v. .i

Y

.5 . i
3353:)?“
‘_ a!

r
n . . v
kin?!
.0535!
.55. a 2:...
1....
.
u..£....::.r. (.1...
r . 4:14: .
n t: . s
j.) .33.».4IM1 '

$1.:
1. I : wit}.f?~u.d
z : i 5:.
3.. 2.4.15 r..¢l
:5:- t .
. l. a; If.
. .3... .
26 .3

. 3:3. . .12.
v. .r-u’l-i...n. yr...
,.

z

 

. . .1 .5 . .. . :
fizz—:9 Eva-Man... «Big-Zilrgrafifi. .1! 2 4
I: 1.5-)... 2.2!: .. It... .
“I s . .x.‘ .. ..
. in“? ..
. .
. . A.
an r A. .
.. .1 n .
.. _ .g
. athl-figrlzthvl
fi. .33....
31. V . .
a: x a r... -.
.r‘hrmwpug («Hum a. . _ thirst? . 9.x.
$1.40.: A "Elma-u “mea.h&h 35 a) 4
‘.“l” u ‘ II. "o- Iln s: .
.3 w i... ,ummnflu...i.n .33, 1 Aw... " r.
I '13-. 'l Afiglm. . .Q‘JWNM
, A . , .
r Q4. 3. .
3.. gnu“ ten: ,4... x <
I ”fig...“ 1! f. . . I .
. . ‘pfizvlgwclg.£ 1- ‘ ‘ D»
3...}.25 . I. X! I! J.” ,
a. 5.1.»... .fi.mu3~a.i§. J? A. .
was . .. Irt. I 9..
3. a u \ ”gnu-t t: lllhnfilLR: It
9:

Slit:
a)? . .
. . .. 1

‘ is:

.3 I5 .I z...

. . w... 3969....

.33.. Wanna... hi
i

if: .2. ..
fir... éfiin .firfi.
: 1...
'1' in! ‘in? J?.’ >
3.3 .:..a...:
.. 5 .
.2 .
' f, $ .
.c, .-......_...........z..
3.: :2. hi.
.3! of... .9 $7.1
.3. .19.?4;
Sam. ,5... no”? 2..
V ..
. x. 1.,h3m....£:.
.; ..
L:
5. 2.
bluné:
2.) .. ..
5......)

. can...“

IV) I) ‘
lu' -. i?
I :9‘.

I) ‘ Ixxl. I.
I 0.9.. it
i If}...
svlala.uu..;lal
’ofiaflntf . .1.
flying.) I
I.
:3
. .31....
anti}...
.2. .
.i r . .
an: . “no“... 9
{ti} .
’12:. It. .
. shiaahttifl ,
. 9.5.12
.1
4 y.
. , 35?...
‘3‘; I:

..c
I

. t,
v. 1.4; . .
.112...

v...

9
6.3.5?
9
3

.1. . c: 7......
2‘12}... .. .
fi~h .1} 3

us!
....9.(ts~::\
3 {bl-.2521;

:11!
. .LI. . Ir!
31:3,..3‘
.I Is...
)5.

v

.7.
:3... 9:... . ‘5

 

Llaniinv
Michigan State
University

 

 

This is to certify that the
dissertation entitled

9403M THE ELEchomc. sTRucwle OF I;
QUIWWZYMS coFAcroQ

presented by

Vim-m SING-H

has been accepted towards fulfillment
of the requirements for

 

Ph-‘D. degreein CHEM'ST‘)’

 

Major professor

QKYW/w 4w

Date/27mg 5:; 9x000

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

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

 

0 70 «316 1515312002

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11m Woes-p.14

PROBING THE ELECTRONIC STRUCTURE OF A QUINOENZYME COFACTOR

By

Vinita Singh

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
PROBING THE ELECTRONIC STRUCTURE OF A QUINOENZYME COFACTOR
By

Vinita Singh

The electronic structure of the tryptophan tryptophyl-semiquinone (TTQ')
cofactor generated by addition of the substrate, methylamine, to methylamine
dehydrogenase (MADH) from Paracoccus denitrificans has been studied by continuous-
wave electron paramagnetic resonance (cw-EPR) and electron spin echo envelope
modulation (ESEEM) spectroscopy. The cw-EPR spectrum of TTQ semiquinone
prepared by substrate addition (N-form) was found to differ substantially from that
observed when the semiquinone was generated by dithionite reduction of the enzyme (0-
form). These differences prompted a detailed study of hyperfine and nuclear quadrupole
interactions of the three I4N atoms of the semiquinone species using ESEEM. Two of
these heteroatoms are derived from the indole and indole-quinone side chains that
comprise TTQ, while the third l4N originates from substrate methylamine. Three-pulse
ESEEM spectra of the CH3l4NH2-reduced sample showed three isolated features at 1.0,
1.5 and 4.3 MHZ, which were absent in the MADH sample reduced with CH3'5NH2.
Analysis of the spectral data for substrate-derived l4N revealed an isotropic hyperfine

coupling of 2.4 MHz and nuclear quadrupole couplings characterized by equ = 1.7 MHz
and 1] = 0.5. The hyperfine and the nuclear quadrupole couplings found for the two 14N

nuclei indigenous to TTQ were: Am, 2.8 and 1.9 MHz; eZqQ, 3.0 and 2.1 MHz and n, 0.3
and 0.7, respectively. Taken together, these couplings provide definitive evidence that

substrate MN is covalently bound to TTQ when the cofactor is in its one-electron reduced

form and that it has an imine-like structure. The intensities of the modulations indicate
that the semiquinone generated by the method recently reported by Zhu and Davidson
(Biochim. Biophys. Acta. , 1998, 1364, 297-300) results in a homogeneous preparation of
the radical. A comparison of the 14N hyperfine and nuclear quadrupole couplings
measured for the N-form semiquinone with those measured previously for O-form
(Warncke, K.; Brooks, H. B.; Lee H.-I.; McCracken, 1.; Davidson, V. L.; Babcock, G. T.;
J. Am. Chem. Soc., 1995, 117, 10063-10075) shows that a significant change occurs in the
highest occupied molecular orbital when substrate nitrogen is bound, and may be related

to the different redox and electron transfer properties of these two semiquinone forms.

To my family

ACKNOWLEDGMENTS

I wish to thank people who have helped me in some way in carrying out the work
presented in this thesis.

First of all I am grateful to my parents for providing me the best education
possible that gave me an opportunity to come to the United States for higher education. I
am thankful to them for giving me the freedom to choose the path of my choice in my life
and for their strong faith in me that whatever I do will be right thing to do.

I am thankful to all the teachers who have taught me so far. My research advisor
Prof. J. L McCracken is one of them for whom I have highest respect and gratefulness. I
started with no experience in EPR methods, but he patiently helped me learn about EPR
and spent time in explaining the working of pulsed EPR spectrometer and in
troubleshooting during experiments. I am thankful to him for his support, encouragement
and help during the course of this work.

I wish to thank our collaborators Prof. V. L. Davidson and Dr. Z. Zhu from
University of Mississippi medical center for supplying the enzyme samples for this
work.

I am thankful to Prof. J. F. Harrison, Prof. P. Piecuch, Prof. N. Jackson, Dr. Jesse
Edwards, Dr. Rudolf Burcl, Chetan Ahuja, John Goodpaster, and Ian Sedrick O. Pimienta
for their help, suggestions and discussion on the molecular orbital calculations.

I wish to acknowledge present and past staff of the department of Chemistry for
their help. I wish to thank Kermit Johnson, Paul Reed, Dr. T. B. Atkinson, and Dr. Eric
Saari for helping me solve computer-related problems. I wish to thank physical

chemistry secretary Vada O' Donell and the secretary for the graduate office, Lisa

Dillingham for their help. I also want to thank John Rugis for making the current pulse
generator for the field-j ump experiments.

I would like to thank past and present members of Prof. McCracken's and Prof. J.
Babcock's research group especially Dr. Vladimir Bouchev, Dr. Marilena Di Valentin, Dr.
C. W. Hoganson, Dr. Georgy Filep, and Constantino P. Aznar III for discussing the
fundamentals of EPR. I also wish to thank Dr. M. Mac, Dr. K. Reidy and Dr. S. Smith. I
would also like to thank all my friends at MSU for their company.

I am grateful to everyone in my family, my parents, my brother, sister, my
grandmother, and my fiancé Karun for their love, support, care and encouragement and for

understanding my changing mood.

vi

TABLE OF CONTENTS

List of tables ............................................................................................................. ix
List of figures ........................................................................................................... x-xiv
List of abbreviations ................................................................................................ xv
Introduction .............................................................................................................. 1-3
Chapter I:
Introduction to Methylamine Dehydrogenase - a bacterial quinoenzyme ....... 4—29
1.1 Quinoenzymes - a different class of redox enzymes ............................................ 4-6
1.2 Methylamine dehydrogenase (MADH) ............................................................... 6-21
1.2.1 Structure ..................................................................................................... 6-13
1.2.2 Biosynthesis of the TTQ cofactor ............................................................. 13
1.2.3 Spectral and oxidation-reduction properties ............................................... 13-15
1.2.4 Function ...................................................................................................... 15-16
1.2.5 Reaction mechanism ................................................................................... 17-21
1.3 Earlier studies on TTQ-semiquinone .................................................................... 21-23
1.4 Usefulness of quinoproteins to higher animals ..................................................... 23-25
List of References ....................................................................................................... 26-29
Chapter II :
Principles of continuous wave and pulsed Electron Paramagnetic Resonance
Spectroscopy ............................................................................................................. 30-48
11.1 Continuous wave Electronic Paramagnetic Resonance (cw-EPR)
11.1.1 The electronic Zeeman interaction ............................................................ 30-35
11.1.2 Nuclear Zeeman interaction ....................................................................... 35-37
11.1.3 The hyperfine interaction .......................................................................... 37-41
11.1.4 Nuclear quadrupole interaction ................................................................. 41-43
11.1.5 Zero field interaction ................................................................................. 43-44
11.1.6 Relaxation effects ...................................................................................... 44-46
11.1.7 Information available from cw-EPR .......................................................... 46-48
11.2 Pulsed EPR .......................................................................................................... 48-75
11.2.1 Two-pulse ESEEM spectroscopy ............................................................ 49-70
11.2.1.1 Formation of Electron Spin Echo (ESE) .......................................... 50-53
11.2.1.2 The nuclear modulation effect ......................................................... 53-60
11.2.1.3 Density matrix formalism ................................................................ 60-70
11.2.2 Three-Pulse ESEEM spectroscopy .......................................................... 71-73

vii

List of References ....................................................................................................... 74-75

Chapter HI :

Instrumentation ....................................................................................................... 76-90
111.1 Resonator .................................................................................................... 84-87
111.2 Gordon Coupler .......................................................................................... 88
111.3 Circulator and Isolator ................................................................................ 89-90

List of References ....................................................................................................... 91

Chapter IV:

Characterization of the TTQ-semiquinone catalytic intermediate of MADH by

ESEEM .................................................................................................................... 92-139

IV.1 Introduction ...................................................................................................... 92

IV.2 Materials and Methods
IV.2.1 Protein purification and preparation ...................................................... 93
IV.2.2 Continuous-wave EPR spectroscopy .................................................... 94
IV.2.3 ESEEM spectroscopy ........................................................................... 94-95

1V.3 Results .............................................................................................................. 95-102
1V.3.l cw—EPR Spectroscopy ........................................................................... 95-97
1V.3.2 ESEEM Spectroscopy ........................................................................... 98-103

1V.4 Discussion ........................................................................................................ 104-134
IV.4. 1. Interactions of nitrogen nucleus ............................................................ 104-105
IV.4.2 Exact cancellation condition ................................................................... 106-107
IV.4.3 Spherical model approximation .............................................................. 108-110
IV.4.4 Theoretical simulations .......................................................................... l 1 1-134

1V.5 Studies on aromatic amine dehydrogenase ....................................................... 134-136

IV.6 Conclusions ...................................................................................................... 137-138

List of References ..................................................................................................... 139-140

Chapter V:

Resolution enhancement of echo detected EPR spectrum of Manganese-ethylene

(dinitrilo) tetraacetate by magnetic field vector jump at X-band ..................... 141-160

V.l Introduction ....................................................................................................... 140

V2 Basic principles of Echo-detected EPR with vector field jumps ....................... 140-145

V.3 Experiments ....................................................................................................... 145-146

V.4 Results ............................................................................................................... 146-150
IV.4. 1. Two-pulse echo detected EPR with field vector jumps ....................... 146-148
IV.4.2 Two-pulse ESEEM experiments with field vector jumps .................... 147-150

V.5 Discussion ......................................................................................................... 151-159

List of references ...................................................................................................... 160

viii

Table IV-l

Table IV-2

LIST OF TABLES

The assignment of the observed frequencies and the principal values of hf
and nqi coupling parameters used in computer simulations for substrate-
derived nitrogen and the two nitrogens from TTQ. ........................... 114

n-orbital unpaired spin densities located on the three nitrogens present in
the N-form semiquinone and comparison with the values determined
previously for the O-form radical. ..................................................... 118

Figure 1-1

Figure 1—2

Figure 1-3

Figure 1-4

Figure 1-5

Figure 1-6

Figure II-l

Figure 11-2

Figure 11-3

Figure 114

Figure 11-5

LIST OF FIGURES

Structures of known quinone cofactors : pyrrolo-quinoline quinone (PQQ),
topaquinone (TPQ), tryptophan tryptophquuinone (TTQ), and lysine

topaquinone (LTQ). ............................................................................... 5
Structure of the oxidized TTQ cofactor of MADH ............................... 7
The tetrameric structure of PD—MADH. ............................................... 10

Orientation of TTQ in the MADH-amicyanin complex showing the
relative positions and distances between TTQ and Cu. Only the side
chains of TTQ and Cu ligands are shown. ............................................. 12

Absorption spectra of different redox forms of MADH. Spectra of the O-
quinone, O-semiquinone, and O-quinol are displayed as solid lines. Spectra
of the N-semiquinone and N-quinol are displayed as dotted lines. All
spectra were recorded in 6.7 M MADH in 10 mM BTP buffer at pH 7.5
and 25 C. (Taken from reference 54) ...................................................... 15

Proposed reaction pathway for MADH divided into reductive (a) and
oxidative (b) steps. ................................................................................. 18

(a) Splitting of degenerate 0t and B electron spin states due to electron
Zeeman interaction in the presence of applied static magnetic field Bo.
(b) An EPR line expected from a transition in (a) for systems with
isotropic g value. .................................................................................... 33

Coupling of the projections of the spin and the orbital angular momentum
in case of two different molecular orientations relative to the applied static
magnetic field B0 ..................................................................................... 34

Energy level diagram showing the effects of electron Zeeman, nuclear
Zeeman and Hyperfine interactions of an unpaired electron (S = 1/2) with
a I = 1/2 nucleus in the presence of an external magnetic field B0. ......... 37

Dipolar interaction of magnetic dipoles pm and 11., due to nuclear spin
angular momentum and electron spin angular momentum respectively. The
length of vectors pm and it. are not drawn to scale. The electron is at the
origin at a distance ‘r’ from the nucleus. The dipoles are aligned with the
external magnetic field and 0 is the angle between the electron-nuclear
vector r and the applied external magnetic field Bo. .............................. 40

Interactions of a quadrupolar nucleus with of four point charges, the
orientation in (b) is of lower energy and is preferred over (a) ................ 42

Figure 11-6
Figure 11-7

Figure II-8

Figure 11-9

Figure II-11

Figure 11-12

Figure III-1
Figure III-2

Figure III-3

Figure III-4

Figure III-5

The pulse scheme used in a two-pulse ESEEM experiment. ................. 49
Two-pulse time domain ESEEM data, dead time = 220nsec ................. 55

Energy level diagram for the interaction of an unpaired electron (S = 1/2)
and a nucleus with I = 1/2. ..................................................................... 56

Vector diagram showing the formation of electron spin echo as a result of a
90°-t-180° mw pulse sequence in a two pulse ESEEM experiment. (a) at
thermal equilibrium the net magnetization M is aligned with the external
magnetic field Bo; (b) as a result of first 90° mw pulse, M is torqued onto
the x'y'-plane of the rotating frame; (0) the spin packets that constitute M
start to dephase due to difference in their precession fiequencies; (d) the
second 180° pulse applied along x', flips the spin packets about x', while
the direction of precession of individual spin packets remains unchanged.
This pulse also introduces branching of transitions; (e) the individual spin
packets refocus to form a spin echo along -y'. ....................................... 58

Energy levels of a quantized system and transitions giving rise to spin
echo. la) and |B) are two electron spin states. Only those levels of the
over—all quantized system for which the transition la,)<—)|Bj) has a
significant amplitudes are shown ........................................................... 62

Vector model showing the formation of electron spin echo in a three pulse
ESEEM experiment. (a) pulse scheme; (b), (c) same events as in figure 11-9;
(d) the second 90° mw pulse along x' converts the transverse
magnetization into longitudinal magnetization and the components start to
dephase after the pulse is removed; (6) the third 90° mw pulse applied
along x' rotates the magnetization components onto x'y' plane;(f) the spin

packets refocus along -y' to give rise to a stimulated spin echo ............. 70
Schematic of pulsed EPR spectrometer setup. ...................................... 80
A series RLC circuit fed by voltage V .................................................... 84

Schematic of folded stripline half-wave resonator mounted in teflon block
as seen from top (a), mw magnetic field lines along the cavity wall couple
to the resonator (b) .............. ' ................................................................... 8 6
Schematic of the folded stripline half-wave resonator probe structure. 87

Symbolic representation of a four port circulator. ................................. 89

Figure IV-l

Figure IV-2

Figure 1V-3

Figure 1V—4

Figure IV-5

Figure IV-6

Figure IV-7

Figure IV-8

Continuous-wave EPR spectra of the semiquinone forms of methylamine
dehydrogenase from P. denitrificans generated by dithionite addition (0-
form) (a), l4N-methylamine addition followed by illumination (b), and 15N-
methylamine addition followed by illumination (c). Spectra were collected
under the following conditions: microwaves power, 0.63mW; microwave
frequency, 9.473GHz; modulation frequency, 100 kHz; modulation
amplitude, 0.1 mT; sample temperature, 103 K; and number of scans
averaged, 10. ........................................................................................... 96

Continuous-wave EPR spectra of the semiquinone forms of methylamine
dehydrogenase from P. denitrificans generated by dithionite addition (0-
form) (dashed line), l4N-methylarnine addition followed by illumination
(solid line). Conditions were same as in figure IV-l. Presence of both the
N-form and the O-forrn in the samples prepared by comproportionation
method lead to the asymmetric spectrum. ............................................. 97

Time-domain three-pulse ESEEM spectra of randomly oriented 14N-
methylamine-reduced semiquinone, Conditions: 1:, 219 ns; Microwave
frequency, 8.815 GHz; magnetic field, 315.5 mT; microwave pulse power,
43W; pulse sequence repetition rate, 4 Hz; events averaged per time point,
10; sample temperature, 4.2 K; number of scans averaged, 9. ............... 99

Frequency-domain three-pulse ESEEM spectra of randomly oriented (a)
l4N-methylarnine—reduced semiquinone; (b) 15N-methylamine-reduced
semiquinone. Conditions were identical to those given in figure IV-3. .. 100

Three-pulse ESEEM spectra of randomly oriented l4N-methylamine-
reduced semiquinone (a); and 1SN-methylamine-reduced semiquinone (b). 1:
= 365 ns. Other conditions were identical to those given in figure IV-3.
................................................................................................................ 102

Three-pulse ESEEM spectra of randomly oriented l4N-methylamine-
reduced semiquinone (a); and l5N-methylarnine-reduced semiquinone (b). 1'
= 450 ns. Other conditions were identical to those given in figure IV-3.
................................................................................................................ 103

Energy level diagram for the interaction of an unpaired electron with a
nucleus of 14N (I = 1) at exact cancellation condition (a); and the expected
spectrum (b). .......................................................................................... 107

Cosine Fourier transformations of numerical simulations of three-pulse
ESEEM data of CH314NH2-generated semiquinone in MADH (a) 1:, 219 ns;
(b) 1:, 365 ns; (c) 't, 450 ns. The simulations were performed using three
coupled l4N nuclei, N(l), N(2) and substrate-derived N, whose hfand nqi
coupling parameters are listed in table IV-l. Other simulation parameters:

xii

Figure IV-9

Figure IV-lO

Figure IV-11

Figure 1V-12

Figure IV-13

Figure 1V-14

Figure 1V-15

Figure V-l

Figure V-2

Figure V-3

gN, 0.4035; magnetic field strength, 315.5 mT. An isotropic electronic g-
factor was assumed. ............................................................................... l 15

A comparison of the normalized three pulse electron spin echo envelope
modulation amplitudes for l4N-methylamine poised MADH semiquinone
measured experimentally (solid line) at 1: = 219 ns with that predicted by
spectral simulation (dashed line) for the same ‘t-value. .......................... 116

Structure of the oxidized form of the model compound of TTQ cofactor of
MADH used in theoretical calculations. Atoms are labeled as, nitrogen in
magenta, oxygen in red, carbon in black and hydrogen in yellow ........... 119

Calculated spin density for the O-semiquinone form of the model
compound. Calculations used semiemperical model and UHF basis sets in
SPARTAN software program; isosurface value, 0.002. Atoms are labeled
as, nitrogen in magenta, oxygen in red, carbon in gray and hydrogen in
white ......................................................................................................... 122

Calculated spin density for the N-semiquinone form of the model
compound. Calculations used semiemperical model and UHF basis sets in
SPARTAN software program. Atom labeling is same as in figure IV-ll
.................................................................................................................. 124

A plot of the electron density and HOMO for the N-semiquinone form of
the model compound showing that the HOMO is spread over the entire
molecule. Atom labeling is same as in figure IV-l 1. ................................ 126

A plot of the HOMO for the O-semiquinone form of the model
compound. Atom labeling is same as in figure IV-l 1. ............................. 128

Continuous-wave EPR spectra of the semiquinone forms of aromatic
amine dehydrogenase from Alcaligenesfaecalis generated by dithionite
addition (O-form) (a), l4N-methylamine addition followed by illumination
(b), and 15N-methylarnine addition followed by illumination (c). Spectra
were collected under the following conditions: microwaves power, 2.0
mW; microwave frequency, 9.484GHz; modulation frequency, 100 kHz;
modulation amplitude, 0.1 mT; sample temperature, 20 K; and ntunber of

scans averaged, 25. ................................................................................... 136
An isotropic energy level diagram at fixed field for Mn(II). .................... 143
Angular dependence of the centers of the sextets in the five fme-structure

transitions of Mn(II) in KN3. Taken from reference 2 ............................. 144
Pulse scheme for two-pulse field jump experiment ................................. 146

xiii

Figure V-4

Figure v-5

Figure V-6

Figure V-7

Figure V-8

Figure V-9

Figure V-10

Figure v-11

Echo detected EPR spectrum with field jump for Cu(II) in frozen aqueous-
ethylene glycol solution. (a) without the application of 13x pulse (b) in
presence of Bx pulse proportional to a voltage of 220V; microwave
frequency, GHz; 1', 500 ns; microwave pulse width = 20 ns. .................. 149

Echo detected EPR spectra of Mn(II)-EDTA in frozen aqueous-ethylene
glycol solution using different microwave pulse-widths (a) 20 ns; (b) 40 ns
(c) 110 ns;. ’1', 500 ns; pulse repetition rate, 60 Hz, microwave power, 40
dBm, temperature, 4.2 K. The Bx tilt field was not applied. .................. 150

Echo detected EPR spectrum with field jump of Mn(II)-EDTA in frozen
aqueous-ethylene glycol solution. (a) without the application of BX pulse
(b) in presence of B,( pulse proportional to a voltage of 250V; microwave
frequency, 9.80 GHz; other conditions same as in figure V-5 ................. 151

Two pulse ESEEM spectra with field jump (a) without the application of
B" pulse (b) in presence of BM pulse proportional to a voltage of 100V
applied. Conditions: microwave frequency, 8.593 GHz; 1?, 250 ns;
magnetic field = 2800 G. ......................................................................... 152

Model of the structure of sexadentate seven-coordinated Mn—EDTA
complex with twofold vertical axis ........................................................... 154

X-band cw-EPR spectrum of Mn-EDTA (0.002 M) (a) at 50° C and (b) at
3° C; microwave frequency, 9.1GHz. Taken from reference 7. (c) Q-band
cw-EPR spectrum of Mn-EDTA (0.01 M) obtained at 20° C; microwave

frequency, 35 GHz. Taken from reference 8 ........................................... 156

Simulation using Simfonia simulation program provided by Bruker
Instruments, with D = 1180G, E = 265G, isotropic g = 1.946; microwave
frequency, 9.104 GHz .............................................................................. 159

Magnetic field positions of the fine structure transitions in a powder
spectrum of a 55Mn(II) complex with axial symmetry as a function of zfs
at X band. D is assumed to be negative. Unprimed letters indicate the
transitions for 0 = 0°; primed for 0 = 90° and C" is for the central
transition at 0 = 41°48'. A, 5/2 <-—> 3/2; B, 3/2 H 1/2; C, 1/2 (-—) -1/2; D, -
3/2 (—9 -1/2; and E, -5/2 <—-> -3/2. Free electron resonance position, g =
2.0023, coincides with C. Taken from reference 12 ................................ 160

Note: Figures 1V-10 - figure IV-14 in this dissertation are presented in color.

AADH
C

Cw
EDTA
ENDOR
EPR
ESE
ESEEM
ESR
FET
FID
FORTRAN
F T

h

Hfi
HYSCORE
F WHM
LabVIEW
MADH
Matlab
M w
NMR
Nqi

PAS

Ppm
TWTA

LIST OF ABBREVIATIONS

Aromatic amine dehydrogenase

The C programming language
Continuous wave

Ethylene diamine tetra acetic acid
Electron nuclear double resonance
Electron paramagnetic resonance
Electron spin echo

Electron spin echo envelope modulation
Electron spin resonance

Field effect transistor

Free induction decay

Formula translator (programming language)
Fourier transform

Planck's constant 'h' divided by 2n = 1.05459x10'34 J .s

Hyperfine interaction

Hyperfine sublevel correlation spectroscopy

F ull-width-half-maximum

Laboratory virtual instrument environment workbench
Methylamine Dehydrogenase

Matrix laboratory (programming and visualization software)
Microwave

Nuclear magnetic resonance

Nuclear quadrupolar interaction

Principal axis system

parts per million

Travelling wave tube amplifier

INTRODUCTION

Many vital bioenergetic processes such as respiration, photosynthesis and many
other oxidation-reduction reactions involve electron transfer. The mechanism and control
of electron transfer within various protein systems, although studied extensively, is still
not well understood. Present study is focused on understanding the electron transfer
process in one such system using methods of electron paramagnetic resonance
spectroscopy. The field of pulsed EPR spectroscopy has made tremendous progress in
the past few years and new techniques are being developed to study more complicated
and challenged spin systems. Paramagnetic manganese is a transition ion of interest
because it acts as a cofactor in many enzymes and can serve as a substitute for a required
Mg(II) cofactor in other enzymes. Present studies explored the possibility of simplifying
the spectrum of a manganese complex by achieving enhancement in the resolution by
application of magnetic field vector jumps.

Chapter I gives a brief introduction to the class of enzymes called quinoenzymes,
and focuses on Methylamine dehydrogenase which is a periplasmic enzyme found in
gram—negative soil bacteria. Its main function is to catalyze the reactions leading to
assimilation of methylamine, the substrate on which these organisms grow. It forms part
of a soluble electron transfer complex where it has been possible to study protein-protein
interactions and structures using spectroscopic and X-ray crystallographic methods.

In chapter 11, principles of Electronic Paramagnetic Resonance (EPR)
spectroscopy are discussed along with the need for the pulsed methods. Various

interactions of the unpaired electron spin with a magnetic nucleus present in its near

vicinity modify the energy levels involved in the transitions and are discussed briefly.
The basic principles of EPR are similar to that of Nuclear Magnetic Resonance (N MR).
A thousand-fold larger electron moment than the nuclear moment makes EPR more
sensitive to its environment, leading to broader spectral widths and faster relaxation times
as compared to NMR.

Although the development of pulsed-EPR techniques faced more technical
limitations than those encountered for the pulsed-NMR, it has overcome the earlier
challenges and due to the availability of faster digital logic components and advances in
microwave technology it is now possible to exploit the potentials of this important
spectroscopic technique. The experimental pulsed-EPR data for this study were collected
on a home-built pulsed EPR spectrometer built by Prof. J. L. McCracken at Michigan
State University in 1990, which has been modified from its original design since then.
The main features of this spectrometer are discussed in chapter 111.

Results from cw-EPR and electron spin echo envelope modulation (ESEEM)
spectroscopic studies on semiquinone forms of the cofactor of MADH are presented and
discussed in chapter 1V. Quantum mechanical principles discussed in chapter 11 are used
to perform simulations of the experimental results and to extract useful information from
the experimental results. Results of Molecular orbital calculations performed on a model
compound structurally related to the cofactor of MADH are presented and they provide
a qualitative look on the distribution of unpaired spin density studied by the EPR

techniques.

Chapter V presents the results of experiments aimed at achieving resolution
enhancement of echo detected EPR spectrum of manganese-ethylene (dinitrilo)

tetraacetate by magnetic field vector jump at X-band.

Chapter I

INTRODUCTION TO METHYLAMIN E DEHYDROGENASE - A BACTERIAL
QUINOENZYME

Electron transfer is central to bioenergetic processes such as respiration,
photosynthesis and many other oxidation-reduction reactions. Electron transfer in
biological systems occurs between protein-bound prosthetic groups separated by long
distances, often greater than 10A, with high efficiency and specificity. The prosthetic
groups of most of the oxidoreductases are enzyme-bound quinones and flavins. These
prosthetic groups and the membrane-bound components of respiratory electron transfer
chains play the important role of coupling the two-electron oxidation of substrates to
single-electron carriers during metabolism and respiration. The mechanism and control of
electron transfer within various protein systems, although studied extensively, is not well
understood. Most electron transfer chains contain membrane proteins that are difficult to
purify, crystallize, and study in solution. Methylamine dehydrogenase (MADH, EC
1.4.99.3), which is a periplasmic enzyme found in gram-negative soil bacteria, along with
its physiological electron acceptor amicyanin, form part of a soluble electron transfer
complex where it has been possible to study protein-protein interactions and structures

using spectroscopic and X-ray crystallographic methods.

1.1 uinoen mes - a different class of redox enzymes
Quinoenzymes are enzymes whose catalytic mechanisms involve quinone-

containing prosthetic groups. The history of the quinoproteins began in the 19605 with

the characterization of the novel prosthetic groups of glucose dehydrogenase1 and
methanol dehydrogenase.2 More than ten years later, it was demonstrated that the
prosthetic group of methanol dehydrogenase was a quinone structure containing two
nitrogen atoms,3 and later it was shown to be pyrrolo-quinoline quinone (PQQ) by X-ray

diffraction analysis.4(figure I-l) A number of other bacterial dehydrogenases were
subsequently shown to contain PQQ, and the term quinoprotein was coined to include all
these proteins.5 About this time, incorrect identification of PQQ as the prosthetic group

of many other enzymes led to considerable confusion that was resolved with the

determination of the structures of other quinone cofactors, Tryptophan
tryptophquuinone (TTQ),6 Topaquinone (TPQ)7 and Lysine topaquinone

(LTQ).8(figure 1-1)

 

O \/\ H H H
_H ('ng N “('3 —C\/\
HOOC ”N \A C”:
1“” —N
CH
I 2
HOOC «Int,
CH
V\ H I 2
N —c —c
H gm
PQQ TPO rro ' LTO

Figure 1-1: Structure of known quinone cofactors: pyrrolo-quinoline quinone (PQQ),
topaquinone (TPQ), tryptophan tryptophquuinone (TTQ), lysine topaquinone (LTQ).

These cofactors can be bound non-covalently or derived from amino acids in the protein
backbone of the enzyme. The only non-covalently bound example is PQQ found in
bacterial enzymes. Tryptophan tryptophquuinone (TTQ) is derived from two
tryptophan residues and occurs in NAD(P)H-independent alkylamine dehydrogenases
from gram-negative soil bacteria. Topaquinone (TPQ) is derived from tyrosine and is the
prosthetic group of the copper-containing amine oxidases found in nearly all forms of life,
mammals, gram-positive and gram-negative bacteria, yeast, fungi, plants, fish, mollusks,
etc. Lysine topaquinone (LTQ) is also derived from tyrosine together with a lysyl
residue, and has been recently found in lysyl oxidase, an enzyme required for proper
cross-linking of collagen and elastin.8 Because these cofactors are quinones, they share
common chemical and biochemical properties. For example, all three are used by

enzymes that oxidize amines and are derived, in a more-or-less direct manner, fiom amino

acyl groups.

1.2 Methylamine dehydrogenase (MADH)

Methylamine dehydrogenase is a periplasmic quinoprotein9 found in several

constitutive and facultative methylotrophic bacterialoa11 This enzyme is induced when
bacteria grow on a medium containing methylamine as the sole source of carbon and

energy.

1.2.1 Structure
The identification of the correct nature and orientation of the redox cofactor of

methylamine dehydrogenase faced many acute difficulties. When MADH was first

described, based on its absorption spectra in the presence and absence of the substrate, its
sensitivity to carbonyl reagents, and its probable Schiff-base formation, it was suggested

that its prosthetic group consisted of a pyridoxal derivative that was acting in a novel

fashion.3a12a13 This group was subsequently thought to be a covalently attached form of

PQQ, but this possibility was later excluded by mass spectrometry of a derivative of the
isolated prosthetic-group peptide,14 and by analysis of X-ray data (at 2.25 A) obtained

for the whole enzyme which suggested a precursor form of PQQ (pro-PQQ).15 The
structure of the prosthetic group was finally solved by McIntire and his colleagues by an

extensive analysis of its semicarbazide derivative by l3C-NMR and mass
spectrometryfi:16 It is 2',4-bitryptophan-6,7-dione or Tryptophan Tryptophyl

Quinone (TTQ; figure 1-2).

Indole ring

Hndole-quinone ring

 

 

Figure 1-2: Structure of the oxidized TTQ cofactor of MADH.

Their proposed structure was consistent with the sequence of the gene for the light

subunit of MADH, indicating that the prosthetic group arose by post-translational
modification of two tryptophan residues.17 The structure proposed by McIntire was

subsequently shown to fit the X-ray data for methylamine dehydrogenase,18 and
resonance Raman spectroscopy confirmed the presence of same structure in a number of
other methylamine dehydrogenases.6a19'21 The two tryptophan residues that are post-
translationally modified are Trp108 and Trp57 and are linked by a covalent bond from C2
of one indole (Trp108) to C4 of the second indole (Trp57), which is also oxidized to an
ortho-quinone at C6 and C7 positions (Trq57). The indole and indole-quinone rings of
oxidized TTQ are not coplanar, but are at a dihedral angle of 38°.22 Crystal structures of

the resting enzyme in MADH at 1.75 A22 MADH-amicyanin binary complex23 at 2.24

A and MADH-amicyanin-cytochrome c551, ternary complex24:25 at 2.4 A are now
available. The native enzyme is void of metal ion cofactors and has an overall H2L2
subunit structure”,25 (figure 1-3) with the exception of MADH from Methylobacillus
flagellatum, which has a HL structure. From the functional point of View, MADH can be
considered as a pair of HL dimers, H ,L1 and Hsz, related by a molecular twofold axis.
The heavy (H) subunits have molecular masses that range from 40 to 50 kD, while the
light (L) subunits have molecular masses in the range of 13-16 kD, depending on the
source of enzyme. The 373 -residue H subunit is disc-shaped and forms a B-propeller
structure composed of seven repeated motifs, each having a four-stranded antiparallel B-
sheet that resembles the letter W. These seven B—sheets are related by pseudo-7-fold

symmetry whose axis passes through the center of the disc and is approximately

perpendicular to the molecular two-fold axis of the tetrameric enzyme. Each L subunit
contains 131 residues folded in a tight structure composed of five B-strands that form two
antiparallel B-sheets, of two strands and three strands each. A covalently bound TTQ
cofactor is present in each of the L-subunits. A channel leading into the active site is
located in the HlLl interface. The active site cavity in L1 subunit containing the quinone

oxygen atoms of TTQ is adjacent to the H1 subunit. In addition to the cross-link

provided by TTQ, the small L subunit contains six disulfide bonds.26 The large subunit
does not contain any cysteinyl groups. The HlLl interface is extensive, covering an area
of about 1600A2. The interface is largely hydrophobic, involving approximately 25
residues from each subunit with additional 20 hydrogen bonds linking the subunits. The
H1L2 interface is slightly less extensive, covering about 1100A2, but has roughly the same
number of hydrophobic and hydrogen-bonding interactions. The HIHZ interface is the
smallest, about 700A2, and contains two salt bridges and 12 van der Waals interactions
from each subunit. There is no contact between L and L2. The crystal structure of the

binary complex of MADH and amicyanin (both isolated from P. denitrificans) shows a
composition HszAmicyanin2.23a25 Each amicyanin is in contact with both the large and

the small subunits of MADH, but its interaction with TTQ-bearing small subunit is more

extensive. The structure of both MADH and amicyanin are virtually the same in the

binary complex as in the isolated molecules.24 Most of the TTQ ring system is deeply
buried within the L subunit. The modified tryptophquuinone side chain of residue
Trq57 is oriented such that it lies in the H-L interface channel with the two carbonyl
groups pointing into the channel region. The Trp108 fragment of TTQ has its phenyl

portion exposed at the surface of L subunit with atoms C5’ and C6’ only 9.3A from the

Pseudo 74 Id
Symmetry Axis

 

  

Pseudo 74 Id
Symmetry xls

Figure L3: The tetrameric structure of PD-MADH.

10

Cu site of amicyanin when the MADH-amicyanin complex is formed (figure 1-4). The 0-
quinol of Trq57 is pointed away from Cu and the distance from its C6 to the Cu is 15.8
A. The cytochrome molecule is also adjacent to the amicyanin, but is separated from
both the H and L subunits of MADH within each HLAC (heavy chain, light chain,
amicyanin, cytochrome) tetramer. The iron of the heme group and the copper of
amicyanin are 24.8 A apart.

Three of the four heteroatoms of TTQ form hydrogen bonds to the backbone
atoms: Trq57 O7 to Asp32 N, Trq57 N1 to Ser30 O/ Gly3l O, and Trp108 Nl’ to
Ala103 O with average bond distances of 3.0, 2.9 and 3.1 respectively in the two B-
subunits. The fourth heteroatom, 06 of Trq57, is hydrogen bonded to side-chains of
Thr122 and Asp76. There are two solvent molecules close to this oxygen, one of which
has interactions with the main-chain carbonyl groups of Asn104 and Ile106 while the
other interacts with the carboxylic group on Asp76. The active-site cavity containing
these two solvent molecules is a closed cavity inaccessible to the bulk solvent and is lined
with several hydrophobic side chains together with the side chains of Asp32 and Asp76.
The most likely route for the substrate to get access to the active site is through a deep
channel containing a tight cluster of highly ordered water molecules between the H and L
subunits in the HL dimer leading to the protein surface. A phenylalanine side chain of the
H subunit (Phe55H) acts as a potential "gate", isolating this large channel from the closed
active-site cavity. Movement of this or other side chain or main chain groups would be
needed for the substrate to gain access to the active site. Phe55H, side chain of Tyrl 19L

from L subunit and a number of neighboring oxygen atoms may also provide a binding site

11

 

L: Trq57 06

L: Trp108 CH2\

Figure 1-4: Orientation of TTQ in the MADH-amicyanin complex showing the relative
positions and distances between TTQ and Cu. Only the side chains of TTQ and Cu
ligands are shown.

12

for monovalent cations that are known to affect the reactivity and spectral properties of

TTQ as well as the oxidative half reaction. 1822,27

1.2.2 Biosynthesis of the TTO cofactor

TTQ biosynthesis likely requires external enzymes because amine dehydrogenases
do not contain metal ions. It has been hypothesized that the first of the two pre-cofactor
tryptophan residues (Trp5 7) in the sequence of the small subunit of the dehydrogenases
is converted to a tryptophquuinone in an extra-enzymatic process. This could be

followed by an enzyme-mediated or a self-catalytic cross-linking with the partner
tryptophan (Trp108) in the sequence.28 Methylamine-utilizing (mau) gene clusters have

been isolated and sequenced from several methylotrophic bacteria.29'31 Studies of these
clusters have indicated that at least 12 genes are involved in methylamine utilization. The
cluster contains the genes for the large and the small subunits (mauB and mauA,
respectively) of MADH. Several of the clusters harbor the gene for amicyanin. In

Paracoccus denitrificans (PD), the cluster is under the control of a LyrR-type
transcription factor.32 It has been suggested that a peroxidase is involved at some stage

in the transformation of Trp to tryptophquuinone and possibly in the

tryptophquuinone/Trp cross-linking process.

1.2.3 Spectral and oxidation-reduction properties

The color given to MADH by TTQ is yellow in dilute solutions and blackish-
green in concentrated solutions. When MADH is resolved into its individual subunits,

only the [3 subunits exhibit a visible absorption spectrum. Anaerobic titration of PD-

13

MADH with substrate or reductant, sodium dithionite proceeds through the semiquinone
intermediate, and reduction of the semiquinone during the second half of the titration was
observed to be quite slow indicating that the semiquinone is stable once formed and
suggested a kinetic barrier to further reduction. The oxidized, reduced and semiquinone
redox states of MADH generated by titration with sodium dithionite have been clearly
characterized with regard to their spectral properties. The UV-visible spectra of the three

redox states of PD-MADH are shown in figure 1-5. The absorption spectrum is

perturbed at alkaline pH values33a34 and by the presence of monovalent cations. A pK

value of 8.2 was obtained from the pH titration of the spectral changes observed for the

oxidized enzyme.34 At neutral pH, the semiquinone is slowly reoxidized under aerobic
conditions, but, the fully reduced enzyme is stable in the presence of oxygen and is
slowly reoxidized by ferricyanide. The overall oxidation-reduction midpoint potential

(Em) value of +100mV for the two electron couple (fiilly oxidized/fully reduced) was

obtained by electrochemical titrations of PD-MADH.35

1.2.4 Function

Quinoproteins may function either as oxidases or dehydrogenases, depending on
the nature of their quinone prosthetic group. Quinoproteins that possess TPQ and LTQ
are oxidases, whereas those, which possess PQQ and TTQ, are dehydrogenases. It is
important, physiologically, that most dehydrogenases do not react with 0; because their
role is to feed electrons derived from the oxidation of substrates, into the respiratory

chain, which would otherwise be prevented. It has been recently shown that when

14

 

0.4

Quinols

0.3 - “

Semiquinones

 

 

 

 

 

 

 

0

3

'9. 0.2

8 O-quinone

.D

<
0.1 - 1 ..
0.0 1 w L : 'T'T

300 400 500 600
Wavelength (nm)

Figure 1-5: Absorption spectra of different redox forms of MADH. Spectra of the O-
quinone, O-semiquinone, and O-quinol are displayed as solid lines. Spectra of the N-
semiquinone and N-quinol are displayed as dotted lines. All spectra were recorded in 6.7
M MADH in 10 mM BTP buffer at pH 7.5 and 25 C. (Taken from reference 54).

MADH is exposed to ultraviolet light in the presence of 02, it is oxidized and exhibits

substrate-dependent steady state oxidase activity.36 The effects of light are completely
reversible and oxidase activity is lost upon removal of light. The light-dependent
oxidation of MADH proceeds via a semiquinone intermediate, which accumulates to near

stoichiometric levels. The absorption of light may provide the energy needed to

15

overcome the thermodynamic barrier for electron transfer from TTQ to 02. This study is
focussed on the dehydrogenase activity of MADH.
As a dehydrogenase, MADH transfers electrons, extracted from substrate

methylamine, into the electron transport chain in the inner membrane of bacteria for the

production of ATP.37'39 It catalyzes the oxidative deamination of methylamine to yield

formaldehyde and ammonia. The net reaction can be written as:

CH3NH2 + H20 + 2Ami(Cu2+) —> HCHO + NH: + H30+ + 2Ami(Cu+)

This is the first step in the assimilation of methylamine, which serves as a carbon source
for methylotrophic bacteria. During this reaction, two electrons are transferred from
substrate methylamine to amicyanin (Ami in the above equation), a type 1 blue copper

protein of molecular weight 11.5kD, which mediates electron transfer from MADH to c-

type cytochromes.37a40'42 The sequential electron transfer pathway is then:
methylamine —> MADH —> amicyanin —> c-type cytochrome (cytochrome c55“) ——>
cytochrome oxidase. MADH from Paracoccus denitrificans (PD), Thiobacillus versutus
(TV) and Methylobacterium extorquens AMI (EX) is not able to reduce c-type

cytochrome in the absence of amicyanin,11 but MADH from Methylophilus

methylotrophus W3Al (WA) can act without the mediation of amicyanin.43
1.2.5 Reaction mechanism

Deuterium kinetic isotope effect and stopped flow kinetic studies of the reductive

half reaction of MADH, showed the absorbance changes upon mixing MADH with

16

methylamine that were best fit to a single-exponential decay. These results suggest that

direct two-electron reduction of TTQ by substrate occurs in the reductive half-reaction.44
As the oxidation of MADH occurs via two one-electron transfer steps, and the reduction
of TTQ is a two-electron process, the overall reaction can be separated into discrete
reductive and oxidative half reactions as shown in figure 1-6.

In the reductive half (figure I-6a), methylamine reacts with TTQ to form an

aminoquinol species. Based on the evidence from the reactions of benzylamines with

MADH, it was proposed by Davidson45 that the reaction is initiated by a nucleophilic
attack of substrate amine nitrogen on the quinone carbon to form a carbinolamine

intermediate; substitution of the amine nitrogen appears to occur at C6 carbonyl carbon of
TTQ.46 Further evidence for this intermediate was provided by resonance Raman

spectroscopy.19'21 The carbinolamine loses water to form an iminoquinone. A
subsequent nucleophilic attack by an active-site base leads to abstraction of a proton from
the irninoquinone intermediate to form a carbanion. The X-ray structure of the enzyme
indicates that the base is probably Asp76. Benzylamines are not full substrates of
MADH, but are competitive inhibitors able to reduce TTQ; analysis of Hammett plots

for their reactions with MADH led to the proposal that a key intermediate in the

reductive half-reaction is the carbanion.45 The involvement of a carbanion in the

mechanism is supported by the results of a detailed investigation of a second TTQ-

containing enzyme, aromatic amine dehydrogenase.47‘49 Hydrolysis of the carbanionic

intermediate leads to release of aldehyde and formation of aminoquinol (figure I-6a).

l7

Reductive half reaction (a)

 

v7 __ \'""7 £20 """ 7 \“7 {20 \""‘7 \‘7
‘— Ho—— ‘— ‘_ == ~<— : “$— =
(O/) \O / \\O N//J \W K \O- {NI/H \O- N/Hg \O-
IN“: Ht I lCHi’hB CH2 HB HO/CH2 EH
CH3 2
NH3
CH3NH2
H20
CU" """" fl \.----./
Cul / \ U / \\O
NH3
Amino-quinol N-semiquinone lmino-quinone Ortho-quinone

Figure 1-6: Proposed reaction pathway for MADH divided into reductive (a) and
oxidative (b) steps.

In the subsequent oxidative half reaction (figure I-6b), two electrons are

transferred sequentially from the substrate—reduced aminoquinol form of MADH to
amicyanin via the formation of an N-semiquinone intermediate.22 The existence of the N-
semiquinone intermediate has been confirmed by kinetic methods,50 and ESEEM
spectroscopy.51 Potentiometric studies on the pH dependence of the redox potentials

for the oxidized/reduced couples of MADH52 have shown that the quinol is singly
protonated while the semiquinone is anionic and deprotonated. This unusual feature
provides stabilization of the N-quinol and N-semiquinone reaction intermediates.
Conversion of semiquinone to the quinone requires two biochemical events, an electron
transfer to amicyanin and release of ammonia from TTQ. These events may occur in a

concerted manner or in a sequential fashion. It was initially thought that the first step in

18

the oxidation of the aminoquinol would involve release of ammonia and production of a

semiquinone, with the second electron transfer to amicyanin producing the oxidized

quinone.1 1’53 It was later shown using rapid scanning stopped-flow spectroscopy54
that the oxidation of the semiquinone by amicyanin occurs by a sequential mechanism in
which oxidation to an iminoquinone precedes hydrolysis and release of ammonia (figure 1-
6). 1n the presence of excess of substrate, methylamine rather than water initiates a
nucleophilic attack at C6 position to release ammonia and to form the next TTQ-substrate
imine adduct. It was also found that the two-electron transfer is linked to transfer of a

single proton, which is involved during the conversion of the fully reduced quinol to N-

semiquinone form (figure I-6b).54 These steps in the reductive and the oxidative half
reaction involve several acids and bases, which could be fulfilled by side chains of Asp76,
Thr122, Asp32 and Tyr119 present near the active site. Since the space in the active site
is limited, it is possible that the same residue is involved in more than one step.

This reaction sequence is just one of many chemical mechanisms employed by

enzymes for the oxidation of amine substrates, other enzymes catalyzing similar
transformations include the flavoproteins trimethylamine dehydrogenase,55 sarcosine
oxidases56 and the copper-containing amine oxidases that utilize topaquinone (TPQ) as

the redox cofactor.57 The reductive half-reaction which involves cleavage of one of the C-
H bonds of methylamine poses a common and major problem to these enzymes, that is
the difficulty in cleaving a stable substrate C-H bond. Stopped-flow kinetics have
demonstrated the presence of two kinetically significant intermediates, a relatively fast
transition due to reduction of TTQ by substrate and a slower transition due to release of

the aldehyde product implying a mechanism involving rate-limiting abstraction of a

19

methyl proton.58 For bovine plasma amine oxidase, a large isotope effect was explained
by a mechanism involving proton tunneling. It is possible that similar quantum

mechanical effects play role in the proton abstraction step in the reaction between

MADH and methylamine. In a recent study59 enzymatic breakage of the substrate C-H
bond by WA -MADH was studied by stopped-flow spectroscopy. The rate of reduction
of the TTQ cofactor was observed to have a large kinetic isotope effect (KIE = 16.8 :t
0.5), and the KIE was independent of temperature. Reaction rates were strongly
dependent on temperature, indicating that the H-transfer reaction was driven by thermally
induced, vibrational motion. From the analysis of the temperature dependence of C-H
bond breakage, it was inferred that extreme (ground state) quantum tunneling driven by
vibrational motion of the protein scaffold was responsible for the transfer of the hydrogen
nucleus. The results demonstrated that classical transition state theory and its tunneling
derivatives were not adequate to describe this enzymatic reaction.

Spectroscopic studies on N-semiquinone showed that the delocalization of
electrons into the ring system of the semiquinone is consistent with the route for

departure of electrons from the indole ring that was proposed on the basis of the X-ray

structure23 and kinetic studies.60 In the MADH-amicyanin complex, the tryptophan
that does not contain the o-quinone (Trp108) lies near the surface, only 9.3 A away from
the amicyanin copper atom. His95, one of the four copper ligands, is on the surface of
the protein and about halfway between copper and TTQ, with the shortest distance
between them being approx. 5.5 A (figure 1-4). It was therefore suggested that this

histidine might mediate electron transfer between the redox centers, thus forming an

electron transfer triad.23 On the basis of st0pped-flow kinetics, an electron transfer

20

pathway was suggested that involves a 3.6 A jump through space from TTQ to the

carbonyl of a proline residue (Pro94) and passage through six covalent bonds to the

copper (a total distance of 14 A)60. An alternative pathway was suggested for electron

transfer in the ternary complex with cytochrome c, which is moderately more efficient but

depends critically on the presence of an intracomplex water molecule.25

1.3 flrlier studies on TTQ-semiquinone

Observation of a metastable N-semiquinone intermediate during controlled in vitro
oxidation of substrate-reduced MADH was reported by Davidson et a1 (1990)61 and

supported by previous pulsed EPR studies.51 For these studies, TTQ semiquinone was
prepared by comproportionation of a 1:1 molar mixture of oxidized and substrate-reduced

TTQ cofactors that resulted from addition of a 0.5 molar equivalent of substrate to the

resting enzyme.62 As a result, enzyme samples contained an approximate 1:1 molar
proportion of N-semiquinone, formed by direct substrate-reduction followed by an
electron transfer and O-semiquinone, formed by direct one electron reduction of the
oxidized TTQ. The N-semiquinone species was not stable in solution at relatively high
protein concentrations. In order to minimize oxidation and disproportionation reactions,
anaerobic conditions were maintained. For comparative ESEEM studies, semiquinone
MADH samples were prepared by comproportionation with 14N- and 15N-methylamine.
The resulting ESEEM spectra clearly showed l4N modulations from methylamine derived
nitrogen. Because the intensities of these l4N modulations were similar to those found for

the heteroatoms of TTQ', the authors concluded that methylamine derived 14N was

21

covalently bound to TTQ' and that NH3 release must take place after, or in concert with

oxidation of the TTQ semiquinone.51

The presence of covalently bound substrate nitrogen atom in TTQ was questioned
by Duine et a1. (1994)63,64 and McIntire et a1. (1994)65 based on their studies of the
effects of alkali metal ions on the EPR signal of MADH semiquinone. These studies
investigated the absorbance changes induced by the addition of a 104-106 molar excess of
the chloride salts of several monovalent cations to the enzyme. Two types of pH

dependent effects were noted: small cations (K+, Na+) caused bleaching of all absorbance

bands in the visible spectrum, whereas large cations (Cs+, (CH3)2NH2+ (CH3)3NH+ and

(CH3)4N+) caused a red shift in the visible absorbance spectrum. With NH4+ and Rb+,
both effects were observed, the former effect at high and the latter at low concentrations.
From the observation that the nature of the effect of NH4+ is concentration-dependent, it
was concluded that MADH has two different binding sites: a high-affinity binding site
that specifically binds bigger monovalent cations and a less specific low-affinity binding

site that preferentially binds smaller cations. From resonance Raman studies it was noted
that the effect of C51", NH4+, and (CH3)3NH+ on the spectrum of oxidized MADH was to

shift the peaks assigned to carbonyl stretching vibrations to lower frequencies. This
downshift was explained on the basis of a possible weakening of the C=O bond by the
presence of a positive charge. Larger shifts were predicted for the more electropositive
K" and Na+ ions, but were not observed. On the basis of above results, these studies
proposed an imine elimination mechanism that ruled out the presence of covalent binding

between methylamine and TTQ. It was concluded that methylamine and monovalent

22

cations share the same binding site on MADH. Hence, it was inferred that methylamine

is bound to MADH as methylammonium ion (CH3NH3+) and is stabilized by the peptide

carbonyl oxygens present in the binding site.64 The fact that the frequencies due to C, N
and H vibrational modes of the TTQ remained essentially unaffected by addition of

cations, was taken as an evidence that the cations did not affect the TTQ structure.

These authors suggested that the l4N-ESEEM detected by Warncke, et a1.51 could have
originated from a through-space magnetic coupling, casting doubt on the mechanistic
conclusions of the earlier work.

Rapid scanning stopped-flow spectroscopy and global kinetic analysis were used

by Davidson, et a1. (1996),50 to demonstrate that the N-semiquinone is a true
physiological reaction intermediate which accumulates during two sequential one-electron

oxidations of reduced MADH by amicyanin. This study was ftu'ther substantiated by

direct detection of the N-semiquinone intermediate of MADH using 15N-NMR. These
authors resolved a 15N-NMR resonance at 54 ppm that was resistant to chemical

exchange and was distinguishable from free- or adventitiously bound l5N as it had shorter

T150 These findings were consistent with the proposed mechanism54 (figure 1-6), which
involves covalent addition of the substrate nitrogen to TTQ and the formation of an

aminoserniquinone after one-electron oxidation of the aminoquinol.
1.4 Usefulness of quinoproteins to higher animals

The usefulness of quinoproteins in nutrition and medicine has been discussed in

the past.66 It is not likely that these quinone cofactors can act as essential nutrients or

23

vitamins for humans. The release of these cofactors from their parent proteins requires
extensive proteolysis and they are released as amino acyl derivatives. Even if these freed
cofactors found their way into the human body, there is no known mechanism to
incorporate such modified amino acids into nascent polypeptides, although it could be
possible that the amino-acyl derivatives could bind noncovalently to an apoprotein. If
human enzymes do contain these quinone cofactors, it is more likely that they are also
made in situ. TTQ also fails to meet the fundamental requirement of a vitamin, that it
should be widely distributed and easily accessible from foods. At present, it is believed
that the accessibility of external TTQ is restricted and has little or no significance for the
well being of humans. It is produced in situ as an integral component in a limited number
of enzymes from small groups of organisms and it is not anticipated that these enzymes
will grow significantly.

It has been proposed that the quinoproteins can be used as biocatalysts to
produce research and industrial amounts of useful chemicals. 1n membrane-free cell
extracts or for pure quinoproteins, large levels of a desired chemical could be produced if a
convenient natural or artificial electron transport chain was used, for example, a pure or

partially pure natural electron acceptor for the quinoproteins, or components of

methylotroph membrane electron transport chains.11:38a57a67 Another possible route is

coupling the protein directly or via mediators to an electrode for electrochemical
regeneration of substrate-reduced enzyme.68'7O Pure alcohol dehydrogenases71”73 and

methylamine dehydrogenase,74 have been used to construct immobilized enzyme-based
electrodes. These enzyme-based electrodes can be used to produce useful chemicals, and

could be designed for use as biosensors to measure levels of specific substances in

24

biological tissues and fluids, in the same way as has been done with the flavoprotein,

glucose oxidase.71975

The role of the semiquinone in catalysis and electron transfer by MADH, and the
mechanism by which this enzyme, possessing a four-electron capacity from two
cofactors, donates electrons to a one-electron carrier, amicyanin , is not yet understood
completely. The present work was directed towards finding evidence in support of the
presence of covalently bound substrate nitrogen in the TTQ cofactor, understanding the
changes in the electronic properties as a result of covalent binding of the substrate

nitrogen into the TTQ ring and getting further insight into the electron transfer process.

25

 

List of References

1)Hauge, J. G. J. Biol. Chem. 1964, 239, 3630-3639.

2)Anthony, C.; Zatman, L. J. Biochem. J. 1967, 104, 960-969.

3)Duine, J. A.; Frank, J .; Weaterling, J. Biochim. Biophys. Acta 1978, 524, 277-287.
4)Salisbury, S. A.; Forrest, H. S.; Cruse, W. B. T.; Kennard, 0. Nature 1979, 280, 843-
844.

5)Duine, J. A.; Frank, J.; Verwiel, P. E. J. Eur. J. Biochem. 1980, 108, 187-192.
6)Mc1ntire, W. S.; Bates, J. L.; Brown, D. E.; Dooley, D. M. Biochemistry 1991, 30, 125-
133.

7)Janes, S. M.; Mu, D.; Wemmer, D.; Smith, A. J .; Kaur, S.; Maltby, D.; Burlingame, A.
L.; Klinman, J. P. Science 1990, 248, 981-987.

8)Wang, S. X.; Mure, M.; Medzihradszky, K. F .; Burlingame, A. L.; Brown, D. E.;
Dooley, D. M.; Smith, A. J .; Kagan, H. M.; Klinman, J. P. Science 1996, 273, 1078-1084.
9)De Beer, R.; Duine, J. A.; Frank, J .; Jones, L. H. Biochim. Biophys. Acta 1995, 662,
370-374.

10)Anthony, C. PQQ and Quinoproteins; Jongejan, J. A. and Duine, J. A., eds., Ed.;
Kluwer Academic Publishers: Dordrecht, 1989, pp 1-11.

11)Davidson, V. L. Principles and Applications of Quinoproteins; Marcel Dekker, Inc.:
New York, 1993.

12) van der Meer, R. A.; Jongejan, J. A.; Duine, J. A. FEBS Lett. 1987, 254, 299-304.
l3)Eady, R. R.; Large, P. J. Biochem. J. 1971, 123, 757-771.

l4)Mc1ntire, W. S.; Stults, J. T. Biochem. Biophys. Res. Commun. 1986, 141, 562-568.
15)Vellieux, F. M. D.; Hol, W. G. J. FEBS lett. 1989, 255, 460-464.

l6)Mclntire, W. S. Methods Enzymol. 1995, 258, 149-164.

l7)Christoserdov, A. Y.; Tsygankov, Y. D.; Lidstrom, M. E. Biochem. Biophys. Res.
Commun. 1990, I 72, 211-216.

18)Chen, L.; Mathews, F. S.; Davidson, V. L.; Huizenga, E. G.; Vellieux, F. M. D.; Duine,
J. A.; Hol, W. G. J. FEBS Lett. 1991, 287, 163-166.

l9)Dooley, D. M.; Brown, D. E. Meth. Enzymol. 1995, 258, 132-140.

26

20)Backes, G.; Davidson, V. L.; Huitema, F .; Duine, J. A.; Sanders-Loehr, J. Biochemistry
1991, 30, 9201-9210. .

21)Dooley, D. M.; Brown, D. E. Principles and Applications of Quinoproteins; Davidson,
V. L., Ed.; Marcel Dekker: New York, 1993, pp 132-140.

22)Chen, L.; Doi, M.; Durley, R. C.; Y., C. A.; E., L. M.; Davidson, V. L.; Mathews, F.
S. J. Mol. Biol. 1998, 276(1), 131-49.

23)Chen, L.; Durley, R. C.; Poliks, B. J.; Hamada, K.; Chen, Z.-W.; Mathews, F. S.;
Davidson, V. L.; Satow, Y.; Huizinga, E.; Vellieux, F. M. D.; Hol, W. G. J. Biochemistry
1992, 31, 4959-4964.

24)Chen, L.; Durley, R. C.; Mathews, F. 8.; Davidson, V. L.; Tegoni, M.; Rivetti, C.;
Rossi, G. L. Protein Sci. 1993, 2, 147-154.

25)Chen, L.; Durley, R. C. B; Davidson, V. L.; Mathews, F. S. Science 1994, 264, 86-89.
26)Christoserdov, A. Y.; Boyd, J .; Mathews, F. S.; Lidstrom, M. E. Biochem. Biophys.
Res. Co. 1992, 184, 1181-1189. .

27)Chen, L.; Mathews, F. S.; Davidson, V. L.; Huizinga, E.; Vellieux, F. M. D.; Hol, W.
G. J. Protein Struct. F unct. Genet. 1992, 14, 288-299.

28)McIntire, W. S.; Christoserdov, A. Y.; Wemmer, D. E.; Lidstrom, M. E. Science 1991,
252, 817-824.

29)Willner,1.; Heleg-Shabtai, V.; Blonder, R.; Katz, E.; Tao, G. J. Am. Chem. Soc. 1996,
118, 10321-22.

30)Lidstrom, M. E.; Christoserdov, A. Y. Microbial Growth on C1 Compounds; Murell,

J. C. and Kelly, D. P., Ed.; Intercept: Andover, UK, 1993, pp 381-400.
31)Christoserdov, A. Y.; Christoserdova, L. V.; McIntire, W. S.; E., L. M. J. Bacteriol.
1994, 1 76, 4052-4065.

32)van Spanning, R. J. M.; van der Palen, C. J. N. M.; Slotboom, D.-J.; Reijinder, W. N.
M.; Stouthamer, A. H.; Duine, J. A. Eur. J. Biochem. 1994, 226, 201-210.

33)Davidson, V. L. Biochem. J. 1989, 261, 107-111.

34)Davidson, V. L.; Jones, L. H.; Kumar, M. A. Biochemistry 1990, 29, 5299-5304.
35)Husain, M.; Davidson, V. L.; Gray, K. A.; Knaff, D. B. Biochemistry 1987, 26, 4139-
4143.

36)Zhu, 2.; Davidson, V. L. Biochim. Biophys. Acta. 1998, 1 3 64, 297-300.

27

37)Tobari, J.; Harada, Y. Biochim. Biophys. Res. Commun. 1981, 101, 502-508.
38)Anthony, C. Bacterial Energy Transduction; Anthony, C., Ed.; Academic Press:
London, 1988, pp 293-316.

39)Anthony, C. Principles and applications of Quinoproteins; Davidson, V. L., Ed.;
Marcel Dekker: New York, 1993, pp 223-244.

40)van Houwelingen, T.; Canters, G. W.; Stobbelaar, G.; Duine, J. A.; Frank, J. J .;
Tsugita, A. Eur. J. Biochem. 1985, 153, 75-80.

41)Husain, M.; Davidson, V. L. J.Biol. Chem. 1985, 260, 4626-4629.

42)Lawson, S. A.; Anthony, C. Biochem. J. 1985, 228, 719-726.

43)Christoserdov, A. Y.; E., L. M.; Mathews, F. S.; McIntire, W. S. J. Bacteriol. 1994,
1 76, 4073-4080.

44)Brooks, H. B.; Jones, L. H.; Davidson, V. L. Biochemistry 1993, 32, 2725-2729.
45)Davidson, V. L.; Jones, L. H.; Graichen, M. E. Biochemistry 1992, 31 , 3385-3390.
46)Huizinga, E. G.; Vanzanten, B. A. M.; Duine, J. A.; Jongejan, J. A.; Huitema, R;
Wilson, K. 8.; H01, W. G. J. Biochemistry 1992, 31, 9789-9795.

47)Hyun, Y. L.; Davidson, V. L. Biochemistry 1995, 34, 816-823.

48)Hyun, Y. L.; Davidson, V. L. Biochim. Biophys. Acta. Protein Struct. Mo]. Enzymol.
1995, 1251, 198-200.

49)Govindraj, S.; Eisenstein, E.; Jone, L. H.; Sanders-Loehr, J .; Christoserdov, A. Y.;
Davidson, V. L.; Edwards, S. L. J. Bacteriol. 1994, 1 76, 2922-2929.

50)Bishop, G. R.; Brooks, H. B.; Davidson, V. L. Biochemistry 1996, 35, 8948-8954.
51)Wamcke, K.; McCracken, J. L.; Brooks, H. B.; Babcock, G. T.; Davidson, V. L. J. Am.
Chem. Soc. 1993, 115, 63-64.

52)Zhu, Z.; Davidson, V. L. J. Biol. Chem. 1998, 273, 14254-14260.

53)van Wielink, J. E.; Frank, J .; Duine, J. A. PQQ and Quinoproteins; Jongejan, J. A. and
Duine, J. A., Ed.; Kluwer Academic Publishers: Dordrecht, 1989, pp 269-278.

54)Zhu, Z.; Davidson, V. L. Biochemistry 1999, 38, 4862-4867.

55)Rohlfs, R. J .; Hille, R. J. Biol. Chem. 1994, 269, 30869-30879.

56)Wagner, M. A.; Jorns, M. S. Arch. Biochem. Biophys. 1997, 342, 176-181.
57)Anthony, C. Biochem. J. 1996, 320, 697-711.

28

 

58)McWhirter, R. B.; Klapper, M. H. PQQ and Quinoproteins; Jongejan, J. A. and
Duine, J. A., Ed.; Kluwer Academic Publishers: Dordrecht, 1989, pp 259-268.
59)Basran, J .; Sutcliffe, M. J.; Scrutton, N. S. Biochemistry 1999, 38, 3218-3222.
60)Brooks, H. B.; Davidson, V. L. Biochemistry 1994, 33, 5696-5701.
61)Christoserdov, A. Y.; Tsygankov, Y. D.; Lidstrom, M. B; Davidson, V. L. Meth.
Enzymol. 1990, 188, 241-246.

62)Davidson, V. L.; Jones, L. H.; Kumar, M. A. Biochemistry 1990, 29, 10786-10791.
63)Gorren, C. F. A.; Duine, J. A. Biochemistry 1994, 33, 12202-12209.

64)Gorren, C. F. A.; Moenne-Loccoz, P.; Backes, G.; Vries, S.; Loehr, J. S.; Duine, J. A.
Biochemistry 1995, 34, 12926-12931.

65)Kuusk, V.; McIntire, W. S. J.Biol. Chem. 1994, 269, 26136-26143.

66)Mc1ntire, W. S. Annu. Rev. Nutr. 1998, 18, 145-177.

67)Anthony, C. Principles and applications of Quinoproteins; Davidson, V. L., Ed.;
Marcel Dekker: New York, 1993, pp 17-45.

68)Hill, H. A. 0.; Hunt, N. 1. Methods Enzymol. 1993, 227, 501-522.

69)Davidson, V. L. Am. Biotech. Lab. 1990, 8, 32-34.

70)Burrows, A. L.; Hill, H. A. 0.; Leese, T. A.; McIntire, W. S.; Nakayama, H.;
Sanghera, G. S. Eur. J. Biochem. 1991, 199, 73-78.

71)Karube, 1.; Yokoyama, K.; Kitagawa, Y. Quinoproteins as biosensors in Principles
and Applications of Quinoproteins; Davidson V. L., E., Ed.; Marcel Dekker, Inc.: New
York, 1993, pp 429-46.

72)Stigter, E. C. A.; de Jong, G. A. H.; Jongejan, J. A.; Duine, J. A.; van der Lugt, J. P.;
Somers, W. A. C. Enzym. Microb. T echnol. 1996, 18, 489-94.

73)Zhao, S.; Lennox, R. B. Anal. Chem. 1991, 63, 1174-78.

74)Loughran, M. G.; Hall, J. M.; Turner, A. P. F.; Davidson, V. L. Biosensors
Bioelectron. 1995, 10, 569-76.

75)Raba, J.; Mottola, H. A. Crit. Rev. Anal. Chem. 1995, 25, 1-42.

29

 

Chapter II

PRINCIPLES OF CONTINUOUS WAVE AND PULSED ELECTRON
PARAMAGNETIC RESONANCE SPECTROSCOPY

11.1 Continuous wave Electronic Paramagnetic Resonance (cw-EPR)

Electronic Paramagnetic Resonance (EPR) spectroscopy is a technique applicable
to systems with net electron spin angular momentum. In this experiment an oscillating
microwave magnetic field is applied to a paramagnetic system in order to invert its
electron magnetic moment in the presence of an external magnetic field. The basic
principles of EPR are similar to that of Nuclear Magnetic Resonance (N MR). A
thousand-fold larger electron moment makes EPR more sensitive to its environment,
leading to broader spectral widths and faster relaxation times as compared to NMR.
Many useful NMR techniques like magic-angle spinning, decoupling nuclei during an FID
are not routine in EPR because of technological limitations. The theory, instrumentation

and methodology of EPR spectroscopy has been described in several books and journal

articles. 1 ‘8 In this chapter principles of cw-EPR and pulsed-EPR methods are discussed

briefly.

[1.1.1fle electronic Zeeman interaction

In a molecule, permanent magnetic dipoles are associated either with electrons or
with nuclei. The magnetic dipoles attributable to the electrons arise from net spin or net
orbital angular momentum or a combination of these. In most of the molecules the
contribution from the spin angular momentum dominates. The magnetic dipole can

interact with the magnetic component of the microwave radiation. When such a molecule

30

is irradiated with a range of spectral frequencies, it normally leads to no absorption
attributable to a magnetic interaction, because the energy levels are degenerate. In the
presence of a static magnetic field, which can also be provided internally by the nuclei of a
molecule (section V. 1 .2), this degeneracy is lifted and absorption attributable to magnetic
dipole transitions may occur. The permanent magnetic dipole moment (ue) arising from

net electron spin angular momentum (S) is given by,

ue=-gfiS=-rh8 [2.1]

The negative sign indicates that for an electron the magnetic moment is antiparallel to the
spin, g has a value of 2.0023 for a free electron and is considered isotropic in the above
expression, B the is Bohr magneton, the magnetic moment for one unit of quantum
mechanical angular momentum. It is defined by B = e h/2me in MKS units, where, e is the
charge on an electron, me is the mass of an electron. y is the gyromagnetic ratio, the ratio
of the magnetic moment to the angular momentum for an electron. According to the
classical electrodynamics, the electron spin will experience a torque equal to ue X B0 in the
presence of an external magnetic field (8.). So, the magnetic moments will take preferred
orientations. The individual spins will precess around the magnetic field axis at angular
frequency (00 = g B B0, also called the Larmor frequency which is roughly 8.3 GHz at
magnetic field of 0.3 T. The energy of a magnetic dipole in a static external magnetic field,

B0, is given by the expression,

W = -He' B0 [22]

31

For a quantum mechanical system the interaction energy of a spin with the
external magnetic field is quantized. The electronic Zeeman Hamiltonian describing the

interaction of electron magnetic moment with the magnetic field is,

gee-Zeeman : '(‘g B S) ' B09 [2'3]

The eigenvalues for the projection of S on to a specified direction, usually the z direction

of a cartesian coordinate system are Msh where MS can have values from -S to +S in

integral steps. So, there will be (ZS + 1) states with different energies and spin
orientations. If the direction of B0 is along z-axis, the Hamiltonian [2.3] can be rewritten

35,

sieve-Zeeman = g fi 8082

and its energy eigenvalues are,

Ee-Zceman = g fl BoMs [24]

For a species with single unpaired electron, M5 = il/2, so there will be two possible
energy levels corresponding to two possible orientations of the electron, parallel and
antiparallel to the external magnetic field (figure 11-1). A transition between these levels
would involve a change in the orientation of the electron magnetic moment. Hence, the
radiation inducing the transition must be polarized such that its oscillating magnetic field
has a component perpendicular to the static magnetic field. The transition energy falls in
the microwave range and is given by,

AB = g 3 Bo [2.51

32

Ea = + gBBo/Z

0!. (M5 = + 1/2)
98
—a- 73—; AE - hV — gBBo 31
EB = - gBBo/Z
.—» B(Ms=-1/2) , _”,
. . Bo 1ncreasrng
B0 = 0 B0 1ncreasrng

Figure II-l: (a) Splitting of degenerate 0t and B electron spin states due to electron
Zeeman interaction in the presence of applied static magnetic field Bo. (b) An EPR line
expected from a transition in (a) for systems with isotropic g value.

As a result of absorption of microwave radiation, the microwave power returned from the
sample to the detector changes and is detected as an absorptive peak. The EPR spectrum
is a plot of intensity against the strength of applied magnetic field and due to the phase
sensitive detection system usually employed, it is recorded as a derivative curve. The
features are easily visible in a derivative mode if the absorption lines are broad.
Difference in the transition energies for different molecules are described by changes in the
g-values, hence g-value is characteristic of the sample. The g value can be dependent on
the orientation of the molecular axis with respect to the external field. This anisotropy in
g-factor can arise from coupling of the spin angular momentum with the orbital angular

momentum. The spin-orbit coupling is described by the Hamiltonian,

veg: J. L - s [2.6]

33

 

where L is the orbital angular momentum and 7» is spin-orbit coupling constant for a
particular shell in particular atom. The spin angular momentum is oriented with the field,
but the orbital angular momentum, which is associated with electrons moving in molecular
orbitals, is locked to the molecular wave function (figure 11-2). The total angular
momentum (J) of electron is the vector sum of spin (S) and orbital (L) components, J = L
+ S, where J has the possible magnitudele - S l, l L - S +1 |,..., l L + S |. When the
effects of orbital moment are small, they are incorporated into the g value thus, making
the g value anisotropic. This corresponds to combining dawn...“ and (37655 and the

overall magnetic moment is given by, u = g B J and the g factor is modified as,

:3+S(S+1)-L(L+l)

 

 

 

 

2 2m + 1) [2'7]
A u,
“I. “1-
ll 1 \
l u X
Bo “5 Mr us
O “
(a) (b) (c)

Figure II-2: Coupling of the projections of the spin and the orbital angular momentum in
case of two different molecular orientations, (a) and (b), relative to the applied static
magnetic field BO. (0) The vector addition of orbital and spin magnetic moments.

34

The modified Hamiltonian is given by,
cilia-gum = B B0 - g - S [2.8]

where g is a tensor. In some cases the g tensor can have axial symmetry, i.e. g“ = gu; g i
= gxx = g”, where z-axis defines the axis of symmetry. In general, g forms a symmetric
tensor with six components, which could be diagonalized by finding its principal axis
system.

Quantum mechanically EPR spectrum can be interpreted as the allowed
transitions between the eigenvalues of the eflective spin Hamiltonian containing all the
terms that modify the EPR transition energies given by equation 2.5. These include
electron-nuclear hyperfine interaction (hfi), nuclear Zeeman interaction, nuclear
quadrupolar interaction (nqi), zero field splitting (zfs) etc. resulting from the interaction
of electron with nearby charges, dipoles and quadrupoles. These effects that modify

electron energy levels are discussed below.

11.1.2 Nuclear Zeemflntqflon

Protons and neutrons both have a spin quantum ntunber of 1/2 and depending
upon how they pair up in the nucleus, the nucleus may or may not have a net non-zero
nuclear spin. If the spins are all paired, then there will be no net spin and the nuclear spin
quantum number I will be zero. When the number of protons and neutrons are not both
even, then the nucleus can have net spin. Nuclear spins may be integral or half integral.

There is a magnetic moment the associated with the nuclear spin angular momentum I,

35

“N = gN fiNI : YN h I, [2.9]

where m is the nuclear gyromagnetic ratio and is characteristic of the isotope, fly is
nuclear magneton, gN is nuclear g-factor characteristic of a specific nucleus. In contrast to
the electron magnetic moment (eq. 2.1), there is no negative sign in eq. 2.9 indicating that
nuclear spin and magnetic moment are parallel. This holds true for protons and most of
the nuclei, but for those nuclei that are one neutron short of being factorizable into an
inteng number of 4He units, e.g. 15N, a negative gyromagnetic ratio, where nuclear spin
and magnetic moment are antiparallel, is found. The interaction energy of a nuclear
magnetic moment with an applied magnetic field is also quantized, and there are (21 + 1)
allowed values for its projection on the field axis, given by -I, (-1 +1),. . ., +1. The axis of
quantization of the nucleus reorients when an electron spin transition occurs. The nuclear
Zeeman Hamiltonian describing the interaction of nuclear magnetic moment with the

magnetic field directed along the z-axis is,

demo-Zeeman = '(YN h I) ' B0 = 'YN h B0 129 [210]

The energies of nuclear Zeeman levels are given by,

EN=-yNhBom1 [2.11]

This would lead to splitting of energy levels (figure 11-3). In a 0.3 T magnetic field,
typical for EPR at X-band frequencies, the nuclear Larmor frequency for protons is about

14 MHz (cf 000, electron, 11.1.1). The nuclear Zeeman effect is much smaller than the

36

electronic Zeeman interaction and typically smaller than the hyperfine interactions. It can

often be neglected during simulations of EPR spectra.

 

 

 

 

 

 

 

 

 

 

M, Ml
-l/2
+1/2 ‘3. -~ A ’i
’o’ ‘s‘ """""" e ‘ :—
.'1 +1/2 E i . .
.t : a tag :*- A... *-
——( l l =\ l l
E 5 :51 =
f.‘ -1/2 ..... ‘1 1:—
-1/2 x - i
~——r ; —>
‘s, ¢ : Bo increasing
+1/2 .... ‘7
gfl Bosz "gNfiN BoIz S ' A ° I

Figure 11-3: (a) Energy level diagram showing the effects of electron Zeeman, nuclear
Zeeman and Hyperfine interactions of an unpaired electron (S = 1/2) with a I = 1/2
nucleus in the presence of an external magnetic field 8,; (b) Resulting EPR spectrum.

11.1.3 The hyperfine interaction

The interaction between the electron spin and neighboring nuclear spins is called
the electron-nuclear hyperfine interaction (hfi). An unpaired electron in the vicinity of a
nuclear spin experiences a local magnetic field generated by the nearby nucleus (Blow).
The transition energy modified by hfi is given by, AE = gB (BC + Blow), the value of B0
needed to excite a transition will depend on Blow. A nuclear spin is quantized in the total
magnetic field it experiences, which is the resultant of B0 and the field due to electron.

For a nuclear spin 1, there are (21 + 1) distinct nuclear spin states and each of these gives

rise to a different local field at the electron. The energy differences between the nuclear
spin states are small, and all the nuclear spin states are equally populated for practical
purposes. Thus, there are (21 + 1) possible values of B0 that can fulfill the resonance
condition and these occur with equal weight. As a result the spectrum is split into (21 +
1) lines of equal intensities.

For the case of an axial hfi, the electron-nuclear coupling can be separated into two
components, the anisotropic dipole-dipole interaction and the isotropic Fermi contact
interaction. When an unpaired electron has access to an s-orbital associated with a
magnetic nucleus, there is a non-zero probability density of the electron and nuclear
moments "contacting" one-another. Because s-orbitals have spherical symmetry, the
resulting hfi is isotropic and is known as Fermi contact interaction. The energy of this

isotropic hyperfine interaction is given approximately by,

Ea. = -2/3 no lv(0)|2 tn- MN [2121

where no is the permeability of vacuum and |W(O)IZ is the probability of finding an electron
at the nucleus assuming that the electron probability distribution is uniform over the
volume of the orbital. Using equations 2.1 and 2.9 to replace magnetic moments by their

operators, the Hamiltonian can be written as,

A

deg: A.,,sr, [2.13]

A,so is called isotropic hyperfine coupling constant expressed in frequency units and is

defined as,

38

Also =. '2/3 ”o g fl gN [3N IW(O)12 [2141

Thus, the isotropic hyperfine coupling is proportional to the probability of finding the
unpaired electron at the nucleus. Hence, magnetic nuclei in paramagnetic molecules can
provide information about the unpaired electron distribution and its wavefunction.

The anisotropic hyperfine interaction arises from the magnetic interaction between
the magnetic dipoles of the electron and a nearby nucleus. This interaction occurs when
the electron is in an orbital with orbital quantum number I > 0. Because the electron
probability distribution for these orbitals is zero at the nucleus, the Fermi contact
interaction is not possible. As these orbitals have non-spherical symmetry, this
hyperfine interaction is anisotropic. For the case of an atom with nucleus at a distance ‘r’
from the electron at the origin, (figure 11-4) the dipolar interaction energy can be obtained
by assuming the electron and nuclear dipoles to be point dipoles. Then, using the point

dipole-dipole model, the energy of a magnetic dipole in the field of the other is,

 

 

Edipolar = “a . ”N _ 3(p'c ' r)(1’lN . r) ' [2.15]
r3 r5

and the Hamiltonian describing such interaction can be written as,
Cfiflipear: -g B gN [Ms - i/r3 -3(s- r)(I- r)/r5] [2.16]

The above Hamiltonian can be expanded in the Cartesian coordinates as,

Gigfipolhr =

        

2 2 2 2
gNZ—r{(r__:__ 3x)1 IS +————-(r 3y>15 +———(r 3Z)IS+

xx 5 yy 5 22
1' 1'

               

—ZSEYZ(IS +Izy—S) —ZS—)EZX(IS +IHS)} [2.17]

39

AzllB0

 

 

 

Figure 11-4: Dipolar interaction of magnetic dipoles rm and 11. due to nuclear spin angular
momentum and electron spin angular momentum respectively. The lengths of vectors tin
and u, are not drawn to scale. The electron is at the origin at a distance ‘r’ from the
nucleus. The dipoles are aligned with the external magnetic field and 0 is the angle between
the electron-nuclear vector r and the applied external magnetic field Bo.

By integrating over the spatial wavefunction [2.17] can be written as,

steam; TxxIxS, + TnyySy + T221282 + Txy(IxSy + Iny) + Tyz(IySz + IzSy) +

sz(lex + Isz) [2-13]

where the coefficients are defined as,

40

 

 

Tit :

u r2 - 3i2 u 3ij
0 .. : -—-——o — 2019

The angular brackets denote average taken over the electronic wave function, since the
electron is not fixed in space. Equation 2.18 can be written in the matrix form as: stamp,
= I - T - S, the anisotropic hyperfine coupling tensor T is traceless and symmetric with
elements Tij- The hyperfine interaction matrix A can be defined as a combination of

isotropic and anisotropic hyperfine couplings,
A = Ago 1 + T [2.20]

where trace of A , Tr{A} = 3 Am.

Information about the geometry of the paramagnetic system, the unpaired electron
spin density and molecular wavefunctions can be obtained from the hfi tensor. When the
molecular motion is fast, as in the case of molecules with low viscosity or at high sample

temperature, the anisotropic hyperfine interaction averages to zero.

11.1.4 Nuclear quadrupole interaction

The nuclear quadrupole interaction is an electrostatic interaction between the
quadrupole moment of the nucleus and the electric field gradient at the nucleus due to the
surrounding electronic charges in an atom or molecule. A nucleus with a nuclear quantum
number I = 1/2 has one net unpaired spin, which imparts a nuclear magnetic moment to
the nucleus. The distribution of charge in such a nucleus is spherical. When 1 _>_ 1, a
nucleus has an electric moment along with the magnetic moment and the distribution of

nuclear charge is non-spherical. Such a nucleus has a quadrupole moment eQ, where e is

41

the charge on an electron and Q is a measure of deviation of nuclear charge distribution
from spherical symmetry. In systems that are studied using EPR spectroscopy, a charge
cloud of electron surrounds the nucleus and an inhomogeneous electric field can result
from asymmetry in the electron distribution and electronic structure of the paramagnetic
species. In such a case, the quadrupole moment will interact with the resulting electric

field gradient and affect the electron spin energy states via electron-nuclear magnetic

coupling.
' "Cl
' " q
+ q + q +q 4' q
' ' q
9 -q
(a) (b)

Figure II-5: Interactions of a quadrupolar nucleus with four point charges, the orientation
in (b) is of lower energy and is preferred over (a).

The quadrupolar moment acts to orient the nucleus in the direction of the largest
electric field gradient (figure 11-5). When the magnetic field axis and the crystal axis are

parallel, the effect of quadrupolar interaction is to displace all the energy levels by a

42

constant amount, which produces no change in the observed transitions. However, when
the two axes are not parallel, there is a competition between the electric field and the
magnetic field and it manifests itself in the EPR spectrum. Since, the axis of quadrupole
moment and the nuclear spin angular momentum are collinear, the Hamiltonian for the

quadrupole interaction can be described in terms of the nuclear spins as,
crave]: x [3 I,2 - I2 + m1,2 - 1,2)1, [2.21]

where K = equ/4, is the quadrupole coupling constant, 11 is the asymmetry parameter
defined in terms of the principal values of the diagonal and traceless electric field gradient

tensor as,
11 = IVxx ‘ Vyyl/sz

The nuclear quadrupole interaction is discussed further in section IV-4.1.

11.1.5 Zero field interaction

The magnetic dipole-dipole interaction between two unpaired electrons can
remove the degeneracy of the electron spin energy levels in the absence of external
magnetic field and is called zero field interaction. In transition metal ion systems (with S
2 1), this term is used to explain any effect that removes the spin degeneracy, including
dipolar spin-spin interactions and spin-orbit coupling. Experimentally it is not possible
to separate the contributions of these two effects to the zero field splitting tensor D. The

zero field Hamiltonian has the form,

076$: é-‘Dé

43

A A

= D SZ+D gaunt:

xx x yy

= ms: - §S(s + 1)] + as,2 - sf) + an“ + D” + Du )S(S +1) [2.22]

where D = Du - (Dxx + Dyy)/2 and E = (Dxx - Dyy)/2. The last term in [2.22] is a constant,
which shifts all the components of the ground state equally. The coefficient of this term
is proportional to the trace of D. This term is usually not included in the spin
Hamiltonian as it is zero for pure spin-spin coupling. This is discussed further in section

V-5.

11.1.6 Relmtion effects
At thermal equilibrium, the relative populations in the electronic energy levels are

given by the Boltzmann distribution,
na/nB = e'AE/RT, [2.23]

where nm and n1} are the populations of the upper (0t) and lower (B) energy levels involved
in an EPR transition, k is the Boltzmann constant and T is the temperature of the sample.
Electrons being Fermions obey F ermi-Dirac statistics, but when interactions between
electron spins are weak, then Boltzmann statistics can be applied to them. This ratio
decreases and the difference in population of two levels increases with decreasing T.
Hence, resolution is expected to improve upon lowering the temperature of the sample.
A population difference between the energy levels allows the absorption of microwave
radiation. In the absence of any relaxation mechanism by which the electron in the higher

energy spin states can decay back to the lower state, absorption of microwave radiation

44

followed by EPR transitions would lead to saturation. But, there are many possible

paths that can lead to spin relaxation. Spins interact with the applied field and with each
other. Thermal motion causes these interactions to fluctuate and a spin experiences a
randomly fluctuating magnetic field. Thus, there is always an oscillating field present
with the right frequency and orientation to induce a transition leading to relaxation. The
major contribution to the energy of the spin system comes from the Zeeman energy
defined by equation 2.4. Major changes in the energy of the spin system due to thermal
fluctuations may therefore be brought by changes in the z-component of magnetization
M2. The relaxation of M2, the component parallel to the external field, is called
longitudinal or spin lattice relaxation and its characteristic time constant (T1) is called spin
lattice relaxation time constant, as it is connected with the exchange of energy between the
spin system and the lattice in which the dipoles are embedded. Quantitatively, T1 is the
time required for the longitudinal magnetization (M2) to decay to l/e of its equilibrium
value M0 after the mw pulse is turned off. The mutual interactions between the dipoles
do not affect the total energy of the system, so in presence of external magnetic field,
which is much larger than the local fields, they do not affect the longitudinal component
but only affect the transverse components Mx and My. This relaxation effect is called
spin-spin relaxation or transverse relaxation. The transverse relaxation time T2 is the
decay time of the transverse components of magnetization. Both spin-lattice and spin-
spin interactions lead to broadening of the EPR signal. Spin-spin interaction varies as a
function of (1 - cos3B)/13, where r is the distance between spins and 0 is the angle between
the field and the symmetry axis of the system. Hence, this broadening effect can be

reduced by diluting the sample. The effect of spin-lattice relaxation can be minimized by

45

decreasing the sample temperature. Various other effects such as electron spin exchange,
rapid chemical processes, electron transfer between a radical and a diamagnetic species can
also cause line broadening. The EPR lineshape can be Lorentzian or Gaussian depending
on the lifetimes of the spin states, for shorter lifetimes it is Lorentzian and for lifetimes
long enough for the spins to experience different local field for different spins, the shape

is Gaussian.

11.1.7 Information available from cw-EPR

EPR is a highly sensitive technique for detecting paramagnetic species and can
provide information about electronic spin state, electronic and spatial structure and
internal dynamics of the system under study. One major advantage of EPR is its extreme
sensitivity to very small amounts of paramagnetic materials. EPR experiments can be
performed on a variety of systems including gases, solutions, powders, frozen solutions,
single crystals, complex biological systems and polymeric matrices having unpaired
electrons. The main information gained from an EPR spectrum is an estimate of the
various terms in the spin Hamiltonian as discussed above. If the spectral features are well
resolved, then the Zeeman and hfi parameters can be found directly from EPR spectrum.
In order to extract maximum information from EPR spectra of a particular system, spectra
are collected at several frequencies and temperature. It is possible to identify an unknown
ion or lattice defect and to distinguish between several states of the same ion.
Considerable information can be obtained about the nuclei in the immediate neighborhood
of the absorbing spin. Sometimes data on relaxation times can provide information about

long-range effects. Diffusion constants, correlation times and the type of hydration can

46

be in principle determined from the EPR spectra of solutions. Chemical bonds in
molecules and crystal may be sometimes characterized by EPR studies. The
concentration of paramagnetic species may be determined. Detailed information about
ferromagnetic, antirnagnetic and ferrimagnetic materials may be obtained.

Conventional cw-EPR experiment is performed at frxed microwave frequency and
the magnetic field is varied to adjust the energy level splitting to the energy of the
microwave quantum. Two commonly used frequencies fall in the range of X-band (8.2-
12.4 GHz) and Q-band (33-50 GHz). The sensitivity of the instrument is proportional
to the square of the microwave frequency and resolution is expected to improve with an
increase in frequency, but there are several limitations on the use of higher frequencies.
With the increase in frequency the sample size decreases, so sensitivity is not as much
greater as predicted from v2. It is more difficult to attain higher field homogeneity
(SBo/Bo) that is required at higher frequencies. For aqueous samples, dielectric absorption
by solvents becomes more serious as the frequency increases and results in decreased
sensitivity. Water, alcohols and other high dielectric constant solvents strongly absorb
microwave power. They can be used when the samples have strong resonance and are
contained in very narrow sample tubes. The best frozen solution results are obtained
when the solvent freezes to form a glass; otherwise, paramagnetic aggregates form, which
lead to a spectrum with broadened lines. Ethylene glycol is often used for this purpose.
The sample tube employed is also important. A quartz tube is preferred, because pyrex
absorbs more microwave power and also exhibits an EPR signal.

The main limitation of EPR is the linewidth changes, which may occur over a

range of approximately 1 us to 1 us. In case of randomly ordered samples EPR spectra

47

are often inhomogeneously broadened and it becomes normally impossible to measure the
hyperfine splittings. The inhomogeneity in the cw-EPR line is caused when the unpaired
electrons in different free radicals in the sample experience slightly different effective
magnetic fields at a given time and hence only a fraction of the spins meets the resonance
condition as the magnetic field is swept. The experimentally observed EPR line is then a
superposition of contributions from a large number of individual components called spin
packets, each slightly shifted from the others. A spin packet is a collection of spins with
individual spin magnetic moments having same Larmor frequencies but different phases.
The inhomogeneity in the effective field at the electron can arise from inhomogeneous
magnetic field or from anisotropic g and hfi interactions that give rise to a distribution of
local magnetic fields. Hence, a need arises to find a method with the ability to resolve

these unresolved hf splittings.

11.2 Pulsed EPR

Electron Spin Echo Envelope Modulation (ESEEM) Spectroscopy is one of the
most useful methods of pulsed EPR. It is a low-temperature, time-domain EPR technique
used for structural determination in paramagnetic systems and is based on the electron
spin echo (ESE) phenomenon. It enables the measurement of weak electron-nuclear
hyperfine couplings that remain unresolved in cw-EPR due to inhomogeneous broadening
of the EPR lineshape. In this method, the time evolution of the spin system is recorded
directly, which allows observation of individual spin packets and hence overcomes the
problem of inhomogeneous broadening. The paramagnetic sample is studied in a constant

magnetic field and is subjected to a series of resonant microwave pulses of varying

48

duration (in nsec range) and power, separated by evolution times, periods when the
microwave pulse is off. Application of a refocusing pulse reverses dephasing due to
inhomogeneity and the spin packets refocus resulting in the formation of a spin echo.
The integrated intensities of electron spin echoes are then measured as a function of time
intervals between pulses. The variation in amplitude or modulation of the spin echo
amplitude is used to measure splittings that are too small to be resolved in the
inhomogeneously broadened EPR line. This method makes it possible to separate the
dephasing arising due to inhomogeneity from the decay of magnetization due to spin-spin
relaxation. The two most common methods of ESEEM experiment are commonly

referred to as two-pulse and three-pulse techniques.

11.2.1 Two-pulse ESEEM spectroscopy

 

 

 

 

 

 

(filx' (70g
Irel<—>l+>l<———>lel
tP “/2 T tPa IT I techo

Figure 11-6: The pulse scheme used in a two pulse ESEEM experiment.

Figure 11-6 shows the pulse scheme used in a two-pulse ESEEM experiment to

generate the spin echo. The pulsed method of EPR has been discussed in various articles

49

in great detail.2’4a5 ,9,10 The formation of the spin echo and the basis of two-pulse and

three-pulse ESEEM methods are briefly discussed below.

11.2.1.1 Formation of Ekctron Spin Echo (ESE)

In the presence of a static external field (Bo) spins precess around the axis of B0
with angular frequency (1)() (section 11.1.1) and a net or equilibrium magnetization M
associated with electron spins is along Bo. Because an EPR transition involves a change in
the orientation of the electron spin, the optimal direction for the magnetic field (B1) of the
mw pulse is most often perpendicular to B0. This time-dependent field can be

represented as,
B,(t) = 2 B1 cos(t1)t) [2.24]

where (1) is the carrier frequency of the mw pulse. The effect of this field can be analyzed
by decomposing it into two rotating components, B, and B1, each of magnitude B], one

rotating clockwise and the other counterclockwise in the plane of B1.

B,(t) = B1 [1 cos((1)t) + j sin(00t)], V [2.25]

B,(t) = B1 [i cos(0)t) - j sin(t1)t)] [2.26]

One of these components will rotate with frequency (0 in the same direction as the
electron spins (section 11.1.1). The other component will have a frequency -2(.0 and at
resonance when to = (00, this component will have little effect, so it may be neglected.

The time dependence of B1 can be removed by introducing a coordinate system x'y'z,

rotating about the axis of B0 at an angular frequency (0.1 1 Hence, in the rotating frame, B,

50

is a stationary field and is defined along x'-axis, B1= Bli. The magnetization M appears
to precess about B0 at a frequency ((1)0 - 0)) which is the Larmor frequency corresponding
to a field of magnitude (toO - 0))/y = ((1)0 - (1))/(tuo/Bo). Thus B is given by,
B = 3.1 + (000 - (1)) Bo/(nok
[2.27]

The magnitude of this field is,
B = [Br2 + (((00 - (”We/(1)0121” [2-28]

and its direction is at an angle 0 from z-axis defined by tan0 = Bltuo/ ((1)o - to) B0. In the
presence of mw pulse, a torque acts on the magnetization according to the Bloch's

equation,
dMi/dt = YBi X Mi, [2.29]

where Bi is the effective static magnetic field experienced by spin packet ‘i’ and is the
resultant of B1 and Bo. Blow is negligible as compared to B. and B0. The magnetization
before and after the application of mw field is obtained from the solutions of the Bloch

equations,

du
— = Bov-vm-u/T
dt “W 2

% = -y Bou + utr) + 7 M28] - vfl"; [2.30]

‘1sz = -v Btv + (M. - MM.

 

51

where u and v are the x'- and y'- components of the magnetization in the rotating frame
and M0 is the magnetization at thermal equilibrium before the application of B1. If it is
assumed that the mw pulse width (tp) is sufficiently small, such that no appreciable decay
of magnetization takes place during tp and the field B] >> B0 (in rotating frame), such that
the effect of B0 can be neglected during the application of mw field, then the effective field
experienced by the spin is just B1 and the magnetization precesses about it at frequency

to] = y B]. Under these conditions the above equations (2.26) reduce to,

flzo
dt

1;; =y MZB, = (0le [2.31]
dMZ
dt

 

= -Y BIV = '(l)1V

At the time t = 0, when the mw pulse is turned on, equilibrium magnetization M0 is along

z-axis and the solutions of [2.31] are,

M2 = Mocos((n1t)

v = My. = Mosin(wlt) [2.32]

These equations [2.32] describe the rotation of magnetization in y'z—plane. When the

pulse is turned off at t = tp, the components of magnetization are,

Mz(tp) = Mocos((r)1tp)

v(tp) = M,.(t,,) = Mosin(tu[tp) [2.33]

52

Thus the effect of the mw pulse is to turn the magnetization vector that was aligned along
B0 by an angle 9111p, given by 0“,], = (ultp. In ESE experiments, pulse widths are chosen to
rotate the magnetization by some predetermined amount, for example, if width and

magnitude of mw pulse are selected such that (ultp = 1t/2, then the magnetization is

flipped along y'-axis,

M Z(tp) = 0

v(tp) = Myt(tp) = M0 [2.34]

Due to inhomogeneity in the external and internal magnetic fields, there will be a
distribution of precessional frequencies of spin packets around the central frequency (00.
Since these spin packets have different offset frequencies they start dephasing
immediately after the pulse. This is called Free Induction Decay (F 1D). The dead time of
the instrument is usually too large for the F [D signal to be detectable. From this FID,
however, an echo called primary echo, can be generated by means of a second mw rt-pulse
(wtp = it) along x'-axis, applied at time t after the first pulse. With the same mw field its
pulse width would be twice that of the 1t/2-pulse. The n-pulse torques the magnetization
vector by 180° about x'-axis. The original precession frequencies and direction of
precession remain unchanged and all the magnetization vectors align along the -y' direction
at time t after the second pulse to build up a ‘spin echo’ which is detected. For any
length of pulses an echo is generated, but maximum echo intensity is obtained with a n/2-

tt sequence of pulses.

11.2.1.2 The nuclear modulation effect

53

The unpaired electron can have anisotropic hyperfine interactions with the
surrounding magnetic nuclei and the NMR transition frequencies can be observed as
modulations on the ESE decay curve (figure 11-7). The dipolar field from unpaired
electron spin (~ [re/13) can be a significant part of B0, for example, it is about 350 G for r =
3 A (figure 11-4). The dipolar field from a nucleus is much smaller, less than 1 G for r >
2.5 A. However, the dipolar field from a nucleus experienced by the electron varies with
time due to the precession of nuclear spin. This leads to modulation of the electron
Larmor frequency at the nuclear Larmor frequency. For the simplest case of a spin
system consisting of an unpaired electron coupled to a single nucleus with 1 = 1/2, the
energy level diagram consists of four levels as shown in figure 11-8. For a totally isotropic
case only the allowed transitions labeled 1<-—>3 and 2694 would occur, but the presence of
an anisotropic hfi of the same magnitude as the nuclear Zeeman coupling may lead to
mixing of the nuclear states. As a result, the "forbidden" transitions, 1(—>4 and 2(—)3 may
also occur to some extent. If the short mw pulse covers the frequency range
corresponding to these transitions then all four transitions will occur simultaneously.

The formation of ESE in a two-pulse experiment can be seen pictorially in figure

H H

11-9. After the first n/2-pulse the electron spins "313 , 524" from allowed transitions and
"514", "523" corresponding to forbidden transitions will be in the x'y'-plane, here "stj"
represents spins undergoing transition 16)]. As 0013, (023 < (r), so 513 and $23 will make
negative angles with respect to y-axis and (1)24, (1)14 > (1), so $24 and 314 will make a positive

angle with respect to y-axis, wij represents the angular frequency of the spin Si] and to is

the angular frequency of rotating frame. Since (1)14 > (024, so 314 will be ahead of $24 and

54

x103
2.5

 

2.0 _

 
    

“—
--- -3

in
1

 

r—I
O
1

Echo amplitude

1

0.5

CD
1

 

 

 

pg -----------------------------

CD

20 2000 4000 6000

Figure II-7: Two-pulse time domain ESEEM data, dead time = 220nsec.

55

 

 

 

 

 

 

 

 

 

 

 

 

 

1h 11 4x 1
1- ‘”a a .

A ll 2 7
‘1’ Ital lVl IVI IUI _mfimall‘ lama)”

é : (1)1300 (024

\‘1 r 5 3

(b)

31 :

th :

(a)

Figure 11-8: Energy-level diagram for the interaction of an unpaired electron (S = 1/2) and
a nucleus with 1 = 1/2.

56

Figure 9: Vector diagram showing the formation of electron spin echo as a result of a
90°-r-180° mw pulse sequence in a two pulse ESEEM experiment. (a) at thermal
equilibrium the net magnetization M is aligned with the external magnetic field Bo; (b)
as a result of first 90° mw pulse, M is torqued onto the x'y'-plane of the rotating
frame; (c) the spin packets that constitute M start to dephase due to difference in their
precession frequencies; (d) the second 180° pulse applied along x', flips the spin
packets about 11', while the direction of precession of individual spin packets remains
unchanged. This pulse also introduces branching of transitions; (e) the individual spin
packets refocus to form an spin echo along -y'.

57

zllBo

 

> y --->y'

x' (3)1: 0 ’ (b) t = tp, m2

We.» N
8
e» ~
8

5’

   

(d)t=tp,m+r+tp,n

"Hug-p N

--..->y'

 

I
x' (e) At time At < t after ((1) x' (D t = 215 + tp,1t/2 '1' tp,1t
+ techo

58

will make greater angle with y'-axis and appear to move clockwise in the rotating frame as
shown in figure II-9(c). Similarly, (023 < (1013, so 323 will lag behind 513. In order to
understand the effect of pulses on the spin packets, we focus on one of the spin packet
513 (solid arrow). After the first free precession period 1:, this packet would make a
negative angle of ((1)13 - (u) with respect to the x'-axis. The effect of the 180° pulse is to
flip the packet 313 by 180° about x'-axis and to introduce two possible transitions, level 1
to 3 and another from 1 to 4. Thus, the spin packet $13 is split into a packet $14 that will
now precess at 00.4, (figure II-9d). This is termed as 'branching of transitions'. As 0014 >
(r) > (1)13, this new packet 5,4 will precess in the direction opposite to 513 from which it is
split, (figure II-9e). After the second free precession period t after the n-pulse, 513 will
refocus to form an echo along -y'-axis, however, at this time 314 will make certain angle
with the -y' axis and thus, will not contribute fiJlly to the echo, (figure II-9f). The
contribution of the component 514 to the echo will be equal to its projection on the -y'
axis, cos{((1)13 - (01.01}, weighted by the probability of branching. Thus, the interference
is modulated with a frequency of (0313 - (1)14) = (03, as t is varied. Its contributions will be
maximum if, ((1)13 - 0014): = n 2n, i.e. when it precesses an integral number of times in time
t, in the rotating frame. In the above treatment it is assumed that the spin-lattice
relaxation time T1 >> I, such that there is no build up of the longitudinal component of
the magnetization (M2) during the time t, i.e. M2 = O. The spin-spin relaxation time (T2)
is of the order of few usec, whereas I is typically few hundred nsec or few tenth of a usec
so, the decay of transverse magnetization is not complete when the second n-pulse is
applied. The requirement on mw pulse length is that it should be shorter than the

shortest modulation period to be observed and the mw field B. must be large enough such

59

that yBl is larger than the nuclear frequencies ma and (1)3. 1f B1 is too small the branching
of transitions will not occur. This can sometimes be useful when different kind of nuclei
with quite different nuclear frequencies are coupled to the electrons. It may be possible
to vary B, such that branched transition for only one nucleus is excited this would
suppress the modulations from the other nuclei.

The theory behind pulsed methods can be described by using density matrix

formalism.

11.2.1.3 Densig matrix formalism

In quantum mechanics the expectation value of an observable depends on the

state, ‘1’ of the system. Various members of an ensemble can be in different states, and

any bulk property (0) measured is an average of various expectation values i.e.,

 

(6) = <‘I’(t)'Ol‘1’(t)> [2.35]

Thus, knowledge about the exact state of the system is a prerequisite for measuring a bulk
property. The density matrix formalism provides a means for calculating properties of
ensembles within the framework of the wavefunctions describing the state of the
microscopic spin system that comprises the ensemble's fundamental element. The
calculation involves solving the equation of motion of density matrix rather than solving
the Schrodinger equation. Since in ESEEM experiments one studies an ensemble of spins,

density matrix formalism can be used to describe these experiments. There are several

articles that describe density matrix1 ’12'1 5 and its use in magnetic resonanceza16 A brief

60

description of the use of this method for deriving the modulation formula for two-pulse

ESEEM experiment will be presented in this section.5

The density operator p describes the projection of one basis set onto another. It

has the form,
60) = I‘i'(t)><‘P(t)l, [2.36]

where ‘1’(t) is the wavefunction describing the state of the system. As any state of a
system can be expanded in terms of a complete set of orthonormal functions 101) with

coefficients ck(t), so ‘P(t) can be expressed as,

‘1’“) = 20km 1%) [237]
k

and the density matrix has the elements

pij = ($1 WOW] > 1238]

So, the density matrix has different forms in different representations but the density
operator p(t) is independent of the choice of basis states. Using equations 2.36-2.38 to

simplify 2.35, it can be shown that,

(o) = Trace{p(t)O} [2.39]

p(t) can be found by solving the equation of motion of the density matrix,

d i
and) - #0010»? (t)] [2.40]

61

where 0’60) is the time-dependent Hamiltonian and the square brackets represent the
commutator of p and (3’6. On integration [2.40] gives the solution for density matrix p(t)

at time t, in terms of its initial value p(0) at time zero,

p(t) = e"°““" pm) 6...... [2.41]

In the case of spin echo experiment the evolution of the spin states during the
periods of nutation, when the mw pulse is on, and free precession, when the mw pulse is
off, can be followed by density matrix formalism. Figure II-10 shows a set of energy
levels and transitions involved in the formation of spin echo. la) and IB) are two electron
spin states and the spread in a and B levels is due to hyperfine structure and

inhomogeneity.

 

log) states ‘ A A

"'1‘ 1 *1

 

 

 

 

 

 

 

 

 

“_f ‘ T7 j
1 B1) states

 

 

V V

Figure 11-10: Energy-levels of a quantized system and transitions giving rise to spin echo.
Ia) and | B) are two electron spin states. Only those levels of the over-all quantized
system for which the transition la,)<—-)|Bj) has a significant amplitudes are shown.

62

The evolution of a and B states during period of nutation and the free precession period
can be followed by using eq. 2.36. The Hamiltonian Oflt) consists of a time-independent
part 0569., describing electronic Zeeman, nuclear Zeeman, electron-nuclear hyperfine,
nuclear quadrupole, zero field interactions etc. during the free precession period and a
time-dependent part @‘61 that describes the interaction of the spins with the resonant mw

field during the period of nutation.

CV60) = 07‘6”, + Wit), [2.42]
06, = @‘C’agmm + GEM-gum + G’Crfi+ 096.4,. + GE; . + [2.43]
diam) = thchoswlt [2-44]

= hmlsx (in rotating frame)

B0 can be neglected during the mw pulses since the pulse energy gBBl is much
larger than the nuclear Zeeman or hfi energies. It might not be a reasonable assumption
for longer pulses (smaller B1) and for nuclei closer than 2 3A (stronger hfi). This
assumption is useful in obtaining simple equations for the echo modulation and some
discrepancies between simulations and experiment may be associated with the partial
failure of this assumption. The mw pulse width is assumed sufficiently small such that it
has sufficient energy to excite all the transitions simultaneously and the effect of
relaxation during the pulses is negligible. The decay of echo due to spin relaxation can
also be neglected during the periods of evolution, because the total echo as a function of

time can be written as a product of a decay function and the echo modulation function,

fecho = fdecay fmod [2'45]

63

The decay function is of the form exp-(t/to)", where to 2 T2 is of the order of usec and
leads to a rapid background decay and is responsible for the reduction of frequency

resolution in two-pulse experiment. The modulation function fmod can be obtained as

described by W. B. Mims5 using the density matrix formalism. The density matrix p(ZT)
at the time of appearance of spin echo is time-dependent so, Tr{p®‘€} will depend only
on the time-dependent part of of given by [2.44] and can be found using equation 2.41.
The echo signal can then be determined as the expectation value of the My component of

the magnetization,

E = n Trtpcehovfl} [246]
where n is a constant of proportionality that depends on experimental parameters, pccho
is the density matrix of the spin system at the time when the maximum echo appears and
of, is the time-dependent Hamiltonian describing the interaction of the spin system with
the microwave pulse. The evaluation of equation [2.46] requires two summations to be
performed: a sum over all [01,) and lBj) states involved in the branching of transition for
each quantized system and a sum over all the individual systems that make up the
inhomogeneous line. (5’63, will have a spread A096, about a mean value 0’60, a, due to

inhomogeneity,

(3’60, g: 01’6“,» A0780, r [2.47]
Provided Add, ris large compared to the spread within or and B manifolds, the two
summations can be factorized. The first sum would give the envelope-modulation
function and the second sum, the form of the spin-echo signal. It is assumed that A6760, t

only affects the spacing between on and B levels but does not contribute towards

64

 

 

inhomogeneous spread within each manifold. During the free precesion periods (3’6 is
time-independent and p(tf) at the end of the pulse can be found in terms of the density
matrix, [30,) at the start of the free precession period using equation 2.37. In the
eigenvalue basis, p(ti) and (>450 are diagonal and at the end of the free precession period
the off-diagonal elements are multiplied by phase factors exp-K000, - (om-)(tf- t,), where to“,
and tom are the nuclear frequencies in or and B electron spin manifolds respectively.
During the periods of nutation, the spin system interacts with the time-dependent

oscillating mw field and the Hamiltonian can be written as,

d€1(t)= afar/(em + em") [2.48]
where u) is the frequency of the mw field and @‘C’yvis a time-independent Hermitian
operator that connects or states with B states, but does not connect states within 01 or B
manifolds. In order to describe evolution of the spin system under the influence of mw
pulses, exp(i:i(u[Sxt) should be expressed in the eigenvalue basis. This matrix consisting
of states loci) and 113]) can be divided into four sub-matrices with 01,01], (1in Bjori and 13ij as

their matrix elements,

 

1011) ”30
(3’6, = [04) 0 create“ [2.49]
IBJ') GEE/1e“ 0

The time dependency of (3’61 can be removed by a similarity transform that is
equivalent to introducing a new frame of reference rotating about z-axis at angular

frequency (1),

65

0’6}: exp(ioz(ut/2) 0’51 exp(-ioztr)t/2) [2.50]

where 62 is a Pauli matrix of dimension (21+l)x(21+1) for rotation about z-axis. Then,

 

1%) 1le
076’, = let.) I 0 076W [2.51]
113]) @591; 0

In this rotating frame, (3‘60 and p can be found by similar transformations,

d6}: exp(ioz(r)t)/2 056., exp(-ioz(ut)/2 [2.52]
p = exp(iozmt)/2 p exp(-iozwt)/2 [2.53]

The final density matrix can be obtained by similarity transformations as,

pechoa) : R-l pecho(0) R [254]
where R = Rm] Rr Rm“ Rv Rtpl and Rm“ are the exponential operators for the first and
second period of nutation and R[ is the exponential operator during the free precession

period. The time 't' in figure 11-6 is approximately tp/2 and is neglected.

Rtpl = exp[i(@‘€o + dfiflpl/h]
Rtpll = €Xp[1(0f€o + Cygjfipu/h] [2.55]
R, = exp[i (3’50 ‘c/h]

Since 6’60 and (3’51 are Hermitian and Rm], Rm", and RT are unitary operators, so R'1 can
be replaced by R1.

The modulation function for the two-pulse ESE can be obtained to be,

66

fm,d(r) = |u|4 + M“ + |u|2|v|2 {2cos((1)a't) + 2eos(tu,,r) ~ cos[(0)a - rope] -

cos[((r)a + wB)t]} [2.56]

where |u| and IV] are the normalized probability amplitudes for the forbidden and allowed

EPR transitions as marked in figure 11-8 and are defined as,

|u| = (2| ASKED/0.50)] = sin[(tp, — (pp/21am M = (1| s,[3)/o.5tul = cos[((pa — tp,)/2]

[2.57]

The angles, (p0L = sin'l[B/(2(1)a)] and (pa = sin'l[B/(2(1)B)] define the axes of quantization for
the or and B spin manifolds respectively. The nuclear frequencies ton] and (”at are defined

as,

0),, = [(0)l — A/2)2 + BZ/4l“2

(DB = [(001 + N2)2 + B2/4]“2 [2.58]
where A and B define the hyperfine coupling terms as,

+ ngN BN(300520 - l) [2.59]

A = Aiso 111‘3

 

g B gN B N(3cosO sin0)
B = h 3
r

 

[2.60]

The angle 0 describes the orientation of the principal axis of the hyperfine tensor with
respect to the external field Bo. Thus, a two-pulse echo amplitude occurs at the
fundamental nuclear frequencies, 0),, and (1)3, and also at their sum and differences, (wa+
(03) and (00a - (1)5). The amplitude of modulation is proportional to square of the product

of the transition probabilities, lulzlvl2 of the classically allowed and forbidden transitions

67

associated with branching. The non-modulated portion of the echo depend on the
individual transition probabilities, In]4 or M4 for the two transitions. Substitution of
equations [2.57] in [2.56] gives another form of echo modulation for the two-pulse

experiment,

Emod(1:) = l - 0.5 k {1 - cos((r)a1:) - cos(o)5‘t) + 0.5cos[((1)a - (1)6)1] +
0.5cos[(ot)0L + wfi)r]} [2.61]
where k = sin2((pa — (pa) = [(1),B/((1)m(1)fi)]2 is called the modulation depth parameter,

(1)l = gNBN BO/h. For large depth parameters, the intensities of the combination lines can

become substantial and may considerably complicate the ESEEM spectrum, especially at
low mw frequencies. The modulation frequencies and depths reflect the hyperfine
couplings between electrons and nuclear spins. 1n the presence of more than one nucleus
contributing to the hfi, the ESEEM function is obtained by taking the product of the

individual modulation functions as,

Em = fa... 1'1 Egan) [2.621

where N is the number of coupled nuclei. Hence, in presence of many nuclei there will
not only be fundamental and combination frequencies associated with each nucleus (Am, =

3:1) but also combinations of frequencies arising from different nuclei. During time I, the

magnetization is in the xy-plane so, its relaxation is governed by spin-spin relaxation (of

the order of usec). As a result, the linewidth of two-pulse ESEEM peaks are severely

broadened.

68

Figure II-ll: Vector model showing the formation of electron spin echo in a three
pulse ESEEM experiment. (a) pulse scheme; (b), (0) same events as in figure II-9; (d)
the second 90° mw pulse along x' converts the transverse magnetization into
longitudinal magnetization and the components start to dephase after the pulse is
removed; (e) the third 90° mw pulse applied along x' rotates the magnetization
components onto x'y' plane;(f) the spin packets refocus along -y' to give rise to a
stimulated spin echo.

69

 

(TE/2)xt (1t/2)x-

[1

(TE/2)x'

 

 

[L

 

 

 

I (—-)<——>
tp T p T tp T + techo
2. 11 B0 2'
l
(1) Q 0)
(tr/2),. E
__> :
M Bl '
—> y. ---.>y'
(a)t=0 I, (b)t=t
X. X. P
2' Z.
l 1
W can»
i (m2)x'
= T?
1
I ‘\
(Q "*1" Tu" . ----->x'
I (c)t=tp+r . (d)t=t+2tp
x' v -z
z'
1 2'
(wt/2),. (J m
—> |
131 5
lam-» ">3" 2""
f” (e)t=T+21p+T f” (Dt=21+2tp+T
x. X' ‘1' techo

70

11.2.2 Three-Pulse ESEEM spectroscopy

Some of the shortcomings of a two-pulse ESEEM experiment can be overcome by
using a three-pulse experiment. The three-pulse experiment employs a tt/2 - 1: - rr/2 - T -
1t/2 pulse sequence (figure 11-1 1). This sequence produces several echoes because it
comprises three pairs of pulses that can produce two-pulse primary echoes in addition to
the expected three-pulse echo, also called stimulated echo, which occurs at a time I after
the third pulse. The stimulated echo signal is recorded as a function of T, the time
interval between second and third pulse, keeping 1: fixed for a given measurement. Two-
pulse echoes are eliminated by means of phase cycling of the mw pulses.

The effect of first mw pulse is same as in the two-pulse experiment discussed
earlier. After the first tt/2-pulse, the components that dephased in x'y' plane during first
free precession period 1: from the magnetization along +y' are rotated and stored along -z
by another 7t/2-pulse, (figure 11-11). If the waiting time T is such that, T2 < T < T1, then
all the transverse magnetization would evolve as longitudinal magnetization that could be
brought into the x'y' plane by another n/2-pulse and refocused to form a stimulated echo
after time 1: along -y' axis. If the waiting time is not long enough for the complete decay of
transverse magnetization, it would lead to formation of an additional echo called Hahn
echo at time I after the second pulse. The amplitude of the stimulated echo depends of 1:
as well as T. During time 1: the magnetization dephases in the x'y'-plane, so the echo
modulation will decay via spin-spin relaxation as in two-pulse spin-echo sequence. The
second rt/2-pulse serves to store the precession frequency offsets (t1) - (1)0) of each spin
packet as a longitudinal magnetization which forms the stimulated echo. As the

magnitude of these longitudinal components are governed by a relatively slow spin-lattice

71

relaxation time, T1, the overall decay rate of the amplitude of the stimulated echo is
reduced leading to better frequency resolution as compared to the two pulse experiment.
An analytical expression for the three pulse modulation fimction for a S = 1/2, 1 =

1/2 system is,

Emmet) = [u[4 + M4 + |u|2|v|2 {eos(tu,r) + cos(tu,,r) + 2 sin2(0)at/2) cos[u),(r + T)] +
2 sin2(tu,,r/2) cos[(1)B(t + T)]} [2.63]
One important feature of the three-pulse echo is that it is modulated only at the nuclear
frequencies (1)Cl and (1),}. The first pulse generates modulation effects at the nuclear
frequencies; the second pulse creates additional modulations at the sum and difference of
the nuclear frequencies. The third pulse causes interference between these sum and
difference frequencies leaving only the nuclear frequencies in the modulation pattern. In

presence of multiple nuclei the echo modulation function [2.62] is modified as,

E(1:, T) = EZ—{fi Ej,(r,T)+ fiE;(r,T)] [2.64]

The three-pulse echo decays as a function of T] which is much longer than T; at
cryogenic temperatures, the decay time in a two-pulse experiment, hence leads to less line
broadening in the frequency domain. Since fewer frequencies are involved and the
modulation occurs over a longer time scale, so three-pulse spectrum is cleaner and better
resolved as compared to a two-pulse spectrum.

In the pulsed experiments, mw pulses of very short widths (~20 us) are required
in order to ensure reasonable spectral width and to keep the effects of electron spin

relaxation during the pulse negligible. For relaxation times in the range of usec, in order to

72

achieve mw pulse-width short enough to be able to neglect relaxation, the requirement for
microwave field is of the order of few mT (~30 G). It is technically difficult to obtain
short and strong pulses. Another major difficulty encountered in a pulsed experiment is
that of 'dead time'. It is the time needed for the high power transmitter pulses in the
resonant cavity to decay far enough such that the echo signal can be observed. As a result

the time domain data starts after the dead time (~ 0.15[.tsec) instead of starting at zero

time (figure 7). Methods have been developed to overcome this problem by shortening
the dead time17'19, improving the detection scheme20 and by developing better methods

of data-analysis.21

73

 

List of References

1)Abragam, A.; Bleaney, B. Electron Paramagnetic Resonance of Transition Ions; Oxford
University Press: Oxford, 1970.

2)Atherton, N. M. Principles of Electron Spin Resonance; Ellis Horwood Ltd.: West
Sussex, England, 1993.

3)Gordy, W. in Techniques of Chemistry, Vol. XV: Theory and Application of Electron
Spin Resonance; John Wiley & Sons: New York, 1980; Vol. XV.

4)Kevan, L.; Bowman, M. K. Modern Pulsed and Continuous- Wave Electron Spin
Resonance; Wiley-Interscience: New York, 1990.

5)Mims, W. B. Phys. Rev. B: Solid State 1972, 5, 2409-2419.

6)Poole, C. P. Electron Spin Resonance; Wiley Interscience: New York, 1967.

7)Wertz, J. E.; Bolton, J. R. Electron Spin Resonance; Hall, C. a., Ed.; McGraw-Hill: New
York, 1986.

8)Carrington, A.; McLachlan, A. D. Introduction to Magnetic Resonance; Chapman and
Hall: London, 1979.

9)Keijzers, C. P.; Reijerse, E. J .; Schmidt, J. Pulsed EPR: A new field of applications:
North Holland, Amsterdam, 1989.

10)McCracken, J. Handbook of Electron Spin Resonance; Poole, C. P. a. F., H.A., Ed.;
Springer Verlag: New York, 1999; Vol. II, 63-89.

11)Rabi, I. 1.; Ramsey, N. F .; Schwinger, J. Revs. Modern Phys. 1954, 26, 167-171.
12)Tolman, R. C. The Principles of Statistical Mechanics; Oxford University Press:, 1938.
13)Slichter, C. P. Principles of Magnetic Resonance; Springer-Verlag: Berlin, 1978.
14)Goldman, M. Quantum Description of High Resolution NMR in Liquids; Oxford
University Press:, 1988.

15)Feynman, R. P. Statistical Mechanics : A Set of Lecture Notes; Benjamin, Reading:
Masschusetts, 1972.

l6)Fano, U. Revs. Modern Phys. 1957, 29, 74-93.

17)Narayana, P. A.; Massoth, R. J .; Kevan, L. Rev. Sci. Instrum. 1983, 53, 624-628.
l8)Davis, J. L.; Mims, W. B. Rev. Sci. Instrum. 1981, 52, 131-132.

74

19)Cho, H.; Pfenninger, S.; Gemperle, C.; Schweiger, A.; Ernst, R. R. Chem. Phys. Lett.
1989, 160, 391-395.
20)Mims, W. B. J. Mag. Res. 1984, 59, 291-306.

75

Chapter III
INSTRUMENTATION

Early pulsed-EPR experiments, performed about forty years ago, showed that the

FID from electron spins was strong enough to be observable. 1 '3 The development of
commercial spectrometers took some time because the technology required for the pulsed-
EPR experiments is demanding. Due to the availability of faster digital logic components
and advances in microwave component design, higher sensitivity and faster response
times can be achieved in modern spectrometers, which nearly meet the idealized
experimental conditions often assumed in theoretical analysis. The design and
performance of the spectrometer is a strong function of the types of samples to be
investigated. The most important criteria in the biological studies are to optimize the
signal-to-noise ratio, to obtain highest sensitivity possible, and fast spectrometer
response time.

Continuous wave-EPR spectra were obtained at X-band on a Bruker ESP300E
commercial EPR spectrometer, using a TEloz EPR cavity, an Oxford ESR-900 continuous
flow cryostat and an [TC 502 temperature controller. The microwave frequency was
monitored by an EIP Microwave frequency counter Model 25B and the static magnetic
field strength was measured using a Bruker ER 035M NMR gaussmeter.

The experimental pulsed-EPR data for this study were collected on a home-built

pulsed EPR spectrometer built by Prof. J. L. McCracken at Michigan State University in

1990,4 which has been modified from its original design since then. Some of the

experiments that can be performed on this instrument include two-, three- and four-pulse

76

ESEEM, ESE-EPR (Electron Spin Echo detected EPR), ESEEM experiments with vector
field jumps, Mims pulsed-ENDOR (Electron Nuclear Double Resonance), Davies pulsed-
ENDOR, and HYSCORE (Hyperfine Sublevel Correlation Spectroscopy) in the
frequency range of 6 tol8 GHz.

The basic components of this spectrometer include a transmitter section capable
of generating short, high-power microwave pulses as well as soft, low power pulses used
to induce echo formation, a sample probe containing the sample, a receiver system for
detecting the echo, a pulse programmer unit for controlling the timing and synchronization
of the transmitter and receiver, and a magnet system. The schematic diagram is shown in
figure III-l.

A microwave synthesizer (Gigatronics model 610) with a bandwidth of 6 to 18
GHz is used to generate cw-microwaves. It has an IEEE 488 interface and is controlled
directly by the data collection program. The cw-microwaves are later converted into
short microwave pulses by PIN diode switches that are directly controlled by a pulse
logic module. The mw output of the synthesizer is about 10mW in power. It is divided
by a directional coupler (Omni Spectra, model PN2025-601810) such that 90% of the mw
power goes through a reference arm and 10% through the microwave pulse-forming arms.
The reference arm has an adjustable phase shifter (ARRA, model 9828A) and serves both
to bias the double balanced mixer (DBM) and as a phase reference for phase sensitive
detection. In the pulse-forming arm the mw radiation goes through an isolator (Innowave,
model 1119IR) to a power divider (Omni Spectra, model PN2089-6209-00), which
divides it further into two independent pulse channels. The isolator prevents the

reflected power from returning to the generator and helps in maintaining frequency

77

stability. These two mw pulse channels enable control of relative phases and the
amplitudes of the mw pulses. Each pulse channel consists of a low-power, high speed
PIN diode switch that creates pulses of desired widths (General Microwave model
FM864-BH), a high-speed 0°/l 80° wide-band phase modulator (General Microwave
model F 1938) that controls the relative phases of the pulses, and an isolator. The PIN
diode switches and phase modulators are controlled by a home-built pulse logic circuit.
One of these pulse channels has an attenuator (ARRA, model P9804-20) and the other
has an adjustable phase shifter. The attenuator is needed to control the relative amplitude
of the pulses in experiments like HYSCORE that require both soft and hard pulses. The
attenuator introduces some phase shift, which is corrected by the adjustable phase shifter
in the other arm. The pulses coming out of these channels are low power pulses of about
100 MW. They are combined by a pulse divider, passed through an isolator to eliminate
any reflected pulses and are then fed into a medium-power GaAs FET amplifier (Avantek
model SWL-89-0437). When GaAs F ET amplifiers are used, some type of receiver
protection, such as diode limiting, is usually used to prevent damaging the amplifier.
Since the pulses of low power are fed to this amplifier, the limiter is not needed. The
pulses coming out of the GaAs FET amplifier pass through an isolator to an attenuator
and are then fed to a pulsed travelling wave tube amplifier (TWTA, Applied Systems
Engineering, model 117). It is possible to achieve a gain of about 40 dBm with TWTA
whereas mw power of about 60 dBm (1 kW) is needed for ESE experiments. So, an
amplifier is used to ensure that the input power to TWTA be at least 20 dBm. To be in

optimal working conditions the TWTA should be driven to saturation. The attenuator is

78

Figure III-l: Schematic of pulsed EPR spectrometer setup Various symbols used in this
diagram represent the following components as :

Various symbols used in this diagram represent the following components as :

Directional Coupler

[fl Isolator

 

Como ‘1’ Phase modulator

W Attenuator
Adjustible phase shifter

Power divider

—>— Galium-Arsenide field effect transistor preamplifier

Rotary vane variable attenuator

 

 

 

Double balanced mixer

79

Circulator

[_
i

Detector I

 

PlN diode

To oscilloscope
for tuning

Limiter

 

 

 

©‘-I

2

Receiver Gate

 

 

 

 

 

 

I To oscilloscope
Twr‘

TWT switch

GASFET

 

 

 

 

 

 

 

 

 

 

0 Control

Pulse Logic
Module

PIN Control

 

 

 

 

[1 ti
_,
_. Bo _
L
>3 2: a)
ReferenccAnn

 

Video
Amp

ewe-[1

Oscilloscope

80

 

is

G)

mw synthesizer

 

 

 

 

 

1ft

 

 

 

 

'1‘0 A B C 1) E
Delay generator

 

 

1

 

‘ Field <—
controller

 

used to make sure that the TWTA is not over driven. The high power mw pulses from
the TWTA pass through a rotary vane attenuator (Hewlett Packard, model X3 82A),
which is used to vary the microwave power going to the sample, via an isolator, which
protects TWTA from reflected power. In case of rotary vane attenuator, the attenuation
is independent of frequency and does not introduce any phase shift in mw. The mw is

then fed to a Circulator (MACOM, model MA8K269) that controls the directional flow
of microwaves into a reflection cavity with a folded stripline half-wave resonators:6 and a

Gordon coupler.6

The microwave receiver utilizes wideband components and a balanced mixer
(RHG model DMZ-18 Ab) in a conventional homodyne arrangement. Signal from the
resonator goes to a directional coupler via a waveguide isolator. It is then fed to a low
noise GaAs amplifier (Avantek model AWT-18635). In order to protect this amplifier
from the high-power excitation pulses, a medium-power, high speed PIN diode limiter
(Innowave, model VPL-6018) and a fast PIN diode switch (General Microwave model
F9114) controlled by the pulse-logic circuit, are used. The PIN diode switch remains
open most of the time and is closed by the pulse-logic circuit only at the time of
appearance of echo. About 1% of the signal from the directional coupler is fed via a
detector to an oscilloscope (Tektronix 620 B digital oscilloscope) for spectrometer tuning.
Electronic components cannot pass microwave frequencies so, the detector is employed
to convert microwave energy to a video frequency. The signal from the amplifier goes
through a bandpass filter (K & L Microwave, model 3H10-2000/18000-0/0), which serves
to remove radio frequency noise introduced by the PIN diode switches, and an isolator

and then is fed to a double balanced mixer (DBM) (RHG, model DM2-l 8 AB) in order to

81

convert to an electrical signal. The mixing scheme employed is a conventional homodyne
arrangement. At the mixer, the signals from the reference arm and the receiver are
combined to generate signals at the sum and the difference of the two input frequencies.
The amplitude of the echo signal is proportional to the difference in amplitudes of the
two input signals and their relative phases. This scheme of mixing the echo signal with a
reference signal at same frequency is known as homodyne detection. The homodyne
detection scheme not only allows for a coherent detection and phase discrimination of the
echo but also provides a method of increasing signal-to-noise ratio. At the output of
DBM the echo signal consists of frequency components centered about 50 MHz. The
amplitude of the difference frequency is next amplified using a video amplifier (Comlinear,
model E220), sampled by feeding to a digital oscilloscope (Tektronix 620 B) for analog to
digital conversion for input to a computer.

The magnetic field is produced by an electromagnet (Walker Scientific, model HF-
12H) and controlled over a range from 50 G to 15000 G by a Hall-effect field controller
(Bruker, model B-H15), interfaced to the computer via IEEE-488 interface device.
Experiments were carried out at 4.2 K, using a home built liquid helium immersion dewar
capable of maintaining constant temperature of 4.2 K for about 15 hours. For most of the
microwave network semirigid and flexible coaxial cable was used to reduce the geometric
constraints of waveguides, but the high-power pulses and emitted echo from the probe
were directed by waveguides to avoid the higher losses inherent in coaxial cables. The
device connections were made with SMA-microwave connectors.

The spectrometer is controlled by a PowerComputing Macintosh computer via an

IEEE-488 interface (National Instruments). Digital delay and gate generators (Stanford

82

research, Model DG535) are used to generate the accurate timing signals which are
transferred to the pulse logic module and serve as a timing base for synchronization of the
echo experiment. It controls the timings required to turn on the high-power TWTA, to
control microwave pulse generation and spacing, to generate accurate pulse delays
required for the experiments, to control the overall repetition rate of the measurements, to
modulate the phase on the alternate echo generating cycles and to supply trigger pulses to
the oscilloscope responsible for measuring spin echo intensities. For best performance it
is essential that all timing pulses and intervals be synchronized to avoid jitter and
instabilities and to insure that all timing intervals be properly aligned. The pulses to the
PIN diode which form 90° and 180° pulses must be positioned such that the trailing edge
of the pulses is aligned with the trailing edge of the TWTA modulator pulse. This is
necessary since the high noise output of the TWTA would otherwise completely mask
the echo. The pulse timing is directly programmed through the computer, which then
performs the entire echo experiment using the pulse programmer and data collection
software.

The data collection program is written in LabVIEW 5.0.1 (National Instruments)
running on a Power Computing model 200 computer. The simulation programs are
written in FORTRAN language. The software for the analysis of the experimental and
simulated data is written in Matlab (Mathworks, Nantick, MA). Numerical simulations

of the experimental data are performed on Sun SparcII work station.

H1.1.l Resonator

83

A resonant cavity is the microwave analog of a series RLC tuned circuit (figure III-
2). The stored electric and magnetic energies inside the resonator determine its equivalent
inductance and capacitance. The energy dissipated by the finite conductivity of the

resonator’s walls determines its equivalent resistance. It is a metallic enclosure with

 

<

Figure III-2: A series RLC circuit fed by voltage V.

dimensions comparable to the wavelength made from high conductivity metal. At
resonance, the cavity can sustain mw oscillations, which form standing waves from
superimposed microwaves reflected from cavity walls.

A given resonator has an infinite number of resonant modes, and each mode
corresponds to a definite resonant frequency. When the frequency of an input signal is
equal to a resonant frequency, a maximum amplitude of standing waves occurs, and the
peak energies stored in the electric and the magnetic fields are equal. The mode that has
the lowest resonant frequency is called the dominant mode. At resonant frequency f0,
(1)0: 21th =l/(LC)m. When the power is turned off, the energy of the RLC circuit

dissipates exponentially with time, approximately as a function of exp(-(t)ot/Q), where Q

84

is called quality factor of the circuit, larger the Q slower the power dissipates. It is
defined as,

Q = 2n (maximum energy stored/ Energy dissipated per cycle) = (9de
In a cw-EPR spectrometer the sensitivity is optimized by choosing a microwave cavity
probe with large Q. The rectangular TE"); mode cavity that was used in cw-EPR
experiments, has Q factor of 5,000. Such large Q factors are unacceptable for pulsed
work because the excessively long ring down time would mask the echo. This situation is
usually compounded by the finite time required by the transmitter TWTA pulse output
to decay to near thermal noise levels. For samples with short phase memory times and
small sample volumes, it is desirable to observe the echo immediately following the last
transmitter pulse. Low Q factors are therefore required to limit the ring down time.
Samples with short phase memory time also require short pulses for echo formation.
Although the pulse areas should approach delta functions, an acceptable pulse width is
generally on the order of 10% of the phase memory time. Most pulsed spectrometer use
probes with Q values ranging between 50 and 500. Such low Q values can be obtained by
using the inherently high Q cavity geometries and over coupling, by loading the cavity
with a high dielectric material, or by loading with metallic blocks. Unfortunately, the
same procedures used for lowering cavity Q also reduce the sensitivity. A better method
to achieve a lower Q while maintaining the sensitivity is to use special microwave
structures which have inherently lower Q than cavities and have large filling factors, the
ratio of the sample volume to the volume in which the microwave field is contained. The
folded stripline half-wave resonator that was used in the ESEEM experiments has a Q

factor of about 800. The loaded Q of the resonator was varied by adjusting the position

85

of a teflon Gordon coupler that determined the coupling between the waveguide and the

resonator. In pulsed experiment it was overcoupled to lower the Q to ~80.

Stripline resonator

The design of folded stripline half-wave resonator is shown in figure III-3 (a). The
resonator is a rectangular strip of brass N2 in length and 5 mm in height bent in the form
of an arc and fits firmly in a hole 5.5 mm in diameter in the teflon block. When immersed
in liquid helium, teflon shrinks more than brass and the resonator is held tightly in place.
The teflon block with the resonator is mounted and is held by a nylon screw inside part
of X-band waveguide with interior dimensions of 1.02 by 2.29 cm, which is effectively a

TEIOZ cavity oriented with its long axis vertically.

Resonator

 

——>Teflon Block

 

:2

‘-—> Brass resonator

 

 

 

 

 

-I-’

 

(a) (b)

Figure III-3: Schematic of folded stripline half-wave resonator mounted in teflon block as
seen from top (a), transverse cross section of the cavity showing the mw magnetic field
lines that couple to the resonator (b).

86

J— Sample guide tube

 

A

X-band waveguide

 

 

e — Gordon coupler

 

A

Brass wedge

 

 

 

 

 

 

,,<' Iris plate
' — EPR tube
Resonator assembly

 

 

 

 

Mounting screw

 

Sliding short

 

Figure III-4: Schematic of the folded stripline half-wave resonator probe structure.

87

This cavity is then screwed to the rest of the waveguide with an iris plate in between.
The resonator couples to the standing wave magnetic field lines running vertical along the
walls of the cavity (figure III-3 b). The bottom of the cavity is made of a screw-operated
sliding short that is initially positioned to provide maximtun coupling of the resonator.

Its position is left unchanged once optimal position is found and all the coupling
adjustments are done through a coupler. An external stainless steel guide tube is mounted
along the axis of the resonator and runs parallel to the waveguide. Sample is contained in
quartz EPR tube, the upper end of which is slipped into a teflon holder. This teflon
holder is attached to a stainless steel rod and is inserted through the guide tube (figure 111-

4).

III.1.2 Gordon coupler

Mw power is coupled into the cavity via a Gordon coupler. A long teflon insert
that tapers as a wedge as shown in figure III-4. The position of this insert can be varied
with the help of a teflon rod whose one end is inserted into the teflon insert and it runs
through the waveguide and has its other end available outside the waveguide for adjusting
its position. By varying the position of this coupler the loaded Q of the resonator can be

changed.

111.1.3 Circulaflr and Isolator
Both mw circulator and mw isolator are nonreciprocal transmission devices that
use the property of Faraday rotation in the ferrite material. A nonreciprocal phase shifter

consists of a thin slab of ferrite placed in a rectangular waveguide at a point where the dc

88

magnetic field of the incident wave is circularly polarized. Ferrite is a nonlinear material
and in presence of a dc magnetic field it exhibits a phenomenon of Faraday rotation and
rotates the plane of polarization of the incident mw. By choosing a suitable length of the
ferrite slab a desired phase shift can be obtained between two directions of propagation
such that the reflected wave can not propagate backwards. At port 1 the polarization of
the mw is altered such that it is directed to the resonator element in the cavity at port 2
(figure III-5). The plane of polarization of the mw reflected from the cavity at port 2 is
such that it is can not propagate to port 1, and is directed to the detector at port 3
instead. Any reflected mw from the detector is directed to port 4 and is terminated to

ground by a 50 Q terminator on reaching port 4.

50 Q terminator

 

’ Mwl.3 ’ To detector

E Reflected power from cavity

To resonator

Figure III-5: Symbolic representation of a four port circulator.

89

An isolator works on the same principles as a circulator. It is in principle a three
port circulator that terminates any reflected power. It is a necessary component in order

to protect devices from reflective mw power.

90

List of References

1)Gordon, J. P.; Bowers, K. D. Phys. Rev. Lett. 1958, 1, 368.
2)Kaplan, D. E.; Browne, M. E.; Cowen, J. A. Rev. Sci. Instrum. 1961, 32, 1182.
3)Wanlass, L. K.; Wakabayashi Phys. Rev. Lett. 1961, 6, 271.

4)McCracken, J .; Shin, D.-H.; Dye, J. L. Appl.Magn. Reson. 1992, 3, 305-316.
5)Bowman, M. K.; Lin, C. P.; Norris, J. R. J. Magn. Reson 1985, 65, 369-374.
6)Britt, R. D.; Klein, M. P. J. Mag. Res. 1987, 74, 535-540.

91

CHAPTER IV

CHARACTERIZATION OF THE TTQ-SEMIQUINONE CATALYTIC
INTERMEDIATE OF MADH BY ESEEM

IV.l Introduction

This chapter presents results from continuous-wave electron paramagnetic
resonance (cw-EPR) and electron spin echo envelope modulation (ESEEM) spectroscopic
studies on the semiquinone forms of the cofactor of MADH. The cw-EPR spectrum of
TTQ semiquinone prepared by substrate addition (N -form) was found to differ
substantially from that observed when the semiquinone was generated by reduction of the
enzyme with dithionite (O—form). These differences prompted a detailed study of
hyperfine and nuclear quadrupole interactions of three 14N atoms of the semiquinone
species using ESEEM. Two of these heteroatoms are derived from the indole and the
indole-quinone side chains that comprise TTQ, while the third l4N originates from
substrate methylamine. Three-pulse ESEEM spectra of CH3'4NH2-reduced sample
showed three isolated features at 1.0, 1.5 and 4.3 MHz, which were absent in MADH
sample reduced with CH315NH2. Analysis of spectral data from substrate-derived 14N
revealed an isotropic hyperfine coupling of 2.4 MHz and nuclear quadrupole couplings
characterized by equ = 1.7 MHz and n = 0.5. The hyperfine and the nuclear quadrupole
couplings found for the two 14N nuclei indigenous to TTQ were: Arse, 2.8 and 1.9 MHz;
equ, 3.0 and 2.1 MHz and n, 0.3 and 0.7, respectively. Taken together, these couplings
provide definitive evidence that substrate 14N is covalently bound to TTQ when the

cofactor is in its one-electron reduced form and that it has an imine-like structure. The

92

intensities of the modulations indicate that the semiquinone generated by the method

recently reported by Zhu and Davidson1 results in a homogeneous preparation of the

radical. A comparison of 14N hyperfine and nuclear quadrupole couplings measured for

the N-form semiquinone with those measured previously for the O-form2 shows that a
significant change occurs in the highest occupied molecular orbital when substrate nitrogen
is bound, and may be related to the different redox and electron transfer pr0perties of

these two semiquinone forms.

IV.2 Materials and Methods
IV.2.1 Protein parification and preparation

The MADH under study was obtained from Paracoccus denitrificans and was
purified as described previously.3 The concentration of MADH was calculated from

known extinction coefficients in lOmM phosphate buffer, pH 7.5.3’5 The extinction

coefficients of the different redox forms of MADH vary with pH and ionic composition

of the buffer.6 MADH was fully reduced with substrate methylamine (1 MADH : 2
methylamine). Then the fully reduced MADH, which exhibits an absorption maximum at

330 nm, was photo-oxidized by exposing it to a source of long-range UV light until the

absorption at 420 nm (due to the semiquinone form) reached a maximum.1 This typically
occurred in 5-10 min. The source for long-range UV light was a Rad-Free RF UV-365
Long-Wave UV Lamp (Schleicher & Schleicher) using an 8W bulb that emitted light
spanning wavelengths from 320-380nm with a peak value at 365 nm. The UV light

source was placed approximately 3 inches away from the sample contained in a quartz

93

cuvette. The effects of light are completely reversible.1

IV.2.2 Continuous-wJave EPR spectrtmpy

Continuous wave-EPR spectra were obtained at X-band on a Bruker ESP300E
EPR spectrometer, using a TB“); EPR cavity, an Oxford ESR-900 continuous flow
cryostat and an ITC 502 temperature controller. The microwave frequency was
monitored by an EIP Microwave frequency counter Model 25B and the static magnetic

field strength was measured using a Bruker ER 035M NMR gaussmeter.

IV.2.3 ESEEM spectroscOpy

The ESEEM spectra were recorded at X-band with a home-built spectrometer7
described in chapter 111. Our spectrometer was modified from its original design in that
signal averaging and subsequent integration of the electron spin echo signals was
accomplished by a Tektronix 620 B digital oscilloscope fed directly by the double-
balanced mixer and mixer preamplifier of the microwave bridge. In the original design a
boxed-car integrator was used for this purpose. The pulse sequence for the three-pulse
ESEEM experiment, rt/2-t-1t/2-T-1t/2-t-echo, used a two-step (0, 0, 0), (7t, 1t, 0)
microwave pulse phase cycling to eliminate two-pulse echo interference. The modulation
was observed by recording integrated spin-echo intensities as a function of T. ESEEM
spectra were recorded at t-values that ranged from 150 to 600 ns. Typical conditions for
data collection were: microwave frequency, 8.815 GHz; pulse power, 43 W; pulse width,
20 ns; sequence repetition rate, 6 Hz; total number of experimental points, 512; time

increment, 20 ns; number of events averaged/point, 10; sample temperature, 4.2 K and

94

each data set represented an average of 9 scans.

ESEEM spectra were obtained by a Fourier analysis procedure that included dead

time reconstruction.8 Numerical simulations of the experimental data were performed on

Sun SparcII work station. Simulation programs were written in FORTRAN and were

based on the density matrix formalism developed by Mims.9 Software for the frequency
analysis of the experimental and simulated data was written in Matlab (Mathworks,
Nantick, MA). The data collection programs for our pulsed EPR spectrometer were

written in LabVIEW 5.0.1 (National Instruments) running on a Power Computing model

200 Power PC.
1V.3 Results
1V.3.1 CW-EPR Spectroscopy

 

Figure IV-l shows cw-EPR spectra of TTQ semiquinone prepared by reduction
of MADH with dithionite (a), CH3‘4NH2 (b), and CH3'5NH2 (c) in frozen buffer. All
three spectra consist of simple derivative line shapes that have no observable hyperfine
structure. The dithionite-reduced (O-form) and the substrate-reduced (N -form)
semiquinone spectra are centered at g = 2.0043 and g = 2.0039, respectively. The peak-
to-peak linewidths of the O-forrn, '4N- and 15N-substrate generated semiquinones were
observed to be 6.81 G, 11.09 G and 11.55 G respectively. The lineshapes for all three of

these TTQ semiquinones are more symmetrical than those found for samples poised by

comproportionation.10 Apparently, the asymmetry observed previously for the

comproportionation products arose from the mixture of 'O-' and rN-' form spectra

present.2

95

 

(a)

dx"/dH

(b)

(C)

 

 

I 1 1 1 1 1 1 L 1 i 1 l 1 1 1 1 I l I 1 1 I 1

335.0 336.0 337.0 338.0 339.0

 

Magnetic Field Strength (mT)

Figure IV-l: Continuous-wave EPR spectra of the semiquinone forms of methylamine
dehydrogenase from P. denitrificans generated by dithionite addition (O-form) (a), 14N-
methylamine addition followed by illumination (b), and lsN-methylamine addition
followed by illumination (c). Spectra were collected under the following conditions:
microwave power, 0.63mW; microwave frequency, 9.473GHz; modulation frequency,
100 kHz; modulation amplitude, 0.1 mT; sample temperature, 103 K; and number of
scans averaged, 10.

96

 

x105

dx"/dH

 

 

 

l l 1 I 1 l 1 l 1 1 1 l l l 1 I 1 l I 1 I L L l 1 1 1 l l

334 335 336 337 338 339 340
Magnetic field (mT)

 

Figure IV-2: Continuous-wave EPR spectra of the semiquinone forms of methylamine
dehydrogenase from P. denitrificans generated by dithionite addition (O-form) (dashed
line), l4N-methylamine addition followed by illumination (solid line). Conditions were
same as in figure IV-l. Presence of both the N-form and the O-form in the samples
prepared by comproportionation method lead to the asymmetric spectrum.

97

1V.3.2 ESEEM Spectroscopy

To understand the differences in electronic structure of substrate- and dithionite-
poised semiquinone, ESEEM spectroscopy was used to probe the hyperfine couplings of
'4N nuclei derived from substrate and from those native to the cofactor. The presence of
the nitrogen derived from the substrate, was revealed by comparing identical samples of
enzyme prepared by initial substrate reduction with l4N-methylamine and by 15N-
methylamine. Distinct differences are expected in the magnetic interactions between the
unpaired electron spin density and the nuclear spin 1 = l 14N nucleus relative to those
with 1 = 1/2 15N nucleus. Typically, three-pulse ESEEM data were characterized by deep
low frequency modulations that damped over a 10 nsec range of T (figure IV-3). A cosine
Fourier transformation of the time domain ESEEM data collected at t = 219 us, for N-
forrn semiquinone poised with l4N-methylamine yielded the spectrum of figure 1V-4(a).
This spectrum shows sharp features in the low frequency region, at 0.5, 1.0, 1.5, 2.0 and
2.5 MHz and a broad feature with a complex lineshape from 4.6-5.8 MHz. To
distinguish peaks that arise from substrate l4N-coupling, data was collected under
identical spectrometer conditions for 15N-methylamine poised semiquinone and is shown
in figure IV-4(b). This spectrum also features deep modulations with peaks at 0.5, 2.0
and 2.5 MHz characterized by narrow lineshapes, along with a broad feature at 4.5-5.8

MHz. The spectra in figure IV-4 are indicative of nitrogen nuclei coupled to a

paramagnetic center near the “exact cancellation condition”.1 1:12 Because 15N is a spin 1

= 1/2 nucleus, its

98

Echo Amplitude

 

x103
2.5 —

2.0 -

1.5_

1.0—

 

 

 

0.5 1 1 1 1 1

 

Time (usec)

Figure IV-3: Time-domain three-pulse ESEEM spectra of randomly oriented l4N-
methylamine-reduced semiquinone, Conditions: 1, 219 ns; Microwave frequency, 8.815
GHz; magnetic field, 315.5 mT; microwave pulse power, 43W; pulse sequence repetition
rate, 4 Hz; events averaged per time point, 10; sample temperature, 4.2 K; number of
scans averaged, 9.

99

 

(a)

 

 

 

5..
y.
4-
3_
r
2_
1_
o_

 

 

 

 

 

 

 

 

 

 

 

 

d.)
'U
3
E-‘ 141%1LL #14; l
< O 2 4 6 8 10
O
.26
a _ (b)
5_
4_.
3...
2_.
1..
O_
-1 lllLélllllllLlllill
0 2 4 6 8 10

Frequency (MHz)

Figure IV-4: Frequency-domain three-pulse ESEEM spectra of randomly oriented (a)
l4N-methylamine-reduced semiquinone; (b) 15N-methylarrrine-reduced semiquinone.
Conditions were identical to those given in figure IV-3.

100

ESEEM contribution is expected to be distinctly different from that of 14N. A
comparison of the two data sets shown in figure IV-4 identifies the narrow peaks at 1.0
and 1.5 MHz as arising from hyperfine coupling to the CH3'4NH2-derived nitrogen.

Figures IV-5 and figure IV-6 show three-pulse ESEEM data collected for
CH3'4NH2- and CH3'5NH2-treated MADH at two other r-values, 365 us (figure IV-S)
and 450 us (figure 1V-6). Because of the t dependence of three-pulse ESEEM amplitudes
(equation [263]), different frequencies are accentuated in these two data sets. For the r =
365 ns data (figure IV-5a, b), there is a significant difference in the amplitude of the 0.5
MHz component, with it being twice as intense with respect to 2.0 and 2.5 MHz peaks
(assigned to TTQ) for the CH3'4NH2 sample (figure IV-5a) as for the CH315NH2
counterpart (figure IV-5b). This observation allows us to assign part of the intensity at
0.5 MHz to substrate-derived l4N. Similarly, peaks at 0.8 and 1.2 MHz are observed for
both 14N- and 15N- methylamine treated samples at t = 450 ns (figs. IV-6a, b) and are
assigned to the second 14N of the TTQ cofactor (N(2)). ESEEM data collected at 1: = 365
and 450 us show a strong peak at 4.3 MHz due to substrate-derived MN. The narrow
linewidth of this modulation component relative to the broad 4.6-5.8 MHz feature
observed for N(l) of the TTQ' (figure IV-4) indicates that the anisotropy in hyperfine
interaction for substrate-derived l4N might be comparatively small. Assignments based
on these observations and the assumption that both TTQ nitrogens (labeled N-l and N-2)
are near "exact cancellation"(section IV.4.2) such that their low frequency lines are

additive, are listed in table IV-l.

101

 

 

 

 

 

 

 

 

 

 

Echo Amplitude
A

3.2 —

2.4 —

1.6 ~

0.8 —

 

 

 

 

 

 

I
Eh
I

1 L l 1 1 L I 1 4L 1 J

0 2 4 6 8 10

 

 

Frequency (MHz)

Figure IV-S: Three-pulse ESEEM spectra of randomly oriented l4N-methylamine-
reduced semiquinone (a); and 15N-methylamine-reduced semiquinone (b). 1: = 365 ns.
Other conditions were identical to those given in figure IV-3.

102

 

 

 

 

 

 

Echo Amplitude

 

 

 

 

 

 

Frequency (MHz)

Figure IV-6: Three-pulse ESEEM spectra of randomly oriented l4N-methylarnine-
reduced semiquinone (a); and 15N-methylarnine-reduced semiquinone (b). 1: = 450 us.
Other conditions were identical to those given in figure IV-3.

103

[V.4 Discussion

Continuous wave EPR spectra of frozen solutions of MADH semiquinones
prepared by reduction with 14N- or 15N-methylamine show lineshapes that are
inhomogeneously broadened due to the presence of multiple hyperfine interactions. The
observed increase in peak-to-peak linewidth of cw-EPR spectra for substrate-reduced as
compared to dithionite-reduced semiquinone is indicative of a different electronic
structure for the two forms. To understand these differences in detail, hyperfine and
quadrupole coupling constants for the nitrogens that give rise to the ESEEM of figures

1V-4, IV-5 and IV-6 must be determined.

IV.4.1 Intemctions of the nitrogen nucleus

The 14N nucleus has a spin 1 = 1 and possesses nuclear quadrupole moment that
is rather small, about +0.016 x 10'24cm2. As a result, 14N equ values are small, seldom
more than 6 MHz. However, the nuclear quadrupole interaction (nqi) tensor is usually
large relative to the nuclear Zeeman interaction (~ lMHz at X-band) and has marked
effect on its modulation frequencies.

The nuclear quadrupole coupling energy is given by,

EQ = ' 6/6 Qij Vij

[4.1]
Where 6 is the electronic charge, V and Q are symmetric 3 x 3 tensors that describe the

time averaged electric field gradient and the nuclear quadrupole moment respectively. The
product of these two tensors will depend on the relative orientation of the two axis

sy Stems. When the axis system of Q is chosen to coincide with that of the spin system,

104

then due to cylindrical symmetry of the nucleus, tensor Q can be defined in terms of a
single parameter 'nuclear quadrupole moment' Q. The electric field gradient tensor V is
symmetric and can also be expressed in its own principal axis system (PAS) to yield a
diagonal traceless tensor, with BZV/axz, an/ay2 and an/az2 as the diagonal elements,
where V is the electrostatic potential at the nucleus. The electric field gradient tensor can

be expressed in terms of two quantities,

BZV/a 2 - azwa 2
qa=azwazz andn =( 3:”; y ) =(qu-qyyvqu [4.21

 

By convention |qu| > quxl > |q”.|, so 1] known as asymmetry parameter, ranges from 0 to
1. It measures symmetry of the electric field gradient around the direction defined by
maximum electric field gradient component (qu), n = 0 indicates axial symmetry. It is
called asymmetry parameter because it is a measure of deviation of the electric field
gradient tensor from axial symmetry. It does not involve any property of the nucleus and
can be interpreted directly in terms of the electron distribution in molecules.

It is necessary to define the orientation of electric field gradient PAS in the
molecular axis system for which three Euler angles are required. In the PAS the

Hamiltonian is given by,
were, = equ/4 [3 I,2 - 12 + 110.3 - 1,2)], [4.3]

The product equ/h is called the 'quadrupole coupling constant', which for a given nucleus

with nuclear electric quadrupole moment eQ, measures the maximum component of the

electric field gradient, eq at the nucleus.

105

For 14N in an electric field gradient, nqi gives rise to three energy levels, and in
zero magnetic field, three transitions are possible with pure nuclear quadrupole

frequencies given by,

v. = equ(3 + n)/4
v. = equ(3 - n)/4 [4.4]

v0 = (equ)n/29

such that v0 = v,. - v. and v+ > v. > v0.

IV.4.2 Exact aancellation

ESEEM spectrum of 14N is often indicative of 'exact cancellation condition'll,12
that arises when its isotropic hyperfine interaction is equal to twice its nuclear Zeeman
frequency, IAisol = 2v“. Under this condition, the hyperfine and the nuclear Zeeman
interactions "cancel" one another in one of the electron spin manifolds (figure 1V-7), such
that the modulation frequencies are dominated by l4N-nuclear quadrupole interaction.
This manifold yields three narrow, low frequency peaks in the ESEEM spectra where the
two lower frequencies add to give the third. They are independent of the direction of
external magnetic field, and thus, lead to the appearance of narrow peaks in ESEEM
spectrum. The other electron spin manifold, where the electron-nuclear coupling and the
Zeeman interaction are additive, typically gives rise to broad, or rapidly damped ESEEM
features with the most prominent spectral component being that between the highest and

lowest energy levels of the manifold. Because the two energy levels involved in this

106

 

 

 

 

 

 

 

 

 

 

Ms M] ,T
+1
, Hi her fre uenc
+1/2 ,4 0 g . q y
’,___.,,‘.--___ ---.____‘ broad line
’I” ‘ -1 K e
’Z \~ ’0
-——*I
\“ -1
“. '1/2 a” 0 “~‘____"’. Lower frequency
1‘. ":9 ’x N 1 lines
‘. +1 x ~ Q
Electronic Nuclear Nuclear uadru ole
Z + + Hyperfine + Q p
eeman Zeeman lnteractron (N Q1)
(3)

 

__l ___/\_

V0 V. V+ qu

(b)

Figure IV-7: Energy level diagram for the interaction of an unpaired electron with a
nucleus of MN (I = l) at exact cancellation condition (a); and the expected spectrum (b).

107

modulation component have mostly m1: 1 or m; = -1 character, this peak is referred to as
a "double quantum" (Aml = +2) feature. The frequency of this broad component at exact
cancellation is roughly four times the nuclear Larmor frequency, or 4 MHz in the case of

MN at X-band. The other two Am] = +1 transitions from this manifold are highly

orientation dependent13 and are not resolved in frozen samples. As long as the condition
for exact cancellation is nearly satisfied, both the components equ and n, of the field
gradient tensor can be obtained from the frequencies of the three sharp 'zero field' or 'pure'

nuclear quadrupole resonance lines in ESEEM spectrum.
equ = 2(v. + v_)/3; and n = 3(v. - v.)/(v+ + v.) [4.5]

Since 15N is a spin 1 = 1/2 nucleus, it does not have a quadrupole moment, and its ESEEM

is determined by hyperfine and nuclear Zeeman interactions only.

IV.4.3 Spherical model approximation

In case of glassy samples the echo signal has contributions from a distribution of
electron-nucleus systems randomly oriented with respect to the external static magnetic
field Bo. Therefore spatial averaging must be performed, which requires integration over
all possible orientations of B0 with respect to a common molecular frame to which the nqi
and hfi tensors have been transformed. In the simulation program used, the quadrupole
tensor is rotated into the PAS of the hyperfine tensor. The Euler angles defining this
transformation are determined by molecular geometry and are independent of the direction

of B0. The overall echo intensity in the PAS of hfi tensor is given by,

108

(Em) = :11; [02" [:E(00,¢o,t)sineodeoc1¢, [4.7]

where 00 and (I)O are the polar angles defining the direction of B0 and t = t or t = t + T for
two-pulse or three-pulse experiments respectively. When the unpaired spin interacts with
more than one nuclear spin, then the resulting modulation is the product of the
modulations due to each nucleus. In such a case a different set of Euler angles will be
needed to transform nqi tensor of each nucleus to the PAS of the hfi tensor. The

modulation is then given by,
N _ 1 2n 1: N _
<E(t) >"EI [0gE,(eo,¢o,t)Stneodeod¢0 [4.8]

Information about the geometry of the nuclei around the paramagnetic center should be
available in order to evaluate the above integral. Extensive computing time may be
required in order to evaluate the above integral. It is often useful to make 'spherical
approximation' as described by Kevan et al.14 to simplify the spatial averaging. When the
dipolar interaction is smaller than the nuclear Zeeman term it can be assumed that the
nuclei are arranged around the electron over a sphere at an 'effective' distance 'r', and 00 is
the angle between r and the magnetic field direction. The dependence on (1)0 can be
removed by shifting to a coordinate system rotating about z-axis. The mutual nuclear
arrangement of the nuclei can be neglected in the averaging procedure, then equation 4.8

becomes,

(E(t)“) —_— 111;]:Ln[E(00,t)]Nsin0°d00d¢0 [4.9]

109

where t = 1: or 1: + T depending on whether it is a two-pulse or a three-pulse experiment.
This equation represents the modulation due to N nuclei at the same orientation, which is
not realistic in many cases. A more appropriate model is to consider the surrounding

nuclei as randomly oriented over the sphere such that,
(Em)N — [ij‘z’cltp ["siut) d6 E(0 0T [4 10]
471: o o o o o o’ '

This approximate spatial averaging method is known as 'spherical model'. It works well
for distance r > 4 A. This approximation can be safely made for cases where there are at
least four or more coupled nuclei.

Our computer simulations made use of this spherical model approximation to the
product rule for assembling the ESEEM spectra of the three magnetically coupled
nitrogens of TTQ. The consequences of this approximation were tested for dithionite-

reduced MADH. For this case, only two nitrogens contribute to the data and they do so

in a nearly equivalent fashion.2 To perform this task we modified our simulation
software to simultaneously consider two I = 1 nuclei whose magnetic axes were related by
a simple Euler angle rotation. If one considers the principal axes of both the hyperfine
tensors to be at an angle of 45° with respect to one another, the ESEEM calculated with
the precise formulation of the product rule yields results that are nearly identical with
those obtained using the spherical model approximations. In the worst case, when both
hyperfine principal axes are collinear or perpendicular, the errors made by using the
spherical model approximation are modest. Because the X-ray crystal structure of

MADH shows the planes of the indole and indole-quinone rings of oxidized TTQ to be at

110

an angle of 38° with respect to each other, use of the spherical model approximation is

justified for our work.

IV.4.4 Theoretical Simulations

The analysis of '4N ESEEM spectra involves 8 adjustable spin Hamiltonian
parameters for each nucleus: three principal elements for the hyperfine tensor; the
quadrupole coupling constant, equ/h, and its asymmetry parameter, T]; and three Euler
angles relating the orientation of the quadrupole tensor’s principal axis system to that of
the hyperfine tensor. Because the substrate-derived 14N and N(l) of TTQ are likely near
exact cancellation, good starting values for equ, 1] and Ago =l/3(All + 2A 1) can be
determined directly. Also, the assumption of an axial l4N-hyperfrne tensor reduces the
number of parameters for each 14N by one. Our simulations are then mostly aimed at
determining the hyperfine anisotropy, Ad", = 1/3(AII - A l), and the Euler angles that relate
the two tensors. These parameters manifest themselves in the spectra of figures IV-4, IV-
5 and IV-6 in that they control the lineshapes of broad peak in the 4-6 MHz region and
influence the ‘t-suppression behavior of the ESEEM spectra. Our numerical simulations
focused on accounting for the frequencies, lineshapes and relative amplitudes of the peaks
and then, on understanding the changes that occurred as t was varied.

Using equation 4.5, the nqi parameters for the substrate-derived l4N nitrogen are
estimated to be: equ, 1.7 MHz; and r], 0.6. Similarly for the N(l) nitrogen of TTQ,
e2qQ = 3.0 MHz and r] = 0.3. When the anisotropic hyperfine interaction is small, Aiso

can be estimated from the frequency of the double quantum peak, vdq, that arises from the

111

electron spin manifold where Zeeman and hyperfine fields are additive, using the

following equation:15

A12... /4 f V. Also + (<32C1Q/4)2 (3 + 1[12) + V: - Via/4 = 0 ‘ [4-11]
Using the values of equ and 11 determined above, Aiso values of 2.1 and 6.1 MHz, and
2.5 and 6.5 MHz respectively, were estimated for the substrate-derived '4N and N(l) of
TTQ. The" higher values of Aiso were neglected because they are far from the exact
cancellation value of Aiso, 1.94 MHz at 3155G and did not give the desired ESEEM
features.

Using these initial estimates for nuclear quadrupole and isotropic hyperfine
coupling constants along with non-adjustable parameters dictated by our experimental

conditions, computer simulations of the data shown in figures IV-4, IV-5 and IV-6 were

undertaken. The ESEEM spectrum of each nitrogen was first simulated separately, and

then combined using the product rule. 1 6 The Euler angles were initially set to zero and
the estimated values of equ and n were varied to account for the positions of the narrow,
low-frequency components whose characteristics are primarily determined by the nuclear
quadrupole interaction. Then, hftensor values were varied to obtain the correct position
and width of the double quantum line. Finally, the Euler angles were varied to account for
the relative intensities of the low-frequency components and shape of the double
quantum peak. As the line shape of the double quantum feature depends on the hf tensor
principal values as well as on Euler angles, they had to be adjusted simultaneously in

order to obtain the best fit to the data.

112

Only two 14N features at 0.8 and 1.2 MHZ from the second indole nitrogen N(2)
of TTQ' are resolved (figure IV-6a). It is possible that these two frequencies represent v0
and v- for N(2) and that v+ could overlap with v- from N(l) at 2.0 MHz. With this
assumption, simulations were carried out using the above procedure. Because the "double
quantum" peak from this nitrogen was not resolved at any of the 1: values studied, we
focused on simulations that showed small vdq amplitudes for this nucleus, presumably
because of larger hf anisotropy than that characterizing N(l) or substrate-derived
nitrogen. The other possibility that vdq’s from both N(l) and N(2) fall in 4.6-5.8 MHz
region, did not predict observed intensities and dependence on ’r.

The principal values of hf and nqi coupling parameters obtained from our
computer simulations for substrate-derived nitrogen and the two nitrogens from TTQ, are
listed in Table IV-l. Figure IV—8 shows the simulated frequency domain ESEEM spectra
calculated using the experimental conditions and r values of 219 (a), 365 (b) and 450 ns
(c). The simulations account well for the frequencies, relative amplitudes and linewidths
of major ESEEM features. The trends in amplitude changes that occur with r are also in
agreement with experiment. The chief shortcomings of our modeling are in accounting for
the details of the double quantum features. Independent information on the symmetries

of the hyperfine tensors as determined from two dimensional ESEEM measurements

(HYSCORE)17 would be beneficial here, but progress on such measurements has been
impeded by long Tlc of the MADH semiquinone (> 1 sec).

Close examination of the data shown in figures IV-4, IV-5 and IV-6 fails to reveal
contributions from substrate-derived 15N. To support this observation, simulations were

performed using l4N-coupling parameters given in table IV—l for the TTQ nitrogens along

113

 

with '5 N hyperfine couplings obtained by scaling the corresponding values measured for

l4N-substrate coupling by a factor g.,N/g.,N = 1.403. The results predicted that no

distinct ‘5 N peaks would be resolved for ESEEM collected using the conditions of figures

IV-3, IV-4 and IV-S. Simulations of 2-pulse data also showed that these measurements

would not reveal ISN contributions in the presence of deep modulation from the TTQ

nitrogens.

Table IV-l: The assignment of the observed frequencies and the principal values of hf
and nqi coupling parameters used in computer simulations for substrate-derived nitrogen
and the two nitrogens from TTQ.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l4N nucleus FEZZEIeVnecfieEilE/IEFE) A l A“ A'so Adip lequ/h 1] 0t [3 y
Substrate 0.5, 1.0, 1.5, 4.3 2.1 3.1 2.4 0.3 1.7 0.5 0° 108° 60°
N(l)-TTQ 0.5, 2.0, 2.5, 4.8-5.1 2.6 3.2 2.8 0.2 3.0 0.3 0° 69° 0°
N(2)-TTQ 0.8, 1.2 1.4 3.0 1.9 0.5 2.1 0.7 0° 0° 0°
N(l) O-form° 0.4, 2.1, 2.5, 4.0-6.0 1.9 2.6 2.1 0.3 3.0 0.2 0° 60° 0°
N(2) O-form. 0.8. 1.8, 3.1. 5.0-6.0 1.2 2.1 1.5 0.3 3.2 0.6 0° 90° 0°

 

* Data taken from reference 2.

Figure lV-9 shows a comparison of echo amplitude modulations for the time

domain data collected at 1: = 219 ns for the MN-serniquinone (solid line), with that

obtained from our simulation (dashed line). The simulation was multiplied by a decay

114

 

Echo Amplitude

 

 

 

 

 

 

 

 

 

 

 

 

 

 

- (a)

l"

_ l l l l L l l

0 2 4 6 8 10
M

'- i l l l l l l l
0 2 4 6 8 10
- (C)

F

+—

L—

l l 1 l 1 l L
0 2 4 6 8 10

Frequency (MHz)

Figure IV-8: Cosine Fourier transformations of numerical simulations of three-pulse
ESEEM data of CH3MNH2-generated semiquinone in MADH (a) 1:, 219 ns; (b) ‘r, 365 ns;
(c)1:, 450 ns. The simulations were performed using three coupled l4N nuclei, N(l), N(2)
and substrate-derived N, whose hf and nqi coupling parameters are listed in table IV-l.
Other simulation parameters: gN, 0.4035; magnetic field strength, 315.5 mT. An isotropic

electronic g-factor was assumed.

115

 

0.9

0.8

0.7

0.6

0.5

Normalized Echo Amplitude

 

 

 

0.4

T+ t ( usec)

Figure IV-9: A comparison of the normalized three pulse electron spin echo envelope
modulation amplitudes for l‘lN-methylamine poised MADH semiquinone measured
experimentally (solid line) at 12 = 219 ns with that predicted by spectral simulation
(dashed line) for the same ‘r-value.

116

function, e“, where 1: was 4.5 us. The depths of modulations in the time domain data
and simulation agree well and show that the semiquinone is nearly pure N-form. The
above 14N-ESEEM analysis has provided two types of Hamiltonian parameters that
report on the electronic structure of the TTQ-semiquinone. The hyperfine coupling
parameters can be directly related to the structure of the highest occupied molecular
orbital (HOMO) of the semiquinone in that the dipolar portion of the tensors are a
measure of the unpaired spin density localized in the pn-orbital of each coupled nitrogen.
When an observed nuclear coupling constant is known to arise from a particular atomic
orbital of the coupling atom in a free radical, the electron spin density of this atomic
orbital can be obtained from the ratio of the measured coupling constant to that for the
corresponding orbital of the free atom. Spin densities of p“ orbitals calculated inthis
manner are accurate only when the Adip is entirely due to the spin density in the particular
p“ orbital with its axis fixed in the crystal. Spin density in the perpendicular p orbitals of
the coupling atom or the motion of the coupling orbital relative to the crystal axis or
Charge on the coupling atom influence the value of Adip and hence affect the spin densities.
In the present case, an estimate of the n-orbital unpaired spin density of each nitrogen in
TTQ is obtained by dividing the Adip values in table IV-l by the corresponding value
meaSured for atomic nitrogen, 47.8 MHz. Table IV-2 summarizes n-orbital unpaired spin

densities located on the three nitrogens present in the N—form semiquinone and compares

1!

them to the values determined previously for the O-form radical.2 All of the calculated p

Spin densities are small showing that our l4N-ESEEM experiments are detecting only a
Small portion of the TTQ° HOMO, and that the HOMO is delocalized over the

framework of the cofactor. However, these numbers do reveal some of the details of

117

electronic structure changes triggered by substrate binding. For O-form semiquinone, the
unpaired p" spin densities on N(l) and N(2) nuclei of the cofactor are equal, consistent
with a more uniform distribution. N-form semiquinone is characterized by an unpaired
spin density on N(2) that is almost three times larger than that found for N(l). The

presence of substrate derived amine has altered the HOMO, increasing the contribution of

N(2) to the MO. while reducing that of N(l).

Table IV-2: n-orbital unpaired spin densities located on the three nitrogens present in the
N- form semiquinone and comparison with the values determined previously for the 0-
form. radical.

 

Serniquinone Electron densities

N- N(l)- N(2)-
Substrate TTQ TTQ
N-form 0.007 0.004 I 0.01 1
O-form* --- 0.006 I 0.006

 

 

 

 

 

 

 

 

* Data taken from reference 2.

Molecular orbital calculations of spin densities and hyperfine coupling constants
wer e performed using SPARTAN and Gaussian 98 programs on N- and O-semiquinone
for-r113 of a model compound structurally related to TTQ (figure IV-lO). This model
CornIDOund has been previously synthesized and characterized by Itoh, et.al.18"20 The
nOta‘lions N1 and N2 in figure IV-10 are the notations used in SPARTAN and Gaussian
98 pro grams, and they are not the same as N(l) and N(2) used in numerical simulations.

In e’Ell‘lier studies molecular orbital calculations and density functional analyses have been

118

 

Fig“ re IV-10: Structure of the oxidized form of the model compound of TTQ cofactor of
H used in theoretical calculations. Atoms are labeled as, nitrogen in magenta,
Oxygen in red, carbon in black and hydrogen in yellow.

119

performed on the tryptophan radical and benzosemiquinone anion radicals providing the

molecular geometries, electron spin densities and approximate hfcoupling values.21'29
In going from these molecules to the model compound of TTQ, due to increased number
of heavy atoms (atoms other than hydrogen) in the molecule, the calculations become
more expensive. The presence of possible hydrogen bonding is also expected to change
these molecular properties to a great extent. The density fimctionals and semiemperical
theory, that can calculate molecular properties to a reasonably good approximation for
these planar molecules, may not yield satisfactory results for the non-planar model
compound of TTQ. This is because all these methods suffer from having poor electron
correlations. For obtaining good quantitative picture and for calculating the values of
isotropic hyperfine and anisotropic hyperfine coupling constants for the three nitrogens
of N-form semiquinone, methods based on correlated coupled cluster theory need to be
used. These methods treat electron correlations to better approximations and are
expected to give more realistic results but are exorbitantly expensive for a large molecule
like model coumpound of TTQ. However, the results obtained from the semiemperical
calculations give a qualitative outlook about the spin densities in this molecule.

The calculations were performed using Hartree-Fock, density functional theory

and semiemperical theory with UHF, B3LYP, 6-31G' basis sets for comparison.30’3 1.
Geometry optimization led to minimum steric energy when the dihedral angle of the two
indole planes was 359° for the N-semiquinone form and 352° for the O-semiquinone
form of the model compound. The EPR spectroscopic measurements on this model

compound had suggested a partial “It-conjugation between the indole and indo-quinone

rings and a non rigid conformation.19 The calculated spin density for the O-form is

120

plotted in figure IV-11, where it can be seen that the extent of the 0.002 e/au3 spin density
contour around N1 and N2 are almost same, and the unpaired spin density is equally

distributed on both the nitrogens of the TTQ ring system which agrees with the results of

previous ESEEM studies.2 Calculations using unrestricted Hartree-Fock model were
found to be less accurate. The calculated spin density for the N-form is plotted in figure
IV-12. Figure IV-13 is a plot of the electron density and HOMO of the N-form showing
that the HOMO is spread over the entire molecule. Figure IV-14 is a similar plot for the
O-form. Using the results of this calculation the N(l) and N(2) nitrogens from theoretical
simulations can be identified. The indole nitrogen has the higher spin density, which
favors the idea that the electron transfer from MADH to amicyanin follows a path
involving this indole ring that is nearest to Cu of amicyanin (figure I-4). An ET from the
indole ring would be more favorable as compared to the other possible path involving
direct transfer from the reduction site at C6. As seen from figure I-4 the most direct
distance between the site of radical generation, the substrate-derived nitrogen, and Cu is
about 16.8 A and a direct ET would involve a large through-space electron jump which is
less favorable because the efficiency of ET through space is relatively very poor.

The alterations of the HOMO for the TTQ semiquinone caused by the presence
of substrate-derived amine are interesting in the light of the different electron transfer

(ET) rates to amicyanin measured for N- versus O-form quinol and semiquinone

intermediates of MADH.32 Electron transfer from the N-semiquinone is slower than
from the O-semiquinone. Conversely, electron transfer from the N-quinol, which yields
the N~semiquinone as a product, is faster than from the O-quinol. The differences in the

ET rates from the N- and O-semiquinone forms appear to be due to differences in their

121

Figure IV-ll: Calculated spin density for the O-semiquinone form of the model
compound. Calculations used semiemperical model and UHF basis sets in SPARTAN
software program; isosurface value, 0.002. Atoms are labelled as, nitrogen in magenta,
oxygen in red, carbon in gray and hydrogen in white.

122

r
I '1."|M
I

   

123

Figure IV-12: Calculated spin density for the N-semiquinone form of the model
compound. Calculations used semiemperical model and UHF basis sets in SPARTAN
software program. Atom labeling is same as in figure IV-l 1.

124

MIN

" I: um

 

-- Mil]

 

125

Figure IV-l3: A plot of the electron density and HOMO for the N-semiquinone form of
the model compound showing that the HOMO is spread over the entire molecule. Atom
labeling is same as in figure IV-l l.

126

 

127

Figure IV-14: A plot of the HOMO for the O-semiquinone form of the model
compound. Atom labeling is same as in figure IV-11.

128

redox potentials.32 The basis for differences in the reaction rates of the N- and O-form
quinols, which yield the corresponding semiquinone forms as products, is more difficult

to interpret because the reaction of the N-quinol is gated by the deprotonation of the

substrate-derived amino group on TTQ33 (discussed later), whereas ET reactions from
dithionite-reduced quinol and semiquinone forms of MADH are rate-limited by the ET
event. In addition to redox potential, two other parameters which determine ET rates are

the electronic coupling (H A 3) related to the distance that the electron must travel and the

reorganizational energy (A) composed of solvational and vibrational components.34 The
covalent incorporation of substrate-derived N into TTQ affects the rate and standard free
energy (AG°) for the ET reaction from the TTQ semiquinone by altering its redox
potential, but it does not alter A and H A 3 associated with ET from TTQ to amicyanin
which are same as in the case of the ET reactions of the dithionite-reduced O-quinol and
O-semiquinone forms. An important result from this study is that the unpaired electron
spin density in the TTQ semiquinone is delocalized over the framework of TTQ. This
suggests that the starting point for the electron transfer reaction may be any point on the
cofactor, including the edge of the Trp108 indole ring, which is closest to amicyanin. The
asymmetry of the spin density raises an interesting question of whether the observed
differences in the HOMO for different TTQ forms would contribute to differences in H A B
for electron transfer reactions from different redox forms. Another interesting question is
to what extent the electronic structure of the TTQ semiquinone, relative to that of the
oxidized and reduced forms, contributes to the 2» associated with electron transfer

reactions that involve the semiquinone TTQ as a reactant or product. The electron

transfer reactions of MADH exhibit relatively large values of A > (2 eV),32 which may be

130

reflective of the kinetic complexity of this reaction. The ET from TTQ to amicyanin may

be a coupled ET in which a rate-limiting ET event is preceded by a relatively rapid but

unfavorable non-ET reaction step.35 For such a reaction, the experimentally determined
A value will include contributions from both the A associated with the ET as well as the
energetics associated with the non-ET step. A possible non-ET step could be a change in
dihedral angle between indole and indolquinone rings in order to optimize the system for

ET. The need for such a change in dihedral angle was also suggested on the basis of

studies with TTQ model compounds,19 and from the molecular orbital calculations and
geometry optimization of the model compound in this study. Such a reaction would be
expected to occur at the expense of large amount of energy and would contribute
significantly towards increasing the A value. These new data will form the basis for future
studies to elucidate any correlation between electron transfer properties and the electronic
properties of this and other organic redox cofactors.

The other group of l4N-Hamiltonian parameters derived from this work are
nuclear quadrupole interaction (nqi) parameters, eZqQ and 1]. Although the 14N nucleus

has a small quadrupole moment, this interaction can ofien be substantial because of the

large electric field gradient created by the atom’s 2p electrons.36 For the tricoordinate,
planar geometries that describe the bonding of the 14N nuclei of the TTQ semiquinone, the
principal axis of the electric field gradient is directed along the 2p“ orbital of the atom,
perpendicular to the three sp2 hybrid orbitals that characterize the sigma bonding. Using
this simple Townes-Dailey description of the interactions, the value of equ, the

quadrupole coupling constant, can be related to the electron occupancy of the nitrogen

2p“ orbital.37 If this orbital houses 2 electrons, the equ value is > 4 MHz. For the

131

indole nitrogen of tryptOphan powder equ is 3.0 MHz, which corresponds to 3 2p“

orbital occupancy of about 1.75.38 The equ values measured for the TTQ nitrogens of
the O-form semiquinone are 3.0 (N (1)) and 3.2 (N (2)), indicating that the occupancy of
the p“ orbital on these heteroatoms is not greatly affected by the protein environment or
the difference between the electronic structure of the cofactor and the amino acid
(tryptophan) model compounds.

The asymmetry parameter 1] is used to describe the degree of rhombicity in the

electric field gradient. 11 ranges from 0, for an axial interaction, to 1 for the completely

rhombic case.36 For tryptophan powders, the indole nitrogen is found to have n = 0.18
indicative of an axial tensor and a measure of the 'equivalence' of the C-N and N-H bonds.
For the O-fonn semiquinone, Warncke, et al. found that n for N(l) was 0.2, identical to
that of the tryptophan model, while the N(2) asymmetry was 0.6 indicating an
asymmetry likely caused by hydrogen-bonding to a peptide. The x-ray crystal structure
of oxidized MADH shows nearly equivalent hydrogen bonding interactions for both the
indole and the indole-quinone nitrogens.39 The results of Warncke, et al. on O-form

semiquinone had shown that only one of these interactions is important for reduced

cofactor.2

The differences observed in MN hyperfine couplings between 0- and N-form
semiquinones are mirrored by changes in eZqQ. For the N(l) nitrogen of N-form TTQ',
both equ and n were determined to be 3.0 MHz and 0.3, respectively, and remain close
to the indole model compound values. The equ values for the N(2) nitrogen dropped

from 3.2 MHz for the O-form to 2.1 MHz for the substrate reduced enzyme. Using the

132

analysis results of Edmonds and Speight, this decrease in equ can be interpreted as a
decrease in the N(2) 2p“ orbital population from ~1.75 to about 1.63. This decrease is
supportive of our ESEEM simulation results for N(2) that attribute the lack of a ‘double
quantum’ peak to an increase in hyperfine anisotropy or 2p“ unpaired electron spin
density. The asymmetry parameter for N(2) was found to be nearly 0.7, essentially
unchanged from the value determined for the O-form.

The ESEEM spectra of the N-form semiquinone also yielded a well resolved set of
resonances due to substrate derived nitrogen. The measured isotropic coupling constant

is 2.4 MHz, in good agreement with value estimated from samples prepared by

comproportionation and studied previously by ESEEM.10 This value compares well to
those measured for the TTQ nitrogens, 2.7 (N (1)) and 1.9 (N(2)), and definitely shows
that the substrate-derived nitrogen is covalently bound to the TTQ semiquinone. The
low value of eZqQ, 1.7 MHz and the deviation of the asymmetry parameter from that
expected for an axial interaction (11 = 0.5) are commensurate with the incorporation of the

substrate nitrogen into the n-system of the TTQ semiquinone and our previous proposal

that the nitrogen is an imine-like species.33a40 The low value of equ dictates that all of
the substrate derived nitrogen’s electrons are involved in bonding interactions, and agrees
well with the equ value determined for the substrate derived nitrogen of the topa-
semiquinone cofactor of copper amine oxidases.41s42 For these systems a combination
of 14N- and 2H-ESEEM, and lH-ENDOR spectroscopies were used to show that the
nitrogen derived from amine substrate was bound to the topa-quinone cofactor as an

imine-like species with one strongly, and one weakly coupled exchangeable proton

accounting for the bonding about the heteroatom.42

133

For MADH, an imine-like semiquinone species with one strongly, and one weakly
coupled exchangeable proton is very consistent with the chemical reaction mechanism that

we have proposed for the deprotonation reaction that gates the electron transfer from the

N-quinol MADH to amicyanin.33 It was proposed that rate-limiting deprotonation of
the substrate-derived amino group on reduced TTQ was required to activate the system
for rapid electron transfer. This deprotonation of TTQ-bound —NH2 requires a
monovalent cation at the active site to stabilize the transient —NH' intermediate from
which electron transfer occurs to yield the =N—H imino-semiquinone product. Thus, in
the semiquinone state the active site residue which removed the proton fi'om —NH2 will
still be in close proximity to the =N—H of the semiquinone and in a position to provide
this same proton for hydrogen bonding to the imino-semiquinone nitrogen. This would

yield one strongly (covalent), and one weakly (H-bonded) coupled exchangeable proton.

IV.5 Studies on aromatic amine dehydrogenase

Aromatic amine dehydrogenase (AADH) is an enzyme that uses the same
tryptophan tryptophquuinone (TTQ) cofactor as methylamine dehydrogenase
(MADH). Like MADH it also catalyzes the oxidative deamination of a distinct class of
primary amines. AADH is most specific for phenylethylamines, and is produced by
Alcaligenesfaecalis and allows growth on phenylethylamine as a carbon source. AADH
uses a blue copper protein, azurin, to mediate electron transfer to cytochromes
homologous to the role of amicyanin in P. denilrificans. Recent studies have shown that

there is spectroscopic evidence for a common electron transfer pathway for these two

tryptophan tryptophquuinone enzymes.43 13C- and l5N-NMR studies of the reaction

134

of AADH with methylamine have also demonstrated that the products of the reductive
half-reaction are an equivalent of formaldehyde hydrate and a reduced aminoquinol form
of the tryptophan tryptophquuinone (TTQ) cofactor which contains covalently bound
substrate-derived N. The substrate-derived aminoquinol nitrogens of AADH and MADH
exhibit 15N chemical shifts, which are separated by approximately 7 ppm. and may reflect
differences in the active-site mediated electrostatic environment of the aminoquinol N.
The aminoquinol of AADH is also less stable against reoxidation than that of MADH,

suggesting that the active-site environments of the respective enzymes influence the

relative reactivity of the aminoquinols.44

The overall structural and functional similarities suggest that AADH and MADH
may have evolved from a common ancestor to yield enzymes with distinct substrate
specificities. A comparison of the experimental data of the N-form (”N- and 15N-
substituted) and O-form of the semiquinone from AADH and those from MADH should
be useful in further understanding the reaction mechanism involved in AADH.

Cw-EPR (figure IV-l 5) and three pulse ESEEM experiments were performed on
the AADH O-form and N-form semiquinones. Unfortunately, the echo intensities for the
AADH samples were weak and a comparison of the three pulse ESEEM spectra for the
14N- and 15N-substituted semiquinones showed no significant difference from that of the
O-form spectrum; only small changes in the intensities of the peaks were observed and
there was absence of any strong peaks due to the substrate-derived nitrogen which were
clearly seen in the MADH data. The reason for this could be that the poising of
semiquinone did not lead to the formation of the N-forms. But EPR-spectra of N-forms

were broadened to a great extent as compared to the O-form, indicating the presence of

135

 

12..

 

 

 

 

3320 3340 3360 3380 3400 3420

Magnetic Field (G)

Figure IV-lS: Continuous-wave EPR spectra of the semiquinone forms of aromatic
amine dehydrogenase from Alcaligenesfaecalis generated by dithionite addition (O-form)
(a), l4N-methylamine addition followed by illumination (b), and 15N-methylamine
addition followed by illumination (c). Spectra were collected under the following
conditions: microwave power, 2.0 mW; microwave frequency, 9.484GHz; modulation
frequency, 100 kHz; modulation amplitude, 0.1 mT; sample temperature, 20 K; and
number of scans averaged, 25.

136

extra coupling. It is possible that since the concentration of N-forms were very small in
these samples, (as reflected in the poor echo amplitude) their signals were obscured by
the l4N-peaks near exact cancellation arising from the indole and indo-quinone nitrogen

atoms of TTQ cofactor.

IV.6 Conclusions

The ESEEM experiments and analyses presented here show that the semiquinone
species obtained by light-induced oxidation of substrate-reduced MADH contains a
covalently bound nitrogen atom that is derived from substrate methylamine. The nuclear

quadrupole coupling constant, equ, measured for this substrate-derived nitrogen is
consistent with that determined for the topa-semiquinone species of amine oxidases41

and is very distinct from the value measured for methylamine by NQR spectroscopy.36
It is likely that this substrate-derived nitrogen exists as imine nitrogen in the TTQ
semiquinone, with the lone pair of electrons that reside in the remaining sp2 hybridized
orbital involved in the formation of a strong hydrogen bond with an available proton-
donor. Furthermore, the match between the depths, or intensities, of the ESEEM
predicted by our simulations and the experimental data provides strong evidence that the

intermediate species prepared using the light-induced oxidation procedure developed by

Zhu and Davidson1 , is pure N-form semiquinone. It is important to note that these

results were obtained using stoichiometric amounts of substrate reacted with enzyme

rather than the 104 to 106 molar excesses that are typical in cation binding studies.6’45a46
The broadening of the cw-EPR lineshape found for substrate-reduced semiquinone

when compared to that of dithionite-reduced enzyme reflects changes in the electronic

137

structure of TTQ° that are a consequence of the substitution of substrate nitrogen for the
C6 oxygen atom. The hyperfine couplings measured in this work for the three nitrogen
atoms of the semiquinone provide details of how the highest occupied molecular orbital of
this chemical intermediate are affected by substrate addition. Further characterization of
this species by proton ENDOR spectrosc0py will provide a detailed picture of the
HOMO and key information for understanding how the electronic structure of the TTQ
semiquinone in MADH effects the rates and mechanisms of electron transfer from

different redox forms of MADH to amicyanin.

138

List of References

1)Zhu, Z.; Davidson, V. L. Biochim. Biophys. Acta. 1998, 1364, 297-300.

2)Wamcke, K.; Brooks, H. B.; Hong-in, L.; McCracken, J. L.; Davidson, V. L.; Babcock,
G. T. J. Am. Chem. Soc. 1995, 117, 10063-10075.

3)Husain, M.; Davidson, V. L. J.Biol. Chem. 1985, 260, 14626-14629.

4)Eady, R. R.; Large, P. J. Biochem. J. 1968, 106, 245-255.

5)Husain, M.; Davidson, V. L.; Gray, K. A.; Knaff, D. B. Biochem. 1987, 26, 4139-4143.
6)Gorren, C. F. A.; Moenne-Loccoz, P.; Backes, G.; Vries, S.; Loehr, J. S.; Duine, J. A.
Biochemistry 1995, 34, 12926-12931.

7)McCracken, J .; Shin, D.-H.; Dye, J. L. Appl.Magn. Reson. 1992, 3, 305-316.

8)Mims, W. B. J. Magn. Reson. 1984, 59, 291-306.

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

10)Warncke, K.; McCracken, J. L.; Brooks, H. B.; Babcock, G. T.; Davidson, V. L. J. Am.
Chem. Soc. 1993, 115, 63-64.

11)Mims, W. B.; Peisach, J. J. Chem. Phys. 1978, 69, 4921-4930.

12)Flanagan, H. L.; Singel, D. J. J. Chem. Phys. 1987, 87, 5606-5616.

13)Colaneri, M.; Peisach, J. J. Am. Chem. Soc. 1992, 114, 5335-5339.

l4)Kevan, L.; Bowman, M. K.; Narayana, P. A.; Boeckman, R. K.; Yudanov, V. F .;
Tsvetkov, Y. D. J. Chem. Phys. 1975, 63, 409-416.

15)Cosgrove, S. A.; Singel, D. J. J Phys. Chem. 1990, 94, 8393-8396.

16)Dikanov, S. A.; Shubin, A. A.; Parmon, V. N. J. Mag. Reson. 1981, 42, 474-487.
17)Dikanov, S. A.; Bowman, M. K. J. Magn. Reson. A 1995, 116, 125-128.

18)Itoh, S.; Ogino, M.; Komatsu, M.; Y., O. J. Am. Chem. Soc. 1992, 114, 7294-7295.
19)Itoh, S.; Ogino, M.; Haranou, S.; Terasaka, T.; Ando, T.; Komatsu, M.; Y., O.;
Fukuzumi, S.; Kano, K.; Takagi, K.; Ikeda, T. J. Am. Chem. Soc. 1995, 117, 1485-1493.
20)Itoh, S.; Y., O. Meth. Enzymol. 1995, 258, 164-176.

21)Hoffman, B. M.; Roberts, J. E.; Kang, C. H.; Margoliash, E. J. Biol. Chem. 1981, 24,
6556-6564.

22)Grafton, A. K.; Wheeler, R. A. J. Phys. Chem. 1997, 101, 7154-7166.

23)Walden, S. E.; Wheeler, R. A. J. Phys. Chem. 1996, 100, 1530-1535.

139

 

24)O'Malley, P. J. Chem. Phys. Lett. 1996, 262, 797-800.

25)O'Malley, P. J .; Collins, S. J. Chem. Phys. Lett. 1996, 259, 296-300.

26)O'Malley, P. J. J. Chem. Phys. A 1997, 101, 6334-6338.

27)O'Malley, P. J. Chem. Phys. Lett. 1997, 274, 251-254.

28)O'Malley, P. J. Chem. Phys. Lett. 1998, 291, 367-374.

29)O'Malley, P. J. Chem. Phys. Lett. 1998, 285, 99-104.

30)Becke, A. D. Phys. Rev. A 1988, 38, 3098.

31)Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785.

32)Bishop, G. R.; Davidson, V. L. Biochemistry 1998, 3 7, 11026-11032.

33)Bishop, G. R.; Davidson, V. L. Biochemistry 1997, 36, 13586-13592.

34)Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265-322.

35)Davidson, V. L. Biochemistry 1996, 35, 14035-14039.

36)Lucken, E. A. C. Nuclear Quadrupole Coupling Constants; Academic Press: London,
1969.

37)Townes, C. H.; Dailey, B. P. J. Chem. Phys. 1949, 17, 782-796.

38)Edmonds, D. T.; A., S. P. J. Mag. Res. 1972, 6, 265-273.

39)Chen, L.; M., D.; Durley, R. C.; Y., C. A.; E., L. M.; Davidson, V. L.; Mathews, F. S.
J.Mol.Biol. 1998, 276(1), 131-49.

40)Zhu, Z.; Davidson, V. L. J. Biol. Chem. 1998, 273, 14254-14260.

41)McCracken, J. L.; Peisach, J.; Cote, C. E.; McGuirl, M. A.; Dooley, D. M. J. Am.
Chem. Soc. 1992, 114, 3715-3720.

42)Warncke, K.; McCracken, J. L.; Babcock, G. T.; McGuirl, M. A.; Dooley, D. M. J.
Am. Chem. Soc. 1994, 116, 4028-4037.

43)Edwards, S. L.; Davidson, V. L.; Hyun, Y.-L.; Wingfield, P. T. J. Biol. Chem. 1995,
270, 4293-4298.

44)Bishop, G. R.; Zhu, Z.; Whitehead, T. L.; Hicks, R. P.; Davidson, V. L. Biochem. J.
1998, 330, 1159—1163.

45)Gorren, C. F. A.; Duine, J. A. Biochemistry 1994, 33 , 12202-12209.

46)Kuusk, V.; McIntire, W. S. J.Biol. Chem. 1994, 269, 26136-26143.

140

CHAPTERV

RESOLUTION ENHANCEMENT OF ECHO DETECTED EPR SPECTRUM OF
MANGANESE-ETHYLENE (DINITRILO) TETRAACETATE BY MAGNETIC
FIELD VECTOR JUMP AT X-BAND

V.l Introduction
The resolution of powder spectra in EPR can often be improved by changing the
microwave frequency and magnetic field strength where resonance occurs. Electron spin

echo detected EPR spectroscopy with magnetic field vector jumps provides an alternative

means for enhancing the resolution of powder spectra.1 In this method specific molecular
orientations are selected for resonance by inducing sudden changes in the direction of the
static magnetic field keeping its magnitude constant, during the time at which refocusing
pulses are applied in two- or three-pulse electron spin echo experiments. The resulting
ESE-detected EPR spectra show improved resolution that should allow more
straightforward analysis of magnetic coupling parameters. This method was used to
explore the possibility of resolution enhancement of the echo detected EPR Spectrum of
Manganese-ethylene (dinitrilo) tetraacetate.

Mn(II) acts as a cofactor in many enzymes and can serve as a substitute for a
required Mg(II) cofactor. 55Mn(II) (S = 5/2, I = 5/2) is a (15 ion and in the high spin state
it has five unpaired electrons. In the absence of an external magnetic field the degeneracy
of its six spin states is partially lifted by a zero-field splitting interaction to yield three
Kramer doublets denoted by MS = i5/2, i3/2, and :t1/2. The origin of this zero field
splitting interaction is the electron-electron dipolar interaction between the five unpaired

Mn(II) valence electrons. Mn(II) complexes that show non-cubic symmetry, have zfs

141

interactions that complicate their EPR spectra, making it difficult to extract structural
information. The complex chosen for this study was manganese-ethylene (dinitrilo)
tetraacetate (Mn-EDTA) complex. This complex is expected to have large anisotropy
due to zfs, which makes it a suitable system to study possible resolution enhancement by
the application of field vector jumps. In the presence of a magnetic field, 55Mn(II) gives
rise to 6 non-degenerate electron spin states that are split into 6 levels by the 55Mn(II)
hyperfine interaction. The resulting 36 energy levels support 30 allowed EPR transitions
(fine structure) according to the selections rules, AMS = i1, Am1= 0, where Ms and m; are
electron and nuclear spin quantum numbers respectively (figure V-l). The laboratory
field competes with the internal field from the Mn nucleus for alignment of spins, and
spacings of the energy levels are dependent on the orientation of the molecular axis in the
laboratory field. Thus, the EPR spectrum is orientation dependent. This orientation
dependence of resonance frequencies of the allowed transitions in the case of isotropic g,

can be expressed as:

13,6,(9) = 130 - D (Ms - 1/2) [3C0320 - 1]
+ (DZ/2B0) [43(s + 1) - 24M,(Ms - 1) - 9] sm29c0s29
4132/83.) [zs(s + 1) - 6M_,(Ms - 1) - 3] 811146
- (AZ/2B0) [I(I + 1) - mlz + m1(2Ms - 1)] + [higher order terms] Gauss
...[5.1]
where, A is the hyperfine coupling, D is the axial component of the zfs parameter
(section V-5), S and I are the effective electron and nuclear spins of Mn, and 0 is the

angle between the principal axis of the zfs tensor and the applied static magnetic field Bo.

142

 

 

 

 

 

 

’I’ ’,_—‘ ~ . . = A
’0’ ’l”
I, 0”
I, 0’
l.’ I”
" +1/2 __
:tS/Z ———-(O ,’ ', :.: -._—_-_—= ll
:1:3/2 ——*.1. .......
i 1 /2 ~ ~ ‘ ~‘~ ~ ~
“‘ ‘ ~~~~s~ -1/2 _—
.‘ °‘*-——‘=.. E
\‘ S‘
s‘ “
‘\‘ ‘\ '3/2 ,
—"'"‘ ‘ ~ - ll
\
\
\\
\
‘\
1, -5/2 ,o’
\.———(~
Zero field Electron Zeeman Nuclear Fine structure
levels Splittings splitting transitions

Figure V-l: An isotropic energy level diagram at fixed field for Mn(II)

143

DEGREES

 

 

 

 

 

 

 

 

O c‘ , e - -
l: \\ l 1’]
" L A 1 ‘ l
\ \ "\ / //
30 \4* \e \- j /
\ \ 1 ,Z'

 

 

 

 

45 i \ A

- 4 c \ I
,5 )1 1.
1 \

1.200

 

 

 

 

 

 

 

0\\

 

\

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure V-2: Angular dependence of the centers of the sextets in the five fine-structure
transitions (as shown in figure V-l) of Mn(II) in KN3. Taken from reference 2.

144

A powder spectrum can be obtained by integrating equation 5.1 over all orientations of
the magnetic field. In such a spectrum if the transition probability for a particular
transition remains constant as the angle 0 is varied, peaks are expected to occur at field
positions corresponding to "turning points" of Brcs(0). These "turning points" are
magnetic field positions where dBreS(0)/d0 goes to zero. Because the intensity I of an
EPR transition is proportional to Sin0/[dB(0)/d0], the turning points can mark positions
of maximum EPR absorption intensity.

I oc Sin0/[dB(0)/d0] [5.2]
The proportionality of I to sinO reflects very large number of systems with axes nearly
perpendicular to the field direction, and by contrast there will be very few systems with
their axes aligned close to the field direction. The angular dependence of the five fine

structure transitions for the case of Mn(II) in KN3 from earlier work of King and Miller is

shown in figure V-2.2 Orientations corresponding to the turning points, for which
dB,eS(0)/d0 = 0, show minimum susceptibility to a variation of magnetic field direction.
The features in between the turning points from one set of Ms transition may mask the
peaks corresponding to the turning points of a different transition that may be of interest
(figure V-2). So, the resolution of a multiline powder spectrum would be considerably
improved if only the features for which dB(0)/d0 = 0, can be accentuated in a magnetic

field jump experiment.

V.2 B_asic principles of Echo-detected EPR with vector fieldiumps
In a one axis field jump experiment, a dc field pulse is applied transverse to the

static field Bo during the application of the refocusing rt-pulse in a two-pulse ESEEM

145

experiment according to the pulse scheme shown in figure V-3. It results in an attenuation

of all the contributions except those from the principal axis directions. For complete

suppression, a three-pulse experiment with two transverse dc pulses is required.1 To
apply echo-detected EPR with vector field jumps there should be anisotropy either in the

electronic Zeeman interaction, the fine structure or in hyperfine interaction term.

 

I T I T
1 1,
Microwave pulses l L——

n/2 pulse 1t pulse primary echo

 

 

 

 

 

 

 

Bx tilt pulse

 

 

Figure V-3: Pulse scheme for two-pulse field jump experiment.

In order to tilt the static magnetic field by a small angle in a field jump experiment,
a transverse field Bx is applied over a time which is short compared to the electron spin
relaxation times T1 and T2, while keeping the magnitude of the static field constant. Since
the intensity I is dependent on the variation of resonant field position with 0 (equation
5.2), the effect of change in 0 on the transitions can be seen by differentiating equation 5.1
with respect to 0,

dB,,,(e)/de :- 6D(Ms-1/2)sin0c030 [5.3]

146

where the higher order terms in D and the hfi have been neglected. Only the orientations
for which dB,eS(0)/d0 = 0, show minimum susceptibility to a variation of magnetic field
direction. This condition is met for the spin systems for which 0 = 0°, 90° or M5 = 1/2;
the resonance condition does not change for small tilts and the refocusing n-pulse can
fully refocus the corresponding magnetization. For the case of M5 = 1/2, some
dependence on orientation is expected to arise due to the second order terms that were
neglected in equation 5.3. For all the other crystallite orientations the resonance
conditions would be changed depending on 0 and the tilt angle. Hence, the features
corresponding to the turning points of B(0) variation with 0 will become pronounced in

the spectrum and the contributions from all other orientations will be attenuated.

V.3 Materials and Methods
Two-pulse echo-detected EPR and ESEEM experiments were performed at X-
band with magnetic field vector jumps on frozen samples of 10 mM Mn-EDTA complex

in 50-50% water ethylene glycol mixture at 4.2K, using a Mims type microwave

transmission cavity3 with an eight-tum ENDOR coil wound around a half wave stripline

resonator. The data were collected on the pulsed-spectrometer described earlier. Current

generated by a home-built pulsed current source4 was passed through the ENDOR coil
producing the desired tilting magnetic field B, whose position in time was controlled by a
delay and gate generator. It was placed symmetrically with respect to the n-pulse (figure
V-3). The directions of the static field, Bo, the microwave field, B], and the tilt field, Bx,

were mutually perpendicular.

147

V.4 M
V.4.l Two-pulse echo detected EPR with field vector jumps

In order to test the instrumentation, two-pulse echo-detected EPR spectra for
Cu(II) in frozen aqueous-ethylene glycol solution were obtained with and without field

jumps (figure V-4). They showed resolution enhancement similar to those observed for

bis-(oxalato) Cu(II) in frozen aqueous-glycerine solution by S. Pfenninger et al.1 For Cu,
I = 3/2 and A“ hyperfine coupling of the copper nucleus gives rise to the two steps
observed in the low-field region. The splitting due to A 1 was not resolved. On
application of magnetic field jump, three distinct steps in the low-field region were
observed. In the high-field region, application of the tilt-field lead to narrowing of the g i
peak. Further experiments were performed on Mn-EDTA in frozen aqueous-ethylene
glycol solution. Use of wide microwave pulses removed any possible modulations arising
from EDTA nitrogens (figure V-5). Although the contributions from arbitrary
orientations can not be eliminated completely, enhancement in resolution was observed
with tilt-pulse generated with dc voltage > 40V as shown in figure V-6. The voltage
corresponds to the amount of current passed through the coils generating the tilt field.
The exact magnitude of the tilt field was not calculated. In two-pulse echo-detected EPR
experiment, the width and the magnitude of the dc field pulse was varied in order to find a

suitable value to perform the two-pulse ESEEM experiment with single axis field jump.

V.4.2 Two-pulse ESEEM experiments with field vector jumps
The two-pulse ESEEM spectrum with single axis field jump on Mn-EDTA

complex in frozen aqueous-ethylene glycol solution is shown in figure V-7. The results

148

 

2500

2000:
1500:
1000:

500_

 

 

20001

Echo Amplitude

1500_
1000,

500_

 

 

 

oAL..1....1....1....1....1....1....1....
2600 2700 2800 2900 3000 3100 3200 3300 3400

 

Magnetic Field (G)

Figure V-4: Echo detected EPR spectrum with field jump for Cu(II) in frozen aqueous-
ethylene glycol solution. (a) without the application of Bx pulse (b) in presence of B,(
pulse proportional to a voltage of 220V; microwave frequency, 9.80GHz; 1:, 500 ns;
microwave pulse width = 20 ns.

149

x 104
10 ' (a)

 

IITIIIIIIIIrITIIIIIT

 

 

L.
P'—
)—
h-
p—
,_.
p—
L.
1—
1—
r—
r—
r—
b-
l—
p-
p-
,—
F
P
b-
p

 

 

x104
10

0°

4:.

O\
jllllllllllllllllllllrr

Echo Amplitude

N

 

llLlllllllllllllLLlLJlllllllllllLllllll

(C)

 

 

x 10“

10

 

Ch
rljllllll‘Illrlrlllllllll

lllLlllllllllllllll44LllLlllllllllllll

0 1000 2000 3000 4000 5000 6000 7000 8000
Magnetic Field (G)

Figure V-5: Echo detected EPR spectra of Mn(II)-EDTA in frozen aqueous-ethylene
glycol solution using different microwave pulse-widths (a) 20 ns; (b) 40 ns (c) 110 ns;. 1:,

500 ns; pulse repetition rate, 60 Hz, microwave power, 40 dBm, temperature, 4.2 K. The
Bx tilt field was not applied.

 

 

150

x104

 

 

 

 

 

 

g l l l l l l l l l l i l l l l l L L l l L l l l l l l l l

E: 1000 2000 3000 4000 5000 6000 7000
< x 10‘°

{=3 1.4 — (‘3)

LL]

1.2
1.0
0.8
0.6
0.4
0.2

 

0

 

 

l L l l l L l l l J 1 l l l l 4 1 l l l l l l l l l l l

 

1000 2000 3000 4000 5000 6000 7000

Magnetic Field (G)

Figure V-6: Echo detected EPR spectrum with field jump of Mn(II)-EDTA in frozen
aqueous-ethylene glycol solution. (a) without the application of Bx pulse (b) in presence
of Bx pulse proportional to a voltage of 250V; microwave frequency, 9.80 GHz; other
conditions same as in figure V-5.

151

 

(a)

ITTITIJIIYIIIIIIIIIITTIIjT—TIIIIIJ

 

 

 

 

 

 

 

 

 

Q)
1::
.2 J l l L l l L l 1 l l I l l l l L l l l l l l l 1T1 l a;
2‘ 0 500 1000 1500 2000 2500 3000
<1 _
o _
'8 ; (b)
m :
T1..1..1.111.111.111.1.1111.1
0 500 1000 1500 2000 2500 3000
Tau (ns)

Figure V-7: Two pulse ESEEM spectra with field jump: (a) without the application of
13x pulse; (b) in presence of B. pulse proportional to a voltage of 100V applied.
Conditions: microwave frequency, 8.949 GHz; 1, 250 ns; magnetic field = 2800 G.

152

were not as expected. The decay time was further reduced in presence of field B, due to
possible increase in available relaxation mechanisms, reducing the frequency resolution

rather than improving it.

V.5 Discussion

The EPR properties of Mn(II) are governed by the distribution of unpaired
electrons around Mn nucleus, which is dependent on the type of ligands present and their
geometrical arrangement about the ion. In the free ion each 3d orbital contains a single
electron, and has a spherical charge distribution about the Mn nucleus. In the presence of
ligands, interactions between (1 electrons and electrons localized on the ligand atoms
remove the degeneracy and the spherical symmetry of the d electron distribution. In a
perfect cubic symmetry of ligand environment either octahedral or tetrahedral, the six-fold
spin degeneracy in absence of external magnetic field is lifted and creates a very slight
cubic zero field splitting. Deviation of electronic symmetry about the ion from octahedral
or tetrahedral case leads to a much larger ligand-induced quadratic zero field splitting.

This is expected to be the case in Mn-EDTA complex, due to non-cubic environment as

seen in the structure determined by analysis of three-dimensional X-ray data.5a6 Mn(II)
is in sexadentate seven-coordinated state (figure V-8) in Mn-EDTA complexes [Mn
(H20)4] - [MnH.EDTA.HzO]2.4 H20 and in MnMn(EDTA).9HzO. In Mn-EDTA
complex, the cis-nitrogen coordination results in a highly anisotropic spectrum. The EPR
spectra of randomly oriented sample of Mn(II) complexes often have features arising due
to both allowed and forbidden transitions. The spectra are best interpreted with the aid

of a spin Hamiltonian. Although the magnitude of zfs parameters D and B may not lead

153

@

/.\ 99
.—Q @’ M" $@

.__ I419 \
01- e 9

e
e
\e%g\0/

Figure V-8: Model of the structure of sexadentate seven-coordinated Mn-EDTA
complex with twofold vertical axis.

 

 

154

directly to structural information, they can be used to account for the positions of all the
features in the spectrum. The major terms in the spin Hamiltonian include the electronic
Zeeman interaction which is the dominant term at X-band; the electron nuclear hyperfine
coupling, which is usually isotropic and typically has a value of 95 G for an octahedrally
coordinated Mn(II) ion; and zfs interaction arising from the interaction of Mn in non-

spherical ligand environment.

The EPR spectra from earlier work by Reed, G. H. and Markham, G. D798 on
Mn-EDTA in solution at Q-band and at 50°C and 3°C at X-band are shown in figure V-9.
The spectrum at X-band is inhomogeneously broadened and does not show the expected
five distinct nondegenerate central electron transitions usually seen in the spectra of
Mn(II) complexes. From the Q-band spectrum, the value of the hyperfine coupling is be
found to be ~95 G. If the small quartic terms in equation [1] are omitted, the spin

Hamiltonian can be written as;
H=BBo-g-S+D [322— 1/3 S(S+1)]+E(Sx2-Sy2)+AI-S [2]

The magnitudes of D and E depend on how strongly the unpaired spins interact.
If D and E are zero then an isotropic absorption line with a g value slightly greater than 2
is usually observed.9 If D and E are finite but small (0.001-0.1 cm") five ESR transitions
are observed. If D or E is large compared to (g [3 Bo), there can be two limiting cases D i
0, E = 0 and D = 0, E at 0. The eigenvalues and eigenvectors of [2] in zero magnetic field
are then three Kramers doublets. In the first case, the lowest Kramers doublet has

effective g values g I = 2, g 1 = 6 and is experimentally observed in systems with a three

fold or higher axis of symmetry with strong axially symmetric electric fields. In the

155

 

(a)

(W

 

 

(b)

dx"/dH

 

 

 

 

(C)

 

 

l l l l l l l l l 1 l l l l l I L l l L l l l l L

 

Magnetic Field (G)

Figure V-9: X-band cw-EPR spectrum of Mn-EDTA (0.002 M) (a) at 50° C and (b) at
3° C; microwave frequency, 9.1GHz. Taken from reference 7. (c) Q-band cw-EPR
spectrum of Mn-EDTA (0.01 M) obtained at 20° C; microwave frequency, 35 GHz.
Taken from reference 8.

156

second case the middle doublet has an isotropic effective g value of 4.29. 1 O This case
occurs so long as the environment of the (15 ion is more symmetric than the over-all
symmetry group of the site, and D need not be zero but shall be small. The small D term

would have the effect of broadening the line at g = 4.29 and the ratio A = E/D, has been

calculated from the observation of such broadening in some cases.11 When both D and E
are large compared with (g [3 B0), g = 4.29 signal can arise if A is 1/3. If A = 0 represents
axial symmetry (D 75 0, E = 0), then an increase in A represents a departure towards
rhombic symmetry, A = 1/3 represents maximum possible rhombic symmetry with
equally spaced Kramers doublets and values greater than 1/3 represents a convergence
towards axial symmetry, with A = 1 representing entirely axial symmetry.

A simulation of the spectrum in figure V-6(a) using Simfonia simulation program
provided by Bruker Instruments is shown in figure V-10. In comparing features of this
theoretical spectrum with the experimental spectrum it is important to note that the
simulation program is based on the perturbation treatment and the perturbation solutions
work fairly well for Q-band spectra (35 GHz) but are more approximate for X-band
spectra because of the relative magnitudes of the Zeeman and zfs terms. Hence, for X-
band, unless the zfs is pretty small, it is difficult to fit the experimental data perfectly.
All of the fine structure transitions need not have the same relaxation time (line width)
and relaxation times may differ for 0 = 0° and 0 = 90° orientations. Additional
complexities may arise from forbidden hyperfine transitions and from rhombic

distortions. Literature data for Mn(II) compounds indicate that D generally does not

exceed 1000G in distorted octahedral environments,12 but similar information is not

available for heptacoordinated manganese ions. The simulation gave an estimate of the

157

isotropic and anisotropic zfs parameters D and E to be 1180-1160G and 265-220G
respectively giving a value of A = 0.22 indicating deviation from axial symmetry, as is
expected for a seven coordinate Mn(II).

Magnetic field positions for centers of fine structure transitions as a function of
the extent of axial distortions are given in figure V-l 1. Diagrams representing solutions to

the Hamiltonian in equation [1] for both axial and rhombic symmetries over a wide range

of zfs are available. 1 3 The values of zfs parameters obtained by simulation fall in the
range of that obtained from this diagram. These values of the zfs parameters also fall in
the range estimated for the case of heptacoordinated site of Mn(II) doped into the

diamagnetic compound MgMg(EDTA).9HzO where Mn occupies a heptacoordinated

site.6

Figure V-6(b) shows the echo detected EPR spectrum obtained in the presence of
tilting field B,. An improvement in the resolution is obtained because features other than
those from the principal axis directions are suppressed. As seen in the figure V-3 central
transition has minimum dependence on 0 so, it is expected to contribute maximum to the
spectrum of figure V-6(b).

Two-pulse ESEEM spectrum were obtained with single axis field jumps on the
Mn—EDTA complex in order to extract information about the hyperfine couplings. The
decreased resolution observed could be due to the fact that the enhancement in resolution
achieved from suppression of contributions from non principal axis orientations was
overcome by decrease of the transverse relaxation time, T2. Transverse relaxation arising
from local magnetic field inhomogenities due to coupling between magnetic dipoles varies

as a function of (1 - c0330)/r3, where r is the distance between spins and 0 is the angle

158

 

 

 

 

1.5-
.1.
1.0-W
0.5- .
4 N
.....-..-.--f-,-.-,..-,-e-.f....,....,
0 1000 2000 3000 4000 5000 6000 7000 8000

[C]

Figure V-10: Simulation using Simfonia simulation program provided by Bruker

Instruments. Simulation parameters: D = 11806, E = 265G, Microwave frequency,
9.104GH2, isotropic g va1ue, 1.946.

159

01

i
:
O
I
:
:
I
:
:
l
:
1
l
i
l
O
l
'c

 

 

l200 . . . .
1260 2260 3260 4260 5260

Figure V-ll: Magnetic field positions of the fine structure transitions in a powder
spectrum of a ”Mn(II) complex with axial symmetry as a function of zfs at X band. D is
assumed to be negative. Unprimed letters indicate the transitions for 0 = 0°; primed for 0
= 90° and C" is for the central transition at 0 = 41°48'. A, 50 <—-) 3/2; B, 3/2 H 1/2; C, 1/2
<—> -1/2; D, -3/2 H -l/2; and E, -S/2 <—-) -3/2. Free electron resonance position, g = 2.0023,
coincides with C. Taken from reference 12.

160

between external field and the symmetry axis of the system. As seen in figure V-3 the
pulse B, is on for a much longer time as compared to mw-pulse B1. At the end of the
refocusing pulse, the magnetizations forming the primary echo, experience field B, for
some time before it is turned off. The presence of an additional field B,, perpendicular to
Bo, may lead to an increase in local field fluctuations and reduction of the ESEEM decay

time.

V.6 Conclusions

Using the magnetic field jump method, some enhancement in the ESE-detected
EPR spectrum of Mn-EDTA was observed. It is also important to obtain a symmetrical
tilt pulse and to be able to place the B,( pulse symmetrically with respect to the
refocusing n—pulse, this would allow refocusing of the magnetization that is defocused due
to the inhomogeneities in the Bx field. To provide the desired enhancement for ESEEM

spectra, probe structures that yield more homogeneous pulsed magnetic fields are needed.

161

List of References

1)Pfenninger, S.; Schweiger, A.; Forrer, J.; Ernst, R. R. Chem. Phys. Lett. 1988, 151, 199-
204.

2)King, G. J.; Miller, B. S. J. Chem. Phys. 1964, 41, 28-40.

3)Davis, J. L.; Mims, W. B. Rev. Sci. Instrum. 1978, 49, 1095-1097.

4)Forrer, J .; Pfenninger, 8.; Sierra, G.; Jeschke, G.; Schweiger, A.; Wagner, B.; Weiland, T.
App]. magn. Reson. 1996, 10, 263-279.

5)Richards, S.; Pedersen, B.; Silverton, J. V.; Hoard, J. L. Inorganic Chemistry 1964, 3,
27-33.

6)Borras-Almenar, J. J .; Burriel, R.; Coronado, E.; Gatteschi, D.; Gomez-Garcia, C. J .;
Zanchini, C. Inorg. Chem. 1991, 30, 947-950.

7)Reed, G. H.; Leigh, J. S. 1.; Pearson, J. E. J. Chem. Phys. 1971, 55, 3311-3316.

8)Reed, G. H.; Markham, G. D. Biological Magnetic Resonance; Berliner, L. J. and
Reuben, J ., Ed.; Plenum: New York, 1984; Vol. 6, pp 73-142.

9)Orton, J. W. Rep. Progr. Phys. 1959, 22, 204.

10)Castner, T.; Newell, G. S.; Holton, W. C.; Slichter, C. P. .1. Chem. Phys. 1960, 32, 668.
11)Wickman, H. H.; Klein, M. P.; Shirley, D. A. J. Chem. Phys. 1965, 42, 2113.
12)Carlin, R. L. Magnetochemistiy; Springer-Verlag: Berlin, Heidelberg, 1986.
l3)Aasa, R. J. Chem. Phys. 1970, 52, 3919.

162

1111111111111111