consequence of this difference will be discussed.
Moreover, water is an inner-sphere ligand to only the type 2 copper site with one

axial and one equatorial DZO. The 2H modulation function of the type 1 copper site is

186

due entirely to ambient deuterons with the nearest metal-to-deuteron distance being

3.7/1.

5.1 Experimental Procedures

Construction of F et3p mutants and production of soluble protein. Fet3p
variants in DZO buffer were prepared similarly as those in H20 buffer as described in
Chapter 4.

CW—EPR. Continuous-wave EPR spectra of the Fet3p and its variants were
obtained at X-band on a Bruker ESP-300E EPR spectrometer equipped with an
Oxford ESR-900 liquid helium cryostat.

ESEEM. Pulsed EPR experiments were performed on a home-built

spectrometer described previously.11 Two-pulse data were collected at X-band using
a reflection cavity that employed a folded stripline resonant element. All experiments

were performed at liquid helium temperature, 4.2 K.

5.2 2H Interaction and Simulation

Water ligation to type-1 and type-2 copper sites was determined using two-
pulse (1112-1—10 ESEEM measurements and the DZO-exhange protocol.9 These

experiments are based on the product rule1112 for two-pulse ESEEM where the
modulation function ( Emod) for a discrete paramagnetic center consists of a product of

the modulation functions of each nucleus coupled to the center.

Em..(IrJz--1.) = Em..(11) Em..(12)- .. Ema.) (5.2)

187

Division of the ESEEM data of DZO-exchanged enzyme by parallel data sets
collected for enzyme in aqueous buffer will yield a quotient function dominated by

the exchangeable deuterons.

— —-Em°d(2H) (5.3)

"‘°“ Emod(‘H) ’
The effect of ambient D20 on the modulation function was also investigated.
Ambient water is treated as statistically distributed in a series of non-coordination

concentric shells with rmin as the distance of water deuterons nearest to the metal
center. 13914 In our calculation, there are ten shells that are incremented in equal
steps of 1/r3. The contribution of ambient water can be approximated to be equal to,

E = V(r,,r)"' x V(r2,r)"2 x V(r3,t‘)"3.... (5.4)

mod,avg

where V(r,z') = 1:8...(0 r,6)d6.13’14 (5.5)
The r1i in the modulation function V represents the number of deuterium nuclei in an
ith shell and is assumed to be equal to (f)(pN)(r3,+i-r3i) where f is a factor whose value
depends on how the copper site is exposed to water and 9N is the mean number of
deuterium nuclei per unit volume which is calculated to be 0.067 D/A3. The resulting
average modulation function, Emodavg may then be incorporated into the modulation

function arising bound water ligands using the product rule.

Equations 5 .4 and 5.5 are derived using spherical model, an approximation to

the equations ,

(E...) = 291V... (5.6a)

(Emod) = zpjl—[VmodJ.j' (5.6b)
j 1'

188

The above equations are modulation functions for disordered systems such as frozen
solutions and powder samples with 5.6a describing one magnetic nucleus interacting
with the electron spin and 5.6b for more than one. In both equations, pi represents the
weighting factor or probability for a spin in the ith orientation with respect to a

reference axis. Equations 5.6a is equivalent to

27m

(EN) = i H V(6)sin 0d0d¢, (5.7a)
0

0

with pl. 2 ism BdBdt/t. Similarly we can write 5.6b as

2n I: n
(Emod") = i j d¢IHV(6)sin ado. (5.7b)
0 0 1

In order to calculate (5.7b), one needs to know the geometry of nuclei around the
paramagnetic center. The geometry can be approximated as spherical with the
magnetic nuclei randomly oriented along the sphere of radius rm. This random
arrangement allows the reversal of arrangement of the order of operation in (5.7b), so

that equation (5.7b) becomes,
(5,10,) = [i 1 81¢] we) sin ado] = [(Vm.)]" (5.8)
0 0

which is essentially equivalent to equation (5.4). In our calculations, the electron spin
is assumed to be quantized along the direction of the z-component of the external
magnetic field and the angle 0 represents the’angle made by vector r, the distance of
separation between the nuclear spin and electron spin, with the applied magnetic

field. The consequence of this assumption will be discussed.

189

5.3 Results

QW—EPR. Figure 5.1A,B, and C are CW-EPR of the wild type, T2D and TlD Fet3
protein in D20 buffer, respectively. The spectra are similar to the spectra observed
for each sample in H20 buffer. Therefore, no significant changes occurred in the

copper sites when the Fet3p samples were deuterated.

TlD FetBQ
The two-pulse ESEEM spectra collected at g i (B0: 3055 Gauss) and ‘r = 250

ns for TlD Fet3p in D20- and HzO-based buffers are shown in Figures 5.2A and
5.2B, respectively. The 2H contribution to these data can be obtained by dividing the
data of spectrum Fig. 5.2A by that of Fig. 5.2B. These ratioed values are given in
Figure 5.2C. Two periodicities are observed in the quotient envelope, one occuring
at ~500 ns corresponding to the second harmonic and the other at ~250 ns

corresponding to the fundamental or Larmor frequency of 2H.

T2D Fet3p
The ratio-envelope 2-pulse ESEEM data obtained at g 1 (B0 = 3090 Gauss) for

the type 1 Cu(II) site in T2D Fet3p is shown in Figure 5.3 C. It is was calculated by
dividing trace 5.3 A over 5.3 B. In contrast to TlD Fet3p, the modulation of the
quotient envelope of T2D Fet3p is shallow. Also, no apparent second harmonic is

observed.

190

Figure 5.1. Continuous-wave EPR spectra of (A) wild-type Fet3, (B) type 2 depleted
(T2D) Fet3 and, (C) type 1 depleted (T1D)Fet3 protein in DZO buffer. For (A),
measurement conditions were: microwave frequency, 9.48GHz; microwave power,
2mW; modulation frequency, 100kHz; modulation amplitude, 4.014G; temperature,
15.6K. For (B), measurement condition were: microwave frequency, 9.48GHz;
microwave power, 2mW; modulation frequency, 100kHz; modulation amplitude, 4.014G;
temperature, 21K. For (C), measurement conditions were: microwave frequency,
9.48GHz; microwave power, 2mW; modulation frequency, 100kHz; modulation
amplitude, 4.014G; temperature, 15.6K.

191

 

A)

 

 

 

 

2500 3000 3500 4000

 

 

 

 

 

 

 

 

 

2000
field (gauss)
B) I l l
2000 2500 3000 3500 4000
field (gauss)
C) T T l
2000 2500 3000 3500 4000

field (gauss)

192

 

Integrated echo

 

 

 

 

0 500 10100 1500 2000 25100 3000
r (us)

 

8)

Integrated echo

 

 

 

o 500 1000 1500 2600 2500 3000
1 (ns)

 

C)

Integrated echo

 

 

 

1) 500 1000 rs‘oo 2000 2500 3000
1:(ns)

Figure 5.2. Two-pulse 2H-ESEEM experimental traces for type-2 copper site in Fet3
protein (TlD). (A) Two-pulse ESEEM data in D20 buffer, measurement conditions: v
= 8.79GHz; microwave power, 44dB; B0 = 3055 G and ‘r = 250ns. (B) Two-pulse
ESEEM for TlD in H20 buffer, measurement conditions are the same as A. (C)
Quotient (A/B) modulation envelope of T1D Fet3p.

193

 

Integrated echo

 

 

 

 

o 500 10100 1500 2000 25100 3000
t (ns)

 

Integrated echo
9
OK

 

 

 

 

0 500 10100 1800 2600 25100 3000
1(ns)

 

C)
L2 2H/‘H

Integrated echo

 

 

 

o 500 10100 1500 2000 2500 3000
t(ns)

Figure 5.3. Two-pulse 2H-ESEEM experimental traces for type-1 copper site in Fet3
(T2D Fet3p). (A) Two-pulse ESEEM data in D20 buffer, measurement conditions: v
= 8.79GHz; microwave power, 44dB; B0 = 3090 G and 1: = 250ns. (B) Two-pulse
ESEEM for T2D in HZO buffer, measurement conditions are the same as A. (C)
Quotient (A/B) modulation envelope of T2D Fet3p.

194

5.4 Discussion

5.4.1 Theoretical 2H-ESEEM Traces for Ambient, Axial and Equatorial Water
For qualitative comparison of the experimental ratioed envelope, the

theoretical modulation functions for equatorial, axial and ambient water are shown in

Figures 5.4A, B and C, respectively. Traces A and B were generated using the

measured principal hyperfine values of axial and equatorial water protons by

Atherton and Horsewill in their single-crystal Electron Nuclear Double Resonance

(ENDOR) study of Cu(HzO)6+2 in Mg(NH4)2(SO4)4-6HZO salt.4 Trace C, the ESEEM
spectrum for unbound DZO, was computed using a purely dipolar axial hf-tensor (A,x
= A”: -0.2 MHz, Au: 0.4 MHz, reff ~ 4.0 A). A principal feature of these theoretical
functions is the strong second harmonic that occurs at a period of 250 ns due to
equatorial water (trace A). This second harmonic is less pronounced for an axially
bound water (trace B) and it is further diminished in the case of ambient water (trace
C).

These theoretical traces were generated using a Fortran ESEEM program,
TPSPNl for isotropic g-factor. The ESEEM program calculates the ESEEM function

using equation that spatially averages all orientations over a unit of sphere,

2m:

(Em) = i H E(9)sin 0d6dtp. (5.9)

An isotropic g-tensor leads to an assumption of the quantization of the electron spin
along the z-direction of the applied magnetic field. This approximation can also be

used for g-tensor with small anisotropy at higher field where the spin tends to align

195

 

  

A) equatorial D20

  
   

B) axial D20

C) ambient D20

W

0 500 1000 1500 2000 2500
t(ns)

 

 

 

Figure 5.4. Theoretical two-pulse 2H ESEEM traces for (A) equatorial, (B) axial and
(C) ambient water. Hamiltonian parameters for ESEEM simulations for equatorial
DzO: Ax, = -0.6, Ayy = -1.2, All = 1.4 MHz, gn = 0.85741, magnetic field = 3055 G, t
= 250ns, e2qQ = 0.22 and 1] =0.5. Parameters for axially bound D20 are the same
except for the hyperfine tensor, Axx = -0.5, Ayy = -0.6, An: 1.1 MHz. For ambient
D20, A,x = -0.2, Ayy = -0.2, A,z = 0.4 MHz which corresponds to an rm of 4A.

with the external magnetic field. Using this assumption, the dipolar contribution to

the hyperfine interaction can be calculated to be

PM, (5.10)
r

196

where gc is the g value at that field. For the copper complex considered in this work,
the high field region corresponds to the g 1 region of the CW-EPR spectrum where
transitions involving all the four Cu(II) nuclear sublevels may contribute to the
ESEEM. To test the validity of this approximation, ESEEM traces generated by an
angle-selection program, AN GTPSPNI, similar to program described in Chapter 4
are compared with traces calculated by spatial averaging over a unit of sphere,
TPSPNl. Figure 5.5 and 5.6 are for case of axial water ligation. Solid lines were
calculated by TPSPNl while dotted lines were calculated by ANGTPSPNl. Figure
5.5 considers only echo modulation due to the hyperfine interaction. The contribution
of NQI to modulation is considered in Figure 5.6. For the dashed lines in Figures
5.5A and 5.6A, 0N = 0°, (1)N = 0° were assumed , while in Figures 5.5B and 5.6B
assumed that 0N = 35°, 11),, 2 0°. A close agreement between TPSPNl and ANGTPSNl
is evident in Figure 5.5 B and Figure 5.6 B. This implies that the Cu-O and O-H
bonds make an angle of 125° and the two O-H bonds, and angle of 110°. Additional
assumption made on the calculation of Figure 5.6B was that the NQI principal axis
system (PAS) is coincident with HFI PAS (i.e, NQI Euler angle 8 = 35° for
counterclockwise rotation along 2.) Comparison for water coordinated equatorially
are described in Figure 5.7 and 5.8 . Again close agreement is achieved when

0H=125° (35° off the Cu-O bond) and HFI and NQI PAS are coaxial.

197

Figure 5.5. Comparison of 2H-ESEEM traces calculated by angle-selection
(ANGTPSPNI, dotted curves) and spatial orientation averaging (TPSPNl, solid curves)
for DZO axially coordinated to copper. All the traces are calculated with Am: 0.15
MHz, reff = 2.8A and no NQI contribution. Additional parameters needed to generate the
dashed curves are diagonal elements of the g-tensor and A-tensor of copper: g,,_= gyy =
2.05, gzz = 2.25, A,“ = Ayy = 35 MHz, and A22 = 586 MHz. Dotted trace 5.4A has 00, t1)D
= 0°,0°. Dashed trace 5.4B has 00, (DD = 35°,0°.

198

Integrated echo

Integrated echo

1.2

0.8

0.6

0.4

0.2

1.2

0.8

0.6

0.4

0.2

 

 

 

 

 

 

 

 

199

A )
. :0: o . .0. I .0. . '0. a
-. g - 3
.0:
o 500 1000 1500 2000 2500 3000
1 (ns)
I I I I I
B)
0 500 1000 1500 2000 2500 3000
1 (ns)

Figure 5.6. Comparison of 2H-ESEEM traces calculated angle-selection (ANGTPSPNI,
dotted curves) and spatial orientation averaging (TPSPNl, solid curves) for DZO axially
coordinated to copper. 2H NQI is included (e2qQ = 0.22 MHz, 1] = 0.50). Dashed curve
5.5A assumes 0,), (DD = 35°,0° and NQI Euler angle coincident with the g PAS (OL = B = y
= 0). Dashed curve 5.5B assumes 09, 9n = 35°,0° and NQI Euler angle coaxial with HFI
PAS (CL = y = 0, B = 225°). All other parameters are the same as in Figure 5.5

200

Integrated echo

Integrated echo

1.2

0.8

0.6

0.4

0.2

1.2

0.8

0.6

0.4

0.2

 

 

A)

 

 

 

500 1000 1500 2000 2500

 

 

 

 

 

3000
1 (ns)
B )
500 1000 1500 2000 2500 3000
1' (ns)

201

Figure 5.7. Comparison of 2H-ESEEM traces calculated by angle-selection
(ANGTPSPNI, dotted curves) and spatial orientation averaging (TPSPNl, solid curves)
for D20 equatorially ligated to copper. All the traces are calculated with Aiso = -1.0 MHz,
reff = 2.5 and no NQI contribution. Additional parameters needed to generate dotted
curves are diagonal elements of the g-tensor and A-tensor of copper: g,x = gyy = 2.05, gzz
= 2.25, A,“ = Ayy = 35 MHz, and A22 = 586. Dotted trace 5.6A has OD, $1) = 90°,0°.
Dotted trace 5.4B has 00, (1),) = 125°,0°.

202

Integrated echo

Integrated echo

 

1.2

0.8

0.6

0.4

0.2

 

A)

 

 

1.2

0.8

0.6

0.4

0.2

500

1000

1500
1 (ns)

2000

2500

3000

 

 

B)

4L

4

O.”
.0
.0
.C

A

 

 

500

1000

1500
1' (ns)

203

2000

2500

3000

Figure 5.8. Comparison of zH-ESEEM traces calculated angle-selection (ANGTPSPN 1,
dotted curves) and spatial orientation averaging (TPSPNl, solid curves) for D20
equatorially coordinated to copper. 2H NQI is included (e2qQ = 0.22 MHz, 1] = 0.50).
Dotted curve 5.7A assumes 0D, (1)0 = 90°,0° and NQI Euler angles, (CL = y = 0, B = 270°).
Dotted curve 5.7B assumes 00, (1)0 = 125°,0° and NQI Euler angles (or = y: 0, [3 = 325°).
All other parameters are the same as in Figure 5.7

204

Integrated echo

Integrated echo

0.8

0.6

0.4

0.2

1.2

0.8

0.6

0.4

0.2

 

 

 

 

 

500 1000 1500 2000 2500
1' (ns)

3000

 

 

B)

 

 

 

501) 1000 1500 2000 2500
1 (us)

205

3000

5.4.2 Frequency Domain Spectra of Ambient, Axial and Equatorial D20
Figure 5.9 shows an energy level diagram for the splitting of electron alpha

and beta spin manifold due to HFI. The diagram neglects the contribution of NQI

which for deuterium (e2qQ = 0.22, n E 0.50)15 is small and negligible. In the alpha
electron spin manifold, the two transition, 0% and tube are (1)ab 5 tube = (0a, This is
true for the beta manifold as well. The nuclear modulation, (1),, and (1)5, manifest
themselves in the frequency domain spectrum as a broad line centered about the 2H
Larmor frequency, (1),). Additional N QI will shift and broaden (1),, and (1),, by an
amount, Atom“. In addition to this fundamental frequency, sum combination peak
with negative phase will be observed. In our system, the nuclear Zeeman interaction
is greater than HFI and NQI so that the sum combination peak is primarily broadened
by the N Q1.

The effect of the increasing the dipolar interaction is shown for the case of
ambient, axial and equatorial D20 in Figure 5.10. Broadening of the fundamental
frequency is accompanied by the faster decay of the fundamental harmonic compared

to the second harmonic (refer to Figure 5.4).

5.4.3 Analysis of T1D Fet3p Data

The data for the type 2 site differ from the model calculations, but show
features in common with all those patterns. The depth of modulation observed for
TlD Fet3 protein is similar to that predicted for equatorially bound water while the

appearance of the sum combination frequency is more in line with that of an axially

206

Figure 5.9. Splitting of mS electron spin level by the nuclear Zeeman and hyperfine
interaction of nearby nucleus with 1:1. Quadrupole interaction is neglected

207

 

 

 

 

 

|1> ,x' 11 I”
0)
|1/2> .51... |0> ____ab v__ |b>
T: 11
|-1> wbc
“4' |c>
|1> ~
d>
_I-1/2__..:.._ 0> med L,
~. (Def
-1> _____ If)

small or negligible contribution of NQI
(Dab: wbc=wa
(Dde: (Def=(t)g

~2MHz

1

~4MHZ

frequency (MHz)
—>

208

12

 

 

  
    

 

 

 

 
    

 

 

 

 

 

 

  

 

 

  

 

 

 

 

 

     
 

 

 

 

 

 

 
  

 

 

 

 

..... . ..... .. ..... ......... , _. ..
....\.. ..... ..\.... .....J.......\ .1. .. 1...... \......a . o.\...-..A-..-..J...--.-\..-...l
. ' , . - ambient. . 4 ' . -

o . - . v - u - 1. r . a . .
_ . ...... . .. .. ....
....k. . 1.,.. .I.... ..5 ..'. .. ....'.. . ..‘. l
. .5......t.......k. ...-l...-..0.---. ....a.... . . .»\.-...-c.-..--I-.... .....l

W. ..... ,_ . ......

..... ., . .. . . . . .. ., . .. , . .. ..

.. , ., ., ..... ..... .,

.. . -. .. .. .. .. . .. . ..... .

. . .\. . ..' . ..\. .t......J. ......)

. ...... ... ...... . .. .

_. ...... . - .

.\ ' ..... .L ...... 1. ..'.. .\ .I

. L A I a

. ....... .. ...... . -. . ..
- .. ..k......‘......1.......\. . A ...In-...-u\.--~a-t.......'.......\......1......1.i . ..‘....-.l....-.-'......85......I
. :2 l v . I I I 1 v a | a J

     

 

 

   

 

  

. ,
..... ...
_.......;...-...j.
..............-.

....I...“ ..g.
.. . -.
.; . ;.
.......... .....-.

2 o _.-,... ... ...... .;_

. . .

. .................
. . .

. . .

1' ,‘. ..'
....................
. . .

t . .

1 ....,......,......§.
.........-..\-

. . .

.. .a ..............
. . .
....,..........-...‘.

 

~10

 

 

     
 
       
      

2

 

 

 

 

 

-
a
.
.

v
-
o

 

,.‘....-‘..,.‘..,.‘....V.,.

4

 

.‘.

 

. a

s
...‘....‘..

 

 

 

4
frequency (MHz

   

v
.....\..

a
.......-.

 

 

 

..‘..'....,.

...........-....... ...... ......
'

.... .....,...... ......
-

   

.....4...

 

.....4...
O
......‘...

é
)

 
   

 

,u.

   

.1.

v
s

....‘..,.~..

v

 

 

 

 

 

 

 

..
p

...Y

...-..

.t.l..-h-J-.|.I-VQ.4......-

 

v
s

1
1

 

 

 

1
l

1
I
o
s
-

1
e
n
s
1

 

 

 

 

.fi...
v

.‘. ..
a

......

 

  

 

10

 

 

 

 

 

 

. . ...... . . . .... . .
. ~
A - 1
r v
I

.‘.. ..
.

.....‘...-...
-

 

 

>,.,..,.‘..,.‘..,.4.

 

 

    
 

 
   
 
 

 

t
»-
..........'
1 t
..s......a
. .
..}......(
. .
..........
. r
.. ..... o

 

 

    

 

 

 

    

    
   
 
  

 

 
   

 

 

 

 

 

. -
.,.‘.....,.,.‘.

   

4......
J.. x
. a
1 u
.‘..._...'..,._
1 -
..............
- -
J.....-.o.....
'

.... ........_.‘$..... ...
. .

...- .r......,..... ...
u a .

........-.... ......_..... ..

 

 

4
frequency (MHz

 

 

Figure 5.10. Broadening of the fundamental frequenc

transform of the simulations shown in Figure 5.3.

29

 

as function of distance of
water deuterons. Traces, (A) equatorial, B) axial and, C) ambient D20, are Fourier

 

 

 

 

@-

c-u"rv-v-a

 

  
    
  

.
vu.A--‘-.‘ --...--\. '-- ...... ..v...- ...-e. .... e -- .. ...... ..-.-- ....... ......
, ’~ ’ -‘ Afeqmtortal DO -‘ ‘ t ‘
....... ._ . . .. ,.....,_... V ..
. .
...._ ... .. .....’.... ‘. . ...... . ..’”.....l.......‘.......l... ...
_____ .. _. .. .. ..... .
..... .\ -- u-o-w‘n-u. uO-uu~a.Jo- . ..‘—finlntlyco-Iolv-Qfil' \v-a-uv
...... . ...,... .. .
.. .-. ..... .y.-. ...}
...... .. . .,
1 \ .\

   
  
  

 

bound water ligand. The damping of the lower frequency, 2MHz, modulation is
matched best by the predicted function for the ambient water simulations. Overall,
the ESEEM data were best fit by a simulated function that takes into account the
contribution of a distribution of ambient water from reff = 4.0A to 8.6A in addition to
one axial and one equatorial water. This simulation is given in Figure 5.11A (dotted
line).

Another way of discriminating the contributions of deuterium coming from
axial water from those due to equatorial water is to ratio the relative amplitudes of the

fundamental 2H frequency (2 MHz) and sum-combination frequency (4 MHz) as

measured from the ESEEM spectra and described by Hegg et al. 16 The ratio
technique is based on the broadening effect of the electron-nuclear hyperfine
interaction on the fundamental frequency of 2H. Because the sum-combination
feature is broadened only by the small NQI of deuterium, it provides an internal
standard for gauging the damping or the linewidth of the fundamental frequency. For
TlD Fet3p, the experimental vD/ZVD ratio is 4.1, well outside the range of ratios given
for several water coordination schemes listed in Table 1 of the paper by Hegg et al..
According to that work, an equatorial D20 should give a vD/ZvD ratio of about 1.5, an
axial D20, 2.5, and a combination of the two, 2.0. The large vD/2vD observed here
could be attributed to the contribution of a large population of ambient water that was
not taken into account by Hegg et al in their calculations.

A tabulation of the theoretical vD/ZvD peak ratios that includes the effect of
ambient water is given in Table 5 .1. Ambient water is assumed to be uniformly

distributed [ni = (4Tt/3)(pN)(rJi+,-r3,)] in a series of ten concentric non-coordinating

210

1.4

 

 

A)

1.2
1
0.8
0.6
0.4 j
0.2
0Lm...i....1....1..L.i....i....i

0 500 1000 1500 2000 2500 3000

I(ns)
B)

Intensity

 

 

l

4 8 12 16 20

p—

frequency(MHz)

Figure 5.1]. Computer simulation (dotted) of (A)time and (B)frequency domain 2H-
ESEEM data (solid) of T1D Fet3p. Calculated ESEEM traces assume a product of
the modulation function of one axial D20, one equatorial D20 and a distribution (ni =

(41t/3)(pN)(r3i+,—r3,)) of ambient D20.

211

Table5.l. Tabulated ratios of the amplitudes of the peaks at the deuterium Larmor
frequency (VD) with the corresponding sum combinations peaks at twice the Larmor
frequency. Peak amplitudes were measured directly from ESEEM frequency. The
distribution of ambient water was modelled assuming the nearest non-coordinating
water to be 4.0A from the Cu center and iii = (41t/3)(pN)(13,+i-r3,).

 

 

 

 

 

 

 

 

 

Model Peak at 2 MHz
Peak at 4 MHz
Type -1 depleted Fet3 (expt.) 4.0
One axial D20 2.8
One equatorial D20 1.7
One equatorial and one axial D20 2.0
one axial/distribution of water 7.5
one equatorial/distribution of water 3.8
One equatorial/one axial/distribution 4.1
of water

 

 

 

 

shells. The ratio, 4.1, that is closest to the experimental peak ratio, 4.0, corresponds
to the modulation function that results from a contribution of one equatorial water,
one axial water and a distribution of ambient water. The effect of one ambient water
on the modulation function is negligible, however, the cumulative effect of a large
population of ambient waters will greatly affect the resulting modulation function so
that the effects of strong dipolar interaction (from bound waters) on the vD/2vD ratios

are partially masked. Therefore, unlike the Cu-substituted deA protein studied by

212

Hegg et al, the type 2 Cu(II) in Fet3p is more open to solvent water. This exposure,

in turn, allows the solvent to have a more pronounced effect on the 2H modulation.

5.4.4 Analysis of T2D Fet3p

The observed damping of the 2 MHz frequency in 2H—ESEEM trace of T2D
Fet3p (Fig. 5.3C) was modest indicating that an axial or ambient deuteron coupling
was possible. Axial D20 or OD‘ coordination can be ruled out based on the observed
depth of modulation and decay of the echo envelope. On the other hand, a single
ambient water could not account for the experimental modulation depth, suggesting
that a distribution of water molecules interacts with the type 1 Cu(II) also.

Simulation (Figure 5.12) assuming a distribution of water with ni = (27t/3)(pN)(r3,+,-r"i)
reproduced the observed depth of modulation. The experimental VD/ZVD ratio is 8.1.
From Table 2, the theoretical \10/2vD ratio that correlates best with the observed ratio
for T2D Fet3p corresponds to the simulation that assumes a contribution of ambient
water distributed according to ni = (2W3)(pN)(r’,+,-r3,).

The simulation indicates that the type 1 Cu(II) in Fet3p has no covalently
bonded water as would be predicted for this type of copper site. This is also
supported by the absence of split peaks in the sum combination frequency of the
proton two-pulse ESEEM of T2D fet3p (Figure 5.13). In contrast, the TlD fet3p has
pronounced shift peak.

The nearest non-coordinating water is about 3.75 (i 0.25) A from the metal
center. The study by Mims, Davis and Peisach on the accessibility of type 1 copper

sites to water suggests that the simulation that reproduced the depth of modulation of

213

.
.
08—
I
P

 

0.4 i

 

0 500 1000 1500 2000 2500 3000

1: (ns)

50~

20_

Intensity

 

 

 

 

 

-20

12 16 20

$—
on

frequency(MHz)

Figure 5.12. Computer simulation (dotted) of (A)time and (B)frequency domain 2H-
ESEEM data (solid) of T1D Fet3p. The ESEEM traces were generated using the
modulation function of a distribution (ni = (21t/3)(pN)(r3,,,-r3,)) of ambient DZO. No
covalent waters were included.

214

Table 5.2. Experimental and stimulated peak ratios of the amplitude of the deuterium
fundamental frequency to its sum-combination frequency for the type 1 center of
Fet3.

 

 

 

 

 

 

 

 

Nearest distance of Peak at 2 MHZ
Model ambient water to Peak at 4 MHz
copper center (A)
Type -2 depleted Fet3 8.1
Axial /distribution of water 4.0 6.2
Equatorial/distribution of water 4.0 4.1
Distribution of water (f=211: /3) 4.0 8.0
Distribution of water (f=21t /3) 3.7 7.9
Distribution of water (f=21t /3) 3.5 7.9

 

 

 

 

 

the type 1 Cu(II) in Fet3p corresponds to a geometric model in which one-half

(surface) of the coordination sphere is exposed to ambient water (hemispherical

model).9 The absence of an OH‘ or water ligation in type-l copper and the limited
access of water solvent to type-1 copper compared to type-2 copper might play a role
in the tuning of reduction potential of type-l copper. Water is a hard ligand and will
favor Cu+2 over Cu“. A more hydrophobic environment would tend to stabilize Cu+1
because of its low charge. The more positive reduction potential would help

facilitate the proposed electron transfer function of this metal cofactor.

5.5 Effect of Distal N-D on the Modulation Function
In both variants, the effect of the exchangeable proton of the remote nitrogen

of the histidyl imidazole was also considered. The hyperfine coupling constant of

215

A)

 

 

 

 

l L ‘— ‘_ A 1 A L A A J

10‘ A AA 15 20 l 25 3o
frequency(MHz)

 

 

L A AA 1 A A J L l A A A A I A A A A I

10‘ 15 20
frequency(MHz)

Figure 5.13 Frequency ESEEM data of T1D and T2D Fet3p in H20 and D20 buffer
system showing only the modulation frequencies due to proton nucleus.

216

the exchangeable lH is predominantly isotropic (Aiso(‘H) S. 5.5 MHz, A,SO(2H S 0.85

MHz).17,18 Assuming that the effective dipolar distance of the this hydrogen is
about 6A, the contribution of the through space interaction to the 2H hyperfine
coupling constant is 0.06 MHz. Thus, the effect of the exchangeable proton of the
remote-imidazole nitrogen to the damping of the fundamental frequency of the 2H is

negligible compared to the contribution of the ambient water.

5.6 Summary and Conclusion

Combining the l“N ESEEM results discussed in chapter 4 and the 2H ESEEM
in this chapter, structures of the type 1 and type 2 copper sites can be proposed.
ESEEM measurements have allowed us to identify some of the type-l and type-2
copper site ligands for the Fet3 protein. The type -2 copper has one strongly bound
histidyl imidazole ligand, while the type-1 has two histidyl irrridazole ligands as
predicted by sequence analysis. Deuterium-exchange experiments show that both
copper sites are solvent accessible and that ambient water must be included in the
ESEEM analysis to account for the observed vD/ZvD ratios. Water is bound to type-2
copper at equatorial and axial positions. For the type-1 center, the deuterium
modulation function is entirely due to the distribution of ambient water.

DZO exchange experiment revealed that type 1 copper site in Fet3p is
restricted to solvent access. Simulation of the two-pulse ESEEM data suggests the
type 1 site assumes hemispherical model in which one-half of the surface is exposed
to ambient to water. This model is consistent with the known protein structures that

include type 1 copper sites. In all of these systems, the copper atom is shielded from

217

direct water coordination by the protein fold and associated side chain ligation. This
shielding is irrespective of the coordination number, which can vary from three to

five depending on the specific type 1 copper-containing protein. Among the
multicopper oxidases this coordination number is either 3 (laccase, Lc)19 or 4 (A0,

th)20'22; the four-coordination in the latter two proteins is due to the thiosulfur of a
Met residue that is replaced by Leu in laccase as it is in Fet3p. As discussed above,
the type 1 Cu(II) site in Fet3p is most similar to that of laccase.

On the other hand, type 1 sites do differ in the extent to which they are buried
in the protein. The typel Cu site of azurin features histidine imidazole ligand whose
distal (NE2) nitrogen is at the surface of the protein exposed to solvent. In azurin,

the removal of this histidine by mutagenesis can be functionally complemented by

addition of exogenous N—methylimidazole23’24 In azurin, there is strong support for

the hypothesis that this imidazole is the site of entry of the electron from the reducing

substrate.25

Structural analysis has shown that the corresponding histidine side chain in
Lacasse, residue 494, is solvent exposed. In contrast, the type 1 Cu(II) site, and the
homologous His1045 in th is completely buried (numbering referenced to the
encoded ATG, not the processed protein). On the basis of sequence alignment and
the known structures of Laccase, th, and A0, the potentially solvent-exposed type 1
Cu(H) ligand in Fet3p is His489. These features of the type 1 sites in Lac, th and
Fet3p are illustrated in Scheme 1, structures A-C, respectively. The ESEEM data on
the type 1 site in Fet3p indicate the coordinating imidazoles are not magnetically

equivalent, perhaps due to differences in H-bonding. Furthermore, the presence of

218

A) \h

N8\H
C489 3 ‘Naégm
..\\\|
H2C\ o
\\
C /
HN’ L499

C) \{E

Scheme 1

219

significant ambient water, most likely distributed asymmetrically, suggests that this
site in Fet3p (structure C)is more similar to the type 1 site in Laccase (structure A)
and more exposed to solvent than the corresponding site in th (structure B). Thus,
the unique activity that th and Fet3p have towards Fe(H) as the reducing substrate
does not obviously depend on a strictly homologous and unique type 1 copper site.
Indeed, given the apparent differences between the Fet3p and th type 1 sites, that
would likely be reflected in differences in reduction potential and electron transfer
rates, the two enzymes may very likely use a different mechanism of electron transfer
from the Fe(II) substrate. If this is true, then the specificity these two enzymes have
towards Fe(H) would have to arise from features of the proteins and not the nature of
the “active sites.” Lindley and co-workers have presented some evidence in support

of a model that this activity is due to a specific Few“ binding site in an Asp/Glu-rich

channel adjacent to Hile45.26 However, this hypothesis has not been
experimentally tested further.

In summary, the 1“N and 2H ESEEM data conform to an asymmetrically
solvent—exposed model of the type 1 Cu(H) site in Fet3p in two ways. First, the distal
N-H of His489 would be in a hydrophilic environment, most likely H—bonded to
solvent. In contrast, the second histidine imidazole at this site, from His413 based on
sequence alignment, would be buried in a non-polar environment that was solvent
inaccessible. This would account for the asymmetry ascribed to the two histidine
imidazoles at the Fet3p type 1 Cu(H) site by the ESEEM analysis outlined above.
Furthermore, the solvation of the HE2 proton on His489 would account for the

presence of the ambient water required by the 2H-ESEEM simulations.

220

The ESEEM measurements were perhaps most revealing about the type 2
copper site in Fet3p. The data indicate that this copper has only one strongly bound,
equatorial histidine imidazole ligand. On the other hand, sequence analysis showed
that Fet3p contained both His residues found at the type 2 sites in the other
multicopper oxidases. Furthermore, the crystal structures of two of them, A0 and
th, provided strong evidence for equatorial coordination by both of these side
chains. The structure of AO also indicated the presence of a single equatorial water
molecule bound at the type 2 Cu(II). In contrast, the ESEEM data are best fit by a
model that has two water molecules bound directly to the type 2 Cu(II) in Fet3p. One
of these waters is bound at an equatorial position while the other appears to be axial.

Although we found only one histidine strongly bound to the type 2 Cu(II) in
Fet3p, a second histidine might be axially coordinated. Even if this were the case,
however, the structure of the Fet3p type 2 site would be markedly different than what
is seen in the crystal structures of its homologs. In both A0 and th, the N-Cu-N
atoms define a plane. Thus, in these proteins the histidine imidazoles are either both
equatorial or both axial; most certainly they are the former. As noted above, the
structure of AO indicates also the presence of a coordinated water that could be either
equatorial or axial, again depending on the axis assignment. Reasonably, the
occupancy by water of an equatorial site might be sufficiently larger than at an axial
one and thus be sufficiently electron dense for detection by x-ray crystallography. A
model that depicts the type 2 Cu(H) coordination in A0 is shown in Scheme 2,
structure A. In contrast, the ESEEM analysis strongly indicates that the type 2 Cu(II)

in Fet3p has one equatorial histidine imidazole, one equatorial water, one axial water,

221

and potentially an additional axial histidine. This is a new view of the type 2 Cu(II)

in a trinuclear cluster. It is illustrated in Scheme 2, structure B.

B)

Scheme 2

222

List of References

1)Mims, W. B. Phys. Rev. B 1972, 5, 2409-2419.

2)Tyryshkin, A. M.; Dikanov, S. A.; Goldfarb, D. J. Magn. Reson. 1993, A 105, 271-
283.

3)Lee, H.-I.; McCracken, J. J. Phys. Chem. 1994, 98, 12861-12870.

4)Atherton, N. M. M01. Phys. 1979, 37, 1349-1361.

5)Gerfen, 6.; Hanna, P.; Chasteen, N. D.; Singel, D. J. J. Am. Chem. Soc. 1991, 113,
9513-9519.

6)McCracken, J .; Friedenberg, S. J. Phys. Chem. 1994, 98, 467-473.

7)Tyryshkin, A. M.; Dikanov, S. A.; Evelo, R. G.; Hoff, A. J. J. Chem. Phys. 1992,
97, 42-49.

8)Reijerse, E. J .; Dikanov, S. A. J. Chem. Phys. 1991, 95, 836-845.

9)Mims, W. B.; Davis, J. L.; Peisach, J. Biophys. J. 1984, 45, 755-766.
10)McCracken, J.; Peisach, J.; Dooley, D. M. J. Am. Chem. Soc. 1987, 109, 4064-
4072.

11)McCracken, J .; Shin, D.; Dye, J. L. Appl. Magn. Reson. 1992, 3, 305-316.
12)Dikanov, S. A.; Shubin, A. A.; Parmon, V. N. J. Magn. Reson. 1981, 42, 474-487.
13)Mims, W. B.; Davis, J. L. J. Chem. Phys. 1976, 64, 4836-4846.

l4)Mims, W. B.; Peisach, J.; Davis, J. L. J. Chem. Phys. 1977, 66, 5536-5550.
15)Waldstein, P.; Rabideau, S. W.; Jackson, J. A. J. Chem. Phys. 1964, 41, 3407-

3411.

223

16)Hegg, E. L.; Whiting, A. K.; Saari, R. B.; McCracken, J .; Hausinger, R. P.; Que,
L. J. Biochemistry 1999, 38, 16714-16726.

17)Van Camp, H. L.; Sands, R. H.; Fee, J. A. J. Chem. Phys. 1981, 75, 2098-2107.
18)Roberts, J. E.; Cline, J. F.; Lum, V.; Freeman, H.; Gray, H. B.; Peisach, J .;
Reinhammer, B.; Hoffman, B. J. Am. Chem. Soc. 1984, 106, 5324-5330.

19)Ducros, V.; Brozowski, A. M.; Wilson, K. S.; Brown, S. H.; Ostergaard, P.;
Schneider, P.; Yaver, D. S.; Pedersen, A. H.; Davies, G. J. Nat. Struct. Biol. 1998, 5,
310-316.

20)Messerschmidt, A.; Rossi, A.; Ladenstein, R.; Huber, R.; Bolognesi, M.; Gatti, G.;
Marchesini, A.; Petruzelli, R.; Finazzi-Agro, A. J. Mol. Biol. 1989, 206, 513-529.
21)Messersehmidt, A.; Ladenstein, R.; Huber, R.; Bolognesi, M.; Avigliano, L.;
Petruzzelli, R.; Rossi, A.; Finazzi-Agro, A. J. Mol. Biol. 1992, 224, 179-205.
22)Zaitzeva, 1.; Zaitzev, V.; Card, G.; Moshkov, K.; Box, B.; Ralph, A.; Lindley, P. J.
Biol. Inorg. Chem. 1996, I, 15-23.

23)den Blaauwen, T.; Kamp, v. d.; Canters, G. W. J. Am. Chem. Soc. 1991, 113,
5050-5052.

24)den Blaauwen, T.; Canters, G. W. J. Am. Chem. Soc. 1993, 115, 1121-1129.
25)van de Kamp, M.; Floris, R.; Hali, F. C.; Canters, G. W. J.Am. Chem. Soc. 1990,
112, 907-908.

26)Lindley, P. F.; Card, G.; Zaitseva, 1.; Zaitsev, V.; Reinhammar, B.; Selin, L.;

Yoshida, K. J. Biol. Inorg. Chem. 1997, 2, 454-463.

224