lllllllllllHill““Hill"”ll”Hill“lllllllllllllllllllllll
3 1293 01563 5877

This is to certify that the

dissertation entitled

Pulsed Electron Paramagnetic Resonance Studies on
Biologically Relevant Ni and Tyrosyl Models

presented by

Gyorgy Filep

has been accepted towards fulfillment
of the requirements for

Ph.D. degfieh, Chemistry

L W?

Major professor

DMe 9 May 1997

042771

 

 

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due. '

1 DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
c:\cIrc\dwedm.pm3-p.1

 

 

PULSED ELECTRON PARAMAGNETIC RESONANCE STUDIES ON

BIOLOGICALLY RELEVANT Ni AND TYROSYL MODELS

By

Gyorgy Filep

A DISSERTATION
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

1997

 

 

 

ABSTRACT

PULSED ELECTRON PARAMAGNETIC RESONANCE STUDIES ON

BIOLOGICALLY RELEVANT Ni AND TYROSYL MODELS

By

Gyorgy Filep

Pulsed EPR methods based on Electron Spin Echo (ESE) phenomena,
such as Electron Spin Echo Modulation (ESEEM) spectroscopy, are capable of
measuring weak hyperfine interactions (of the order of a tenth of a Gauss). In
order to expand the capabilities of ESEEM, its 2D variant, Hyperfine Sublevel
Correlation Spectroscopy (HYSCORE) was proposed in 1986. Our work has
been aimed at exploring the application of HYSCORE in disordered samples
of biologically relevant model systems.

The nitrogen ligand hyperfine couplings of Ni(III)(CN)4(I-IZO)2‘ have
been measured using a combination of isotopic substitution, orientation
selection, and 1- and 2-dimensional Electron Spin Echo Envelope Modulation

techniques. The shapes of the contour lines obtained from HYSCORE
experiments were analyzed, using the graphical method developed by
Dikanov and Bowman (J. Magn. Reson. 1995, Series A 116, 125.) for samples

prepared with C15N‘. The results show an axially symmetric hyperfine

 

 

Z

1,

 

 

interaction with IA“ I=1.93 MHz and IAi I =1.06 MHz (for 15N). The cyanide
14N nuclear quadrupole coupling is characterized by a quadrupole coupling
constant, ezqzzQ= 3.67 MHz and an asymmetry parameter, n=0.09, with the
principal axis of the NQI tensor being along the C-N bond.

The very slow relaxation of tyrosyl radicals at low temperatures makes
pulsed EPR experiments on them rather difficult to perform, because usually
we cannot use sampling rate higher than ca. 10 Hz. We have remedied this
problem by adding 5 mM (3d3+ to the sample. HY SCORE data were analyzed by
the graphical method of Dikanov and Bowman mentioned above. The
principal values of the HFI tensor are found to be (-O.8, -3.1, —3.6) (MHz),
which are reasonably close to the values found by Warncke and McCracken
using 1D ESEEM (I. Chem. Phys. 1994, 101, 1832).

The electronic structure of Ni(III)(H_ZGB)(terpy) has been studied using
1D ESEEM combined with isotopic substitution. To sort out the contributions
of the various nuclei, specific isotopic labeling (with 15N) has been used at the
middle pyridine ring. The ”product rule” of ESEEM allows one to observe the
modulation from this nucleus by dividing the data from the non-labeled
sample by that from the 15N-labeled one. The results show no magnetic
interaction from this nitrogen. This indicates very low unpaired electron spin

density, which may suggest that this nitrogen does not coordinate at all, in

contrast to the geometry proposed previously.

 

 

 

 

 

Draga Sziileimnek

This Dissertation is dedicated to my beloved parents,

 

Téth Réza and Dr. Filep Gyb‘rgy

 

 

 

Acknowledgment

First of all, I would like to express my greatest respect toward my
mentor and teacher, prof. John McCracken, who has given me excellent
professional guidance and permanent support during my years at MSU.

I also thank my colleagues, Hong-In Lee, Kurt Warncke, Michelle Mac,
Kerry Reidy and Vladimir Bouchev, for their help, support and numerous
valuable and open discussions. Special thanks to Andrew Ichimura for his

friendship and for being a great colleuague and tennis partner.

Special thanks to some of the members of the community of Eastern
European students for their friendship and the fun we had together.
Especially to Vladimir F. Bouchev (Vlado), Evstatin T. Krastev (Ati),

Branislav Blagojevic (Bane), Marian G. Marinov (Chado), Alexander S. Volya
(Sasha). I feel privileged to know all of you. Wish you a happy life and see
you back home !

My deepest respect goes to my friend, Dalila G. Kovacs, who always
stood by me, mostly in difficult times. I thank my good friend Szasz Barna for
his care and understanding. I also thank my former wife, Gabriella Székely,
for her support.

I could never have finished this long journey without my parents’

continuous support and faith in me. Thank You, Mom and Dad, for your

endless love.

 

 

3.4

 

 

 

TABLE OF CONTENTS

LIST OF TABLES

LIST OF FIGURES

LIST OF ABBREVIATIONS
INTRODUCTION

Chapter 1:

Chapter 11:

Chapter 111:

Principles of Electron Spin Echo
Spectroscopy

1.1. Introduction

1.2. Formation of two-pulse ESE

1.3. Three—pulse echo (stimulated echo)
1.4. Four-pulse echo

1.5. Semiclassical description of ESEEM
1.6. Two—pulse ESEEM

1.7. Three-pulse ESEEM

1.8. Four-pulse ESEEM

19. Density matrix description of ESEEM
References

HYSCORE

11.1. Introduction to HYSCORE

11.2. HYSCORE data analysis

11.3. Experimental aspects

11.4. HYSCORE on disordered samples
References

Ni-containing Hydrogenases

111.1. Introduction
111.2. Evidence for the Involvement
of Ni in Hydrogenases
111.3. Structural Components of
Hydrogenases - General Model
111.4. Hydrogenase Activity
111.5. EPR Spectra from Hydrogenases
111.6. Coordination of the Ni-site
111.7. Interaction Between Nickel
and [Fe-S] Clusters
111.8. Redox Properties of Hydrogenases
111.9. The Role of Ni as a Binding Site
111.10. Summary

vi

10
12
15
20
21
22
23
27

29

34
39
45
48
59

61

61

62

62
64
65
69

73
73
80
82

 

Chapter IV:

Chapter V:

Chapter VI:

Chapter VII:

 

References

Synthetic Models for the
Ni—site of Hydrogenases

1V.1. Introduction

IV.2. Polynuclear Ni—complexes

1V.3. Mononuclear Ni-complexes

1V.4. Complexes Based on Tetracyano—
nickelate(111)

1V.5. Monopeptide Complexes of Ni(111)

1V.6. Bis(peptido) Complexes-of
nickel(111)

1V.7. Ni(111) Mixed Ligand Complexes

References

ESEEM Study on
Ni(III) (triglycinate)(terpy)

V.1. Abstract

V.2. Theory of EPR of Ni(111) Complexes
V.3. Sample Preparation

V.4. Results and Discussion

References

Structural Characterization of
Bis(aquo)tetracyano-nickelate(III),
Using One- and Two-dimensional
Pulsed EPR Methods

V1.1. Abstract

V1.2. Introduction

V1.3. Materials and Methods
V1.4. Results and Discussion
V1.5. Conclusion
References

HYSCORE Study on a Tyrosyl
Model System

V11.1. Abstract

V112. Introduction

V113. Functional and Structural
Properties of Radical Enzymes

V114. Example: Galactose Oxidase (GOase)

vii

85

89

89
9O
94

97
99

101
103
108

110

110
111
119
120
133

134

134
135
137
140
153
155

157

157
158

160
161

 

 

V11.5. EPR Spectroscopic Studies of

Tyrosyl Radicals 162

V11.6. Sample Preparation 168

V11.7. Results and Discussion 169
References 187

Appendix 190

 

 

viii

 

 

LIST OF TABLES

TABLE HI-1 Redox schemes proposed for the Ni-Fe cluster in
hydrogenases. 78

TABLE VII-1 Hyperfine tensors of tyrosyl radicals. 165

 

' at

 

Figure 1-1

Figure I-2

Figure 1-3

Figure 1-4

Figure 1-5

LIST OF FIGURES

A linearly polarized oscillating magnetic field, Bx(t) can

be considered as the vectorial sum of two circularly
polarized fields that are rotating in opposite directions,
namely, clockwise (BL) and counter-clockwise (BR). 7

Vector diagram to explain the formation of the

2-pulse spin echo. (a) the pulse sequence;

(b)~(c) the first pulse tilts the magnetization (M) onto

the (xy) plane; (d) the spin packets that comprise lVI
dephase due to their different precession frequencies;

(e) the second pulse inverts each spin packet around

the x—axis; they continue to precess in the (xy) plane;

(f) they refocus along y and reform M [from Ref. (8)] 9

The inhomogeneously broadened EPR lineshape.

(a) Bo, avg is the mean value of the resonant magnetic

field, ABo is the linewidth of the portion of the

spectrum excited by a pulse. (b) the magnified View of the
portion excited by the pulse, with the resonant lines
corresponding to individual spin packets. (00, (0,, 0)].

are the precession frequencies of the spin packet
magnetization components. 11

Vector model to explain the formation of the 3—pulse echo.
(a) the pulse sequence; (b), (c) same events as in Figure 1—1.
((31) the second pulse torques the magnetization along 2;

(e) the magnetization components precess about 2;

(f, g) the surviving magnetization components

(directed along 2) are flipped onto the (xy) plane by

the third pulse; (h) the spin packets refocus along y,

forming the echo [from Ref. (8)]. 13

Illustration of the formation of the 4—pulse spin echo.
(a) Position of spin packet 8MZ after the second pulse;
(b) the inverting rt—pulse reorients 8M2;

(c) the third pulse torques 5MZ onto the (xy) plane;

(d) spin packets start precess about 2; (e) by the time t
after the last pulse various spin packets assume
the position shown, which creates the echo along

 

«.i

 

Figure I-6

Figure I-7

Figure II-1

Figure II-2

Figure II-3

Figure 11-4

the —y axis. 14

Energy level diagram of an S=1 / 2, 1:1 / 2 spin system
in the solid state, described by the Hamiltonian
of Eq. [1-3]. 17

Schematic illustration of the origin of spin echo
modulation. The upper left insert shows the scheme

of the four~level energy level diagram of Fig. 1—5,

while the upper right insert shows the spin packets

in the inhomogeneous line. A), B), C) show

the branching of the transitions, which leads to

the modulation phenomenon (see text). 18

Scheme of obtaining a HYSCORE contour plot.

(21) Measurement of the amplitude of the 4—pulse echo

as a function of t2 and t1. (b) The resulting 2D

time-domain data matrix, with only 33 of all the

128 tl-slices shown. (c) 3D representation of the frequency—
domain data (HYSCORE magnitude spectrum) obtained

by 2D Fast Fourier Transformation of the time domain

data matrix and taking the absolute value of the resulting
complex matrix. (d) Contour plot of

the above spectrum. 31

Schematic of the expected peaks in HYSCORE
spectrum of 8:1 / 2, 1:1 / 2 spin system [from Ref. (4)]. 38

Computer simulated spin echoes generated by

a sequence of four non—ideal pulses

(1:400 ns, t1=2us, t2=1.1us); (a) without phase-cycling;
(b) with phase-cycling according to the cycling scheme

of Table 11—1 [from Ref. (4)]. 38

Graphical representation of Eq. [II—12], calculated
with the following parameters: coA/27t=2.00 MHz,
mP/21t=1.33 MHZ, (DD/21t22.66 MHz.

(a) The time domain data matrix, S(t1, t2), with only
13 of the 128 slices shown. (b) (+, +) quadrant of

the absolute value spectrum, S(0)1, (02),

xi

 

 

(c) (—, +) quadrant of 8(0)], (02). 41

Figure 11-5 HYSCORE absolute value spectrum of irradiated
malonic acid [from Ref. (10)]. 46

Figure II-6 HY SCORE absolute value spectrum of irradiated
succinic acid. (a) Width of the it / 2 pulses 10 ns,
that of the it-pulse 20 ns; (b) width of it / 2 pulses 10 ns,
that of rt—pulse 14 ns (power of the n-pulse is +3dB
higher than that of it / 2 pulses) [from Ref. (10)]. 49

Figure 11-7 (a) Simulated powder ENDOR spectrum for
an S=1 / 2, 1:1 / 2 system. Parameters: gn=2.261 (31F),
AI l=7.1 MHZ, A 1:24 MHZ, B0=3200 G.
(b) The corresponding simulated 3—pulse
ESEEM spectrum. 50

 

Figure 11-8 Schematic HYSCORE contour plots arising from
an S=1 / 2, 1:1 / 2 spin system with a small anisotropic
hyperfine interaction. (a) (01> Al I / 2, (b) 031< Al I/ 2. 52

Figure II-9 Simulated HYSCORE contour plots arising from
axial HFI from an 521 / 2, 1:1 / 2 system. Parameters:
031/ 27t=3 MHZ, A6022 MHZ, Gaussian linewidth:0.1 MHZ,
T=O.4 MHz (a), T208 MHZ (b), T212 MHZ (c),

T=1.6 MHZ (d). 53

Figure II-10 Simulated HYSCORE contour plot arising from a
rhombic hyperfine interaction in an 821/ 2, 1:1/ 2

spin system. Parameters: 031/ 271523 MHZ, A. :2 MHZ,

ISO

T=1.2 MHZ, 8:0.7, Gaussian linewidth=0.1 MHZ. 58

Figure III-1 Schematic general structure of (Ni, Fe) hydrogenases
[from Ref. (13)]. 63
Figure III-2 X-band EPR spectra of D. gigas hydrogenase at
different stages of incubation under H2. Spectra taken
at 77 K. A: Native enzyme. B—E: Evolution of

 

Figure 111-3

Figure III-4

Figure III-5

Figure III-6

Figure IV-1

Figure IV-2

Figure IV-3

Figure IV-4

Figure IV-5

the EPR spectra upon increasing the time of
incubation under H2. [from Ref. (17)]. 66

Various X-band EPR spectra displayed by the
D. gigas hydrogenase. (a) Ni-A at 105 K;

(b) Ni-B at 105 K; (c) Ni-C at 30 K;

(d) as (c) recorded at 8 K; (e) as (c) after
illumination, at 30 K; (f) reduced with H2 and

treated with CO, 105 K. [from Ref. (17)]. 67
The structure of the Ni-Fe cluster in the D. gigas
hydrogenase. 71

EPR redox titrations of the D. gigas hydrogenase.
Full squares: g=2.02 signal, 4 K; open squares:
Ni—A signal, 77 K; full circles: Ni—C signal, 20 K;

 

open circles: Ni-C signal, 4 K. 75
Scheme describing the redox chemistry associated

with the Ni—Fe cluster in hydrogenases. 76
Schematic structures of model complexes referred to

in the text. A: [{Ni(BME—DACO)FeCl}2(u—Cl)2], where
(BME—DACO)IrI2 = N,N’-bis(mercaptoethyl)—1,5-diaza—
cyclooctane; B: [{Ni(dmpn)}3Fe]2+, where (dmpn)H2 = N,N’—
dimethyl—N,N’-bis(2—mercaptoethyl)—1,3—diaminopropane;
C: [Ni2(memta)2], where (memta)H2 =

(HSC2H4)2NC2H4SMe; D: [Ni2{P(2—SC6H4)3}2]';

E: [Ni2(SC4H9)6]2’ ; F: [Ni2(S-2,4,5-iPr3C6H2)5]‘.

From Ref. (3). 91

An example of an alkyl-thiolate Ni(11) complex

which undergoes S—based oxidation. Electrochemical
oxidation of 21 affords a Ni(11) thiyl radical. Chemical
oxidation with 12 yields 22. Oxidation of the monomeric
cyanide derivative of 21 with O2 leads to 23

[from Ref. (8)].

93
Summary of redox properties of synthetic
tetrathiolate and thiolate/amidate Ni—complexes 96
Structure of [Ni(111)(CN)4L2]'. 98
General structure of Ni(111) monopeptide complexes. 100

xiii

'EJ“

 

Figure IV-6

Figure IV-7

Figure IV-8

Figure V-1

Figure V-2

Figure V-3

Figure V-4
Figure V-5
Figure V-6

Figure V-7

Figure V-8

Figure VI-1

Figure VI-2

Structure of [Ni(111)(H_2GG)2]‘.

EPR spectra of the monopeptide Ni(111) complex,
[Ni(111)(H_2GAG)(H;,_O)2]+ (a), and of its bipy (b)
and terpy (c) adducts.

Structure of [Ni(111)(H_2Aib3)(HZO)(CN)]' (A)

- and [Ni(111)(H_2Aib3)(CN)2]' (B);

where Aib=aminoisobutyrate.

Energy level diagram for d-orbitals in various
ligand fields [from Ref. (8)].

Oblate (a) and prolate (b) shape of quadrupolar
nuclei. Schematic representation of the interaction
of the quadrupolar nucleus in a uniform electric
field (c), and in a field gradient (d).

Energy level diagram for S=1/ 2, 1:1/ 2 spin system
at ”exact cancellation”.

The synthesis of terpyridine.
CW-EPR spectrum of Ni(111)(H_2G3)(terpy).
Proposed structure of Ni(111)(H_2G3)(terpy).

Three—pulse ESEEM on Ni(111)(H_2G3)(terpy):

traces (a), (c) and (e) are from the non—labeled complex,
traces (b), (d) and (f) are from complex containing 15N
in the middle pyridine ring. (a), (b) taken at g=2.175;
(c), (d) at g=2.100; (e), (f) at g=2.014.

ESEEM functions obtained after dividing
the all-”N traces by the corresponding 15N traces;
(a) g=2.175; (b) g:2.100; (c) g=2.014.

CW-EPR spectrum of Ni(111)(CN)4(HzO)2'.

(a) HYSCORE contour plot obtained from the g i region

of the Ni(111)(C15N)4(HZO)Z' EPR spectrum. (b) Plot
of the square of the correlating frequencies of the six

102

104

106

113

116

118

121

122

123

126

130

141

points on the ridge, marked by +’s. The slope and intercept

of the resulting straight line are explicit functions

xiv

 

 

 

Figure VI-3

Figure VI-4

of the HFI parameters, A150 and T. 144

(a) HYSCORE contour plot obtained at the g L edge

of the EPR spectrum of Ni(111)(C15N)4(H20)2‘.
Experimental conditions: resonance frequency,

v0: 8.862 GHZ; static magnetic field, Bo=2890 G;

1:200 ns; 128x128 echo amplitudes were collected;

time increment was 50 ns (b) The result of the
corresponding numerical simulation. Simulation
parameters: Larmor frequency of 15N, vL=1.25 MHZ;
isotropic HFI constant, AgQ=135 MHZ; anisotropic

HFI constant, T=+0.29 MHZ corresponding to an effective
dipole-dipole distance, reff=3.10 A (Aiso=1.65 MHZ with
T=-0.29 MHZ gives the same result); Gaussian linewidth,
0.12 MHZ. (c) HYSCORE contour plot at gl ,. v0=8.905 GHZ;
3023170 G; 1:200 ns; size of data matrix, 128x128;

time increment between pulses, 50 ns. (c) Simulation
using vL=1.35 MHZ; the other parameters are the

same as in (b). 146

 

(a) 3-pulse ESEEM spectra of Ni(111)(CN)4(HzO)2' at g i.
Experimental conditions: Bo=2866 G; 12164 ns;

v0=8.838 GHZ; (b) Result of the corresponding numerical
simulation using the following parameters: AiSO=0.90 MHZ
(scaled for 14N from Ai50=1.35 MHz of 15N by the ratio of the
nuclear g-values); reff=3.10 A; polar and azimuthal angles
specifying the orientation of the principal axis of the axial
HFI tensor with respect to the g, I axis, 1t/ 2, 0;

E32un=367 MHZ; 1120.09; Euler angles that rotate the
principal axis system (PAS) of the NQI tensor into the

PAS of the g—tensor, 0, TC / 2, TC / 2. (c) 3—pulse ESEEM

spectra of Ni(111)(CN)4(1-IZO)2'at gI 1. Experimental conditions:
Bo=3l35 G; v0=8.838 GHZ; 1:149 ns. (d) Simulated spectrum.
Simulation parameters for HFI and NQI are

the same as for (b). 150
Figure VII-1 Summary of various properties of metallo-radical

enzymes. 159
Figure VII-2 The active site of Galactose Oxidase. 163
Figure VII-3. A proposed catalytic cycle of Galactose Oxidase. 163

XV

.3
l‘
ligl

 
 

.r...

 

 

Figure VII-4 Spin—density distribution of tyrosyl radicals. 167

Figure VII-5 CW-EPR spectrum of the 3,5 2H—tyrosyl radical
generated by UV irradiation in a 12 M LiCl frozen
solution, with 5 mM Gd3+ added. 170

Figure VII-6 Spin—echo detected EPR of the 3,5 2H-tyrosyl sample,
containing 5 mM Gd“. 171

Figure VII-7 (a) Time-domain HYSCORE data (128x128 data points)
from 3,5 2H-tyrosyl, with 12200 ns; (b) the corresponding
HYSCORE frequency—domain contour plot. B0=3129 G,
v=8.910 GHZ. 172

Figure VII-8 Definition of the position of the external
magnetic field vector (B0) in the Principal Axis System
of the (rhombic) hyperfine tensor. 173

Figure VII-9 3-pulse ESEEM spectra of 3,5 2H—tyrosyl radical with
the corresponding simulations; (a) experimental, 1:293 ns.
(b) simulated; (c) experimental, 1:438 ns. (d) simulated.
Simulation parameters: the principal values of the HFl
tensor (MHZ), AXX=O.8, Ayy=3.1, AZZ=3.6; nuclear Larmor
frequency, VL=2.1 MHZ. 176

Figure VII-10 (a) HYSCORE contour plot of 3,5 2H-tyrosyl (same as
Figure VI—7b); (b) the corresponding frequency—domain
simulation performed with the MATLAB program
”hyslineRom.m” using the following parameters:
nuclear Larmor frequency, VL=2.1 MHZ; A150=2.6 MHZ,

T=0.8O MHZ, asymmetry parameter of HFl tensor,
620.53; tau-value, 1:200 ns. 179

Figure VII-11 The t-suppression effect in our experiments.
HYSCORE spectra of the 3,5 2H-tyrosyl radical taken
at various values of t: (a) 200 ns; (b) 300 ns; (c) 400 ns;
(d) 500 ns; (e) 600 ns. 183

 

 

 

 

 

LIST OF ABBREVIATIONS

 

Aib ...... Aminoisobutyrate

bipy ...... Bipyridine

CW ...... Continuous wave

C. Vinosum ...... Chromatium Vinosum
(bacterium)

1D ...... One dimensional

2D ...... Two dimensional

D. gigas ...... Desulfovibrio gigas (bacterium)

dien ...... Diethylene triamine

DEFENCE ...... Dead-time free ESEEM by
nuclear coherence transfer
echoes

en ...... Ethylene diamine

ENDOR ...... Electron nuclear double
resonance

EPR ...... Electron paramagnetic
resonance

ESE ...... Electron spin echo

ESEEM ...... Electon spin echo envelope
modulation

EXAFS ...... Extended x-ray absorption

fine structure

FORTRAN ...... Formula translator (program-
ming language)

FT ...... Fourier transform

FW ...... Formula Weight

G3 ...... Triglycine

GOase ...... Galactose Oxidase

H4EDTA ...... Ethylene diamine tetra-
acetic acid

HFI ...... Hyperfine interaction

H-203 ...... trigyline with deprotonated
amide groups

HYSCORE ...... Hyperfine sublevel correlation
spectroscopy

MATLAB ...... Matrix laboratory (program-
ming and visualization
software package)

M W ...... Microwave

NAD ...... Nicotinamide adenine
dinucleotide

NHE ...... Normal hydrogen electrode

NMR

...... Nuclear magnetic resonance

. xvii

 

 

 

 

NQI ...... Nuclear quadrupole

interaction
PAS ...... Principal Axis System
PGHS ...... Prostaglandine H
Synthase
PS 11 ...... Photosystem II
RNR ...... Ribonucleotide Reductase
SCE ...... Saturated calomel electrode
terpy ...... 2,2':6’,2" terpyridine
XANES ...... X-ray absorption near edge
spectroscopy

XAS ...... X—ray absorption spectroscopy

 

 

xviii

 

 

 

INTRODUCTION

Electron Spin Echo Envelope Modulation (ESEEM) spectroscopy has
become a valuable tool to study weak electron-nuclear hyperfine interactions
(HFI) in paramagnetic samples. The analysis of ESEEM in randomly oriented
samples is complicated by the fact that the lineshape is determined not only
by the principal values of the g- and HFI— tensors but also by the orientation-
dependent amplitude factorak.

Chapter I of this Dissertation covers the basic principles of Electron
Spin Echo (ESE) methods, emphasizing the main characteristics, advantages
and drawbacks of various one-dimensional ESEEM methods.

To facilitate the interpretation of ESEEM spectra, two-dimensional (2D)
versions of the method have recently been introduced. The most widely used
2D ESEEM is Hyperfine Sublevel Correlation Spectroscopy (HYSCORE), which
allows one to obtain the principal HFI tensor components via direct analysis
of the shape of the 2D correlation patterns rather than via numerical spectral
simulations.

Chapter II is devoted to the description of HYSCORE. It presents the
principles of the method, followed by general features of 2D data analysis.
Then the experimental aspects of HYSCORE are detailed, with an emphasis
on disordered systems.

Electronic structure of Ni(III) complexes has been of interest since the

discovery of the Ni EPR signal in hydrogenase enzymes. Analysis of the weak

 

 

 

 

HFI between the unpaired electron of the Ni(III) and the magnetic nuclei of
the ligands yields structural information on the Ni center.

Chapter III is a brief review of the relevant properties of hydrogenase
enzymes. It focuses on electrochemical and spectroscopic (mainly EPR)
results. Questions concerning the coordination of the Ni site in hydrogenases
are discussed, using mainly the D. gigas hydrogenase as a prototype.

Chapter IV summarizes the literature on the most relevant Ni
compounds that have been proposed as structural or functional models for
the hydrogenase nickel site.

Chapter V first gives an introduction to the EPR spectroscopy of Ni(III)
compounds. It explains the connection between structure and quantities
measured by EPR, such as g—values, HFI parameters, Nuclear Quadrupole
Interaction (NQI) parameters. The second part presents our study on a Ni(III)
complex, Ni(III)(triglycinate)(terpyridine). The results indicate an unusual
electron distribution around the Ni center.

Chapter VI presents a HYSCORE study on another Ni(III) model
compound, tetracyano nickelate (111). It demonstrates that important
hyperfine information can be obtained fairly easily with the aid of this 2D
method.

Amino acid radicals have recently been discovered as essential
participants in the catalytic mechanisms of some enzymes. Tyrosyl radicals
have been intensively studied by EPR, Electron Nuclear Double Resonance

(ENDOR) and ESEEM spectroscopies.

 

 

 

 

 

 

 

Chapter VII gives a short introduction to metallo-radical enzymes,
including the review of previous EPR investigations. This is followed by the
presentation of our HY SCORE study on a tyrosyl model system. It
demonstrates (for the first time) that HYSCORE can effectively be applied to

systems with rhombic hyperfine tensor.

 

 

 

 

 

 

 

Chapter I

Principles of Electron Spin Echo Spectroscopy

 

LLJMLQClmLiQn

Electron Paramagnetic Resonance (EPR) spectroscopy has established itself as
an indispensable method in studies of the structure of paramagnetic materials
(1-3). EPR in a liquid phase is able to measure isotropic hyperfine interaction
(HFI) of magnitude as low as 0.1 G. For paramagnetic species stabilized in
disordered solid matrices, i.e. polycrystals, glasses and frozen solutions, the
resolution of EPR is reduced to a few Gauss, owing to inhomogeneous

spectral broadening, which masks the details of EPR spectra. In these cases the
hyperfine information can be extracted only if the splittings are larger than
the inhomogeneous width (usually 10-100 G). To solve structural problems,
however, it is often necessary to obtain information on weak HFI.

, In order to make EPR applicable in these cases, various modifications
of the method have been proposed. The most widely used of them is Electron
Nuclear Double Resonance (ENDOR) spectroscopy (2, 3). ENDOR yields the
NMR Spectrum of the paramagnetic system by detecting the change in the
amplitude of the saturated EPR signal of the sample upon sweeping the
frequency of an additional radiofrequency field. This method can often
provide HFI and Nuclear Quadrupole Interaction (NQI) parameters of the

order of a few Hertz.

 

 

 

 

 

u.
”1

 

 

 

Other developments in EPR aimed at resolving small hyperfine splittings in
solid samples are the variety of pulsed EPR methods, which are primarily
based on Electron Spin Echo (ESE) phenomena (4-8, 12). As compared to
Continuous Wave (CW) EPR, which is a frequency domain spectroscopy, ESE

is a time domain method, in which the evolution of the spin system is
directly recorded. In ESE experiments the spins are probed by short (z 10—100

ns), high energy resonant pulses, with the static magnetic field being kept
constant. Therefore, only a small portion of the inhomogeneously broadened
EPR spectrum is excited in the ESE experiment. The response of the
paramagnetic sample to these pulses is a spontaneous emission of microwave
energy'called the spin echo, whose magnitude as a function of separation
between two of the pulses is measured. The weak hyperfine interactions may
manifest themselves as a modulation of the spin echo envelope (Electron
Spin Echo Envelope Modulation, ESEEM). The periods of these modulations
are related to the nuclear transition (NMR) frequencies of the sample.

In this Chapter a qualitative introduction of the ESE phenomenon is
given. Vector models explaining the formation of two, three and four pulse
echoes are presented, followed by a semiclassical picture of the ESEEM

phenomenon.

1.2. Formation of Two—pulse ESE

The simplest pulse sequence to generate a spin echo consists of two

consecutive pulses. Two pulse echoes produced by nuclear spins first were

 

 

 

 

hi ‘

observed by Hahn in 1950, using the pulse sequence TC/ 2-‘C --1t/ 2 (9), where 712/ 2

is the turning angle of the pulse. Later, an improved sequence (1c / 2-1 -7t) was

introduced by Carr and Purcell (10). The formation of the two pulse echo will
be explained by a simple magnetization vector model.

When the paramagnetic material is placed in a static external magnetic field
I30, the magnetic moments of the electron spins will form precession cones
either parallel or antiparallel with respect to E0. For electrons the [3 spin state
has the lower energy, therefore, the higher population (according to

Boltzmann’s Law). This results in a total macroscopic magnetization (M),

 

parallel to E0. During an ESE experiment a linearly polarized oscillating

 

microwave electromagnetic field is applied, whose magnetic component is
perpendicular to I30 and can be formulated as Ex(t)=2filcos((ot). The effect of
this field is most conveniently analyzed by decomposing it into two circularly
polarized components that rotate opposite to one another, denoted ER and
EL(Figure I-1). They can be expressed as

ER=B1[icos(wt)+jsin(cot)] [I-1a]

EL=B1[icos((ot)-jsin(u)t)]. [I-1b]
We notice that under resonance conditions ER will rotate at approximately

the angular velocity of the precession of the electron spins, while EL will
rotate in the Opposite direction and will have little effect on the resonance

8Xperiment, thus will be neglected for further discussions.

 

h;

 

 

 

 

 

 

   

Figure I-1. A linearly polarized oscillating magnetic field, Bx(t) can be
considered as the vectorial sum of two circularly polarized fields that are

rotating in opposite directions, namely, clockwise (BL) and counter—clockwise

(BR)-

 

 
 
 
 

 

‘1
sq

 

 

 

Let us adopt a reference frame, rotating about the laboratory axis 2 (the

direction of the static magnetic field) at angular velocity 0) that is equal to the
carrier frequency of the pulse. In this frame BR appears to be stationary and
we may choose it to lie along the x axis. When the microwave (MW) pulse is
applied for a period of time tp, the B1 field torques the magnetization M about

the x-axis by an angle @mp,

 

GmngeBltpzwltp' [1‘2]

where ge is the gyromagnetic ratio of the electron. We choose tp so that

 

(imp=900 for the first, 1800 for the second pulse. The evolution of the spins
during the experiment is demonstrated in Figure I—2. Following the first

pulse, the magnetization lies in the plane perpendicular to z (c). M consists of
a number of spin packets, i.e. groups of spins that have different local

magnetic environments, therefore different precession frequencies. Because

of this they will fan out during the free precession period 1, leading to the

disappearance of the macroscopic magnetization, M (d). Spin packets labeled
2, 3 and 4 have precession frequencies larger than (no, so they develop a
positive phase angle (with respect to the ~y axis) during the free precession
period, ’c. Packets with an effective local field smaller than B0, for example

packets 6, 7 and 8, will accumulate a negative phase angle. The arrows in

Figure I~1 indicate the direction of precession of these packets. The second

 

h;

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

t=2r

Figure I-2. Vector diagram to explain the formation of the 2—pulse spin echo.
(a) the pulse sequence; (b)-(c) the first pulse tilts the magnetization (M) onto

the (xy) plane; (d) the spin packets that comprise M dephase due to their
different precession frequencies; (e) the second pulse inverts each spin packet

around the x-axis; they continue to precess in the (xy) plane; (f) they refocus

along y and reform M [from Ref. (8)]-

L—

 

 

 

10

pulse torques the magnetization of every Spin packet about the x axis by 1800
(e). Since the spin packets continue precessing at the same angular velocity,
they rephase along the +y axis after time t following the second pulse (f). This
refusing of the magnetization components leads to an emission of MW
radiation by the sample and is called the spin echo. When a short MW

irradiation of fixed wavelength (pulse) is applied only a portion of the

inhomogeneous line is excited(Fig. 1-3). The excitation bandwidth (Av) is

determined by the length of the pulse, namely, Av=1/tp. This can be

understood quantitatively from taking the Fourier transform of the finite

wave train of length tp, and is discussed elsewhere (11). For a fixed flip angle,

 

one may increase tp by decreasing B1 (decreasing the MW power), making the
excitation bandwidth of the pulse smaller (making the pulse more selective);
or vice versa, tp may be decreased by increasing 8,, making the pulse less

selective. In the latter case limitations are imposed by the performance of the

 

MW electronics (how fast switches one can have) as well as by the threshold

of the MW power (limit of the amplifier).

1.3. Thr - 1 echo timulat d E ho

The 3-pulse sequence (it/ 2—t -TC / 2—T-7t/ 2) creates a stimulated echo at time 1:

after the third pulse (9). The dynamics of the spin system is shown in Fig. 1-4
[from Ref. (8)]. The first pulse rotates the magnetization about the x axis (b), so
it lies along the -y axis (c). Then dephasing of the magnetization in the xy

plane fOllows, in the same way as described above for the primary echo (d).

 

 

—¥—

 

 

 

 

11

ABC

 

 

 

BO,avg

 

b)

(0i 0)}

4“ V)»

 

 

 

 

 

 

___.

Figure I-3. The inhomogeneously broadened EPR lineshape. (a) Bang is the
mean value of the resonant magnetic field, ABO is the linewidth of the

portion of the Spectrum excited by a pulse. (b) the magnified View of the

portion excited by the pulse, with the resonant lines corresponding to
individual spin packets. (00, mi, mi are the precession frequencies of the spin

Packet magnetization components.

 

 

—;

 

 

12

The second pulse rotates the y components of the Spin packets along the z
axis (e). If the condition T2e<<T<T1e holds (T2e and T18 are the transversal and

longitudinal relaxation times of the electron, respectively), the magnetization

in the (xy) plane will fully decay during the free precession period T and only
the magnetization along the z axis survives (f). At time T the third it/ 2 pulse
is applied which transfers this polarization pattern to magnetization in the
(xy) plane (g), which, after time 1 following the third pulse refocuses along the

y axis, yielding the stimulated echo (h).

1.4. Four-pulse Echo

The four-pulse echo sequence is basically the stimulated echo sequence with

one additional mixing pulse (a rt-pulse) inserted between the second and third
n/ 2 pulses (Figure H-la). The role of this pulse is to invert the magnetization,
while interchanging spin packets belonging to the opposite electronic
manifolds. We focus on a Spin packet BMZ located along the —Z axis after the
second Tc/Z-pulse (Figure I-5, (a)). By applying the rt-pulse with the B1 field
along x axis, 8MZ is inverted onto the +2 axis (b), then the last n/ 2—pulse

torques the spin-packet onto the -y axis (c), where it starts precessing about B0
immediately after the fourth pulse. In analogy to the formation of the

stimulated echo, the total magnetization along the —y axis is maximized at

time 'c after the fourth pulse.

 

 

 

 

13

 

 

 

 

 

 

 

 

 

 

   

all t = (1%” t = (”TY t = 21”

 

Figure 1-4. Vector model to explain the formation of the 3-pulse echo. (a) the
pulse sequence; (b), (c) same events as in Figure H. (d) the second pulse
t0rques the magnetization along z; (e) the magnetization components precess
about 2; (f, g) the surviving magnetization components (directed along 2) are

.flipped onto the (xy) plane by the third pulse; (h) the Spin packets refocus

along y, forming the echo [from Ref. (8)].

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

14
(a) Z M Z
5M2 A
X ,___ X
aMz‘i
(C) . (i)
-y 6
5M
l
X x

 

 

Figure I-S. Illustration of the formation of the 4-pulse echo. (a)‘Position of

spin packet 6Mz after the second pulse; (b) the inverting it-pulse reorients
5M2; (c) the third pulse torques 5M2 onto the (xy) plane; ((1) spin packets start

precess about 2; (e) by the time I after the last pulse various Spin packets

assume the position shown, which creates the echo along the -y axis.

 

 

:u
w

it

 

15

 

I.5. Semiclassigal Description of ESEEM

One can gain qualitative understanding of the modulation phenomenon
when combining the classical picture of spin echo formation with the
quantum mechanical treatment of an unpaired electron spin coupled to a
spin 1/2 nucleus (12). The Spin Hamiltonian for such a spin system (8:1 / 2,
1:1/2) consists of electronic Zeeman, nuclear Zeeman and electron-nuclear

hyperfine interaction terms and it is given (in angular units) as

jg . .. . . . .

Where AZZ=A=Al ,cosz®+Aisin2@, AXZ=B=(A. ,-Ai)cos(~)sin®, (osngB0 / h and

m1=ganBo/ h (12). Al , and A i are the principal values of the axially symmetric

I-IFI tensor, which can be described in terms of a Fermi contact coupling, A

iso'

and a dipole-dipole coupling term, T=ggnBBn/r3, With A. I=Aiso+2T and AfAiso'
T. The angle (9 describes the orientation of the principal axis of the hyperfine
tensor with respect to the laboratory field. The remaining terms are g, the
electron g-factor; gn, the nuclear g~value; B, the Bohr magneton; B“, the

nuclear magneton. The Hamiltonian matrix is easily constructed in the
uncoupled basis set consisting of electron and nuclear spin product states,
lms, mI>. Diagonalization to yield the eigenvalues and eigenvectors of Eq. [I—3}
can be managed independently for the two electron spin manifolds as the

only term in the operator that gives rise to off-diagonal elements is that

involving Ix.

 

 

 

 

 

.5?

16

The results are summarized in the energy level scheme of Fig. 1-6, where the
normalized probability amplitudes for the EPR transitions marked lul and

lvl are given by

0.5gBB1 2
_ <1l§Xl3> _ (Pa "(pfi
Iv | -—_0-5gl5131 — cos[——————2 ] [14b]

The angles (Pa and (,0B define the axes of quantization for the a and b electron

spin manifolds, respectively, and are given by SincpazB/ 2(a)0L and simpfizB/ 2mg.
In general, all four of the possible EPR transitions for the S=1/2, 1:1/2 spin
system of Fig. I-6 are allowed. These transitions will be excited Simultaneously

in a pulsed EPR experiment, provided that the MW pulses have sufficient
bandwidth.

To illustrate the modulation effect the rotating reference frame used in
the explanation of Spin echoes will again be applied (Figure L7). The four

partially allowed transitions are designated by letters A, B, C and D. The
discussion will focus on transition A. After the first (715/ 2) pulse, the
transverse magnetization that corresponds to transition A will precess more

slowly than (no ((DA<oso). Thus in the rotating coordinate system it will acquire a

negative phase angle, (mA-mofi during the free precession period. Application

of the second pulse will excite transition A, and because the MW pulse excites
all transitions simultaneously, the corresponding semi-forbidden transition C

will be excited as well. So, as shown in Fig. I-7b, the effect of the second pulse

 

———

 

 

 

 

 

 

 

l7

 

 

 

 

 

 

 

 

 

state, described by the Hamiltonian of Eq. [1-3].

 

 

 

Eigenvectors:
|1> ‘ é I u) |l> = cos(qia/2) l+ +> + sin((pa/2) l+ ->
. ‘ a __
|2> T f ‘ l |2> :2 -sin(cpa/2) |+ +> + COS(Cpa/2) |+ ->
In! Ivl lvl lul u) S
|3> l l L f L 13> :: COSfCPg/z) l- +> + sin(cpfi/2) l- ->
JL
(1)
a = 5
l4> ~ I l ‘i a 14> : -sin(cpfi/2) l— +> + cos(<pB/2) l- ->
(iii/(To ]— A/2)2+ 132/4
(of/(31+ A/2)2+ B774

Figure I-6. Energy level diagram of an 5:1 / 2, 1:1 / 2 Spin system in the solid

———

 

 

 

 

 

 

 

18

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l |1>
(I)
t l '3>
A D
(Du
. A D C B
B C
“’0
A) B)
X' X
timcrafier
n/Z ulse -

P Y' 1: pulse y, (o

c

(0‘ 0A
mo
C) X.
we
timer A 9
(”A

 

Figure 1-7. Schematic illustration of the origin of spinecho modulation. The
upper left insert shows the scheme of the four-level energy level diagram of
Fig. I-6, While the upper right insert shows the spin packets in the

inhomogeneous line. A), B), C) Show the branching of the transitions, which

leads to the modulation phenomenon (see text).

 

 

 

 

 

 

 

19

is not only flipping the magnetization component (DA, but also inducing
”branching” of the spin packet into two portions, precessing at (0A and (0C. Spin

packet C has a precession frequency greater than 030, and while Spin packet A
will refocus along Y after time t following the second pulse, C will be at an
angle 9 with reSpect to Y (Fig. I-7c), due to the difference in their precession

frequencies. The echo amplitude is given by the projections of all the spin

packet magnetizations onto axis Y. While A contributes fully to the echo, the

contribution of C can be written as c-cosG, where c is the length of the
magnetization vector C. One can immediately see that (9: l mA—ooc 11: or (9203,11,

where too, is the hyperfine frequency (NMR frequency) in the 10t> electron spin

manifold (the difference between levels 13> and 14>). If we designate the
portion of magnetization Split from Spin packet A upon the second pulse (i.e.

the ”branching”) as k, we may write the expression for the echo amplitude as

 

E(t)=k'cos(o)0;1:)-E0 [1.5]
'where'Eo is the full echo amplitude (i.e. amplitude observed with no
modulation). Equation [1-5] shows that the echo amplitude as a function of t
indeed is modulated by (no, and dependent on the modulation depth k (which

is determined by the extent of the ”branching"). The same reasoning may be

expanded to the other pair of transitions B and D, which are simultaneously
excited during the pulse experiment. Thus, we expect to have a),3 as a

modulation frequency as well. The quantitative quantum mechanical

 

 

—;

 

gem 011

110111112
helm

the

“13101

 

20

treatment outlined later shows that the echo amplitude is modulated by

frequencies (00,, (DB, (Dad-COB and com-o)B (13, 14).

1.6. Two-pulse ESEEM

In this experiment the amplitude of the primary echo is measured as a
function of pulse separation, ”C. The echo decay is primarily determined by

spin-spin relaxation (characterized by relaxation time T 2). The weak HFI with

nearby nuclei manifests itself as modulation of the echo envelope. The total

 

time dependence of the echo amplitude thus can be expressed as
E(T)=Edec(T)'Emod(T) [L6]
where Emod is the modulation, Edec is the decay of the echo envelope.

The exact expression for the modulation function for S=1/2, 1:1/2 spin system

(13) is

 

Emod(r)-_-1 - E[2—2coscoat-2cosmflt+cos(o)u+a)fl)t+cos((oa—cofi)t] [TI—7]

Thus, the echo intensity is modulated by the basic nuclear transition
(ENDOR) frequencies, 000,, (0,3, and their combinations, ma+mw coming. These
modulation frequencies can be resolved by Fourier transform of the full echo
envelope function E(t). Note that the combination frequencies have opposite

phase to the basicfrequencies (cf. Eq. [I—7]), thus negative components also

appear in the 2-pulse ESEEM spectrum. Prior to Fourier transform it is

 

 

 

customary to perform a dead-time reconstruction first suggested by Mims (15)

 

 

—~—

 

. a1
”.11

21

to reduce the line-shape distortion introduced by the instrumental dead-time

(usually ca. 150 ns).

1.7. Thregpulse ESEEM

In the three pulse ESEEM experiment usually T is varied while I is kept
constant. Thus the amplitude of the Spin echo is measured as a function of T
(or T+‘C). The modulations can be observed for longer periods than in the 2-
pulse ESEEM (T1>>T2), which results in a better frequency resolution. With
the introduction of t‘=t+T, the explicit form of Em0d(t, T) for the 3-pulse echo

(13,14) is
Emu, T)=1-S—{[1-cos((oat)][1~cos(u)BT‘)]+[1-COS(CDBT)Ill'CosfwaT‘m [I- 1

This equation clearly shows two important features of the 3-pulse

experiment. One that only the fundamental hyperfine frequencies contribute

to the modulation. The other that depending on the actual t, certain

 

frequencies can be enhanced, while others can be partially or fully suppressed.
This leads to blind spots in the 3-pulse ESEEM spectra, which results in

lineshape distortion in disordered systems (6, 14, 18).

1.5. Four-pulse ESEEM

Two basic versions of the 4—pulse echo experiment have been used: the two-
dimensional 4-pulse ESEEM (HYSCORE) and the 1D version. HYSCORE is

dealt with in Chapter 11. Here we mention a few characteristics of the 1D 4-

 

 

_;

 

 

 

 

 

 

 

 

 

 

 

22

 

pulse ESEEM. This is carried out in such a way that the two time variables (t1

and t2) of the pulse sequence (Figure II-la) are incremented simultaneously,
so that t1=t2=T / 2, while I is kept constant. In this case the modulation

formula, Eq. [II-8] takes the following form (1648):

S(t, T)=1— E [C0+2Cacos[ 92—“- (I+T)]+2CBcos[ 95‘?- (t+T)]

+2Cc{c2cosl £051 (T+T)]-SZCOS[ 2);— (r+T)] ]] [I-9]

where C0, C0,, C5, 5, c, k, (0+, (0_ are defined by Eq. [II-9]. It can be seen that apart

from a scaling factor of 1/ 2, the frequencies of the two-pulse and four-pulse
ESEEM are the same. The four-pulse sequence, however, has the advantage
that during the variable time intervals nuclear spin coherence evolves which
decays much more Slowly than electron Spin coherence (6, 8). This results in a

pronounced reduction of line widths (i.e. in a better spectral resolution) and is
particularly important in studies of sum peaks ((0,) in spectra of disordered

systems (17, 21-23), where the frequency of these peaks is closely related to the
dipolar part of the hyperfine coupling (19).

A novel application of the 4—pulse sequence has been proposed by Ponti
and Schweiger (20) to exploit undistorted ESEEM spectra. This experiment is
termed DEFENCE (deadtime free ESEEM by nuclear coherence transfer
echoes) and is shown to yield the distortion free absorption powder lineshape.

A DEFENCE spectrum essentially is the projection of the two—dimensional

HYSCORE spectrum on its (i)2 axis.

 

 

 

-~.~..a .

the 11
53 €th
limp:

tlper'm

119 He

depth

 

 

 

 

 

 

 

 

 

23

Lg. Densig Matrix Description of ESEEM

The density operator formalism has proven particularly useful to describe the
evolution of the spin ensemble in magnetic resonanCe (2, 24, 25). It can be

shown easily that the expectation value <A> of an observable represented by
operator A can be obtained as

<A>(t)=Tr {13(0 - A} [I-IO]
where on) is the matrix representation of the time-dependent density
operator. The ”equation of motion” for f) can be derived from the time-

dependent Schrodinger equation, and takes the form:

 

d. 1..
— t=——-H,, 1-1
dtm) iht p] l 1]

where the Hamiltonian Fl may be time-dependent. In a spin-echo experiment

the expectation value of interest is the magnetization along the y axis, which
is proportional to Sy. The normalized echo amplitude in the two-pulse
experiment is then:

____ Tr{i>(21') - SY}

. . 1-12
Trlp(0) - Sy} [ 1

 

E(t)

The Hamiltonian in the electron spin echo experiment consists of a time-

d6pendent and a time-independent part,

IAiror = fio + film [1—13]

 

 

 

 

 

 

 

 

24

 

where F10 describes the static interactions in the sample (electronic and

nuclear Zeeman, nuclear hyperfine and quadrupolar interaction), while FlI
expresses the interaction between the Spins and the MW pulse. Here we
outline the derivation of the modulation formula for the two-pulse ESEEM,
following Mims (13).

The time-dependence of F11 can be removed by transforming the problem

into a reference frame rotating at the MW carrier frequency, to. The entire

derivation will be cast in this rotating coordinate system. The evolution of

the density matrix can be expressed as

 

not) = R - 13(0) - R“, [1.14]
where

R = R, . R2P . R, . R,,., [1-15]
with

Rt = exp[—iF10t/h], [1-16]

RiP = exp[—ifITOTt,P Ni]. [1.17]

Rt and R1,, are the propagators during free precession and nutation periods, tip

is the duration of the ith pulse. It is convenient to carry out these matrix
manipulations in a representation in which F10 is diagonal ("interaction
r8presentation”) (13). The unitary transformation which diagonalizes H0 is

represented in a sub-matrix form as

 

r )1: 2

litre E
lllhe u
malice

are [he

 

25

,. Ma 0 . M; 0
£0 :( 0 MBJ'H0'[ 0 Mg]; [1-18]

where Ma and MB are matrices of dimension 21+1 (I is the nuclear spin). H0 is
the matrix representation of the Hamiltonian in a state space with basis
vectors 10L, mI>, and 1 [3, mI> (InI is the nuclear quantum number). In this
representation F10 is block diagonal (high-field approximation). Thus, Ma and

M3 are matrices diagonalizing these blocks. In the ”interaction

representation” the matrices of our interest will take the following form (26):

,
Rt = [Pt 0 J, [I-19]
0 Qt
R z [3 E ”M [I-20]
”’ 2 —iMT E ’
o —iM _
RZP = [_th O )1 [I 21]
. o —iM/2] [1-22]
Y _ iMl/Z o ’
. /aE 0 LB
mm = k 0 ME), 1 l

where E is the einheit operator, M: M; -M[5, a and b are the initial populations

. + -
0f the on and [3 electronic spin states, respectively. Pt'r and Qt are d1agonal

matrices, with elements Psl=€XP(-lwmt) and Qii*=exp(—1wfi,t), where 00a. an - B.

are the nuclear transition (ENDOR) frequencies in the CL and [3 electronic

 

 

 

 

26

manifolds, respectively. Computing the echo amplitude using Eq. [I-3] leads to

(13):

 

13(1) = Re {Tr(Q1MTPlMQ,MTP,M)}.

21+l

When 1:1/2, M has a simple form (13):

V 11
Mi . *1
—u V

and the normalized echo modulation function will take the form

Em = Ivl“ +1ul“ +1le ‘1le -{2ws<wai> + 2008(0313“) —

—cos[(ma — £05m — cos[(o)Cl + wB)r]}
This is the familiar formula for the two—pulse echo for 821 / 2, 1:1/2

71)-

[1-24]

[1-251

[I-26]

(cf. Eq. [1—

 

 

27

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

Atherton, N. M. Principles of Electron Spin Resonance, Ellis-
Harwood PTR Prentice Hall: New York, 1993.

Pilbrow, J. R. Transition Ion Electron Paramagnetic Resonance
Clarendon Press: Oxford, 1990.

Keijzers, C. D.; Reijerse, E. J.; Schmidt, J. Pulsed EPR: A New Field
of Applications, North Holland: Amsterdam, 1989.

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

Schweiger, A. Angew. Chem. Int. Ed. Engl. 1991,30, 265.

Dikanov, S. A.; Tsvetkov, Yu. D. Electron Spin Echo Envelope
Modulation (ESEEM) Spectroscopy CRC Press: Boca Raton, 1992.

Ponti, A.; Schweiger, A., Appl. Magn. Reson. 1994, 7, 363.
Hahn, E. L. Phys. Rev. 1950,80, 580.
Carr, H. Y.; Purcell, E. M. Phys. Rev, 1954, 84, 630.

Arfken, G. Mathematical Methods for Physicists , 3rd edition,
Academic Press: San Diego, 1985; pp. 801—803.

McCracken, J. In Handbook of Electron Spin Resonance. V0111.

P0016. C. P.; Farach, H. A., eds, in press.
Mims, W. B. Phys. Rev. 1972, B 5, 2409.

Mims, W. B. Phys. Rev. 1972, B 6, 3543.

 

 

 

 

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

28

Mims, W. B. J. Magn. Reson. 1984, 59, 291.

Gemperle, C.; Aebli, G.; Schweiger, A.; Ernst, R; R. J. Magn. Reson.
1990, 88, 241.

Tyryshkin, A. M.; Dikanov, S. A.; Goldfarb, D. J. Magn. Reson.
1993, Series A 105, 271.

Schweiger, A. In Modern Pulsed and Continuous-Wave Electron
Spin Resonance, Kevan, L.; Bowman, M. K., eds., Wiley—
Interscience: New York, 1990; Chapter 2.

Reijerse, E. J.; Dikanov, S. A. J. Chem. Phys. 1991, 95, 836.
Ponti, A.; Schweiger, A. J. Chem. Phys. 1995, 102, 5207.

Lee, H-I.; McCracken, J. J. Phys. Chem. 1994,98, 12861.

Dikanov, S. A.; Spoyalov, A. P.; Hfittermann, J. J. Chem. Phys.
1994, 100, 7973.

Dikanov, S. A.; Burgard, C.; Hfittermann, J. Chem. Phys. Lett.
1993, 212, 493.

Cohen-Tannoudji, C.; Diu, B.; Laloé, F. Quantum Mechanics, John
Wiley & Sons: New York, 1977. p. 295ff.

Slichter, C. P. Principles of Magnetic Resonance, 3rd ed.,
Springer Verlag: Berlin, 1989. p. 157ff.

Lee, H.-I. Ph. D. Dissertation, Michigan State University, 1994.

 

 

 

 

 

:tstBt techniqueS Pm"
r" correlating pairs 0f
1?th and determining.
-5 variant of 2D ESEI
seed on the threepulse
21R! pattems ( functions
Era Tramst'onn ( Fl) wi
serum in which, for a S
W210,0).(m,,0t (0, (
mum is symmetrical r
m11135 FT spectrum, all
l«ttplitude and a suppres
ESEEM is absent.

The disadvantage 1
utthEprimary echo d e ca

:13
aresult, the resolutim

Orders of magnitude, 95F

h‘u '
. 3mg autoregression

he larger linewidth, as l

Chapter II

HYSCORE

2D ESEEM techniques provide a means of disentangling complicated ESEEM

spectra, correlating pairs of ESEEM lines belonging to a certain hyperfine

 

coupling and determining the relative sign of different hyperfine couplings.

The first variant of 2D ESEEM spectroscopy pr0posed by Merks and de Beer (1)

 

is based on the three-pulse sequence. It involves collecting sets of stimulated

ESEEM patterns (functions of T) at different values of 1. Subsequent 2D

Fourier Transform (FT) with respect to 1: and T yields a two-dimensional

 

spectrum in which, for a 8:1 / 2, I21 / 2 spin system, the following peaks are
expected: (0, 0), (ma, 0), (0, 030,), (mg, 0), (0, (DB), ((00,, (DB), (00$, (00,); note that the

spectrum is symmetrical relative to the diagonal, (2)1402 (1, 10). In the

modulus FT spectrum, all six peaks with non-zero coordinates have the same

amplitude and a suppression effect typical of the one—dimensional stimulated

ESEEM is absent.

 

The disadvantage of this 3-pulse 2D ESEEM, however, is that the time
of the primary echo decay is much shorter than that of the stimulated echo.
As a result, the resolution along the different frequency axes may differ by
orders of magnitude, especially in single crystals. This problem can be solved
by using autoregression methods for the spectral analysis along the axis with

the larger linewidth, as in Ref. (2).

29

_

 

 

 

 

 

Extreme these drawbac
333,9 sequence was P“
7:52:19 correlation bent
33;; Sign. manifolds. Tl
“first. srecrrsscez’y N F
repulse stimulated ec‘nc
:1. first. t. :s varied v
:eiemain data set. Thei
1;:6 ll-l'ba. The linewid
iZ'CItiC spin-lattice relax
RECORE spectrum) is o
ficnbe represented as
since there are three t

listihite the axis of the

dusted according to 3pc

Na
ill.

l” this Chapter, a t
represented. First, a gel
important features of the

ll'iew
0t -
the experiment.

isordered samples is di<

3O

 

To overcome these drawbacks of the 3—pulse 2D ESEEM a new method using a
four-pulse sequence was proposed by Hofer et a1. (3) (Figure lI-la), which also
reveals the correlation between hyperfine sublevels belonging to the opposite
electron spin manifolds. Therefore it was termed hyperfine sublevel

correlation spectroscopy or HYSCOREIn this experiment the amplitude of

 

the 4-pulse stimulated echo is measured as a function of the two time periods

t1 and t2. First, t2 is varied while t1 is kept constant to obtain one ”slice” of the

 

time—domain data set. Then t1 is incremented until all the slices are collected
(Figure II-lb). The linewidth in both dimensions is determined by the

electronic spin-lattice relaxation time, T1. The frequency-domain data

 

(HYSCORE spectrum) is obtained by 2D Fourier Transformation (Figure II-lc)

and can be represented as a contour plot (Figure II-ld). In the four-pulse
. sequence there are three time variables (1, t1 and t2). Two of them (t1 and t2)

constitute the axis of the 2D ESEEM spectrum and the third one (I) has to be

adjusted according to spectral properties and desired information (see Section

II.3).

 

In this Chapter, a brief overview of some relevant aspects of HYSCORE
are presented. First, a general description of the method is given, then some
important features of the 2D data analysis are examined, followed by a short
review of the experimental aspects. Finally, the application of HYSCORE to

disordered samples is discussed.

 

 

_

 

 

 

 

{are ll-l. Scheme of 0N
is amplitude of the «H
linedomain data malt
fegrsenration oi the treq‘
Stdmn) obtained by 2K
in matrix and taking th

intour plot of the abort

 

 

31

 

Figure II-1. Scheme of obtaining a HYSCORE contour plot. (a) Measurement

 

of the amplitude of the 4-pulse echo as a function of t2 and t1. (b) The resulting
2D time-domain data matrix, with only 33 of all the 128 tl-slices shown. (c) 3D
representation of the frequency-domain data (HYSCORE magnitude

spectrum) obtained by 2D Fast Fourier Transformation of the time domain

data matrix and taking the absolute value of the resulting complex matrix. (d)

 

Contour plot of the above spectrum.

 

 

 

Ev

n
0
.nul...
flu
rl.
a
p
mu
P

 

 

 

 

 

TC/Z

ll

rt/2

it (a)

rt/2 A—

 

 

l Preparation

   

a

l

O

<——-t1

 

Evolution

it

 

1000 2000

 

t2——>-<—- '1:

Detection

Mixing

  

3000 4°00 5°00 6000 7000

t2 (HS)

 

 

 

illll

33

 

V 1 (MHZ)

 

 

O.

 

 

m-n
.4

Aal-

 

 

34

11.1. Introduction to HYSCORE
The HYSCORE modulation formula for the S=1/2, 1:1/2 case has been
derived by Gemperle et a1. (4). Nonselective, ideal 1t / 2 and 1t pulses were
considered. The spin system is described by the Hamiltonian

it = (0582 + IAS - (DIIZ [II-1]
where (nszgeBeBo/ h (isotropic g-tensor) is the electron Larmor frequency,
wrzganBo/ h is the nuclear Larmor frequency, A is the (anisotropic) hyperfine
tensor. By neglecting the Sx and SY terms (high field approximation), the two
nuclear transition frequencies, (Do and (DB, associated with electronic spin states

m5=+1/2 and msz-l / 2, respectively are given by

co,=( IA+I)1/2 [II-2a]
o,=( iA_i )1/2 [II-2b]
with
A,=:A/2 + 03,13. [II-3]
ude of

E is the 3x3 unit matrix and l is the unit vector along Bo. The amplit

the four-pulse echo is obtained in the following form, using the density

matrix approach described in Chapter I:

A " —l -1 -l -1
8(t1, t2)=Tr{ SXU(t)Pp/2U(t2)PpU(t1)Pp/2U(t)Pp/2SZ P,,,U(t) P,,2U(tr)

P;1U(t,)-1P,;,‘,U(t)-1}, [II—4]

With the prOpagators U, P for the free precession and nutation periods,

respectively:

 

 

 

 

35
U(t)=exp{-i.‘i€t} [ll-5]
P,,,=exp{-i(n/2)§x} [no]
P.=expi-i<n>éxi [II-7]

An evaluation of Equation [II-4] was performed by the algebraic computer

program MACSYMA(4, 5) and lead to the result

 

 

k
S(t1, t2)=1- Z{C0+Ca[cos(wat1+ 933“: )+cos((nat2+ (DST )]+
OJBT (9B1
+CB[cos(thl+ -2——)+cos(oofit2+ T )]+

(1)+’t)+
2

 

(o t
+Cc[c2cos(mat1+mfit2+ ——;——)+c2cos(coflt1+(oat2+

 

0) T to t
+szcos((oat1-0)Bt2+ j—Hszcosmatrwetz-t “3“)“ [IFS]

with amplitudes
C0=3-cos(wBt)-cos((oar)—szcos(w+r)—c2cos((njr)

(our

wort
~COS

2 )

 

 

(o t
Ca=c2cos((nut- ——§—— )+SZ((D|3”C+

u) t 0) t (1) T
Cfizczcosmut- ——2l1-— )+szcos(u)at+ %— )-cos( —~—2B——)

x . (1) T
Ccz-Zsin( 9421-— )sm( -—§-— ), [ll-9]

the sum and difference frequencies (13,—tonne),3 and (o_=oo,,-(DB, amplitude factors

szzsin2(8/ 2) and c2=cosz(8 / 2), and the depth parameter k=4sin2(8 / 2)cosz(5 / 2). 8

 

is the angle between the two effective magnetic fields at the nucleus,

 

 

 

 

 

 

 

 

"2 ilk. 0
1:1", lllllf

:, he 2D

 

36

corresponding to the two electron spin alignments (mszil/Z) and can be

expressed as

In): 1(0),, + (DB) '

sinZ(5/2)= 4 [II-10a]
(90,0)B
l

cosz(8/2)= [ll-10b]
than)£5

The constant term C0 in Eq. [II-8] depends on I only and does not contribute to
the echo modulation. The second and third terms proportional to Ca and CB
contain the modulation frequencies (0,1 and c0fl either in the t1 or t2 time
domain. In the frequency domain this will result in peaks at (O, 0)“), (com, 0), (0,
(DB), (tan, 0). The last term with coefficient Cc contains cosine functions with
both time variables simultaneously in the argument. It produces cross peaks
in the 2D spectrum at ((00,, (DB), ((05, 0),). Thus, for a S=1/2, 1:1/2 spin system, the
HYSCORE spectrum, 5(0),, 0),) will consist of six peaks as shown schematically
in Fig. 11-2 by the solid squares. Two pairs of axial peaks with frequencies (0,,

and 0),, appear along the axes (x)1 and 0),, and form two identical 1D spectra. The

two cross peaks are related by reflection symmetry relative to the diagonal

men)? Note that since ideal pulses are assumed, peaks along the diagonal do

not occur. In real experiments the pulses are never ideal and diagonal peaks

do occur (Fig. II-2, empty squares). From Eq. [II-8] it can be inferred that the

 

 

O

 

 

 

 

37

 

peaks appear as a mixture of absorption and dispersion type signals (6) with a

variable phase shift that depends on T. Phasing is normally impossible and it

is advisable to plot the magnitude spectrum. The peak amplitudes in this case

(4), are expected to be

 

 

 

A((oa, O)=A(O, ma): El czcos(0)B‘C- (0:1 )+SZCOS((DBT+ mgr )-cos( (0:1) I [II-11a]

k (1)51: 2 (OBI (9B1
A((op, 0)=A(0, (05): El czcos((oa'c-—§—)+s cos(mat+—E——)-cos(——-)| [II-11b]

0) T
A(ma,(ofi)=§| sin(9—§£)sin(—2L)| (2c4+254)1/2. [II-11c]

Depending on the actual values of (nu, ma, t, 32 and c2, blind spots may occur, at
which the intensity of some peaks drops to zero. If the condition of complete
excitation (o)1 >> (00,, coB, where (01 is the MW field strength of the pulses) is not

fulfilled, the two allowed and two forbidden transitions are no longer equally
excited. This has the following consequences: (a) the Hamiltonian of Eq. [II-1]

must be taken into account also during the pulses (7, 8). (b) The coherence

exchange between the two nuclear transitions (Du and (oB is incomplete for
non-ideal mixing rc-pulse; this leads to the appearance of two diagonal peaks

((90,, 60a), ((1)3, mp). (c) The peak intensities are no longer described by Eqs. [II-11]
and the formula for the modulation depth parameter, k, is not valid
anymore. (d) All possible two-, three— and four-pulse echoes are created (Fig.
H-Ba), of which, however, the unwanted ones can be eliminated by phase-

cycling (4) (Fig. II-Sb). In our HYSCORE eXperiments a 4—step phase-cycling

 

 

 

 

 

 

 

 

 

:nre Ill

2:“, Igor

 

 

fl

Uq £1
1.,“ 1 i

E

 

 

 

 

 

 

 

k

Figure Il-2. Schematic of the expected peaks in HYSCORE spectrum of S=1/ 2,

1:1/ 2 spin system [from Ref. (4)].

 

 

H
M
( l (H _) HM)
‘ (4+) (_+0+) (0+0) (~++) (M) (+++)
FID (‘H 'l t
a
(W)
(- +-) (' ”I
H
t
b A T
0 0.5 l 1.5 2 2.5 3 1"]

Figure 11-3. Computer simulated spin echoes generated by a sequence of four
non-ideal pulses (1:400 ns, tlzzps, t2=1.1us) ; (a) without phase-cycling; (b) with

phase-cycling according to the cycling scheme of Table 11-1 [from Ref. (4)].

 

 

 

 

39

 

was used (see Chapter VI). It should be mentioned that the 3—pulse stimulated
echo always coincides with the 4—pulse echo and cannot be eliminated by
phase cycling. Since cross peaks between nuclear transitions for the same m5
state do not appear, HYSCORE may not only be used to assign peaks to
particular sites but it also allows one to determine the relative sign of the

hyperfine coupling constants within each site (4).

[1.2. HYSCORE data analysis

As mentioned above, the HYSCORE correlation spectrum, 8(0)], (02) is

obtained via a 2D Fourier Transform (FT) of the time-domain data, S(t], t2). 2D

FT simply means taking the Fourier Transform (9) of S(t1, t,) with respect to t2,

leading to S(t1, col) , then the FT of S(t], (0,) is performed with respect to t1,

resulting in 8(0)], 032). The order of FT’s with respect to t1 and t2 may be

interchanged. The complex Fourier Transformation of the data set can be
performed in different ways leading to different spectral representations as
well as different spectral densities (10). Therefore, it may be of interest to give
a brief introduction to certain relevant aspects of 2D FT.

Amplitude— and Phase—Modulation of 2-D ESEEM Signals

In a 2D experiment, a signal observed during a time t2 is modulated as a
function of a previous time interval t,. This history can either affect the phase
or the amplitude, or both (6). If we consider non-quadrature detected signals,
as is normally the case in ESEEM experiments, an exemplary 2D time-domain

Signal can be described as

 

m

 

D .0

S(t1, t2)=cos(0) At1)'COS((DDt2+(DPt1), [II—12]
which is represented in Fig. H—4a. For the sake of simplicity, all relaxation and
amplitude factors have been neglected. Also, constant phase shifts arising

from the spectrometer deadtime are not important here and are therefore
omitted. The signal S(t,, t2) contains a frequency (DD detected during t2 and is
modulated by the two frequencies (0A and (DP as a function of t1 where (0A affects
the amplitude while (01, modulates the phase of the signal. A FT performed in
the t2 domain will generate a real- and imaginary part spectrum F(032)=A(coz) +
iD(co2), containing the absorption (A) and dispersion (D) line-shapes,
respectively. For a phase modulated signal, this assignment is only valid for

the first slice in the t1 domain (t1=0). The signal S(t1, (02) then becomes:
S(t1, (1)2)=cos((o,\t])-exp(i(opt1)'F(w2=wD) =
aexpamfimpnl) + exp(i(wp-coA)tl)‘F(w2=(oD). [II-13]
From this equation we can derive the special cases of exclusive amplitude or
phase modulation. If the signal S(t2) is only modulated in amplitude (coP=0),

no information about the sign of the modulation frequency is present. Thus,

a complex FT in the t, domain results in a pair of lines at the positions (00A, (DD)
and (-0)A, (on). As only the real part signal in t2 is recorded, the complex FT in

the (2)z domain also generates two lines at +(DD and -(DD and the calculation of

%

 

 

 

5‘

echo (Imp (n u )

 

echo amp. (a.u.)
O

5(0)], (oz).

 

41

 

 

.7 6000

- 5000

— 4000

- 2000

- 1000

 

| l i l l l
0 1000 2000 3000 4000 5000
t2 (as)

Figure II-4. Graphical representation of Eq. [II-12], calculated with the

L;

 

6000

following parameters: (DA/211:2.00 MHz, cop/27t=1.33 MHz, wD/2n=2.66 MHz. (a)
The time domain data matrix, S(t,, t2), with only 13 of the 128 slices shown. (b)

(+, +) quadrant of the absolute value spectrum, S((o,, (02), (c) (—, +) quadrant of

 

42

 

N
l

 

0@'9——*. “D

 

 

 

0.5 -

 

 

2 2.5 3
_ :1 (MHz)

(C) .

‘WD

 

 

 

43

the 2D-FT results in 4 lines, each of which contain the same information. In

the case of an amplitude modulated real part signal, it is therefore sufficient

to consider only the frequency range (l)1 > 0, (l)2 > 0. If only the phase
modulation is present ((1) A=0), the complex FT in the t1 domain preserves the

sign of the modulation frequency resulting in a single line at (031,, (02).

If both amplitude and phase are modulated, the second FT again results
in a pair of lines, now at the positions (co “(01,, (DD) and (cop-03A, (00). In contrast
to the amplitude modulation, we now have to consider positive and negative

frequencies in the a), domain, while it is still sufficient to deal only with the

positive 032 frequencies (Figure II-4b, c). In general, the line shapes obtained

after a 2D—FT are a mixture of absorption and dispersion lines (6). Therefore,
the absolute value (magnitude) spectrum defined by

5(a)], w2)=[Re2(w,, (1),) + Im2((o], (1)2)1“2 [ll—14]
is normally used for evaluation and presentation purposes.
2D-FT of HYSCORE Data
Here we examine only those parts of the HYSCORE modulation formula (Eq.

[II-8]) which produce cross peaks after 2D-FT, namely:
S(t1, t2)=C[c2cos(o)at,+wpt2+<1>+)+c2cos(oaBt1+(oat2+<I>+)+
+szcos(o)ut,-coBt2+<I>_)+szcos(thI-(out2+¢_)] [11.15]
The amplitude factor C and the phase constants (I)i are independent of t1 and

t2 and depend on t only. The t1 dependence we are interested in acts only on

L_f—-

 

 

 

44

the phase of the signal. A most interesting situation arises from the factors c2
and 52 defined by Eqs. [II-10]. If we consider weak anisotropic HFI we can

approximate the nuclear transition frequencies by
(ow = lo)I i- A/2l [II-16]
where A denotes the hyperfine coupling for a particular crystal orientation, (:01

is the nuclear Larmor frequency. Using [II-16] we get

 

 

 

A2
(1)12 — T
c2 = : and 52:0 [II-17]
(00,003

fora)I > A / 2 (relatively weak HFI), while

 

 

 

 

A2
(of —- —-—
2 4
c2 =0 and s =——-—-————— [II-18]
(00,005

for (:01 < A/ 2 (relatively large HFI). FT of Eq. [II-15] with respect to t2 then
results in ’

S(t1, (02)=exp(io)at1)'F(m2=-(DB) + exp(imflt1)'F(w2=o)a) [II-19a]
for (01> A/2, and

S(t1, wz)=exp(~i(oat1)-F((o2=(n5) + exp(—i(oBt1)'F(m2=0)u) [II-19b]
for c01< A / 2.

The contribution from the time independent phase shift in Eq. [II—15] has been

omitted. The sign of the phase modulation in t1 depends here on the relation

 

0f the hyperfine coupling to the nuclear Larmor frequency. Carrying out the

 

 

FT in the second dimension and calculating the absolute value spectrum

 

 

—~—

 

\
‘
. .2
.
§

.

.
§
5

I
..
it
. (I
‘l

c
v.
c

w _

 

 

 

 

45

results in a pair of cross peaks at (cow (DB), ((1)3, mu) if (01> A / 2, while in the case
of a)I < A/ 2, the cross peaks are in the negative (01 region, at (£00,, (0,.) and («98,
(0,1). This is verified by an experimental example shown in Figure II—5 [from
Ref. (10)], showing both (+, +) and (—, -) regions of the HYSCORE spectrum of
the malonic acid radical. At this particular crystal orientation A5200], while
A2>2wr 80, A1 is expected to give rise to cross-peaks in the (+, +) region, while
cross peaks corresponding to A2 are expected to be in the (-, +) region. This
indeed is observed, except that cross peaks that correspond to A2 also appear
in the (+, +) region, though not as intensely as in the (-, +) region. This is due
to the fact that Eq. [II-16] is an approximation, only. With the exact expressions
for the nuclear transition (ENDOR) frequencies, factors 52 and c2 will not drop
to zero. Therefore, there will always be a contribution from a positive phase ,

modulation for every negative phase modulation in t1 and vice versa. It is

also interesting to note that the diagonal peaks, which are due to an
incomplete inversion by the n-pulse (4), are mainly confined to the positive
(01 region. Thus, the modulation that leads to the diagonal peaks is a phase

modulation in t, with a positive sign.

11.3. Experimental aspects

In the HYSCORE experiment, two important parameters have to be adjusted

to achieve optimum results. The first step is to find a I value so that the

46

A2

 

    

   

   

     
   

       

 

 

  

    

       
    

  

   

  

  

A

   

   

\_.

  

 

.A AA AA ,A o».
A... .. . . ; ..... A a.
.AAA . AA. A... A. AA... . A AA
A. .._._A.A... A. ...A..A A ..A .
. .AAAAMAAAAAAAAAA. :AA __._AA.AA A.A...A.A AAAA. m
.. gags: RA. AAAAA“... A {A Ayn—AA M.

1c aci

Figure II-5. HYSCORE absolute value spectrum of irradiated malon

[from Ref. (10)].

 

 

 

 

 

 

-4.

g u
_u"

«.‘ ‘

47

 

suppression effect of the cross peaks is minimized and in the second step, the
n-pulse is adjusted in width and amplitude to obtain the best inversion of the

stimulated echo.

A method of finding the Optimum t value is to record a series of 3-pulse
experiments with increasing I (separation between the first and second

pulses) then Fourier Transform them with respect to T and calculate the
absolute value spectrum. Choose I for the HYSCORE at which the least
suppression is observed in the frequency region of interest (10). In single
crystals with a limited number of lines there may be several I values fulfilling

this condition. In disordered systems, however, it is important to use the

 

smallest I at which all lines are present to minimize line-shape distortions. In

practice, I values in the range of 120-200 ns are used (10, 14). This approach is

justified by the formula for the amplitude of the cross peaks in the absolute

. value mode (Eq. [II—11c]), which may be rewritten in the following form,
A(ma, (on): 1;-[(1—coswaI)(1-cosmBI)]1/2(2c4+234)1’2. [11-20] ,

This shows that the cross peak amplitude is mainly determined by the
amplitudes of the corresponding line—pair measured in the 3—pulse 1
experiment. Therefore, if one pair of lines are suppressed in the B-pulse
experiment, the corresponding cross peaks in the HYSCORE experiment
should also be suppressed. This is not exactly observed in the experiments,

however (10).

 

—A

 

 

 

 

 

.L: 1
HA

SPEC

{is

48

The next aspect of the Optimization procedure is to increase the amplitude
ratio of the cross peaks to undesired diagonal peaks. It has been pointed out

(4) that the diagonal peaks result from an incomplete mixing by the inverting
rt-pulse. For perfect mixing this pulse should be much shorter than the II/ 2-

pulses. In the case of relatively high frequency proton lines, this condition
cannot be fulfilled because of the limited MW power available. The degree of
inversion obtained for certain pulse Widths and amplitudes can be estimated

from the appearance of the HYSCORE echo. Fig. II-6 compares the results of

two experiments (10). In (a) the it and 1t / 2 pulses had the same amplitude and

 

their width was 20 ns and 10 ns respectively. Under these conditions, the

diagonal peaks dominate the spectrum. The spectrum in (b) was obtained

with a pulse width of 14 ns fOr the it-pulse and 10 ns for the TC / 2 pulses.
Correspondingly, the power of the n—pulse was 3 dB higher than that of the

n/Z-pulses, which leads to significant improvement of the ratio of cross peak

 

amplitudes toldiagonal peak amplitudes.

114° HY RE on disordered sam l s

In disordered samples (glasses, frozen solutions, powders) the analysis of 1D
ESEEM spectra may become complex. Part of the reason for this complexity is

illustrated in Figure II—7, where part (a) shows the simulated ENDOR powder

 

spectrum for an S=1/2, 1:1/2 spin system with an axial hyperfine tensor. In

this case the principal values of the tensor can readily be determined from the

 

   

 

49

A... __._—_—
- , _._A A...
_A AAA—-A’M‘ ‘

 

 

 

 

0 92mm 5°

Figure II-6. HYSCORE absolute value spectrum of irradiated succinic acid. (a)

Width of the n/ 2 pulses 10 ns, that Of the “’P‘flse 20 ns; (b) Width Of 1t/ 2 pulses
10 ns, that of n-pulse 14 ns (power of the n-pulse is +3dB higher than that of

1t/ 2 pulses) [from Ref. (10)].

 

   

 

 

 

FiSure II-
Shim P

The Com

 

 

(a)

l

 

 

l l

 

 

 

0 2 4 6 8 10 12
frequency (MHz)
- (b)
.
V F l l l 1 .I
O 2 4 6 8 10 I2

frequency (MHZ)
Figure II-7. (a) Simulated powder ENDOR spectrum for an S=1 / 2, 1:1 / 2
System. Parameters: gn=2.261 (311’), A, I=7.1 MHz, A 52.4 MHz, B0=3200 G. (b)

The corresponding simillated 3-pulse ESEEM spectrum.

 

 

 

 

respect“
”it in 1
m
ElL90tro
W is

IFOgran

exPress
may h

A

 

51

frequencies of the turning points and lineshape singularities (12). The
corresponding ESEEM simulation is shown in Fig. II—7b. Here the lineshape
features are absent, because the intensity of the ESEEM is determined by the
product of the orientation dependent coupling parameter (which gives rise to
the classical powder lineshape seen in (a)) with the modulation depth
parameter, k, which is also orientation dependent; k goes to zero at the
canonical orientations, so the turning points and singularities are missing.
Due to the absence of these characteristic features, computer simulation is
typically required to extract the principal values of the HFI tensor from
ESEEM. Some of these limitations can be overcome by HY SCORE. The
contour plots of the cross peaks carry information about the hyperfine tensor
(11, 14). This is illustrated in Figure II-8, where (a) and (b) correspond to the
weak and strong coupling cases,

respectively. The contours will be parallel and perpendicular to the diagonal
only in the case of small anisotropy in the HFI, where the nuclear transition

(ENDOR) frequencies may be approximated by Eq. [ll-16]. When the

anisotropy is increased (greater T), the cross peaks develop a curvature. This

effect is demonstrated in Fig. lI-9. These simulations were done by a MATLAB

Program, ”hysline1.m” (see Appendix).

Recently, Dikanov and Bowman have derived simple analytical

exPresflons for cross-peak contours for disordered S=1/ 2, 1:1/2 systems (13).

They hElVe proposed a simple graphical method to obtain the principal

' ' . ' st,
elements of the hyperfine tensor. Here we outline their approach Fir

 

 

 

 

/-\

(b)

figure 11.3. Schematic
spin system with a 5m

“l /2.

 

 

52

 

 

   

(b)

 

 

Figure 1158. Schematic HYSCORE contour plots arising from an 821/2, 1:1/2

Spin System with a small anisotropic hyperfine interaction. (60 (91> A. . / 2, (b) 0),

<A../2.

plr)
Riff,

.1!)
my,

figure 11-9. Simulated

§:11/1,1:1/25ystem- Pl

linewid

th:0.l MHZ, T:
Id) Simulation was Pf

lppendix).

 

53

 

(a)

/

12(MHz)
(A)

 

 

 

 

/

12(MHz)
(A

 

 

OLrb'AA—J—f—b—
2 4 5 6

o 1 a
l1(MHz)

Figul‘e II-9. Simulated HYSCORE contour plots arising from axial HFI from an

S=1/2/1=1/2 system. Parameters: (01/ 2n=3 MHz, Aiso=2 MHz, Gaussian

linewidth=0,1 MHz, T=O.4 MHz (a), T=0.8 MHz (b), T=1.2 MHz (c), T=1.6 MHz

ormed with the MATLAB program ”hysline1.m” (see

(d). Simulation was perf

Appendix).

6

p
{v

.0.
C. :27:

 

6

 

 

a...
AuIEK.

 

 

f2(MHz)

 

12(MHz)
00.

 

 

(0 )

 

5

 

 

 

 

54
1 2 3
11(MHz)
1 2 3
t1(MHz)

 

 

 

 

 

 

W an axial HF! ta
W part, with pl
zlthe unique principa

twitter as

raw flag-arm I2
mareA, =Afl+2Ti

angular units). Eqs. 1
fitting point of the de

idweai the ENDOR f1

Iiich means that in tb

bequencies form two a

0i the HFI parameters

Ga:

Thissuggests a straigl

uperimental 2D spec

 

 

7'» lllP'T' * fi fi

55

consider an axial HFI tensor, consisting of an isotropic component, Aiso and
anisotropic part, with principal values (-T, -T, 2T). If 6 is the angle between B0

and the unique principal axis of the HFI tensor, the ENDOR frequencies can

be written as
030,2: 0), ,azcosze + (oiazsinZG) [II—21a]
(1)32: (i), .BZCOSZQ + (Diffsin2 [II-21b]
where (i), ,a(m=-(i)1 iA, l/2, (23“,,(mz-(i)I i A J./ 2; the principal values of the HFI

tensor are A, I=Aiso+2T and A l=Ai50-T, and (i)I is the nuclear Larmor frequency

(in angular units). Eqs. [II-21] are equivalent to Eq. [1] in Ref. (20), which is the

 

starting point of the derivation. It can easily be shown that the relation

between the ENDOR frequencies has the following form:
(w.)2=Qa(wp)2+Ga [II-22a]
(we)2=QB(ma)2+Gp [ll-22b]

which means that in the 2D HYSCORE contour plot the correlated (no, and (”is

 

frequencies form two arcs, whose geometrical parameters, Q, and QB depend

0f the HFI parameters Aiso and T. It can be shown that

 

T + 2A. 0 i 4m,
:: 448: H-
0‘1“” T + 2A,“, i 40), I 23a]
+ 2 2 2

 

“(‘5’ — T + 2A3o i 4a)I
This suggests a straightforward way to extract Aiso and T from the

eXperimental 2D spectrum: obtain a set of correlated tea and (DB frequencies

 

 

———

 

 

mkmpmkcon
mittewillhavea
lllorAmandTwill yi

lhisanalysiscan

isotropic part [-T(1+d

(0,2 = (01,2:

Iiflettm are the EN
mutation of B, with I
hemmed as Iowa
”8‘95 Specifying the c
lltHFl tensor. Eqs. [I]
mwfillm

Where (0w) designate
plane and are defined

“hoof“
Note that Eq. [II-25] h
the axial case, except 1

25] corresponds to ar

lncrementing (D betw

 

 

 

56

from the cross peak contour, plot (an vs. 0052 (or vice versa), the resulting

straight line will have a slope of Qam and an intercept of Gum). Solving Eqs. [11-

23] for AL.o and T will yield two equivalent sets of parameters (13).

This analysis can easily be extended to a rhombic HFI tensor, with an
anisotropic part [~T(1+d), -T(1—d), +2T]. The ENDOR frequencies in this case
are

(nu2 = wzazcosz® + (nyazsinzasinZCI) + (nxazsinzecoszd) [II—24a]
(0‘52 = szZCOSZG + myflzsinzesinZCD + (DXBZSiIIZGCOSZCD [II-24b]
where (ow) are the ENDOR frequencies corresponding to the canonical
orientation of B0 with respect to the principal axes of the HFI tensor and can
be expressed as mia(B):-(oliAfi/ 2; i=x, y, z. (9 and CD are the polar and azimuthal
angles specifying the orientation of B0 in the principal axis system (PAS) of
the HFI tensor. Eqs. [II-24] can be rewritten in the following form,
wanizliwzaanz ‘ commfkosze + (Oratnizlim [II-25]
where comm) designate the effective EN DOR frequencies in the transverse (xy)
plane and are defined as (compare to Eq. [10] in Ref. (13)),
wTa(B)=wxa(B)2C052(D + wyamfsinzdl [11‘26l
Note that Eq. [II-25] has the same form as Eq. [1] of Ref. (13), which describes
the axial case, except for the (D-dependence of (9mm Thus at a certain (1), Eq. [II-

25] corresponds to an axial tensor, yielding an arc-shaped correlation pattern.

Incrementing (1) between 0 and n/ 2 produces a series of arcs, which form a

 

 

 

 

 

 

and cross Peak
ma result of a nun
modicum" (see A

Despite its grnwi
illhas been utilized
rogers from those of
ittosynfliefic reaction
iiSEEM frequencies 1
‘lll Analysis of the co
bfldine complexes of
attributing to the hyn
previously. Pilbrow’s g
method (19). It is base
Mose the modulatit

tie intensity of cross 1
Enhanced. Héfer has i
disordered samples it
possible I is the most
appears to be useful ‘1
Eliminate the suppre:

Used HYSCORE to a

broadening effects of

 

 

 

 

 

57

horn-shaped cross peak pattern. This is illustrated in Figure II—10, which
shows a result of a numerical calculation done by the MATLAB program
”hyslineRom.m” (see Appendix).

Despite its growing popularity, HYSCORE has had few applications so
far. It has been utilized to isolate the ESEEM contributions of weakly coupled
nitrogens from those of the more strongly coupled ones in a bacterial
photosynthetic reaction center (15) and also to correlate unambiguously pairs
of ESEEM frequencies belonging to certain 14N couplings in a Fe-S protein
(16). Analysis of the contour line shapes in HYSCORE of imidazole and
histidine complexes of VO2+ was carried out by Dikanov and coworkers (18),
contributing to the hyperfine information obtained on these complexes
previously. Pilbrow’s group have proposed a new six-pulse 2D ESEEM

method (19). It is based on. augmenting the five-pulse sequence proposed to
increase the modulation depth (21) with an additional rt-pulse. As a result,
the intensity of cross peaks relative to that of diagonal peaks is significantly

enhanced. Hofer has examined the suppression effect in HYSCORE in
disordered samples in detail (14). He has shown that generally the shortest
possible I is the most beneficial. Thus HYSCORE with remote detection (15)
appears to be useful in some cases, and proposed a 3D version of ESEEM to
eliminate the suppression effect completely (14). Poppl and coworkers have

used HYSCORE to analyze the inhomogeneous and homogeneous line-

broadening effects of proton ESEEM signals in single crystals (17).

 

 

 

 

 

 

'g(MD-I-)
can

 

Figure “-10. Simulate
hyperfine interaction
MHZ, Aim=2 MIi-zl T:

Simulation was calcl

Appendix).

 

 

 

 

 

 

 

 

 

58
6 If _fi —1 —f l
5- u-
4— -
i“
23- l
20.7
2- -
1‘ i
o 4 1 1 l L
0 1 2 3 4 5 6

f1 (MHz)

Figure II-10. Simulated HYSCORE contour plot arising from a rhombic
hyperfine interaction in an S=1/ 2, 1:1/ 2 spin system. Parameters: w,/ 21::3
MHz, Aiso=2 MHz, T=1.2 MHz, 620.7, Gaussian linewidth-=01 MHz.

Simulation was calculated with the MATLAB program ”hyslineRom.m” (see

, Appendix).

 

 

 

 

 

 

 

‘4.

“—

p

>-

pa

:0

5:

L Merle,RP-l.;d

Barkhuijsen, H.;
Rtson. 1982, 50,

Hbfer, P.; Grup]
1986, 132, 279.

Gemperle, C.; A
38, 241.

MACSYMA, Int

Ernst, R. R.; B(
Magnetic Reson
Oxford, 1987.

Barkhuijsen, H
Reson. 1985, 61

Astashkin, A. '
Chem. Phys. Le

Arfken, Mathe

Press: San Die:

Hofer, P. In 1
EMARDIS-QI
pp. 1-15.

 

l

 

10.

"fimw“ 7

59

References

Merks, R. P. J.; de Beer, R. I. Phys. Chem. 1979, 83, 3319.

Barkhuijsen, H.; de Beer, R.; de Wild, E. L.; van Ormondt, D. I. Magn.
Reson. 1982, 50, 299.

Hofer, P.; Grupp, A.; Nebenfiihr, H.; Mehring, M. Chem. Phys. Lett.
1986, 132, 279.

Gemperle, C.; Aebli, (3.; Schweiger, A.; Ernst, R. R. I. Magn. Reson. 1990,
88,241. '

MACSYMA, Interactive Computer System, Mathlab Group, MIT, 1983.

Ernst, R. R.; Bodenhausen, (3.; Wokaun, A. Principles of Nuclear

Magnetic Resonance in One and Two Dimensions, Clarendon Press:
Oxford, 1987.

Barkhuijsen, H.; de Beer, R.; Pronk, B. J.; van Ormondt, D. I. Magn.
Reson. 1985, 61, 284.

Astashkin, A. V.; Dikanov, S. A.; Kurshev, V. V.; Tsvetkov, Yu. D.
Chem. Phys. Lett. 1987, 136, 335.

Arfken, Mathematical methods for physicist, 3rd Edition, Academic
Press: San Diego, 1985; pp. 794-823.

Hofer, P. In Electron Magnetic Resonance of Disordered Systems
EMARDIS-91; Yordanov, N. D., Ed.; World Scientific: Singapore, 1991;
pp. 1-15.

 

 

 

 

 

5. Kiss, H.; Ran
1995, 99, 436.

a

Shergill, I. K; I
1995, M51), 16.

 

q

. Poppl, A. ; Bétt
214.

13. Dikanov, S. A.
Chem. Soc. 199

19. Song, R.; Zhor
Phys. Lett. 199

20. Cho, H.; Pfer
Chem. Phys. L

21. Gemperle, C.
565.

 

 

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

60

Shane, I. J.; Hofer, P.; Reijerse, E. J.; de Boer, E. I. I. Magn. Reson. 1992,
99, 596.

Blinder, S. M. I. Chem. Phys. 1960, 33(3), 748.
Dikanov, S. A.; Bowman, M. K. I. Magn. Reson. 1995, Series A 116, 125.

Hofer, P. I. Magn. Reson., 1994, Series A 111, 77.

Kass, H.; Rautter, J.; Bonigk, B.; Hofer, P. ; Lubitz, W. I. Phys. Chem.
1995, 99, 436.

Shergill, I. K.; Joannou, C. L.; Mason, I. R.; Cammack, R. Biochemistry
1995, 34( 51 ), 16533.

Poppl, A.; Bottcher, R.; Volkel, G. I. Magn. Reson. 1996, .Series A 120,
214.

Dikanov, S. A.; Samoilova, R. 1.; Smieja, I. A.; Bowman, M. K. I. Am.
Chem. Soc. 1995, 117, 10579.

Song, R.; Zhong, Y. C.; Noble, C. J.; Pilbrow, I. R.; Hutton, D. R. Chem.
Phys. Lett. 1995, 237, 86.

Cho, H.; Pfenninger, S.; Gemperle, C.; Schweiger, A.; Ernst, R. R.
Chem. Phys. Lett. 1989, 160, 391.
Gemperle, C.; Schweiger, A.; Ernst, R. R. Chem. Phys. Lett. 1991, 178,
565.

 

 

 

11mm

to the identification
tibial ittatsively
mmdcautyti

httmo major groups
he that contain onl
Dmifaoibyio vulgaris
(llwhile the Desulfo
Wotype of the (Ni,

Enzyme is to catalyze
Sapmsed by the f(

Understanding the <
gain insight into the
hydrogen oxidizing
the study of the me
efficient means of I

In this Chap
are summarized. F

described briefly. ',

 

 

Chapter III

Ni-containing Hydrogenases

III.1. Introduction

 

Since the identification of the Ni—centers in hydrogenases, these enzymes
have been intensively studied from various types of biochemical approach:
molecular and catalytic properties, spectroscopy and genetics. There are at
least two major groups: the Ni-Fe-S type (with the subgroup Ni-Fe-S-Se) and
those that contain only [Fe-S] clusters. The hydrogenase from the bacterium
Desulfovibrio vulgaris is the most extensively studied Fe-only hydrogenase

(1), While the Desulfovibrio gigas hydrogenase may be considered as a

PTOtOtype of the (Ni, Fe)-hydrogenases (2). The function of the hydrogenase

enzyme is to catalyze the production or consumption of hydrogen gas, which

is expressed by the following equation:

HZwZH+ + 2e‘ [III-1]

Understanding the catalytic mechanism of these enzymes is of importance to

gain insight into the H-metabolism of certain bacteria (sulfate reducing,

hydrogen oxidizing, etc). Since hydrogenases are strikingly effective catalysts,

the study of the mechanism of action is also motivated by the need for

efficient means of producing hydrogen gas (3)-

In this Chapter some relevant results of previous hydrogenase research

' ' es are
are summarized. Redox and spectroscopic properties of these enzym

described briefly. Experimental approaches in the study of the Ni-center in

61

 

 

 

life) hydrogewses a
more of the Ni bind

monthwgigasl

W
lrhistory of the ider
ngrmd through se‘
litigants eutrophus 1
worth (4). In 1981, it
not eutrophus was t
ispresent in hydrogei
(m was found to be
filehydmgenases fro:
rtported by Lancaste
aSSigIment was cont
figrtals were observe
lll).ln1984 a secon:

rcported in the NA]

MW
Some general featu

conceptual model C

crystallographic st]

 

 

 

 

 

62

(Ni, Fe) hydrogenases are presented. Results and questions concerning the
structure of the Ni binding site are discussed, with the emphasis on EPR

studies on the D. gigas hydrogenase.

' 111.2. Evidence for the Involvement of Ni in Hydrogenases

The history of the identification of Nickel in certain hydrogenases has

progressed through several stages. In 1965 it was observed that the bacterium
Alcaligenes eutrophus (strains H1 and H16) required Ni for their autotrophic
growth (4). In 1981, it was demonstrated that the synthesis of hydrogenase in
the A. eutrophus was dependent on the presence of Ni. It was shown that Ni

is present in hydrogenase. In 1973, the activity of hydrogenase from Nocardia

 

opaca was found to be stimulated by nickel (5). In 1981, nickel was found in

. the hydrogenases from various bacteria (6). The EPR signal from Ni(III) was
reported by Lancaster in 1980 (7) in membranes of methanogenic bacteria. The
assignment was confirmed by isotopic substitution with 61Ni (8). Similar
signals were observed in hydroganases from D. gigas (9, 10) and C. oinosum
(11). In 1984 a secondary role for nickel in the binding of subunits was

reported in the NAD linked hydrogenase from N.opaca (12).

III.3.trutura1 m nnts fHdr na — nralMdl

Some general features of the hydrogenase molecule are summarized in the

conceptual model of Fig. III-1. This picture is mainly based on the x-ray

 

crystallographic structure of the D. gigas hydrogenase determined recently

 

 

   

Figure 1114. Schem

(1311-

 

 

63

Electron acceptor/
Hydrogen ase donor protein
Electron transfer pathway

3:

      
   

 

H'I-

Proton
channel

Docking site for
acceptor/
donor protein

Hydrogen~
activating site

       
 
 

:l . fife/A
NV , .. . .. J.;-1V Small
“ism-_.msdhrgaw ' “wwe;‘-~Jt‘et~1‘-‘¥-"~'*” subunit
Large
subunn

Figure III-1. Schematic general structure of (Ni, Fe) hydrogenases [from Ref,

(13)].

 

 

 

 

it) Most hydrogenase

MC center, see b

  
  
  
  
  
  
   

an Wm
shining active site
palatial docking site
actions The prop
aim acid residues I
sang a small, diffusi

hannel.

From this point on Vt
hydrogenase from E
and best characterize
refer to this enzyme
three states (15). Tht
enzyme after isolati
inactive toward hyd
hydrogen/ tritium e
to the EPR signal I)
10%) is in the 'read

signal Ni-B) is also

 

 

64

 

(14). Most hydrogenases consist of two subunits. The active site (a Ni-Fe
dimetallic center, see below) is buried inside the large subunit. The electron
pathway comprises three Fe-S clusters, arranged in a line from the Ni-
containing active site to the surface of the small subunit, where there is a
potential docking site for a c—type cytochrome, which acts as a source or sink of
electrons. The pr0posed proton channel consists of a chain of H+-carrying
amino acid residues leading to the surface of the large subunit. Hydrogen,
being a small, diffusible molecule probably does not require any special

channel.

111.4. Hydrogenase Activigg

From this point on we shall focus our discussion on the water soluble
hydrogenase from Desulfovibrio gigas, which is the most extensively studied
and best characterized of all Ni containing hydrogenases. All statements will

refer to this enzyme unless specified otherwise. This enzyme can exist in

 

 

 

three states (15). The predominant form (typically over 90%) of the oxidized
enzyme after isolation is termed the ’unready’ state. This is completely
inactive toward hydrogen as shown by hydrogen uptake (15) and

hydrogen/ tritium exchange studies ( 16). This state of the enzyme is assigned
to the EPR signal Ni-A (see below). Most of the rest of the enzyme (less than
10%) is in the ’ready’ state. This form of the enzyme (giving rise to the EPR

Signal Ni—B) is also inactive but can rapidly be reduced to the active state by

 

 

 

mtg reducing agents
:r‘mdy' state but 510

figure III~2 sho

nae Ni signals are
fippearance of the
font), followed by

lnEPR silent state i

 

npically observed w
irrelops (C, D). Afte
htained (B). At low

4Sldusters can be 01:

manuals
EPR spectra recorde
activation (2, 18) art
state was correlated
The hyperfine patte
thatNi-A arises fre
to room temperatu

(g: 2.33, 2.16, 2.02:

 

I if' 7
"""i.uii'ilit

 

65

strong reducing agents. The ’active' state is produced rapidly by reduction of
the 'ready’ state but slowly by reduction of the ’unready’ state.

Figure III-2 shows EPR spectra representing a typical sequence of events
detected during the reduction of D. gigas hydrogenase exposed to a H2
atmosphere. The spectra were recorded at 77 K. At this temperature EPR
signals of Fe—S clusters are not observable (they become dominant below 30 K,
where Ni signals are not discernible due to saturation). The first event is the
disappearance of the g=2.02 signal arising from the [3Fe-xS] cluster (not
shown), followed by the disappearance of signals Ni-A and N i-B (Fig. III-2A).
An EPR silent state is attained, at which a low intensity radical type signal is
typically observed whose origin is unknown (B). Next, the Ni—C signal
deve10ps (C, D). After a long incubation under H2 another EPR silent state is

obtained (B). At low temperatures (below 15 K) EPR signals from reduced [4Fe-

 

 

 

48] clusters can be observed at this stage (2).

 

111.5. EPR Spectra from Hydrogenases

EPR spectra recorded for the D. gigas hydrogenase at various states of
activation (2, 18) are shown in Fig. III-3. As already mentioned, the ’unready’
state was correlated with a signal named Ni-A (g: 2.31, 2.23, 2.02) (Fig. III-3a.).
The hyperfine pattern observed upon substitution with 61Ni (1:3/ 2) indicates
that Ni-A arises from a single Ni ion (8, 10). The signal remains detectable up
to room temperature. The ’ready’ state appears to correlate with signal Ni—B

(g: 2.33, 2.16, 2.02), which is present to a small proportion (less than 10%) in

 

 

 

 

E.P.R- RELQ'IIUE INTENSITY

Figure III-2. X-band

incubation under E

0f the EPR spectra '
Ref.(17)].

 

 

 

 

66

 

 

 

 

E. P.R. RELQ'IIUE INTENSITY

 

 

 

HRGNETIC FIELD ( T )

Fi - - .
gure III 2. X band EPR spectra of D. gigas hydrogenase at different stages of
I l I ' o

of the EPR spectra upon increasing the time of incubation under H [from
2.

Ref. (17)].

 

 

 

 

 

 

 

Figure III-3. Vario

hydrogenase. (a) Ni

recorded at 8 K; (e)

   

 

67

(d1

 

 

 

lillJJJlllllllJJLlllllJllJlllJJll

280 300 320 340
MRGNETIC FIELD (mTl

 

Figure III-3. Various X-band EPR spectra displayed by the D. gigas
hydrogenase. (a) Ni-A at 105 K; (b) Ni-B at 105 K; (C) NC at 30 K; (d) as (c)
recorded at 8 K; (e) as (c) after illumination, at 30 K; (f) reduced with H2 and

treated with co, 105 K. [from Ref. (17)].

 

 

 

is‘aolated enzyme.

till) oxidation state

 

A third type of
luiobserved and is :
”'mmdy state, the act
studies on the Ni-C 5;
oxidation and reduc
he enzyme (2, 19, 20)

An important 1
(11). Irradiation with
lithe EPR spectrum
oordination site. Th
11) K in the dark. Tl
kinetic isotope eflect,
monoxide (CO) is a
H,. Addition of CO

Spectrum assigned t

Species of C. oinosu

identical to that of

 

 

 

 

68

the isolated enzyme. Both Ni-A and Ni-B are postulated to represent the
Ni(HI) oxidation state (see below).

A third type of EPR signal, termed Ni-C (g: 2.19, 2.14, 2.02), has also
been observed and is shown in Fig. III—3c. During the slow activation of the
’unready’ state, the activity correlated with the appearance of Ni-C (18). Redox
studies on the Ni-C species (see below) show that it disappears upon both
oxidation and reduction, thus it represents an intermediate oxidation state of
the enzyme (2, 19, 20).

An important property of this Ni-C species is that it is light sensitive
(21). Irradiation with visible light at temperatures below 100 K causes a change
in the EPR spectrum (Fig. III-3e), which indicates a structural change in the Ni
coordination site. This process is reversed by raising the temperature to about
200 K in the dark. The rate of this photochemical reaction shows a strong
kinetic isotOpe effect, being nearly 6 times slower in D20 than in H20. Carbon
monoxide (CO) is a strong inhibitor for most hydrogenases, competitive with
H2. Addition of CO to the hydrogen reduced enzyme produces a transient
Spectrum assigned to a carbonyl species (22) (Fig. III-3f). Irradiation of the CO
species of C. vinosum hydrogenase produces an ‘EPR spectrum almost

identical to that of the irradiated Ni-C species (Fig. III-3e).

 

 

 

 

 

      
   
  
  
 
  

pinstripe of the
liltNi‘C being the ’
again distorted
lint Ni-B and Ni-C
[timid detect we

0. gigs hydrogenase
mible to the solv

Which indicates wea

EPR studies c
Alhracht et a1. (25) i
lltis number is Sign
coordination sphere
(EXAFS) spectrosco
hydrogenases, respe

The understa
mhanced by the re
the hydrogenase fr
Subunits of the bet
Various redox cofa

entirely within th

 

 

 

69

 

111.6. Coordination of the Ni-site

The lineshape of the Ni EPR signals all reveal a rhombic g-tensor anisotropy,
with Ni-C being the ’most axial’. The g values (g1, g2 > g3, while g3 E 2) may

suggest a distorted octahedral geometry for the Ni coordination sphere. The
Ni-A, Ni—B and Ni—C signals show no hyperfine splitting. However, pulsed
EPR could detect weak hyperfine interaction with one nitrogen nucleus in the
D. gigas hydrogenase (23). This ESEEM study also reveals that Ni ion is
accessible to the solvent only when the protein is in its Ni-C form. On
replacement of H20 by D20 a small decrease in the linewidth of the Ni-C
signal (5 G) was observed in C. vinosum and D. gigas hydrogenases (21, 24),
which indicates Weak interaction with exchangeable protons.

EPR studies on 3P’s-enriched Wolinella succinogenes hydrogenase by
Albracht et a1. (25) indicated hyperfine interaction with one sulfur nucleus.
This number is significantly lower than three or four sulfurs located in the Ni
coordination sphere, estimated by Extended X—ray Absorption Fine Structure
(EXAFS) spectroscopy for M. thermoautotrophicum and D. gigas
hydrogenases, respectively (26, 27).

The understanding of the structure of the Ni site has been greatly
enhanced by the recent publication of a crystal structure at 2.8 A resolution of
the hydrogenase from D. gigas by Volbeda et a1. (14). It shows that the two
subunits of the heterodimer are associated with each other, and that the
various redox cofactors are widely separated. The Fe-S clusters are located

O

entirely Within the small subunit, while the Ni site lies entirely within the

 

 

 

 

 

 

 

bye subunit. The Fe-1
Mlle-43} clusters a
'rear fashion spaced 1
h Ni center lies abo
hiest from the Ni 5
ipnds is a histidine
mngement of the F4
bot from the Ni cer
One of the mo
revelation that the N
and Fe (Figure III-4).
m the metal analysi:
(distinct from Ni) an
(appropriate for a fi
ligated by the four
these cysteines are
two cysteines bridg
tanbe described as

four cysteine resid

appears to be five-

modeled as H20 m

 

7O

 

large subunit. The Fe-S clusters found in the D. gigas enzyme are comprised
of 2 [4Fe—4S] clusters and a [3Fe-4S] cluster. These clusters are arranged in a

linear fashion spaced ca. 10 A apart with the [3Fe-4S] cluster in the middle.

 

The Ni center lies about 10 A from the proximal [4Fe-4S] cluster. The [4Fe-4S]

 

farthest from the Ni site (the distal cluster) is unique in that one of the Fe
ligands is a histidine imidazole. The surface exposed histidine plus the linear
arrangement of the Fe-S clusters suggest an electron transfer pathway leading
to or from the Ni center.

One of the most interesting aspects of the crystal structure is the

revelation that the Ni center is actually a dimetallic cluster composed of Ni

 

and Fe (Figure 111-4). The assignment of the second metal as an Fe was based

 

1 on the metal analysis, the strong anomalous scattering of Cu—K“ radiation

(distinct from Ni) and the electron density associated with the metal center
.(apprOpriate for a first row transition metal) (14). The dimetallic cluster is
ligated by the four conserved cysteines found in the large subunit. Two of
these cysteines are bound as terminal ligands to the Ni center. The remaining
two cysteines bridge between the Ni and Fe atoms. The structure of the Ni site

can be described as a highly distorted trigonal-pyramidal arrangement of the

 

four cysteine. residues. The Fe center lies ca 2.7 A from the Ni atom and

 

appears to be five-coordinate with three exogenous ligands, which were

 

modeled as H20 molecules. Recent IR data, however, identified bands in the

 

 

 

 

 

 

Cy853o

Figure III-4. The structure of the Ni-Fe cluster in the D. gigas hydrogenase.

 

 

 

 

 

Mm" region of the
ill) (73).

the iniormatior
hint the crystals e

motheenzymd
tnoontributedto a

nstallographic struc
haired from X-ray .
nm a number of di
mymes from phott
sulfate reducing bac‘
performed (30). The ‘
(XANES) data provi
(ll).

The EXAFS d
Practice of S~scattei
tiltibit evidence for
test by a set of four

roseopersicina., Whit

(32).

 

 

 

 

 

 

 

 

72

2000 cm'1 region of the spectrum that suggest triply bonded species such as CN'
or CO (28).

The information obtained from the crystal structure suffers from the
fact that the crystals employed were a mixture of at least three different redox
states of the enzyme (50% SI, 36% A and 14% B as defined below). This may
have contributed to a large amount of disorder in the Ni-Fe site. The
crystallographic structure of the Ni site can be compared with information
obtained from X-ray Absorption Spectroscopy (XAS) on redox poised enzymes
from a number of different species (29). A recent comparison involving
enzymes from photosynthetic bacteria (T. roseopersicina. and C. vinosum),
sulfate reducing bacteria (D. gigas and D. desulfuricans) and E. coli has been
performed (30). The Ni K-edge X-ray Absorption Near Edge Spectroscopy
(XANES) data provide a useful predictor of coordination number/ geometry
(31).

The EXAFS data from all of these hydrogenases is dominated by the
presence of S-scattering atoms at 222(1) A. Oxidized enzymes frequently
exhibit evidence for O, N ligation, but the reduced enzymes are most often fit
best by a set of four S—donor ligands. One exception is the enzyme from T.

roseopersicina., which exhibits evidence of O, N ligation at all redox 19‘7915

(32).

 

 

 

 

 

 

indition to signals
hEPRsignals from
hingame the sign
tiltspectrum below
tremble and readil
ndMagnetic Circula
titlel3l'e45] type.
No magnetic i
tile-45] clusters are t
hitemperatures is
was to be a redu

indicates that the 1‘81

tadlitate electron tr:

Mama:
llhas been found tl‘
Nemst type relatior
beltave quite differt
The Em/pH value f4
involves one proto

of the maximal act

 

 

73

111.7. Interaction Between Nickel and |Fe-S| clusters

In addition to signals arising from nickel, many of the (Ni, Fe) hydrogenases
give EPR signals from oxidized [Fe-S] clusters (6, 17, 33). In the case of D. gigas
hydrogenase the signal is almost isotropic, at g=2.02, which dominates the
EPR spectrum below 30 K. At this temperature the Ni-A signal is barely
discernible and readily saturated. It has been demonstrated by Mossbauer (34)
and Magnetic Circular Dichroism (35) spectroscopies that the [Fe-S] cluster is
of the [3Fe-4S] type.

No magnetic interactions of the Ni-A or Ni-B species with the oxidized
[3Fe-4S]‘ clusters are detected. On the other hand, splitting of the Ni-C signal at
low temperatures is observed.(Fig. III-3d) (2). The interacting paramagnet
appears to be a reduced [4Fe-4S] cluster (20). This spin-spin interaction
indicates that the redox centers are located close enough in the enzyme to

facilitate electron transfer.

1118. Redox Pr erties fH dr na es

It has been found that the rate of H2 production by hydrogenases follows a
Nernst type relationship with the applied redox potential. Hydrogenases
behave quite differently, some being pH dependent, others pH independent.
The Em/ pH value for the D. gigas hydrogenase indicates that the redox process
involves one proton and one electron (Em is the redox potential at which half

of the maximal activity is observed).

 

 

 

 

 

 

 

he E, values must
redox centers pres
initial correspondii
«hates to the situe
inns are equal. Thee
'ilit redox titration"
sides is followed as
teed stepwise by
mylne (2, 20, 36, 37
gigs hydrogenase is
~70 mV midpoint pa
with a redox proces:
wasshown to be pH
with the Nernst eqt
associated with the
develops at a poten
680 mV and comp
The redox cl
Iiigure Ill~6. The p
Electrode (NI-IE). '1

understood irreve‘

and/ or incorporat

 

 

74

 

These EIn values must be related in some way to the midpoint potentials of

the redox centers present in the enzyme. Midpoint potential is defined as the
potential corresponding to the inflection point of the S-shaped titration curve
and refers to the situation when the amounts of the oxidized and reduced
forms are equal. These midpoint potentials can be obtained by the method of
"EPR redox titration”. This means that the intensity of the EPR signal of the
species is followed as a function of the redox potential of the system, which is
changed stepwise by adding the reducing agent (dithionite or H2) to the
enzyme (2, 20, 36, 37). The result of such an experiment performed with D.
gigas hydrogenase is displayed in Fig. III-5 (9, 17, 38). The 3Fe center titrated at
-70 mV midpoint potential. The disappearance of the Ni-A signal is associated
with a redox process with the midpoint potential of -220 mV. This process
was shown to be pH dependent (-60 mV per pH unit) and the data were fitted
with the Nernst equation for a one-electron reduction (9). Redox processes
associated with the Ni-C signal were also studied by titrimetry (2). Ni-C
develops at a potential of about -270 mV, reaches maximum intensity at about
-380 mV and completely disappears below -450 mV.

The redox chemistry associated with the Ni site is summarized in
Figure III-6. The potential values are relative to the Normal Hydrogen
Electrode (NHE). The slow conversion of form Ni-B to form Ni-A is a poorly
understood irreversible process, which may involve a conformational change

and / or incorporation of Oz. The presence of O atoms in the oxidized forms

 

 

 

 

 

 

 

 

        

Figure III-5. EPR r:
g=2 02 Signal, 4 K.
K; open circles: Ni-

ul‘v

5 I'llilo a.“ tn: .3.

A '8 flu.-
’.FII~I..I.IL.L \I‘ZflVI“

 

 

 

 

 

)

SIGNQL INTENSITY

R.
(arbitrary units

E.P.

75

 

 

 

 

 

’d’x\ x“ T
I, 9' \ ‘ a
y p \ Q
I I, 8 t .\ o
1 I \ \
’l 6' \\ \
O \
I. e \ \
3 9'8," “ Q
/ Q i \
I I, \ \\
I O\ \
'1‘ \ \\
‘I . o ‘ \ r
I \
M g8~ ‘2
' ‘ c‘g v¥.%—'/ W..- w. ' 1
“400 " 350 " 306 " 200 0 200

REDOX POTENTIQL ( MU )

Figure III-5. EPR redox titrations of the D. gigas hydrogenase. Full squares:

g=2,02 signal, 4 K; open squares: Ni-A signal, 77 K; full circles: Ni-C signal, 20

K; Open circles: Ni-C signal, 4 K [from Ref. (17)]-

 

 

 

 

 

 

 

 

 

coUl
0nd

(5"?

Om
{5‘ II

9?

p—d—l

Figure III-6. Schem

(inter in hydrogEI

 

 

 

E9J1D..A.ox
coupled epr
oxidized. inactive

 

 

1L+150 mV

 

 

Lerner;

g = 2.31 , 223. 2.02
oxidized, inactive

 

 

i—w

 

 

Farm St "
epr silent

 

 

 

 

Form C u
g = 2.19, 2.14. 2.02
reduced, inactive

 

 

76

 

 

£093 8 ox
coupled epr
oxidized. inactive

 

1L+150 mV

 

 

Format.

9 = 2.33, 2.16. 2.01
oxidized. inactive

 

immv

 

 

Form 8! r
epr silent

 

1i-330mv
’ .

 

 

Form C r
coupled epr
reduced, active

 

 

 

 

epr silent
fully reduced

 

 

 

cluster in hydrogenases.

Figure III-6. Scheme describing the redox chemistry associated with the Ni-Fe

 

 

 

 

 

 

 

 

 

tabeen directly obs
W2 (39). Oxidativ
hnative species Ni
ozyme (40). At mor
rated that is couph
sigrral oi the oxidize
hateelectron to p
oistmoe of active a
imported by a stud
B(41). When form 1
form A is observed
and the resulting 8]
produced upon oxi
Sisamples produce

Several redo
proposed, example
provide simple ex]
of5=1l2 Ni EPR s
the Ni center, whe
+3 to 0). Scheme A
in Ni, and that tl

by thiolate ligatic

 

—. in?" -

 

 

77

has been directly observed via line broadening of EPR signals in the presence
of 17O2 (39). Oxidative titrations of the C. vinosum hydrogenase revealed that
the native species Ni-A and Ni—B are not the fully oxidized forms of the
enzyme (40). At more positive potentials (+150 mV) a new 821/ 2 center is
created that is coupled to the EPR signals Ni-A and Ni—B as well as to the EPR
signal of the oxidized [3Fe-4S] cluster. Both species Ni—A and Ni-B are reduced
by one-electron to produce an EPR Silent Intermediate (SI) (41, 42). The
existence of active and inactive conformations of the enzyme is further

supported by a study of the reductive interconversion of species Ni—A and N i-

 

B (41). When form A is reduced at 4 0C to the SI level and reoxidized, only
form A is observed as a product. However, when the reduction is carried out
and the resulting SI state is allowed to equilibrate at 30 0C, form B is also
produced upon oxidation. A similar temperature dependence is observed for

SI samples produced from form B.

 

Several redox schemes to explain these observations have been
proposed, examples of which are summarized in Table HI-l. Schemes A and B
provide simple explanations for the sequential appearance and disappearance
of S=1/ 2 Ni EPR signals. Scheme A utilizes one-electron redox chemistry at
the Ni center, whereas Scheme B reduces Ni through four redox states (from
+3 to 0). Scheme A has the advantage that it utilizes only two oxidation states
for Ni, and that the unusual oxidation state involved (+3) is the one favored

by thiolate ligation.Scheme B implies unprecedented redox chemistry for a

 

 

.mUmmEONOthwr— r: uwumS—u wmlmz gt uCu _umeCCuC gEmr—um ~8va .HA: mus—mafia.

 

 

 

78

 

 

 

Even .5sz 80m .3sz AOVMZ QCMZ

 

 

25s it
SE .892 Sam ”82 8.2 892 SN .2 .m 53. n m o snow
8ch .892 Ed 82 €sz fizz . 2.8 a
060m .892 Sam .82 252 :52 SN .8 .m «3 ..... w m 8.5m
SE .252 SE .62 Suez 25% mod .mg .13. n m < 8.5m
Q o m a. l
323%. 2239.4 mfimfim 2223“. Emma. Sc M R 382%.

@358onch E $353 9132 9: cos. pmmomoud mmEmaum xopmm HA: mum—4:.

 

 

 

style Ni center that
nnpms the Ni(III/
opened by about 2
bprovided by a mo
reduction of Ni(III)
'lhe redox act
lleasiremmt of the
redox levels provid
dersity on the Ni c.
ashiit of ca. 2 eV ir
instigated was th
reduced samples g
inscription of the
However, EXAFS (
function of redox 1
three electrons to g
delocalized over 5
Another po
reSponsible for thr
(Scheme C) or Ni(‘
Vary the Ni oxida

coupled antiferro

silent states. The:

 

 

 

79

single Ni center that must involve considerable structural change in order to

compress the Ni(III/II) and Ni(II/ I) redox couples, which are normally
separated by about 2 V, into a z250 mV range. This kind of structural change

is provided by a model system where Ni-S bonds lengthen by 0.14 A upon
reduction of Ni(III) to Ni(II) (43).

The redox activity of the Ni center has been examined by XAS.
Measurement of the differences in the Ni K-edge energy between the various
redox levels provides a sensitive method for detecting changes in the charge
density on the Ni center (44). For a one-electron metal-centered redox process,
a shift of ca. 2 eV in the edge energy is expected. In none of the cases
investigated was the edge energy shift between the fully oxidized and fully
reduced samples greater than 1.5 eV, ruling out Scheme B as a realistic
description of the redox chemistry. Scheme A cannot be ruled out entirely.
However, EXAFS data reveal no significant changes in the Ni-S distances as a
function of redox poise. Thus, the redox chemistry, which requires at least
three electrons to go from the oxidized to the reduced enzyme is likely to be
delocalized over several atomic centers in the Ni-Fe cluster.

Another possibility is that the Fe atom in the Ni-Fe cluster is primarily
responsible for the redox chemistry (Table III-1). These schemes feature Ni(I)
(Scheme C) or Ni(III) (Scheme D) to account for the EPR signal, but do not
vary the Ni oxidation state. Instead, the S=1/ 2 Ni centers are proposed to be
coupled antiferromagnetically to S=1/2 Fe centers to account for the EPR

silent states. These schemes have several advantages over the ones involving

_;

 

 

 

 

 

adusively Ni. First,
redox center in biolc
ha in that the strut
Midi. They are also
lobar). Nevertheles
states for Ni and he
couple, they also j
hyperfine splittings
lender in labeled
imply transfer the

A third poss
nthe redox proces
and not a particula
dithiolato dinickel

ligation (44—46).

111.9. The 39!: of [
Another aspect of
the use of Ni as tl
C0). Such a role i
photochemical re

hyperfine coupliI

 

 

80

exclusively Ni. First, Fe has a facile redox chemistry and is Widely found as a
redox center in biology. Second, these schemes are consistent with the XAS
data in that the structure and charge density of the Ni center need not change
much. They are also consistent with a role for Fe as a CO binding site (see
below). Nevertheless, both Scheme C and D still suggest unusual oxidation
states for Ni and have many of the problems found in Scheme A and B. For
example, they also provide no compelling explanation for the absence of 57Fe
hyperfine splittings or line broadening in the EPR spectra arising from the Ni-
Fe center in labeled hydrogenase samples (34). In essence, Scheme C and D
simply transfer the redox processes from Ni to Fe.

A third possibility is that the thiolate ligands are ”intimately involved”

in the redox process and that the redox chemistry is a property of the cluster
and not a particular metal center._This possibility is supported by studies of p.-

dithiolato dinickel(II) model complexes that also feature terminal thiolate

ligation (44-46).

111.9. Th R l f Ni as a Bindin Site

Another aspect of many hypothetical reaction mechanisms for hydrogenase is
the use of Ni as the binding site for substrate (H2 or H') and inhibitors (e.g.
CO). Such a role is suggested by two properties of the enzyme: the
photochemical reactivity of Ni—C (see above) and the observation of 13C

hyperfine coupling in the Ni-C signal at 77 K in the presence of 13CO (39).

 

 

 

 

 

—‘—

However, recent in
afar H2, or at leas

The photoche
eta ligand photodr
mdianism involve
obmation of the l
plotodissodation o
ENDOR and XAS s
Impersicina (47). ‘
hthe enzyme con
with a coupling co:
anda more strong
exchangeable with

exchangeable prot

 

reappear. This ob:
and recombinatio:
C0 as a ligand gi‘
Cdistance and a
EXAFS spectrum
with a ligand 31'
the hydrogenase
second coordina

EXAFS spectrui

 

 

81

 

However, recent investigations suggest that Ni may not serve as the binding
site for H2, or at least is not the only binding site.

The photochemical behavior. of the Ni-C species has the characteristics
of a ligand photodissociation and recombination reaction. One possible
mechanism involves an S=1/2 Ni-H’ complex, which is consistent with the
observation of the large isotope effect mentioned earlier. The
photodissociation of a Ni-H species has been examined by a combination of
ENDOR and XAS studies of the reaction using hydrogenase from T.
roseopersicina (47). The 1H—ENDOR spectrum reveals two sets of resonances

in the enzyme corresponding to protons that were not solvent exchangeable
with a coupling constant of 12-14 MHz (assigned to cysteine B-CHzprotons),

and a more strongly coupled (20 MHz) proton resonance arising from protons
exchangeable with D20. Exposure to light causes the resonance due to the
exchangeable proton to vanish, and annealing causes the resonance to
reappear. This observation is consistent with the photochemical dissociation
and recombination of a Ni-H species, but do not exclude other mechanisms.
CO as a ligand gives rise to a rich EXAFS spectrum that arises from a short M-
C distance and a second coordination sphere O atom. Examination of the
EXAFS spectrum of a crystallographically characterized Ni—CO complex (47)
with a ligand environment similar (with respect to EXAFS) to that found for
the hydrogenase Ni center clearly showed the short Ni-C interaction and the
second coordination sphere O atom. No such features can be observed in the

EXAFS spectrum of hydrogenases. These observations implicate the Fe center

 

 

 

 

 

 

 

ititN'rPe cluster

(caplet, because

form C, the EPR si
isnot a reactive in
presence of water,
of H2. It is more 1i
an enzyme substrai

the Ni-Fe cluster.

Mammary
The crystal structi
heterodimetallic
Which have been
reassessed. The

hYdrogenases ob

 

 

82

 

of the Ni-Pe cluster as the binding site for the CO ligand. This conclusion is
further supported by recent Mossbauer studies of reduced C. oinosum
hydrogenase (30).

Because CO is a competitive inhibitor, the Fe atom is also a likely
binding site for H2 or H‘. This suggestion is supported by the discovery that
the unique IR features associated with the Fe site are also found in Fe-only
hydrogenases (48). An Fe-H2(H') complex is also consistent with the observed
isotope effect. On the other hand, it is unlikely that species Ni-C is an H2 or H'
complex, because when H2 is carefully removed from a sample containing
form C, the EPR signal Ni-C is stable indefinitely (49). This proves that form C
is not a reactive intermediate, since the hydride should convert to H2 in the
presence of water, while an H2 adduct should dissociate at low partial pressure
of H2. It is more likely that form C is an enzyme product complex, rather than
an enzyme substrate complex; in other words it has protons associated with

the Ni-Fe cluster.

111.1(2. Summary

The crystal structure obtained from D. gigas hydrogenase reveals a
heterodimetallic Ni-Fe active site. The interpretation of the Ni EPR data,
Which have been based on the assumption of an isolated Ni center, have to be
reassessed. The basic structural features seem to be similar in several

hydrogenases obtained from a variety of bacteria. The largest difference is the

 

 

 

 

 

 

 

thiolate ligands. A
by IR studies, in w
on the redox potei
supported by the i
tidithiolato and it
substrate and inhi
he existence of N
hl’drogenase imp]
Given that
hYdtogenases is
hat one of the
iITeversibly inac

fact that Fe-only

 

 

83

 

presence of O, N ligation in many enzymes as opposed to the D. gigas
hydrogenase, in which the active site shows pure S—coordination.

The redox chemistry of hydrogenases has been probed by‘several
techniques, none of which provide unambiguous evidence for Ni centered
redox chemistry. ENDOR spectrosc0py reveals that the resonances associated
with cysteinate methylene protons exhibit the same coupling constant in
forms A, B and C, suggesting that all three of these forms feature the same
oxidation state of the Ni-Pe cluster, in agreement with XAS experiments.

Several explanations for the redox chemistry of the cluster are possible
including redox activity at the Fe center and/ or redox processes involving the
thiolate ligands. A participation of Fe in the redox chemistry is also suggested
by IR studies, in which the frequencies of bands from the Fe ligands depend
on the redox potential. Mechanisms involving thiolate redox chemistry are

supported by the redox behavior of a series of dinickel model complexes with
tt-dithiolato and terminal thiolate ligation. Studies on the role of Ni as a

substrate and inhibitor binding site fail to provide unequivocal evidence for
the existence of Ni-H or Ni-CO complexes. Recent studies on C0 complexes of
hydrogenase implicate Fe as a potential binding site for CO.

Given that one of the outstanding properties of the Ni—containing
hydrogenases is their stability to irreversible oxidation by 02, it seems possible
that one of the functions of Ni is to modify the catalytic site so that it is not
irreversibly inactivated upon exposure to 02- This notion is supported by the

fact that Fe—only hydrogenases are much better catalysts of H2 redox chemistry,

 

 

 

 

 

 

 

 

84

but are irreversibly deactivated by 02. The stability toward 02 seems to be

increased further by the substitution of a terminal Ni-cysteinate ligand by

selenocysteinate.

 

 

 

 

Lancaster, 1.
Albracht, S.

Cammack, ]
1982,142, 28

. Moura, I. ].

DuVamey,
LeGall, I. E

Albracht, S
1983, 724,3

Schneider,
Cammack

Volbeda,
Fontecilla

Fernande
Biophys.

Hallahan
Biochim.

 

 

10.

11.

12.

13.

14.

15.

16.

85

References

Huynh, B. H.; Czechowski, M. H.; Kruger, H.-].; DerVartanian, D.
V.; Peck, H. D., In; LeGall, J. Proc. Natl. Acad. Sci, USA 1984, 81, 3728.

Teixeira, M.; Moura, 1.; Xavier, A. V.; Huynh, B. H.; DerVartanian,
D. V.; Peck, H. D., Jr.; LeGall, J.; Moura, I. J. G. ]. Biol. Chem. 1985,
260, 8942.

Benemann, I. R. Proc. 10th World Hydrogen Energy Conf. (Cocoa
Beach, Florida; 1994).

Bartha, R.; Ordal, E. I. 1. Bacteriol. 1965, 89, 1015.

Aggag, M.; Schlegel, H. (3. Arch. Microbiol. 1973, 88, 299.
Cammack, R.; Fernandez, V. M.; Schneider, K. In The Bioinorganic
Chemistry of Nickel, Lancaster, J. R., In, Ed., VCH: New York, 1988;
Chapter 8.

Lancaster, J. R. PEBS Lett. 1980, 115, 285.

Albracht, S. P. J.; Graf, E.-G.; Thauer, R. K. FEBS Lett. 1982, 140, 311.

Cammack, R.; Patil, D. S.; Aguirre, R.; Hatchikian, E. C. FEBS Lett.
1982,142, 289.

Moura, I. I. G.; Moura, I.; Huynh, B. K.; Kruger, H.-].; Teixeira, M.;
DuVarney, R. C. ; DerVartanian, D. V.; Xavier, A. V.; Peck, H. D., In;
LeGall, J. Biochem. Biopys. Res. Commun. 1982, 408, 1388.

Albracht, S. P. J.; Kalkman, M. L.; Slater, E. C. Biochim. Biophys. Acta

1983, 724, 309.

Schneider, K.; Schlegel, H. (3.; Iochim, 1<. Eur. J. Biochem.1984,138, 533.

Cammack, R. Nature 1995, 373, 556.

Volbeda, A.; Charon, M—H.; Piras, C.; Hatchikian, E. C.; Frey, M.;
Fontecilla-Camps, I. C. Nature 1995, 373, 580.

Fernandez, V. M.; Hatchikian, E. C.; Cammack, R. Biochim.
Biophys. Acta 1985, 832, 69.

Hallahan, D.; Fernandez, V. M.; Hatchikian, E. C.; Cammack, R.
Biochim. Biophys. Acta 1986, 874, 72.

 

 

 

1'. hloura, l. l. C-

Chemistru of N
Chapter 9.

‘3 Fernandez, \'.

Biochim. Blopl

'i Kojima, N; Ft

l\'.' H.; Walsh

‘1 Cammack, R.

Biochim. Biop

‘1. van der Zwa.

I‘i

FEBS Left. 19

van der Zwa
Biochim. Bio;

- Chapman, A

FEBS lett. 1‘.

‘1- DerVartania

(4.

>4

_xo'

Qulm. 1985,

, Albracht, 8.

BMW, R;
116.

Lindahl P

I .

Walsh. c. -

‘* Scott, R. A

LeGall, ] 1

'I

‘ Bagley, K.

W- H. Bioc

EidshESS,
Chemistry
Chapter 4

Mammy,
Zh' ln TTt
K'I EdS.; p

 

86

Moura, I. I. G.; Teixeira, M.; Moura, I.; LeGall, I. In The Bioinorganic
Chemistry of Nickel, Lancaster, I. R., Jr., Ed., VCH: New York, 1988;
Chapter 9.

Fernandez, V. M.; Hatchikian, E. C.; Patil, D. S.; Cammack, R.
Biochim. Biophys. Acta 1986, 883, 145.

Kojima, N.; Fox, I. A.; Hausinger, R. P.; Daniels, L.; Orme-Iohnson,
W. H.; Walsh, C. Proc. Natl. Acad. Sci. USA 1983, 80, 378.

 

Cammack, R.; Patil, D. S.; Hatchikian, E. C.; Fernandez, V. M.
Biochim. Biophys. Acta 1987, 912, 98.

van der Zwaan, I. W.; Albracht, S. P. J.; Fontijn, R. D.; Slater, E. C.
PEBS Lett. 1985, 179, 271.

 

van der Zwaan, I. W.; Albracht, S. P. I.; Fontijn, R. D.; Roelofs, Y. B. M.
Biochim. Biophys. Acta 1986, 872, 208.

 

Chapman, A.; Cammack, R.; Hatchikian, C. E.; McCracken, J.; Peisach, I.
PEBS Lett. 1988, 242, 134.

 

24. DerVartanian, D. V.; Kruger, H.-I.; Peck, H. D., Jr.; LeGall, I. Rev. Port.
Quim. 1985, 27, 70.

25. Albracht, S. P. J.; Kroger, A.; van der Zwaan, I. W.; Unden, G.;
Bocher, R.; Mell, H.; Fontijn, R. D. Biochim. Biophys. Acta 1986, 874,
116.

26. Lindahl, P. A.; Kojima, N.; Hausinger, R. P.; Fox, I. A.; Teo, B. K.;
Walsh, C. T.; Orme—Iohnson, W. H. I. Am. Chem. Soc. 1984, 106, 3062.

27. Scott, R. A.; Wallin, S. A.; Czechowski, M.; DerVartanian, D. V.;
LeGall, J.; Peck, H. D., In; Moura, I. I. Am. Chem. Soc. 1984, 106, 6864.

28. Bagley, K. A.; Duin, E. C.; Roseboom, W.; Albracht, S. P. J.; Woodruff
W. H. Biochemistry 1995, 34, 5527.

I

 

29. Eidsness, M. K.; Sullivan, R. J.; Scott, R. A. In The Bioinorganic
Chemistry of Nickel, Lancaster, I. R., Ir., Ed.; VCH: New York, 1988;
Chapter 4.

30. Maroney, M. J.; Allan, C. B.; Chohan, B. S.; Choudhury, S. B.; Gu,

Zh. In Transition Metal Sulfur Chemistry, Stiefel, E. 1.; Matsumoto,
Kw Eds; ACS Symposium Series 653, 1996; Chapter 4.

—

31 Colpas, G. 1.; h

S;Suflx 8.1”}:

3‘. Bagn'nka, C.;

05,3576.

SE Cammack, R.;

Transport and
Holland: New

3:. Kriiger, H.-].;

Derl’artanian
lloura, I. I. C

‘5 Johnson, M. l

Derl’artaniar
Chemistry, X.

‘1 Lissolo, T.; l

liege, R.—M.

5;)

g3

fi-S

_._
,_4

it;

_ES

3

Teixeira, M,‘
l-; Peck, H. I
130, 481.

van der Zw.
Albracht, S,

Surems’ K.
M; Duin, E
4980.

Albracht, S
ROberts, R.
Kruger. H.

Mammy.
Curbiell R

' ZCOlPas, G

29. 4779.

fi

87

Colpas, G. J.; Maroney, M. J.; Bagyinka, C.; Kumar, M. ; Willis, W.
S.; Suib, S. L.; Mascharak, P. K.; Baidya, N. Inorg. Chem. 1991,30, 920.

Bagyinka, C.; Whitehead, I. P.; Maroney, M. I. I. Am. Chem. Soc. 1993,
115, 3576.

Cammack, R.; Lalla-Maharajh, W. V.; Schneider, K. In Electron
Transport and Oxygen Utilization, Ho, C., Ed.; Elsevier North
Holland: New York, 1982; pp. 411-415.

Kruger, H.-].; Huynh, B. H.; Ljungdahl, L. 0.; Xavier, A. V.;
DerVartanian, D. V. ; Moura, I.; Peck, H. D., In; Teixeira, M. ;
Moura, I. I. (3.; LeGall, I. I. Biol. Chem. 1982, 257, 14620.

Johnson, M. K.; Zambrano, I. C.; Czechowski, M. H.; Peck, H. D., Ir.;
DerVartanian, D. V.; LeGall, I. In Frontiers in Bioinorganic
Chemistry, Xavier, A. V., Ed., VCH: Weinheirn, 1986; pp. 36—44.

 

Lissolo, T.; Pulvin, S.; Thomas, D. I. Biol. Chem. 1984, 259, 11725.

 

Mege, R.-M.; Bourdillon, C. I. Biol. Chem. 1985, 260, 14701.

38. Teixeira, M.; Moura, 1.; Xavier, A. V.; DerVartanian, D. V.; LeGall,
J.; Peck, H. D., Ir.; Huynh, B. H.; Moura, I. I. G. Eur. I. Biochem. 1983,
130, 481.

39. «van der Zwaan, I. W.; Coremans, I. M. C. C.; Bouwens, E. C. M.;
Albracht, S. P. I. Biochim. Biophys. Acta 1990, 1041, 101.

 

40. Surerus, K. K.; Chen, M.; van der Zwaan, I. W.; Rusnak, F. M.; Kolk,
M.; Duin, E. C.; Albracht, S. P. I. Muenck, E. Biochemistry 1994, 33,
4980.

41. Albracht, S. P. I. Biochim. Biophys. Acta 1994,1188, 167.

42. Roberts, R. M.; Lindahl, P. A. I. Am. Chem. Soc. 1995, 117, 2565.

43. Kruger, H.-].; Holm, R. H. I. Am. Chem. Soc. 1990, 112, 2955.

44. Maroney, M. J.; Pressler, M. A.; Mirza, S. A.; Whitehead, I. P.;
Gurbiel, R. J.; Hoffman, B. M. Adv. Chem. Ser. 1995, 246, 21.

45. ”Colpas, G. J.; Kumar, M.; Day, R. O.; Maroney, M. J. Inorg. Chem. 1990,
29, 4779.

 

 

g Mirza, 5. A.; PIC
lllng. Chem. 191

Manley, M. I.

t wander Spek, T
Bagley, K A.; 81
Biochem. 1996, .'

l. Bamndeau, D. ‘
1994, 116, 3442.

 

 

 

88

Mirza, S. A.; Pressler, M. A.; Kumar,- M.; Day, R. O. ; Maroney, M. J.
Inorg. Chem. 1993, 32, 977.

Whitehead, I. P.; Gurbiel, R. J.; Bagyinka, C.; Hoffman, B. M.;
Maroney, M. I. I. Am. Chem. Soc. 1993, 115, 5629.

van der Spek, T. M.; Arendsen, A. F.; Happe, R. P.; Yun, S. ;
Bagley, K. A.; Stufkens, D. J.; Hagen, W. R.; Albracht, S. P. I. Eur. I.
Biochem. 1996, 237, 629.

Barondeau, D. P.; Roberts, L. M.; Lindahl, P. A. I. Am. Chem. Soc.
1994, 116, 3442.

 

 

 

 

Syntl

Syrtetic compoun
to active site of hy
mine and fun
noidered as a m
lshrws similar
introduces the
similar to that of
structural models,
as functional moc
Before X-ra
gigas hydrogenm
the (Ni, Fe) hydr
llrere are still u:
While the mechi
the oxidation st
binuclear cluste
results in hyper

line broadeninl

peculiar unpai

 

Chapter IV

Synthetic Models for the Ni-site of Hydrogenases

 

IV .1. Introduction

Synthetic compounds that mimic the chemical and / or physical properties of
the active site of hydrogenases have contributed to the understanding of the
structure and function of the catalytic site. A particular Ni-complex may be
considered as a model if it fulfills at least one of the following conditions: (a)
it shows similar spectroscopic properties to those of the enzyme; (b) it
reproduces the redox properties of the enzyme; (c) it has catalytic activity
similar to that of hyrogenases. Compounds with properties (a) are called
structural models, while those which have properties (b) or (c) are referred to
as functional models (1-3).

Before X-ray crystallographic data showed that the active site of the D.
gigas hydrogenase was most likely a N i-Fe cluster (4), the catalytic center in
the (Ni, Fe) hydrogenases had been postulated to be a mononuclear Ni-site.
There are still unanswered questions concerning the nature of the active site,
while the mechanism of the catalysis is unclear. One fundamental problem is
the oxidation states of the ions involved and electronic structure of the

binuclear cluster. Interesting, that while 61Ni enrichment in the protein
results in hyperfine splitting in the EPR signal (a 20 G), no splitting or even

line broadening is observed upon 57Fe enrichment (5), which suggests a

peculiar unpaired electron distribution involving mostly the Ni metal and

89

 

 

 

 

 

whirled cystein}
reamed nafi'
mispecies (e.g.
llllS far on the assu

pamgnet, must 1
11011) complexes II

studies of which c

 

Only two thiolate

namely the tetran
DACO)H2 = N,N‘
llNi(dmpn)}3Fe]2‘
mercaptoethyl)-1
to compare the 1
complexes with
lack of structur:
drawing firm c4
Ni ~ Fe distance
distances obsel

organization 0

 

F ll' Int“ 7 7

 

90

coordinated cysteinyl S atoms. Another problem is the relationship between
the determined native structure of the active site and the structure of the

l reduced species (e. g. Ni-C). Certainly, EPR and other spectroscopic data, based
thus far on the assumption of the active site being an isolated Ni(III) or Ni(I)

l paramagnet, must be reinterpreted. On the other hand, some mononuclear

‘ Ni(III) complexes may still be considered as relevant models, spectroscopic
studies of which could provide useful information on this unusual oxidation
state. Synthetic modeling strategies in the future will probably focus on

binuclear Ni-Fe or Ni-Ni thiolate complexes.

IV.2. Polmuclear Ni complexes

 

Only two thiolate-bridged Ni-Fe complexes have appeared in the literature,
namely the tetranuclear species, [{Ni(BME-DACO)FeCl}2(u-Cl)2], where (BME-

DACO)H2 = N,N’~bis(mercaptoethyl)-1,5-diazacyclooctane (Figure IV ~1A) and

llNi(dmpn)}3Fe]2*, where (dmpn)H2 = N,N’—dimethyl—N,N’-bis(2-

 

mercaptoethyl)-1,3-diaminopropane (Figure IV-lB) (6, 7). While it is tempting
to compare the Ni—S and Fe—S bond lengths in these and other Ni and Fe
complexes with those of the Ni-Fe cluster in the hydrogenase structure, the
lack of structural data on relevant Ni(III) model complexes prevents us from
drawing firm conclusions about the electronic character of the Ni-Fe site. The

Ni ~ Fe distance of 2.7 A is at the low end of the 2.6 - 3.9 A range of Ni - Ni
distances observed for synthetic [Ni2(u-SR)2]2+ complexes (1) and suggests pre-

Organization of the Ni-Fe center for the inclusion of a third bridging ligand.

 

—

 

Figure IV-1. Sche
It [{Ni(BME-DA‘
bis(rnercaptoethj
(dmpn)H2 = NJ
C: [Ni2(rnernta)2

SCsHalal217 E: [b

 

 

Ni
Cl Ni N'
Sew"Ir‘e““"‘Cl""F'tzt-s-gl 93‘
Ls’ ‘cr’r NOS-We «m
u . .
N‘N ’SN
-'- ruse-"N" 3" "
3/ 1 g\ _l g .1
as... / ”an .. ._
C “1" RSan NI‘SR "MSU“ an
I I» h h
N! «A: VNi WI
(S’q/ \: E F
v
D

Figure IV-1. Schematic structures of model complexes referred to in the text.
A: llNi(BME-DACO)FeCl}2(u-Cl)2], where (BME-DACO)H2 = N,N’~
bis(mercaptoethyl)—1,S-diazacyclooctane; B: [(Ni(dmpn)]3Fe]2+, where
(dmpn)H2 = N,N’-dimethyl—N,N’-bis(2-mercaptoethyl)-1,3—diaminopropane;
C: [Ni2(memta)2], where (memta)H2 = (HSC2H4)2NC2H4SM8; D: [Ni2{P(2~

SC.H,),},]',- E: [Ni2(SC,H9)6]2‘ ; F: [Ni,(S-2,4,5-iPr3C.H2)sl'- [From Ref. (3)]-

 

 

 

 

 

 

llllhe number of 1
WW COOI'd
the (1). In particu
Manning], whe:
irhgure lV-2) exhi
133V versus the l
nhrce species (N
oxidized form is c
pm of monon
mggested that oxi
bound sulfur ato
electrochemical or
(21 in Figure IV-2
Simulations only
oxidation of 21 b'
derivative of 211
samples of a do:
[Nif's]+ electroni
By contra
related [Ni(pdn
achemically re

a valence—delm

in behavior be

 

92

A large number of dimers with two thiolato bridges and Ni(II) ions with
square-planar coordination have been reported, some of which are redox
active (1). In particular, Maroney and coworkers have reported that
[Ni2(memta)2], where (memta)H2 = (HSC2H4)2NC2H4SMe (Figure IV-lC and 21
in Figure IV-2) exhibits a chemically reversible one-electron oxidation at E1,2 =
-0.35 V versus the ferrocene/ferrocenium (+0.05 V vs. SCE) to form a mixed-
valence species [N i2(memta)2]+ (8). Moreover, the EPR spectrum of this
oxidized form is comparable to the Ni-A and Ni—B signals, unlike the EPR
spectra of mononuclear Ni(III) thiolates (9). Maroney et al. also have

suggested that oxidation of the Ni site in the hydrogenase occurs at a Ni-

 

bound sulfur atom as opposed to the Ni-center (8, 10). They have shown that
electrochemical oxidation of the dimeric alkyl thiolate-ligated Ni(II) complex
(21 in Figure IV-2) affords a radical in which, according to their EPR
simulations only ca. 20% of the spin density resides on the Ni. Chemical
oxidation of 21 by I2 results in 22, while oxidation of the monomeric cyanide
derivative of 21 by O2 yields 23 (Figure IV-2). Experiments with 61Ni-enriched
samples of a closely related oxidized product indicate a valence-delocalized
[Nizz'sr electronic structure (11). A

By contrast, Kruger and Holm have reported that the structurally
related [Ni(pdmt)2], where (pdmt)H2 = 2,6-bis(mercaptomethyl)pyridine shows
a chemically reversible one-electron reduction at E1 ,2=-1.21 V vs. SCE to form
a valence-delocalized species [Ni21'5(pdmt)2]' (12). The reason for the difference

in behavior between these two complexes is unclear, and it is therefore

 

 

 

 

 

H,C$\_

Figllre lV-2, An
lbaSEd oxidatioi
Chemical OX'rdat

lemme of 21

 

 

 

 

93
"mos l +
l \—\Na/\ “8 I
u "‘
E CS'NIUIIISS\\W'N1‘N7 I Ila”: "$878
5 V V
2 2
\4’6.
- e'
/

 

 

 

 

 

’ s I ' ' p02 1 '
RNWNHCN ' 02 RN—h‘l—CN
s s
r _ r -
2 3

 

Figure IV-2. An example of an alkyl-thiolate Ni(II) complex WhiCh undergoes
S-based oxidation. Electrochemical oxidation of 21 affords a Ni(II) thiyl radical.
Chemical oxidation with I2 yields 22. Oxidation of the monomeric cyanide

derivative of 21 with O2 leads to 23 [from Ref. (8)].

 

;_;gssible to predic
jg hydrogenase.
Perhaps the r
tintedelocalized
rafters (13). 1t con
s::.:51-2.2so A 2
zurdinated by one
2.373 In and
Impound demon
an model for th
he parent complt
Spurned by arms
Also relev.
and lNillS-ZAS-r
With distorted te
ligands, respecti

N111) thiolates

Mm“ and Cow
monomefic let

oxidized in air

 

94

impossible to predict the redox chemistry for the Ni-Fe thiolate complex of
the hydrogenase.
Perhaps the most relevant structural model for the Ni-Fe center is the

valence-delocalized complex, [Ni2{P(2-SC6H4)3}2]', described by Millar and co-
workers (13). It contains a double thiolato bridge, [Ni22‘5(u-SR)2]3*, with Ni - (u-

S) = 2251-2260 A and Ni'"Ni = 2.501 A. The square-pyramidal Ni ions are also
coordinated by one basal and one apical thiolato ligand (Ni - Sba = 2.115, Ni -

Sp,p = 2.373 A) and one basal phosphano ligand (Figure IV—lD). While this
compound demonstrates the viability of a binuclear [Ni(u—S)2Fe] cluster, it is a

poor model for the reductive activation of the hydrogenase Ni—Fe site, since
the parent complex [Ni2(P(2-SC6H4)3}2]' is a dimer of square-planar Ni(II) ions
spanned by arms of the tris(2-mercaptophenyl)phosphano ligands.

Also relevant models are the complexes [Ni2(SC4H9)6]2' (Figure IV-lE)
and [Ni2(S-2,4,5-iPr3C6H2)5]' (Figure IV-1F), which are dimers of Ni(II) ions
with distorted tetrahedral coordination bridged by two and three thiolato
ligands, respectively (14). The redox behavior of these and other tetrahedral

Ni(II) thiolates is irreversible, however.

IV.3. Mononuclear Ni complexes

Millar and coworkers have synthesized and structurally characterized a
monomeric tetrathiolate Ni(II) complex, [Ni(nbdt)2]2', which is readily

oxidized in air to form the correSponding stable Ni(III) complex at biologically

—

 

iamt redox poten
no can be attains
{moment by incr
{firs r13, 16). The
can and the ch

:35: stabilizing en

 

hare anionic and

 

coworkers (17
:3 be particularly
nil follow later
roperties of a se'
orthis group of c
coworkers have 1
Srrrthes'rzing rea
rich coordinatio
structurally cha
incorporate a re
site in the hydr
ligated Ni(Il) o

glaSsy Carbon

H2 ll9).

__fi

95

relevant redox potentials (9). Holm and coworkers have demonstrated that

 

 

Ni(III) can be attained in a sulfur-containing multidentate ligand
environment by incorporating carboxyl or amidate groups adjacent to the
sulfurs (15, 16). They have investigated the relationship between Ni(III)
stability and the character of its local. ligand environment and found that the
most stabilizing environment is one in which the atoms coordinated to the
Ni are anionic and polarizable. In the comprehensive studies by Margerum
and coworkers (17) deprotonated oligopeptides (amidates) have been shown

to be particularly effective in stabilizing Ni(III) (more details on Margerum’s

 

work follow later in this Chapter). Holm has studied the electrochemical

properties of a series of thiolate/amidate complexes (16). The redox properties

 

' of this group of compounds are summarized in Figure IV-3. Kovacs and
coworkers have focused on functional model systems by designing and
synthesizing reactive mononuclear Ni complexes which contain Ni in a S-
rich coordination environment (2). Mascharak and coworkers have
structurally characterized two mononuclear Ni thiolate complexes that
incorporate a reactive site which is able to mimic some functions of the active
site in the hydrogenase (18). Crabtree et al. have discovered a labile sulfur~
ligated Ni(II) complex, which promotes H / D exchange and a macrocyclic N-

ligated Ni(II) complex, which displays proton-coupled electron transfer at a
glassy carbon electrode and that ultimately results in the reaction 2H+ +2e'——>

H2 (19).

 

-O.|

-O,2

vs. SCE

E1/2,

Figure lV.3_ SL

llllOlate / amlda

fi

96

LOW POTENTIAL NI COMPLEXES

 

 

 

 

 

 

 

 

 

 

 

 

*O'Zl. 0 ti:
«o .3 ~— 6R1}
#01 ~
0 1H12
6523
. -.-004*—-"
-O.| q ".“0 09 ‘—
M. ‘ IO
M- N
, l/
w “0 2 -o 20 J °9~>T/ “WW7
0 ’ t-o.24 ~————— "2 ”My”, 0,.
(D o N\ I
. -O 3 . T lex Mr I
g ' 1-032 A s s J E”"‘<é§>"'<és
,- ' : r—‘i '
0.34 a s-'

6‘: ‘ o N N 0 Na:
s. 3’
I.” -O.4

-o.5l

~O.61

-O.7

’08}-

 

 

Figure IV-3. Summary of redox properties of synthetic tetrathiolate and

thiolate/amidate Ni—complexes [from Ref.(16)l-

 

 

 

new

bill, complexes 0
sfinesized by Mar
1;: site in hydroger
sable {half-life of ‘
"$5913 min, Cf. R
sizsdtution is sirr
{fiese complexes 2
The prepar

caribe oxidized I‘t
llNit'HItCNW
can be replaced l

llecua feature a

tenagonally elor
u“paired electr<
the equatorial l:

mleractions the

V1) Samples I]

llle CW EPR S

HFI, l'ESlllllng

broadening ir

 

 

97

IV.4. Complexes based on tetracyano nickelateflII)

Ni(III) complexes of general structure shown in Figure IV—4 have been
synthesized by Margerum et al. (20, 21) and served as magnetic models for the
Ni site in hydrogenases for the following reasons: (1) the Ni(III) is relatively
stable (half-life of tetracyano-Ni(III) in solution at low pH values is reported
to be 28 min., cf. Ref. (17)); (2) hyperfine splitting generated by 61Ni isotopic
substitution is similar to those observed in proteins; (3) the EPR spectra of
these complexes are similar to the protein’s Ni signals Ni-A, Ni-B and Ni-C.
The preparation of these complexes starts from K2[Ni(II)(CN)4], which
can be oxidized readily by K23205 or excess HOCl, yielding
K[Ni(III)(CN)4(H20)2] (20, 21). The axially coordinated water molecules then

can be replaced by other ligands, yielding mixed ligand complexes, whose EPR
spectra feature axial g-tensor with g 981 I, g, ,52.00. This is consistent with a

tetragonally elongated ligand field with a (dzz)1 ground state (Chapter V). The
unpaired electron being mainly in the dzz orbital creates low spin density in
the equatorial plane of these complexes, leading to small HFI with the CN‘
nitrogens, not resolved in the CW EPR spectrum. These hyperfine
interactions may readily be measured by pulsed EPR techniques (see Chapter
VI). Samples made with 13CN‘, on the other hand, show large HFI splitting in
the CW EPR spectrum. Axially coordinated nitrogen also gives rise to large

HFI, resulting in HFI splitting only in the gH region of the spectrum and line

broadening in the g l region.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

98.

 

Figure IV-4. Structure of [Ni(III)(CN)4L2]'.

 

 

 

 

 

 

 

IV-S. These com

oxidation of the
deprotonated pt

help to stabilize

EPR stud
of these Ni(III)
molecules by b
EPR spectrum;
was observed,
positions (26).

pyridine, imi

 

 

 

 

99

 

lV.5. Monopeptide Complexes of Ni(III)

As early as 1971, Margerum and coworkers discovered that Ni(II) tetraglycine

 

reacts spontaneously with O2 to give a species absorbing at 350 nm and an
intermediate that oxidizes iodide to I2 (22). The intermediate was later
proposed to be a Ni(III) complex, based on its EPR spectrum and redox
behavior (23). A detailed study (24) showed that the rate of 02 uptake was
proportional to the rate of decomposition of Ni(HI)(H_3G4) (H_3 denotes the 3
deprotonated amide groups in the peptide, G refers to glycine), and that Ni(III)
catalyzed the reaction.

Tripeptides, tetrapeptides, pentapeptides and the corresponding peptide
amides form Ni(III) complexes with the general structure shown in Figure
IV-S. These complexes are readily prepared by electrochemical or chemical

oxidation of the corresponding square-planar Ni(II) complexes (23-25). The

deprotonated peptide nitrogen groups (amidate) are very strong o—donors that

 

help to stabilize trivalent oxidation states of Cu as well as of Ni.

EPR studies gave strong evidence for the elongated tetragonal structure
of these Ni(III) complexes (26). Replacement of one of the axial water
molecules by NH3 results in a three line (1:1:1) splitting in the gH region of the
EPR spectrum; at higher concentrations of NH3 a five line splitting (1:3:5z3zl)
was observed, consistent with two equivalent NH3 molecules at the axial
positions (26). Monopeptide Ni(III) complexes form adducts also with

pyridine, imidazole, and azide ion (N3) (27). Extended X—ray Absorption Fine

 

 

00539

Figure IV

 

 

 

 

100
TZ
H o
o 2 /9
a H /CH2———C{(\—l /, Y

\\ /

. C:-=N- —-—--—--N
/ // \ // \
HZC // Ni?” ///CH2
l/ \ //
‘— —x<’C

HzN “““““ \
I (-l\ $0

H20 \4

X

 

 

A Ni(III)(H_2G3), x = o
B Ni(III)(H _3G4) -, X = NCHQCOO’

C Ni(III)(H , 3G4a), x = NCH2CONH2

Figure IV-S. General structure of Ni(III) monopeptide complexes.

 

 

 

leahires gl l>gy
monopeptide cc
rather than e101
shown by the 15
complex forms,
Species conven
this complex is
The inductive
thus the liganl

high-spin six-i

 

 

 

101

Structure (EXAFS) data also confirm the assignment of tetragonally distorted

six coordinate structure for these Ni(III) complexes (17).

IV.6. Bistpeptido) complexes of nickel (III)
At high pH and with excess diglycine (GG) Ni(II) forms a blue complex that

corresponds to the formula Ni(II)(H_1GG)22' (28). X-ray crystallography reveals
a six-coordinate structure, which can be described as a tetragonally distorted
octahedron with a compressed N'-Ni(II)-N' axis (29). This complex can be
Oxidized electrochemically to form the corresponding Ni(III) complex with a
violet-black color (Figure IV-6) (30). The EPR spectrum of this Ni(III) species

features g| I>gi, which is in marked contrast to the EPR spectra of the

monopeptide complexes (Chapter V), due to the tetragonal compression

rather than elongation of the octahedral ligand field. Interesting behavior is

shown by the Aib2 complex of Ni(II) (Aib=aminoisobutyrate). Initially a blue

complex forms, but as the pH is increased this high-spin Ni(II)(I-I_1Aib2)(Aib2)‘

Species converts to anorange low—spin Ni(II)(H_1Aib2)22' complex (17). When

this complex is oxidized it forms the tetragonally compressed Ni(III) species.

The inductive effect of —CH3 groups in Aib2 increases the bond strength to Ni

thus the ligand field strength will now be sufficient to convert the normally

high'spin Six-coordinate Ni(II) to a low-spin form. The strong donor groups

0
in the Aib2 ligand also stabilize the +3 oxidation state. As a result, the E value

 

 

 

 

 

 

102

 

 

  

(H-2GG)2]3

Figure IV-6. Structure of [Ni(III)

 

 

 

     
    

this complex is

Slillt indefinitely

Bymixed ligan
which at least c
other than wat
lmidazole, azic
tripeptide and
adducts vary 1
pyridine (17).

Bipyric
mono(tripept
(31). The EPI
the bipy con
the gH regic

 

103

of this complex is remarkably low (+0.34 V vs. SCE), and Ni(III)(H_IAib2)2' is
stable indefinitely at pH 8-12 at room temperature.

Tripeptides form stable low-spin square-planar complexes with Ni(II)
and do not readily add a second tripeptide ligand. On the other hand,
mono(tripeptido)Ni(III) complexes react readily with a second tripeptide to
form bis(tripeptido) nickel(III) complexes. Hence, the mode of preparation of
the bis complexes is to first oxidize the mono-peptido-Ni(II) to the

corresponding Ni(III) complex and then add excess tripeptide.

IV.7. Ni(III) Mixed Ligand Complexes

By mixed ligand complexes we mean those Ni(III)monopeptide complexes, in
which at least one of the two axial coordination sites is occupied by a ligand
other than water. Axial ammonia adducts have already been mentioned.
Imidazole, azide (N3) and pyridine also form axial adducts with Ni(III)
tripeptide and tripeptide amide complexes. The stability constants of the
adducts vary from 1100 to 50 M‘], in the order: imidazole > NH3= N3'>
pyridine (17).

Bipyridine (bipy) and terpyridine (terpy) will react with
mono(tripeptido)-Ni(III) complexes to form stable mixed ligand complexes
(31). The EPR spectra of these complexes are given in Figure IV-7. The EPR of
the bipy complex shows a triplet, while the terpy adduct a quintet of peaks in

the gll region due to one and two axially coordinated nitrogens, respectively.

 

 

 

 

 

 

Figur
[Ni(IIIX

 

 

104

2700 2990 BIO 3300

[ fii *1 f ‘7 fii I T 1 T fl

 

 

 

 

 

 

r_ . a n i . r r . . . 1 r
2700 2900 3l00 3300
H. GAUSS

Figure IV-7. EPR spectra of the monopeptide Ni(III) complex,
[Ni(III)(H_2GAG)(HZO)2]‘ (a), and of its bipy (b) and terpy (c) adducts.

[from Ref. (17)]

   

 

 

 

Ethylenediamir

hipeptideasmc
heyact asl

coordination sit

 

frozen solution
iidicative of tw
conditions
molecules as
Cyanide i
monoadduct, N
EPR spectrum 5
useda large Sp.
splitting is grea
isotropic HFI Ct
ligand orbital (
of the unpairei
new species to
Species shown
ligated nitrog
CISN' is used
one of the Ai
the z-axis. Tl

shown in Fig

 

105

Ethylenediamine (en) and diethylene triamine (dien) can add to a Ni(III)
tripeptide as monodentate ligands at low pH, coordinating axially. At pH 7
they act as bidentate ligands, occupying additionally one equatorial
coordination site, replacing the C00‘ (carboxylate) group. Above pH 11, the
frozen solution EPR of the dien complex shows a quintet in the gll region,
indicative of two axially coordinated nitrogens. This means that under these
conditions dien acts as a tridentate ligand, replacing both axial water
molecules as well as the equatorially coordinated carboxylate group (31).

Cyanide ion reacts with Ni(III)(H_2Aib3)(HzO)2 to form an axial
monoadduct, Ni(III)(H_2Aib3)(HzO)(CN)' (Figure IV-8A). The frozen solution

EPR spectrum shows a single peak in the gll region. When 13CN' (1:1/2) is
used, a large splitting occurs in both g i and g I regions (17). The hyperfine

splitting is greater than that observed for the 14NH3 adduct, because the
isotropic HFI constant (Aiso) is larger and because the greater 5 character of the
ligand orbital (sp for CN‘ versus sp3 for NH3) allows for a stronger interaction
of the unpaired electron with the nucleus. Addition of more CN” causes a
new species to form that has an EPR spectrum typical for an elongated Ni(III)
species showing a triplet splitting in the g” region, indicating one axially
ligated nitrogen donor. This spectrum does not change when either ”CN' or
CISN' is used to make the sample. Thus, the triplet splitting must come from
one of the Aib3 nitrogens and the CN' ions are no longer coordinated along
the z—axis. These observations can be explained by assuming the structure

shown in Figure IV-8B (17). Excess cyanide results in the displacement of the

 

 

 

 

 

Figui

 

 

106

N

 

c ,c p
0 \(-) H30 >C‘C if"

\

 

 

Nim(H_2Aib3)(H20)CN'

 

 

 

Ni'"(H-2Aib3)(CN):-

Figure IV-8. Structure of [Ni(III)(H_2Aib3)(HZO)(CN)]' (A) and

[Ni(III)(H_2Aib3)(CN)2]‘ (B); where Aib=aminoisobutyrate.

   

 

:eptltle and the
teprepared by l
:gnpiex present

grades become

 

107

peptide and the formation of Ni(III)(CN)63‘. Hexacyanonickelate(III) can also
be prepared by the oxidation of Ni(II)(CN)42‘ followed by addition of CN‘. This

complex presents an example of dynamic Iahn-Teller distortion, where the six

cyanides become equivalent because of vibrational interchange (20).

 

 

 

 

 

 

 

Halcrow, l
Koyacs, l -
Halcrow,

Volbeda, .
Fonticella

Kruger, t
DerVartaI
G; leGal

Mills, D.
Darensbc

Colpas, t

Kumar,
1989,11]

Fox, 8.; I
Maroney
Transitic
Stiefel, E
1996.

Choudh
l. Inorg.

Krijser.

FTanolic
6587.

SilVer,
Kn'iger
Kn'iger

Marge

NiCkEl;

10.

11.
12.
13.
14.
15.

16.

17.

 

108

References
Halcrow, M. A.; Christou, G. Chem. Rev. 1994, 94, 2421.
Kovacs, I. in Advances in Inorganic Biochemistry 1994, Vol. 9, 173.
Halcrow, M. A. Angew. Chem. Int. Ed. Engl.1995, 34(11), 1193. '

Volbeda, A.; Charon, M.-H.; Piras, C.; Hatchikian, E. C.; Frey, M. ;
Fonticella-Camps, I. C. Nature 1995, 373, 580.

Kruger, H.-I.; Huynh, B. H.; Ljungdahl, P. 0.; Xavier, A. V.;
DerVartanian, D. V.; Moura, I.; Peck, H. D., In; Teixeira, M.; Moura, I. I.
G.; LeGall, I. I. Biol, Chem. 1982, 257, 14620.

Mills, D. K.; Hsiao, Y. M.; Farmer, P. J.; Atnip, E. V.; Reibenspies, I. H.;
Darensbourg, M. Y. I. Am. Chem. Soc. 1991, 113, 1421.

Colpas, G. I.; Day, R. 0.; Maroney, M. J. Inorg. Chem. 1992, 31, 5053.

 

Kumar, M.; Day, R. 0.; Colpas, G. J.; Maroney, M. I. I. Am. Chem. Soc.
1989, 111, 5974.

Fox, 8.; Bang, Y.; Silver, A.; Millar, M. I. Am. Chem. Soc. 1990, 112, 3218.

Maroney, M. J.; Allan, C. B.; Chohan, B. S.; Choudhury, S. B.; Gu, Z. in

Transition Metal Sulfur Chemistry; ACS Symposium Series Vol. 653,
Stiefel, E. I.; Matsumoto, K., Eds.; ACS Publication: Washington, D. C.

1996.
Choudhury, S. B.; Pressler, M. A.; Mirza, S. A.; Day, R. 0.; Maroney, M.
J. Inorg. Chem. 1994, 33, 4831.

Krfiger, H.-I.; Holm, R. H. Inorg. Chem. 1989, 28, 1148.
Franolic, I. D.; Wang, W. Y.; Millar, M. I. Am. Chem. Soc. 1992, 114,
6587.

Silver, A.; Millar, M. J. Chem. Soc. Commun. 1992, 948.

Kruger, H.-I.; Holm, R. H. I. Am. Chem. Soc. 1990, 112, 2955.

Kruger, H.-].; Peng, G.; Holm, R. H. Inorg. Chem. 1991, 30, 734.
r S. L., in The Bioinorganic Chemistry of

Margerum D' W'; Anhke d.; VCH: New York, 1988; Chapter 2.

Nickel; Lancaster, I. R., Ir., E

 

1,.

Baidya, M

Efros, L. l

Chem. 199

‘2 Pappenha

‘1. Wang, Y.

Chem. 19E

Paniago,
1971, 1427

~. Bossu, F.

Bossu, F.
l. Inorg.

Bossu, F.

v).

' lappln, l

1630.
Murray,

Brookes,

Freeman
Jacobs, E

Pappen
Inorg. C

 

 

18.

19.

20.

21.

23.

24.

25.

26.

27.

28.

29.

30.

31.

109

Baidya, M.; Olmstead, M.; Mascharak, P. K. Inorg. Chem. 1991, 30, 929.
Efros, L. L.; Thorp, H. H.; Brudvig, G. W. W.; Crabtree, R. H. Inorg.
Chem. 1992, 31, 1722.

Pappenhagen, T. L.; Margerum, D. W. I. Am. Chem. Soc. 1985, 107, 4576.

Wang, Y. L.; Beach, M. W.; Pappenhagen, T. L.; Margerum, D. W. Inorg.
Chem. 1988, 27, 4464.

Paniago, E. B.; Weatherburn, D. C.; Margerum, D. W. Chem. Commun.
1971, 1427.

Bossu, F. P.; Margerum, D. W. I. Am. Chem. Soc. 1976, 98, 4003.

Bossu, F. P.; Paniago, E. B.; Margerum, D. W.; Kirksey, S. T., Ir.; Kurtz, I.
L. Inorg. Chem. 1978, 17, 1034.

Bossu, F. P.; Margerum, D. W. Inorg. Chem. 1977, 16, 1210.

Lappin, A. G.; Murray, C. K.; Margerum, D. W. Inorg. Chem. 1978, 17,
1630.

Murray, C. K.; Margerum, D. W. Inorg. Chem. 1982, 21, 3501.

Brookes, G.; Pettit, L. P. I. Chem. Soc. Dalton Trans. 1975, 2106.

Freeman, H. C.; Sinclair, R. L. Acta Crystallogr. 1978, B34, 2451.

Jacobs, S. A.; Margerum, D. W. Inorg. Chem. 1984, 23, 1195.

Pappenhagen, T. L.; Kennedy, W. R.; Bowers, C. P.; Margerum, D. W.
Inorg. Chem. 1984, 23, 1345.

 

 

 

 

 

 

twin 5N) has 1
ring of the terp
observe the mc
labeled sample
Surprisingly, it
may suggest tl

geometry prop

 

 

 

Chapter V

ESEEM Study on N i(III)(triegcinate)(terpy)

V.1. Abstract

The goal of this investigation is to gain information on the electronic
structure of Ni(II[)(H_2G3)(terpy) (Figure V-6). This complex features four
equatorially coordinated nitrogens, which are weakly coupled to the Ni(III)
center. To sort out the contributions of these nuclei specific isotopic labeling
(with 15N) has been used, by substituting the nitrogen in the middle pyridine
ring of the terpyridine chelate. The ”product rule” of ESEEM allows one to
observe the modulation from this nucleus by dividing the data from the non-
labeled sample by that from the 15'N--labeled one. The result, quite
surprisingly, indicated no contribution from this particular nitrogen. This
may suggest that this nitrogen does not coordinate at all, in contrast to the

geometry proposed earlier [see Ref.(16)].

110

 

 

 

 

 

 

 

to make compli
magnitude of t]
all Ni(III) studi
Opposite reason
conditions. Ni(
and those with
elongated octa
The g-tensor

The anisotrop1

The g—tensOl'

 

111

V2. Theory of EPR of Ni(III) Complexes

EPR spectra of Ni are observable in at least three redox states: Ni3+ (d7), Ni2+
(d8) and Ni+ (d9) ions are all paramagnetic in at least some complexes. the d7
and d9 ions are classical Kramers systems, which must be EPR detectable
under appropriate experimental conditions. Ni2+ has an integral spin (S=1)
therefore it is not guaranteed to have a doubly degenerate ground state in the
absence of external magnetic field.

Ni(III) is isoelectronic with Co(II) and Fe(I), but is far from being
identical to them in its behavior. The larger positive charge on the Ni tends
to make complexes with negatively charged ligands tighter. This increases the
magnitude of the crystal field terms and favors the low-spin state. As a result
all Ni(III) studied thus far have low—spin electron configuration. For just the

opposite reason Fe” prefers high-spin ground state under all reasonable

conditions. Ni(III) complexes fall into two major classes; those with 81 > gI I

and those with gH > 81° The former have been recognized as tetragonally

elongated octahedral complexes, the latter as square planar complexes (1, 2).
The g-tensor l

The anisotropy of the g-tensor of d7 ions results from the admixture of the d-
orbitals primarily by spin-orbit coupling, which contributes orbital angular
momentum to the electronic Zeeman interaction. In the absence of the spin-
orbit coupling the orbital angular momentum would be quenched out (3).

The g-tensor can be calculated by using appropriate ligand-field terms and

 

 

 

 

 

 

 

 

system with the
the diagram in
electron primal
g~valueo in this

ELI gn=2'

ABM

represents the

Thus, the squa

 

 

112

additional matrix elements provided by the spin-orbit coupling. The
Hamiltonian can be written in terms of the free ion orbitals that are
eigenstates of L2 (z-component of the angular momentum). If the electronic
Zeeman term is much smaller than either the field or spin-orbit terms (which
is always true for fields of ca. 3000 G), only the lowest doublet need be
considered in the calculation. If the ligand-induced separations of the d
orbitals are much greater than the important spin-orbit matrix elements,
perturbation theory can be used to calculate the g—values. This method has
widely been used in an attempt to relate the observed g-tensor to the

symmetry of the complex (4—7).

 

Figure V-1 shows a simple energy level diagram for a low-spin d7
system with the relative energies of the d orbitals in different ligand fields.
The diagram indicates that in a square-planar arrangement the unpaired
electron primarily resides in an orbital which is mainly dxy in character. The
g-values in this case are predicted by the following expressions (9): gxngyy=2_

2?» 2?»
’ fguf=2f
AB AE

xy—xz

 

 

, where it is the spin-orbit coupling constant, AE

xy-xz

xy—(xz—yz)

 

represents the energy difference between the xy andxz unperturbed states.
Thus, the square-planar geometry leads to gll > g.

In the case of a tetragonally distorted octahedral ligand field the

unpaired electron is located in the dzz orbital.

 

 

22
‘ xz-yz

xz,yz:”_-i

xy—l-

Tetrago
compre

Figure V-l. I

‘1 l

l 13
,/ x2_y2
22 \\ ”///
\\\ III/r
\\ . ”,
‘>=::<:
l/ll \\\_L___ \\
’l/X\\ I I 22
__LL_._/”
/
x2, yz __._—LLfl= \\\ I’l’
\“~=L;’=t_=—n-_-<’
//’ \‘\“‘
/// ‘T‘&:\\
ll/ \\
xv —-“——”
\
\zh'L—u: x2, yz
Tetrogonolly Octohedrot Tetrogonolly y Square
compressed elongated planar

Figure V-l. Energy level diagram for d-orbitals in various ligand fields [from

Ref. (8)].

 

 

 

 

 

 

The measured

 

the Ni(III) cen
CW EPR.

The 0th
electron and ti
l'lsuperhyperf
and usually 11
measured by
parameters, e

constant, rept

 

 

114

6?»
AEz

Z -XZ

The g—values now are gxx=gyy=2- , 822:2, so 81 > g.. (9). This kind of g-

tensor has been the most commonly observed for mononuclear Ni(III)
complexes and in hydrogenases.

The nuclear hyperfine interaction

Two kinds of hyperfine interactions (HFI) have been studied in Ni(III)
complexes. One is the HFI between the unpaired electron and the “Ni
nucleus (I=3/2). These measurements require “Ni enrichment in the sample.

The measured HFI tensor yields information on the electronic structure of

 

the Ni(III) center. These interactions are large and can readily be measured by
CW EPR. ‘ I

The other kind of HFI is the magnetic coupling between the unpaired
electron and the magnetic nuclei (such as 1H, 2H, 14N, etc.) of the ligand
(”superhyperfine interactions”). These are of small magnitude (few Gauss)
and usually not resolved in frozen solution EPR spectra. They can be
measured by ENDOR or ESEEM. In general, the HFI tensor provides three
parameters, each of which carries structural information. Am, the isotropic
constant, reports on the unpaired electron spin density as well as on the
nature of the molecular wavefunctions. The anisotropic portion of the HFI,
which we designate T, provides geometric information, since its magnitude is

proportional to 1 /r3, where r is the effective distance between the electron and

 

 

 

 

nucleus. The

gamergnetic spe

linear Quadruj
itclei with 121
:agnetic mome
elector field (Fig
electronic struct
quadrupole par
nuclear spin (it
field determine
moment in tur
field gradient.
luadrupole m.
and may mani
measured dire
llil- Owing g
$le atlgular

lerms 0f mud.

115

  

the nucleus. The rhombicity parameter, 8, is related to the symmetry of the

paramagnetic species (10).

Nuclear Quadrupolar Interaction (NQI)
Nuclei with 121 possess a quadrupole moment, which is collinear with the

magnetic moment. The quadrupole moment interacts with a non-uniform
electric field (Figure V-2). The electric field gradient is determined by the
electronic structure of the paramagnetic species, therefore, measuring the
quadrupole parameters can yield useful structural information (3, 11—13). The
nuclear spin (through its magnetic moment) is quantized along the effective
field determined by the magnet and nearby electrons. The quadrupole
moment in turn orients the nucleus in the direction of the largest electric
field gradient. This competition between the magnetic moment and
quadrupole moment alters the energy of the nuclei placed in a magnetic field
and may manifest itself in the result of the EPR experiment. NQI can also be
measured directly, using Nuclear Quadrupole Resonance (NQR) spectroscopy
(12). Owing to the collinearity of the quadrupole moment with the nuclear
spin angular momentum, the Hamiltonian of the N01 can be expressed in

terms of nuclear spin operators (3, 11),

jeNQl = K133 " 12 ‘l‘ ”(fit ”" fill],

 

 

 

Figure V-Z. ‘

 

116

h
I

 

 

 

 

 

 

-9 w

T r

(C) (d)

cit :p

7+L

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure V-2. Oblate (a) and prolate (b) shape of quadrupolar nuclei. Schematic
representation of the interaction of the quadrupolar nucleus in a

uniform electric field (c), and in a field gradient (d).

 

 

 

 

While K is the Cl1

gander define
cleric field gra
The effect
n shows a s
exactly twice th
leads to the col
appearance of p
be obseryed f1
appear as shar
feature located

about the hyp

117

  

2

€le

where K is the quadrupole coupling constant, Kz—Zfi—fin is the asymmetry

parameter defined as n: | qxx-qyyl / qu, where qjj are the principal values of the
electric field gradient tensor.
The effect of JACNQI on ENDOR and ESEEM is illustrated in Figure V-3,

which shows a special case for S21 / 2, 1:1, where the isotropic nuclear HFI is
exactly twice the nuclear Larmor frequency. This ”exact cancellation” (14)
leads to the collapse of energy levels in one spin manifold and results in the

appearance of pure quadrupole frequencies in the ESEEM (or ENDOR) spectra.
The observed frequencies (3) are v+=(3+n)ic, v,=(3-n)K, v0==2nK (Figure V-3) they

appear as sharp peaks. The other electron spin manifold gives rise to a broad

feature located at higher frequency (15); this peak contains all information

about the hyperfine interaction.

 

 

 

 

Flgure V-3.

 

 

118

1>

+1/2 > f
I I o > VDQ

 

 

 

 

 

 

 

 

. l -1 >
Va A

l 0 > ’_ v

|-1/2 > v- t
t

l 1 >
Electronic Nuclear Nuclear
Zeeman Zeeman Hyperfine Quadrupole

FiSure V-3. Energy level diagram for S=1/2, 1:1/2 spin system at ”exact

cancellation”.

 

13 Sample Pn
he procedure
hmrahon of l
it mg of Ni(C
fast Also pref
:yg‘iicine (Sigr
:2. 0.1 M “OX(
‘OXONIS” is a
mol. All the
Combine 5 m
at beaker eq
llNaOH soh
NaOH). The
f0lmallOn of
the amide p]
Pltparation Q
terPyridine (
ml of aceto]
ml of the N
NaOH “Or
to mElke Ni

Sillllllon of

119

V3. Sample Preparation

The procedure follows the one published in the literature (16).

Preparation of Ni(II)(H-2G3): Make up 0.1 M Ni(II) stock solution by dissolving
366 mg of Ni(C104)2.6H20 (Aldrich, FW=365.7O g / mol) in a 10 mL volumetric
flask. Also prepare 0.14 M triglycine (G3) solution by dissolving 265 mg
tryglicine (Sigma, FW=189.2 g/mol) in a 10 mL volumetric flask. Also make a
ca. 0.4 M ”0XONE” stock Solution, using 1.23 g ”0XONE” in 5 mL water.

”0XONE” is a trademark of Aldrich for 2KHSOS. KHSO4. K2804, FW=614.78

g/mol. All these solutions should be freshly prepared.
Combine 5 mL of the Ni(H) stock solution with 5 mL of the G3 solution in a 50

mL beaker equipped with a magnetic stirrer and a pH electrode. Slowly add 1

M NaOH solution to it to raise the pH to ca. 10 (it takes about 1 mL of the
NaOH), The color of the solution changes to deep yellow, indicating the

formation of the Ni(II)(H_2G3) complex (H_2G3 denotes triglycine with two of

the amide protons lost). The Ni concentration in this solution is 45 mM.

Preparation of Ni(III)(H_2G3)(terpy): Prepare 0.3 M solution of 2,2’:6’,2”

terpyridine (terpy) by dissolving 699 mg terpy (Sigma, FW=233 g/ 11101) in 10

mL of acetone. Mix 300 uL of this solution with 3 mL of ethylene glycol. Mix 1

mL of the Ni(II)(H_2G3) stock solution with 1 mL of H20 and 0.1 mL of 0.1 M

NaOH (for pH adjustment). While stirring, add 300 uL of ”OXONE” solution

t0 make Ni(III)(H_2G3)(H20)2 in situ. Then quickly add the ethylene glycol

solution of the terpy just prepared. Put an aliquot of this solution into an EPR

 

 

 

 

 

 

 

she, freeze it v
all, with a 2-fc
Preparation of t
acetyl pyridine
(arrests of thre
introduce LTN

‘xaoac (c;

oi he product

he Clo-EPR
Spema 0f the

small HFI co
Spectrum disy
idtntical Witl
cllllSlSlem W
V1). The res
SPectrum re:

(Figure V-6)

of

120

tube, freeze it with liquid N2. The sample has a Ni concentration of ca. 10
mM, with a 2-fold excess of terpy.

Preparation of the ”N-labeled terpy: the synthesis of terpy starting from 2-
acetyl pyridine is described in Org. Synth. 1986, Vol. 64, 189. The procedure
consists of three major steps summarized in Figure V-4. Our goal is to
introduce 15N only into the middle ring. This can be done conveniently using
15NH4OAC (Cambridge Isotopes) in step B of the synthesis. The identification

of the product was done by using 13C and 1H-NMR and mass spectrometry.

V.4. Results and Discussion

 

The CW-EPR spectrum of the Ni(H_2G3)(terpy) is shown in Figure V-5. The
Spectra of the all-“N and of the 15N—labeled samples are identical, because the

Small HFI coupling is masked by the inhomogeneous line broadening. The

Spectrum displays a quasi-axial HFI tensor with g i=2.154, gl |=2.011 and is

identical with the one published in the literature (2, 16). These g-values are

consistent with a tetragonally distorted octahedral ligand field (see Section

V-1)- The resolved hyperfine splittings observed in the gll region of the

SPECtrum result from the two axially coordinated (equivalent) nitrogens

(Figure V-6). This strong coupling is due to the fact that the unpaired electron

of

 

 

/

N

O

 

 

Figure

121
1) i-BuOK/THF /|
————'—’ ‘ SMe
2) cs, N \
3) Mel C SMe
SMe
\
1) flip-com, r-auOK / |
e _.—-——-———-> | N
2) NH40Ac. AcOH \ N
/
omen4 \ I
/ N
NIClz-GHZO \ IN

V-4. The synthesis of terpyridine.

 

i_m._1__

”1.....u. hummus“. . maxim. .1; Nero a.»
t. .........i:.. .

\...i.

 

 

 

122

.....

......

 

 

.aaeeexeuemxaxz do 65:83 35.30 .m-> mama

.
.....

......

.......
.......
.....

.....

 

3:62.. .1”: c. 925.;
Ill.

(£5 0.98 «.2 .(ZUDC

30¢ {Oilcv 23¢ d5

.8. :8!a° 3...:

    

( htgo Ktl.>¥.

dmeéWmemfww. __ .

 

 

 

 

35...:

 

 

123

H C o
3 \ N a
O\ -—- ——-C\l'l
\f _._. N’I‘I ________ \.\. N
Cy /’ \

 

Figure V-6. Proposed structure of Ni(III)(H_2G3)(terpy).

 

 

 

 

 

urn mainly
as land low
coordinated n
Effhl. The s
because the It
lllt pulse, sc
simultaneous
to sort out th
replacing one
function of 5.
individual In
correct only )

disordered 5.

Where Ell) p
i“ nitrogen.
lelPl'. the ot

ESEEM iunr

ll‘here f4 is
position, f4.

NQI 0f lSN

124

 

Ni(III) mainly resides in the dzz orbital, creating high spin density along the z
axis (and low spin density in the equatorial plane). The equatorially
coordinated nitrogens in turn are weakly coupled and can be studied by
ESEEM. The strongly coupled axial nitrogens do not contribute to the ESEEM,
because the magnitude of their HFI exceeds the excitation bandwidth of the
MW pulse, so the corresponding branching transitions cannot be excited
simultaneously (Chapter I). One approach to analyze complex ESEEM data is
to sort out the contributions from the various nitrogens by specifically
replacing one 14N by 15N. This is made possible by the fact that the modulation
function of several nuclei is the product of the modulation functions of
individual nuclei ["product rule”, see Ref(17)]. Though this rule strictly is
correct only for 2-pulse single crystal ESEEM, it has been found applicable to

disordered samples to a good approximation. In our case,
4
E(t)= I]1 fi (c)

where E(t) is the experimental ESEEM functiOn, fl. is the modulation from the
ith nitrogen. We have prepared two samples, one with ordinary (non-labeled)
terpy, the other with terpy labeled with 15N in the middle ring. Dividing the
ESEEM functions we obtain

Ed,v(t)=f,/f,’
Where f4 is the modulation function from the 14N nucleus at the middle
position, f,” is the corresponding 15N modulation function. Due to the lack of

N01 of 15N (1:1 / 2), f; features very small modulation depth, so Ediv to a good

 

 

 

 

 

eproximation
analyzed by I)
figure ‘
ESEEM data. '.
if) have b
31' rel-(f) hay
respectively.
\ and "N d
ESEEM traces
he concludec
to be equate
ot'Nillll)(fl1
measuremen
not shown),
lliltogens, p:
This 1
directly Coo
tenter in tet
be to Pr0po
Ni(III) bilt i
“lllOgens v

Unlikely, ill

C0"lPloces

 

125

approximation is identical to the 14N modulation function, and can be
analyzed by numerical simulation.

Figure V-7 summarizes our results by showing 3 pairs of 3-pulse
ESEEM data. Traces (a), (c), (e) arise from non-labeled samples, while (b), (d)

and (f) have been obtained from the 15N—labeled sample. Data (a)-(b), (c)-(d)
and (e)-(f) have been collected at g=2.175 (g i), g=2.100 (gin) and g=2.014 (gI |),

respectively. No detectable difi‘erence can be seen between the corresponding
14N and 15N data. This has been confirmed by dividing the corresponding
ESEEM traces, as described above. The results are shown in Figure V-8. It can

be concluded that the nitrogen of the middle pyridine ring of terpy (proposed

 

to be equatorially coordinated) has no significant contribution to the ESEEM
of Ni(III)(H,2G3)(terpy). Identical conclusions have been drawn from

measurements performed at 11 and 13 GHz resonance frequency range (data
not shown). The ESEEM in this complex is apparently due to the other three

nitrogens, probably primarily to the two negatively charged amidate groups.

 

This result is rather unexpected, eSpecially knowing that equatorial, not
directly coordinated CN‘ nitrogens show significant coupling to the Ni(III)
center in tetracyano-nickelate(III) (Chapter VI). A possible explanation would
be to propose that the middle nitrogen of terpy is not coordinated directly to
Ni(III) but assumes a bidentate coordination by terpy, involving the two axial
nitrogens with the middle pyridine ring rotated away from the Ni(III). This is
unlikely, however, since terpy usually acts as a tridentate chelate, making

Complexes of extremely high stability. The other possible explanation

 

 

 

Figure V-7.

 

 

126

 

Figure V-7. Three-pulse ESEEM on Ni(III)(H_2G3)(terpy): traces (a), (c) and.(e)
are from the non-labeled complex, traces (b), (d) and (f) are from
complex containing 15N in the middle pyridine ring. (a), (b)

taken at g=2.175; (c), (d) at g=2.100,' (e), (f) at g=2.014.

 

 

 

ODDfia HQEE OEUO

 

127

 

 

 

3“”; tau: 183 no (2%)
consent: freq.:8.7960 6H2
f11¢§N1(fi§)T14.(9).05 f101622890.00 6
dato:10-27-1993 PlpJ‘Itl: 40 Hz power: «.0 dan
ntant t1w213 no events/pt: 34 team: 4.20 K
[1" IIIIIIIIjIIIll]I]IIIITIIIIIIIIITII‘IIIIIITIIII

1139 ' -
- 4
0)
U L In
:3
.p
fl - d
H 7:: -
a '- N-
E
(O
r- ..
(J
.C ..
U l-
G)
5 ”unsunlrinlnul....hurl“..11...I....1...L1.

 

 

 

O 1 2 3 4 5 5 7 B 9 10

tau+T (usec)

 

 

 

 

 

 

gents
tau: 162 no ( 2%)
comnt‘ 6" from: 8.8009 GHz
9110:!“ teams. (9) .11 field: 2893.23:
date: 10-29—1993 ropmnto: 40 Hz pour. 48.
start the: 212 no events/9cm tom.:4.20 K
Tlilllllllllllllllllllllllllillllllllllllllllllllllul
1593 - b
. r _
U r-
3
p -
** P
fl -
D.
E F
m -
o P _
‘c t-
L)
m d
I llllllllllllllll
:SEB ll 11 l: 11 ll 11 III II 11 I: III In ll 'Eslll ‘7 E; S3 1()

0 1 2 3 4 5
tau+T (usec)

 

 

 

115
31

0
CU
fl HQEM 0

U

QUE

878

1
0

EU

« H060 O

u

003

 

 

 

 

128

37".qu
(gig; teu:236 ne ( Bit.)

 

 

 

comment: freq.:8.7960 6H2
f11e:N1(63)T14.(9).09 f1e16:2990.00 6
dete:10-27-1993 nep.rete:40 Hz pouer:44.0 one
etent t1ee:286 no evente/pt:34 teeo.:4.20 K
IlTll—IIIIIIITT]IIIIIIIIT—lflI—TPTWT—[1llll[[11]]1TTITT
87B - - ~
r- -i
a)
13 _ _
:3
4J
a - ‘
r1
CL _ -
E5
to
c) _
c -1
o h
a)
b I!
h-
1 bllllIllllllll[lllllllllllllllllllllllllllllllllllll

 

 

0 1 2 3 4 5 6 7 8 9 10

tau+T (usec)

 

 

 

 

 

 

 

3 towaasne (3”)
comment: M freq.:8.8009 on:

‘ . G
f1]e:N1(63)T15.(9).09 f101d:i:9: :2.
dete::o-29-1993 rep.rete:40 Hz power: .
etert t1ee:286 ne evente/pt:34 teen..4.20 K
jjllllillllIIIIIIIIIIIlllllllllllillllllllllllllllll-J

1190c d
r
1
P
0)
13 P 1
23
u 1
** _
H ’1
C1
E r-
m *w d
o P _
L:
U h-
u) 4
h:
llllllllllllllll
31 Fillllllllllllllllllllllllllllllll7 8 9 10

O 1 2 3 4 5 5
tau+T (usec)

 

31

mo
CU
« H 050 0

0

MD:

.
00

0C

HQEU

ud

003

 

 

 

129

 

 

 

 

 

 

 

 

 

 

 

 

 

 

consent 9:: teu: 151 no ( ZIL)
' foeq.:8.7960
otent t1ee:201 l1! avgnt./pt:34 to”; 4.20 K
137011IIlllIII]IIIIIIIIIIIIIIIIIIIIIITIIIIIIIIII[lIIl{I
m .1
U h
3
v l
co-I -
H CI!
0.
E " ~
to
O " ' if- - _‘v -
.C
U " -1
a)
P- a:
-5 _. u
l1111111111114.lllLlJlllllllJlllllllllllllllllll'lll
O 1 2 3 4 5 6 7 B 9 10
tau+T (usec)
37- 2.011,
teu: 150 no (UL)
consent: 5“ freq.:8.8009 9H2
f11e: N1 (63) T15. (9) .14 field: 3127.00 6
date: 10-29-1993 repmete: 40 Hz power: 46.0 dBo
otert t1ee: 200 no evento/pt: 34 teno.:4.20 K
IIII IIII IIII IIII'IITWTTII III IIIT IIII IlII r
1089 I I I I I I '1 I I I ..
D- -l
0.)
U -
3 ..
4.:
+1 . -
H
‘1 .
E P
IO
0 ' 5"" w .A ~
5 ..
U I-
a.)
31 I‘llIllllllllllllllllllllllllllllllllllllllllllljllll
O 1 2 3 4 5 6 7 8 9 1O

tau+T (usec)

 

 

102i

 

Um
UUDUHHQEM 0C

Figure v.3,

 

 

 

 

 

{ @ teu:325 ns

cement: . freq.: 8.8010 GHz

file: Ninet. (9) .07 field: 2891.00 G

date: 11-5-1993 repmete: 40 Hz power: 46.0 dam

start time: 375 ns events/pt: 34 temp.: 4. 19 K

Illj IIII IIII IIII IIIllllIIIIIIIIIIIIIIIIIIIIIIIII -
1021 . PM!” ' ' I ((1)
0.)
U _ _
D
4..)
.fl __ -
H
0. ..
E I-
r0
0
C _
U h
(D
5 llllllllllllllllllllllllllllllllllllllllllllllllll

 

 

 

0 1 2 3 4 ‘5 6 7 8 9 10

tau+T (usec)

Figure V-8. ESEEM functions obtained after dividing the all-”N traces by the

corresponding 15N traces; (a) g=2.175; (b) g=2.100; (c) g=2.014.

1065

um
JJAHQEQ 05
GO .

 

10

um
UJUfiHQEm 0:
m

 

 

comment:
f11o:N1Ret.(9).08
dotoz11-6-1993

131

I'uJ‘

__a’

rep.reto:40 Hz

teuz152 no
freo.:8.8010 6H2
f1old:2992.00 G
Dower:46.0 dam

 

 

 

 

 

 

 

 

 

 

 

 

 

otort t1lne:202 no events/pt: 34 team: 4.19 K
IIII IIII I'IIT TIII mil IIII IIIW IIII IIII III I
1055_ I I I I I I 1 I I *T (b)
' 1
a) -
13
3 ’ i
4;
.H
H ' 1
C1
E? ' -
(O
o I- .1
1:
L) ” 1
Q)
[ '1
#- OJ
1
lllllJlllllllJlJllllLliULll[JULIJIJIIJLIJIIIJIJ
O 1 2 3 4 5 6 7 8 9 10
tau+T (usec)
9) tou:226 no
comment: froo.:8.8010 8H2
f11e:N1Rot.(9).12 f1e16:3127.00 6
dotez11-6-1993 reo.rote:40 Hz power:46.0 08m
otort t1ee:276 no ovonto/pt:34 tonn.:4.19 K
IjII IIIT1FIT WTFI TTFFWTTI Illj TIIT IIII [III
0: t '
ID
3 l"
4.:
'H r- q
H
£1
E - ..
'° l
C) .
I:
L) F -
a)
O . d
lllll[144111.14]IllllllLJJJlllllllllJJlllJllJ’lllljtx
O 1 2 3 4 5 6 7 8 9 10

tau+T (usec)

 

 

 

 

 

 

 

 

that the spi
equatorial ter
igands. The s
the case of
between the l

on Ni(III) ele

132

is that the spin density in the equatorial plane is too small at the site of the
equatorial terpy nitrogen, in spite of the presence of two anionic amidate
ligands. The substantial spin density (therefore strong hyperfine interaction)
in the case of equatorial CN’ may be due mainly to back-bonding interaction
between the ligand and the dzz orbital of Ni(III). The lack of experimental data

on Ni(III) electronic structure prevents further conclusions being drawn.

 

 

 

 

 

 

 

Salemc
Jr, Ed.)

Marge
Nickel,

Slichte
V erlag

Lappir
1630.

Bossu,
Desid:
Bemsl
Jacobs

Maki,
1964, .
Cordj
York,

AtheI
PTR :

Luck.
Press

Char
Mim
Gerh

Papr
In0ré

Miir

Bioc.

10.

11.

12.

13.

14.

15.

16.

17.

133

References

Salerno, J. C. In The Bioinorganic Chemistry of Nickel, Lancaster, J. R.,
Ir., Ed., VCH: New York, 1988; Chapter 3.

Margerum, D. W.; Anliker, S. L. In The Bioinorganic Chemistry of
Nickel, Lancaster, I. R., Ir., Ed., VCH: New York, 1988; Chapter 2.

Slichter, C. P. Principles of Magnetic Resonance, 3rd ed., Springer
Verlag: Berlin, 1989.

Lappin, A. (J.; Murray, C. K.; Margerum, D. W. Inorg. Chem. 1978, 17,
1630.

Bossu, F. P.; Margerum, D. W. ]. Am. Chem. Soc. 1976, 98, 4003.
Desideri, A.; Raynore, J. B. J. Chem. Soc. Dalton Trans. 1977, 19, 2051.
Bernstein, D. K.; Gray, H. Inorg. Chem. 1972, 11, 3035.

Jacobs, S. A.; Margerum, D. W. Inorg. Chem. 1984, 23, 1195.

Maki, A. H.; Edelskin, N.; Davison, A.; Holm, R. H. I. Am. Chem. Soc.
1964,86,4580.

Gordy, W. In Techniques in Chemistry, Vol. XV: Theory and
Applications of Electron Spin Resonance; West, W., Ed.; Wiley: New
York, 1980.

Atherton, N. M. Principles of Electron Spin Resonance, Ellis—Harwood
PTR Prentice Hall: New York, 1993.

Lucken, E. A. C. Nuclear Quadrupole Coupling Constants, Academic
Press: New York, 1969.

Chapter VI of this Dissertation.
Mims, W. B.; Peisach, I. I. Chem. Phys. 1978, 69, 4921.
Gerfen, G. J.; Singel, D. J. I. Chem. Phys. 1994, 100, 4127.

Pappenhagen, T. L.; Kennedy, W. R.; Bowers, C. P.; Margerum, D. W.
Inorg. Chem. 1985, 24, 4356.

Mims, W. B.; Peisach, J. In Advanced EPR: Applications in Biology and
Biochemistry, A. Hoff, Ed.; Elsevier: New York, 1988; Chapter 1.

 

 

 

 

Structural I

The nitrogei
measured u
and l- and

techniques.
correlation:
graphical m
Prepared w
Interaction .
stimulated

determine I
the electror
OClahedral

along the C

Chapter VI
Structural Characterization of Bis(aquo)tetracyano-nickelate(III), using one-

and two-dimensional pulsed EPR methods

V1.1. Abstract

The nitrogen ligand hyperfine couplings of Ni(III)(CN)4(HZO)2' have been
measured using a combination of isotopic substitution, orientation selection,
and 1- and 2—dimensional Electron Spin Echo Envelope Modulation
techniques. The shapes of the contour lines obtained from hyperfine sublevel
correlation spectroscopy (HYSCORE) experiments were analyzed, using the
graphical method developed by Dikanov and Bowman [Ref (25)] for samples

prepared with C15N'. The results show an axially symmetric hyperfine
interaction with lAll | =1.93 MHz and I A i I =1.06 MHz (for 15N). Conventional

stimulated echo experiments on 14N—containing samples were done to
determine the nuclear quadrupole parameters and gain further insight into
the electronic structure of a Ni(III) complex in a tetragonally elongated
octahedral ligand field. The cyanide 14N nuclear quadrupole coupling is

characterized by a quadrupole coupling constant, ezqzzQ= 3.67 MHz and an
asymmetry parameter, 11:0.09, with the principal axis of the NQI tensor being

along the C-N bond.

134

 

\l,2. lntrod
liydrogena
of microorg
mechanism
biologists l
means of p
based on tl
hydrogenas

Base
Spectroscor
0f the HZ-b
show a cor
isotopic su
and have t
catalyticall
Chemically
llnready‘ 1
hOLlrS; Nil
reduced to
Observed e

rhombic E

135

V1.2. Introduction

Hydrogenases are enzymes responsible for hydrogen metabolism in a variety

of microorganisms by catalyzing the reaction H2 4:) 2H” + 2e' (1-5). The
mechanism of the catalytic cycle presents an interest not only from the
biologist's point of View but also in the scope of a search for an efficient
means of producing hydrogen gas (6). These enzymes have been classified
based on their metal content as 'iron only', Ni—Fe, and Ni—Fe—Se
hydrogenases, with the Ni—Fe hydrogenases representing the largest group.

Based primarily on Electron Paramagnetic Resonance (EPR)

 

spectroscopic data, the Ni-site in these enzymes has been postulated to be part
of the Hz-binding site (1). At ambient redox potentials Ni-Fe hydrogenases
show a composite of two EPR signals, which have been shown by 61N i
isotopic substitution studies to originate from a Ni-based paramagnetic center

and have been termed Ni-A and Ni-B (7). This oxidized state of the enzyme is

 

catalytically inactive. To become active, the protein must be reduced
chemically or by incubation under H2. The Ni-A signal corresponds to the
'unready' form of the enzyme, whose complete activation may take several
hours; Ni-B has been assigned to the 'ready' state, Which can be quickly
reduced to the active form. Upon reduction, a new EPR signal termed Ni-C is
Observed and has been attributed to a hydride complex of the Ni-center. The

rhombic EPR signals Ni—A, Ni-B and Ni-C differ considerably, which

 

 

 

indicates m:
.\’i upon rec

This I
ESEEM) sti
was accessil
same study
nucleus in t
basis of the
postulated t

A rec
hydrogenas
chuter cont
data togeth
59 hydroge
Shows the t
Cysteiny].5l
CI’Steine 51(
second met
molecules j
Significant
revealed a]
for OXidiZe

the nltIOge

136

indicates major structural changes in the coordination environment of the
Ni upon reduction and subsequent substrate binding.

This was demonstrated by an Electron Spin Echo Envelope Modulation
(ESEEM) study on the hydrogenase of D. gigas showing that the nickel site
was accessible to solvent only when the protein was in its Ni—C form (8). The
same study showed weak magnetic hyperfine coupling with a nitrogen
nucleus in the vicinity of the Ni in both the Ni-A and Ni—C forms. On the
basis of the nuclear quadrupole interaction (NQI) parameters it was
postulated that this nitrogen might be due to a histidine moiety.

A recent X-ray crystallographic study on the native structure of the

 

hydrogenase from the bacterium Desulfovibrio gigas suggests a binuclear
cluster containing Ni and Fe at the active site (9). Consideration of the X-ray
data together with the results of EXAFS studies on this enzyme and on Ni-Fe-
Se hydrogenases (10-12) led to a proposed structure for this binuclear site that

shows the two metal ions separated by about 2.7 A and bridged by two

 

cysteinyl-sulfur ligands. The terminal ligands for the Ni are also provided by

 

cysteine side chains resulting in a four-sulfur coordination for the metal. The
second metal ion is most likely Fe and is bound by three putative H20
molecules in addition to the bridging cysteines. The X—ray data also show a
Significant amount of disorder at this site and the EPR spectra of the crystals
revealed an 85% / 15% composite of the Ni-A and Ni-B signals, typically found
for oxidized, inactive proteins. Although the Xfray data show no evidence for

the nitrogen-based Ni ligand suggested by previous ESEEM studies of Ni~A

 

 

 

 

and Ni-C, t
be hydroge
binuclear o
The
Ni-center 5
relevant sti
with the Ir
(HFI) and :
nitrogens c
methods. 1
hydrogena
with g; > E
those obse
the axially
PTEViously
ligand hyy
can be use

hydTOgen;

V1.3. Mat
Bistaquo)

add (HOI

137

and N i-C, the side chain of a highly conserved histidine residue, His 72, may
be hydrogen bonded to one of the bridging cysteinyl sulfur atoms of the
binuclear center (9), giving rise to the observed echo modulation.

The lack of knowledge of the electronic structure of the hydrogenase
Ni-center serves to motivate the systematic spectroscopic investigation of
relevant structural model compounds (5, 7). The present work is concerned
with the measurement of nitrogen electron-nuclear hyperfine interaction
(HFI) and nuclear quadrupole interaction (NQI) parameters of the ligand
nitrogens of Bis(aquo)tetracyano nickelate(II[) by 1- and 2-dimensional ESEEM

methods. Although this compound lacks the binuclear structure of the

 

hydrogenase Ni-site, it provides a good magnetic model for this metal center,
with g i > gl ,, 8i I=2.01, and 61Ni hyperfine coupling constants consistent with

those observed for the protein (13-16). The HFI parameters of the protons of
the axially coordinated water molecules in this complex were determined

previously (17, 18). These results will contribute to a foundation upon which

 

ligand hyperfine couplings obtained by ESEEM and ENDOR spectroscopies

can be used to gain insight into the electronic structure of the Ni-Fe

 

hydrogenase active site.

V1.3. Materials and Methods

BiS(aquo) tetracyano nickelate(III) has been prepared according to literature
procedures by oxidation of Ni(II)(CN)4, using a 10-fold excess of hypochlorous

acid (HOCl) (13, 19). After the addition of HOCl the sample was mixed with

 

 

equal volum
immediately
at the sampl
Cit-EPR sp
obtained frc
Cont
Varian E4
instrument
carried out
frequency (
to calibrate
Iespectivel
Hu
dimension
hyperfine
(21,22). n
Shape is d
an 8:] n,

0f the cor

 

138

equal volume of ethylene glycol to enhance glass formation, then frozen
immediately in 4 mm outer diameter quartz EPR tubes. Ni(III) concentration
of the sample was estimated to be 2mM, as judged from the intensity of its
CW—EPR spectrum. KCISN for preparation of the 15N-containing samples was
obtained from Cambridge Isotope Laboratories, Ltd.

Continuous Wave (CW) EPR spectra were obtained on an X-band
Varian E-4 spectrometer. Pulsed EPR studies were performed on a home-built
instrument described in detail previously (20). The pulsed experiments were
carried out at 4.2 K in a cryogenic immersion dewar. An EIP Model 25B
frequency counter and a Micro—Now Model 515B NMR gaussmeter were used
‘ to calibrate the microwave frequency and magnetic field strength,
respectively.

Hyperfine Sublevel Correlation Spectroscopy (HYSCORE) is a two
dimensional ESEEM technique that reveals the correlation between the
hyperfine frequencies of the opposite electron spin manifolds as cross peaks
(21, 22). For disordered samples HYSCORE contour spectra show ridges whose
shape is determined by the hyperfine (HFI) and NQI parameters (23, 24). For
an S=1 / 2, I=1 / 2 spin system the quantitative relationship between the shape
of the contour lines and principal values of the HFI tensor has been derived
(25).

A four-pulse (7r / Z—r-rt / 2-t1-1r-t2-1c / 2) sequence was used to collect the 2-D

time-domain HYSCORE data set. A second microwave pulse channel was

employed for the n-pulse so that its width and amplitude could be controlled

 

 

 

 

 

 

 

independent
was 32 ns, tl
being 8 time
appeared to
and maximi
step phase c
eliminate u

in th
measured a
Then t1 is ii
collected (5.
128x128 p0
lourier Tr;
Points, wit
resulting 0
map (See C
using M A“
also Writte

Cus
Out on the
data Was 1

each Poss}

 

139

independently. The length (full width at half maximum) of the 112/2 pulses
was 32 ns, that of the rc-pulse was 16 ns, with the peak power of the n—pulse

being 8 times that of the It / 2 pulses. The application of the narrow Ir-pulse

appeared to be necessary to achieve efficient mixing of the hyperfine states
and maximize the intensity of the cross peaks in the spectrum (22, 26). A four
step phase cycle, +(0,0,0,0), +(rt,1t,0,0), +(0,0,rc,0), +(1I:,1t,1r,0), was used to
eliminate unwanted spin echoes.

In the HYSCORE experiment the amplitude of the 4—pulse echo is
measured as a function of t2, yielding one row (or "slice") of the data matrix.
Then t1 is incremented and the subsequent slices of the time-domain data are
collected (see Chapter II). Time domain data presented in this paper are of size
128x128 points. The correlation contour plots have been obtained by 2-D Fast
Fourier Transformation of the time-domain data with zero-filling to 512
points, without dead-time reconstruction. The absolute values of the
resulting complex matrix may be represented as a 3—D plot or as a contour
map (see Chapter 5). These calculations and the associated graphics were done
using MATLAB. The HYSCORE simulation program called ”hysline1.m” was
also written in MATLAB and is included in the APPENDIX.

Customary 3-pulse (stimulated echo) ESEEM experiments were carried
out on the 14N-containing sample. Numerical simulation of the time domain
data was based on the diagonalization of the spin Hamiltonian matrix for

each possible orientation of the external field with respect to the molecule

 

 

 

 

 

hen summ
irequency t
function w.
similar to t
pulse exper

graphical tr

ill. Resul‘
The CW -El
identical to
axial g-tens
Splitting. H
nitrogen ar
inhomogen

Fig.
sa“lPle obt
It the EPR
perPendiCt
ESEEM 8pc
to be alOng
SN HFl ter

CharlCtEris

140

then summing all the contributions to obtain the powder average. To get
frequency domain information (ENDOR like spectra), the modulation
function was Fourier transformed using a dead-time reconstruction routine
similar to that described by Mims (27). The simulation program for the 3-
pulse experiments was written in FORTRAN and linked to MATLAB's

graphical toolbox for output display and manipulation.

VI.4. Results and Discussion

The CW-EPR spectrum of a frozen solution sample of Ni(III)(CN)4(HzO)2' is

identical to that reported previously (13) (Figure VI-l.). The spectrum reveals
axial g-tensor anisotropy with 852-201 g, l=2.01 and no resolved hyperfine

splitting. Hyperfine interactions between the unpaired electron and the
nitrogen and hydrogen nuclei of the ligands are masked by the
inhomogeneous broadening of the resonance lineshape.

Fig. VI-2a shows the 4-pulse HYSCORE spectrum of the 15N-containing
sample obtained at a static magnetic field value corresponding to the g i edge

of the EPR spectrum. In this case all molecules with their gll axes
perpendicular to the external magnetic field vector will contribute to the
ESEEM spectrum (28). Taking the principal axes of the cyanide 15N HFI tensors
to be along their corresponding Ni-C bonds, all possible orientations of the

15N HFI tensor will be sampled. The observed contour lineshape is

Characteristic of an axial HFI tensor (24, 25). The inhomogeneous broadening

 

 

 

 

 

 

l

 

 

 

 

 

Figure VI-l. CW—EPR spectrum of Ni(III)(CN)4(HZO)2'.

 

 

due to the
peaks only
homogene
directions
diagonal i:
Dikanov a
fermi cont

These auth
HYSCORE

{onsets of

Where V (I a

8Pin manifr

Wlth Vi beir

that If One ‘

HYSCORE

 

142

due to the spread of the nuclear frequencies results in a broadening of the
peaks only in the direction perpendicular to the diagonal, while the
homogeneous broadening mechanism widens the lines equally in all
directions (29). Thus, the linewidth of the cross-section parallel to the
diagonal is purely due to homogeneous broadening. The formulas derived by

Dikanov and Bowman (25) were used to extract the HFI parameters, i.e. the

Fermi contact constant (A. ) and the dipolar interaction constant, sz—Egl.

150
1'

These authors have shown (cf. Chapter II) that the expected lineshape of a
HYSCORE contour for an S=1 / 2, I=1 / 2 spin system with an axial HFI tensor
consists of two arcs, described by the equations:
Va2=QaVoz+Ga [VI-1a]
vfi2=QBvOf+Gfl [VI-1b]
\ where Va and VB are the nuclear coupling frequencies in the or and B electron

spin manifolds, respectively and

T + 2Aiso 3; 4vL
T + 2AM i 4vL

 

ani) -_- [VI-2a]

i2vL (4v: - A? + 2T2 — Aiso - T)

180

“‘9’ ” T + 2Aiso 2|: 4vL

 

[VI-2b]

with vL being the nuclear Larmor frequency. It is readily seen from Eqs. [VI-1]

 

that if one selects points along one of the two correlation ridges of the

HYSCORE contour map and plots the square of frequencies v2 versus the

 

—;_

 

 

 

 

square of t
intercept 0

along the r
fig. Vl-Zb.
71.317. Sol
Cc yields 1
litres
frequency-
3bl. Both p
well with r
lnderaende
position. F
obtained a
absorptior
ll? Static ;
Spectrum (;
inhomOger
directiOn T
Perpendicr
exPerimen
HFI t911801

the 1W0 PE

143

square of the corresponding vl’s, a straight line with a slope of Q0, and
intercept of Ga should be obtained. In Fig. VI-2a we show six points selected

along the ridge, marked by +. The corresponding (v2)2 vs. (v1)2 plot is shown in
Fig. VI-2b. The slope of the resulting straight line is -0.251, the intercept is
+1317. Solving equations [VI-2a] and [VI-2b], using these values for Q0, and

G0‘ yields two possible sets of HFI parameters: Aiso=i1.35 MHz, T=i0.29 MHz;

A150=i1.65 MHz, T=T—‘O.29 MHz. To confirm these values we have carried out a

frequency-domain numerical simulation of the HYSCORE contours (Fig. VI-

 

3b). Both parameter sets give rise to identical simulation results that agree
well with experiment (Fig. VI-3a). To distinguish between these sets, an
independent HYSCORE measurement was made at a second magnetic field
position. Fig. VI-3c shows the contour plot of the HYSCORE spectrum
obtained at a magnetic field value corresponding to the g.. edge of the EPR

absorption envelope. In this case only molecules with their g” axis parallel to
the static magnetic field vector, Bo, contribute to the "single crystal like"
Spectrum (28). Indeed, the HYSCORE contour plot shows no significant

inhomogeneous broadening, that is, no extra peak broadening in the

direction perpendicular to the diagonal (29). Because the orientation of B0 is

perpendicular to the hyperfine axis of the nitrogens at this field value, this

exPfi‘riment only measures A i=AiSO-T, i.e. the perpendicular component of the

HFI tensor. The observed coupling (the separation between the projections of

the tWO peaks on either axis) is 1.05 MHz, which fixes the actual HFI

 

Figure VI-2

 

 

144

 

Figure VI-2. (a) HY SCORE contour plot obtained from the g l region of the

Ni(III)(C15N)4(H,_O)2- EPR spectrum. (b) Plot of the square of the
correlating frequencies of the six points on the ridge, marked by
+’s. The slope and intercept of the resulting straight line are

explicit functions of the HFI parameters, Aiso and T.

 

 

 

ANNEEV m ~m>v

 

145

 

0))

v1 (MHz)

 

 

 

 

O 5-
0.4
0 3
0.2
01-

ANNEZV mac

 

4.8

4.4

4.0

3.6

3.2

(v1)2 (Msz)

146

Figure VI-3. (a) HY SCORE contour plot obtained at the g i edge of the EPR

spectrum of Ni(III)(C15N)4(HzO)2-. Experimental conditions:

resonance frequency, v0: 8.862 GHZ; static magnetic field,

30:2890 G; 1:200 ns; 128x128 echo amplitudes were collected;

time increment was 50 ns (b) The result of the corresponding

 

numerical simulation. Simulation parameters: Larmor
frequency of 15N, vL=1.25 MHz; isotropic HFI constant, Aiso=1.35

MHz; anisotropic HFI constant, T=+O.29 MHz corresponding to

an effective dipole-dipole distance, refF3.1O A (AEO=1.65 MHz

with T=-O.29 MHz gives the same result); Gaussian linewidth,
0.12 MHz. (c) HYSCORE contour plot at g ,. v0=8.905 GHZ;
Bo=3170 G; 1:200 ns; size of data matrix, 128x128; time increment

between pulses, 50 ns. (c) Simulation using VL=1.35 MHz; the

other parameters are the same as in (b).

$5.): «.2
FIVE o>

 

 

 

147

 

v2 (MHz)

 

 

 

 

 

 

 

 

 

 

 

2.5 I 3

 

 

Ann—1:): c>
ANEEV o>

 

 

148

 

 

 

 

 

© n
T Qém
a

s a m

0 -

ans: ~>

2.5

1.5

v1 (MHz)

 

 

 

(d)

1.5

V1 (MHz)

 

 

2.5 -

0.5 -

2.5

 

 

 

 

 

 

paramete
dipole ap
is shown

N1
into the e
paramete
coupling
angles of
principal
the electr
parameter
nitrogen -
taken at \
”N-ESEE
Fig» VI-4.
Study We

the ratio

 

 

149

parameters at A30=i1.35 MHz and T=i0.29 MHz. According to the point-

dipole approximation T corresponds to reff=3.10 A. The simulated contour plot
is shown in Fig. VI-3d.

Nuclear Quadrupole Interaction (NQI) parameters provide an insight
into the electronic distribution about quadrupolar nuclei (e.g. 14N) (30). The

parameters that fully describe the NQI tensor are ezqzzQ, the quadrupole
coupling constant, 11, the asymmetry parameter, and a,b,c, the three Euler

angles of the rotation matrix, which transforms the NQI tensor into the

principal axis system (PAS) of the g-tensor.The qii are the principal values of

 

the electric field gradient tensor, i=x, y, z with Zqfi=0. The asymmetry

parameter is defined as n=(qxx-qyy) / qzz. The NQI parameters of the CN‘
nitrogen were obtained by simulation of a series of 3-pulse ESEEM spectra
taken at various fields and t-Values (separation of the first two pulses). Two

14N-ESEEM spectra together with the corresponding simulations are shown in
Fig. VI-4. The hyperfine coupling parameters obtained in the CISN’ HYSCORE
study were used for the 14N-ESEEM simulations after apprOpriate scaling by

the ratio of the nuclear g-factors. The NQI parameters determined for the

Ni(III)(CN)4(HzO)2' were equZQ=3.67 MHz, n=0.09 and a,b,c = 0, 1t/ 2, 113/2. The z-

 

axis of the NQI tensor (direction of the largest electric field gradient) was
found to be along the ON bond.
The NQI constants of the organic nitriles are in the range of 3.70-4.27

MHz, while the asymmetry parameters fall between 0.0046 and 0.183 (30).

 

 

 

 

 

 

150

Figure VI-4. (a) 3—pulse ESEEM spectra of Ni(IH)(CN)4(HZO)2' at g l.

Experimental conditions: Bo=2866 G; 1:164 ns; vo=8.838 GHz; (b)
Result of the corresponding numerical simulation using the

following parameters: Aiso=0.90 MT-Iz (scaled for 14N from

 

Aiso=1.35 MHz of 15N by the ratio of the nuclear g-values);
reff=3.10 A; polar and azimuthal angles specifying the orientation

of the principal axis of the axial HFI tensor with respect to the gl ,
axis, 1t/2, 0; e2q12Q=3.67 MHz; 11:0.09; Euler angles that rotate the
principal axis system (PAS) of the NQI tensor into the PAS of the
g-tensor, 0, n/ 2, n/ 2. (c) 3—pulse ESEEM spectra of
Ni(III)(CN)4(HZO)2'at g, I. Experimental conditions: 80:3135 G;
v0=8.838 GHZ; 1:149 ns. (d) Simulated spectrum. Simulation

parameters for HFI and NQI are the same as for (b).

 

 

 

 

 

T

 

IELT

1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L...
4 ' s :10
frequency (MHz)
0 4 5 6 9 1O
frequency (MHz)

 

 

 

152

 

 

 

 

 

 

 

 

 

)—

l l l I l I

 

 

2 4 6 a
frequency (MHZ)

10

 

 

 

 

 

 

l I

 

 

I l 1

1 2 3 4 5 6 7
frequency (MHZ)

1O

 

 

 

 

For exai
MHZ an
expectec
N‘Ql pal
electron
equQ=3
the five-
distribul
Symmeti
sample,

located 1

asymme

153

For example, the parameters for HCN in the solid state at 77 K, are ezqzzQ=4.02
MHz and 11200085 (31). Due to the axial symmetry of HCN the value of n is
expected to be close to 0. Wang and de Boer have determined the HFI and

NQI parameters of coordinated CN‘ in Fe(CN)§', using pulsed EPR and

electron nuclear double resonance (ENDOR) techniques (32). They obtained
ezqzzQ=3.84 MHz, n=0.00. Fe(CN)? is a low-spin 3d5 complex (S=1/ 2), where

the five-electron configuration can be treated as a positive hole equally

distributed between dxz, d and dx2_y2 orbitals. These orbitals are located

yz'
symmetrically with respect to the CN‘ ligands, thus n=0 is expected. In our
sample, however, the presence of the Ni(III) ion with the unpaired electron
located in the dzz orbital breaks this axiality, resulting in a relatively large

asymmetry parameter for cyanide (33).

V1.5. Conclusion

This work details the application of HY SCORE to a disordered sample with
large g-tensor anisotropy. The results demonstrated the power of this 2-D

ESEEM method in that the HFI coupling constants could be accurately
determined without numerical simulation. For Ni(III)(CN)4(H20)2' the

orientation selection of the pulsed experiment due to the Ni(III) g-anisotropy

allows one to make unambiguous assignment of the HFI parameters. Further

insight into the electronic structure of the Ni(III) complex may be obtained

14
from the Nuclear Quadrupole Interaction (NQI) tensor. The N NQI

 

 

parame
an ezqZZ

distribu

154

parameters determined by numerical simulation of 3—pulse ESEEM data show
an ezqzzQ value typical for CN’ and an asymmetry parameter that reflects the

distribution of electrons in the d-orbitals of the Ni ion.

 

11.

13.

14.

16.

155

References

Lancaster, I. R., Jr., Ed. The Bioinorganic Chemistry of Nickel; VCH:
New York, 1988.

Kolodziej, A. F. Prog. Inorg. Chem. 1994, 41, 493.

Kovacs, I. A. Adv. Inorg. Biochem. 1994, 9, 173.

Hausinger, R. P. Biochemistry of Nickel; Plenum: New York, 1993.
Halcrow, M. A.; Christou, G. Chem. Rev. 1994, 94, 2421.

Benemann, J. R. Proc. 10th World Hydrogen Energy Conf. (Cocoa
Beach, Florida; 1994).

Moura, I. I. G.; Moura, 1.; Teixeira, M.; Xavier, A. V. in Nickel and its
Role in Biology; Met. Ions Biol. Systems 1988, 23, 285.

Chapman, A.; Cammack, R.; Hatchikian, C. E.; McCracken, J.; Peisach, J.
PEBS Lett. 1988, 242, 134.

Volbeda, A.; Charon, M-H.; Piras, C.; Hatchikian, E. C.; Frey, M.;
Fontecilla-Camps, I. C. Nature 1995, 373, 580.

He, S. H.; Teixeira, M.; LeGall, J.; Patil, D. S.; Moura, I.; Moura, I. I. (3.;
DerVartanian, D. V.; Huynh, B. H.; Peck, H. D., Jr. ]. Biol. Chem. 1989,
264, 2678.

Eidsness, M. K.; Scott, R. A.; Prickril, B. C.; DerVartanian, D. V.; LeGall,
J.; Moura, 1.; Moura, J. J. G.; Peck, H. D., Jr. Proc. Natn. Acad. Scz. U. S. A.

1989, 86, 147.
Sorgenfrei, 0.; Klein, A.; Albracht, S. P. J. FEBS Lett. 1993, 332, 291.
Pappenhagen, T. L.; Margerum, D. W. I. Am. Chem. Soc. 1985,107, 4576.

KOjima, N .; Fox, I. A.; Hausinger, R. P.; Daniels, L.; Orme-Johnson, W.
H-,‘ Walsh, C. Proc. Natl. Acad. Sci. U. S. A. 1983, 80, 378.

Albracht, S. P. J.; Graf, E. G.; Thauer, R. K. FEBS Lett. 1982, 140, 311.
Moura, J. I. G.; Moura, 1.; Huynh, B. K.; Kruger, H. J.; Teixeira, M.;

DuVarney, R. C.; DerVartanian, D. V.; Xavier, A: 018, 131632;; H' D" Jr.;
LeGall, J. Biochem. Biopys. Res. Commun. 1982, 4 , -

 

 

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

 

156

Lee, H-I.; McCracken, I. I. Phys. Chem. 1994, 98, 12861.
McCracken, J.; Friedenberg, S. I. Phys. Chem. 1994, 98, 467.

Wang, Y. L.; Beach, M. W.; Pappenhagen, T. L.; Margerum, D. W. Inorg.
Chem. 1988, 27, 4464.

McCracken, J.; Shin, D. H.; Dye, I. L. Appl. Magn. Reson. 1992, 3, 305.

Hofer, P.; Grupp, A.; Nebenfiihr, H.; Mehring, M. Chem. Phys. Lett.
1986, 132, 279.

Gemperle, C.; Aebli, G.; Schweiger, A.; Ernst, R. R. I. Magn. Reson. 1990,
88, 241.

Shane, I. J.; Hofer, P.; Reijerse, E. J.; DeBoer, E. I. I. Magn. Reson. 1992,
99, 596.

Hofer, P. I. Magn. Reson., 1994, Series A 111, 77.

Dikanov, S. A.; Bowman, M. K. I. Magn. Reson. 1995, Series A 116, 125.
Hofer, P. In Electron Magnetic Resonance of Disordered Systems
EMARDIS-91 ; Yordanov, N. D., Ed.; World Scientific: Singapore, 1991 ;
pp. 1-15.

Mims, W. B. I. Magn. Reson. 1984, 59, 291.

Hoffman, B. M.; Martinsen, J.; Venters R. A. I. Magn. Reson. 1984, 59,
110.

Poppl, A.; Bottcher, R.; volkel, G. I. Magn. Reson. 1996, Series A 120,
214.

Lucken, E. A. C. Nuclear Quadrupole Coupling Constants, Academic
Press: New York, 1969.

Negita, H.; Casabella, P. A.; Bray, P. I. I. Chem. Phys. 1960, 32, 314.
Wang, D. M.; deBoer, E. I. Chem. Phys. 1990, 92, 4698.

Townes, C. H.; Dailey, B. P. I. Chem. Phys. 1949, 17, 782.

 

 

 

m
lhe vet
example
because
limitatic
data poi
Cd} to l
HYSCOl
Bowmar
(0.8, -3.:
Warncke

the app]:

HYSCOt
brief intr
which is
Then a n

 

 

 

 

Chapter VII

HYSCORE Study on a Tyrosyl Model System

VII.1. Abstract

The very slow relaxation of tyrosyl radicals at low temperatures (4.2 K, for
example) makes pulsed EPR experiments on them rather difficult to perform,
because usually we cannot use sampling rate higher than ca. 10 Hz. This
limitation makes HY SCORE impossible to use, due to the large number of
data points to be collected. We have overcome this problem by adding 5 mM

Gd“ to the sample.

 

HYSCORE data were analyzed using the graphical method of Dikanov and
Bowman [Ref. (33)]. The principal values of the HFI tensor are found to be
(-0.8, -3.1, -3.6) (MHz), which are reasonably close to the values found by
Wamcke and McCracken using 1D ESEEM [Ref (31)]. The work demonstrates

the applicability of the 2D ESEEM to cases of rhombic hyperfine interactions.

The work presented in this Chapter presents an application of
HYSCORE to a randomly oriented sample with rhombic HFI tensor. First, a
brief introduction to the biologically essential tyrosyl radicals is presented,
which is followed by a summary of the general properties of radical enzymes.
Then a review of the most relevant EPR spectroscopic studies is given. Lastly,

We present the results and discussion of our HYSCORE study.

157

 

oxidize.
\‘HE (7,
catalysis
thiyl rac
to oxidi;
to have
been fol

R
non-spe
radicals
has bee]
aging pi
to be thi
blOlOgiq
Class fla

Primarii

 

 

158

VII.2. Introduction

Amino acid radicals are known as essential participants in the reactions of a
growing number of enzymes (1, 17). Tyrosyl radicals have been found in a
variety of enzymes: in Ribonucleotide Reductase (RNR) (2), Photosystem 11
(PS [1) (3), Prostaglandin H Synthase (PGHS) (4), Galactose Oxidase (GOase) (5)
and in amine oxidases (6). Of the 20 amino acids, tyrosine is the easiest to
oxidize. In aqueous solution at pH 7 it has a redox potential of +930 mV vs.
NHE (7); this may partially account for its widespread occurrence in enzyme
catalysis. Cysteine has a similar redox potential (8). It can be oxidized to its
thiyl radical form but is rapidly oxidized further by 02. Tryptophan is also easy
to oxidize; it has a redox potential of +1.05 V vs. NHE (7) and has been shown
to have an important function in cytochrome c peroxidase (9). Glycine has
been found to give rise to a radical in pyruvate formate lyase (10).

Radicals in biology can be classified in two groups: radicals-involved in
on-specific chemistry (which is ultimately deleterious to the cell) and
adicals that are essential for certain enzymes. Destructive radical chemistry
as been recognized for a considerable time and has been associated with the
ging process and cell death (11). Radicals originating from 02 are considered
0 be the principal reactants in these processes. Radicals essential for proper
iological function have been recognized somewhat more recently. In this
lass flavin, quinone and chlorophyll radicals have been discovered first (12),

rimarily because they are relatively stable due to their low redox potentials

 

 

 

 

8X0

del

6X0

6X0

Figure '

 

159

etaIlo-Rad'cal mes

1.Families

- Glycyl/Thiyl radical enzymes

examples: pyruvate formate lyase, anaerobic RNR, clostridial diol
dehydratase;

- BIZ-dependent radical enzymes

examples: glutamate mutase, ethanolamine ammonia lyase;
~02-dependent radical enzymes

examples: galactose oxidase, RNR;

2.Functional and Structural Principles

- Metal catalyzes radical formation, radical abstracts H atom from

 

substrate

- Metal cluster/radical/substrate usually in close physical proximity

3.Example: Ordependent radical enzymes

 

 

 

 

LSubstrate
H”, e' T
r—"—":""——'
l radical , H-atom source

  

.—

OZ/HZO activaci—ng
metal site

T

Thermodynamic sink

Figure VII-1. Summary of various properties of metallo~radical enzymes.

 

 

(200 tc

SPGCll'O

the RN
their dc
during
trappin
excepti.

radicals

M
Figure
eItzyme
Cu, Co
active c
to that t
referrec'
divided
depend
C0mmo:
Functio.
in turn

Though

 

 

160

(—200 to +400 mV vs. NHE) and thus can be detected by conventional
spectroscopic methods. \

The first catalytically essential amino acid radical, the tyrosyl radical in
the RNR from E. coli , was only discovered in the 19705 (13). The reason for
their delayed discovery lies in that they typically occur as transient species
during the catalytic cycle and more sophisticated spectroscopic and / or
trapping methods are required to detect and characterize them. There are
exceptions, such as the Y122 in RNR or YD in PS II, both of which are stable

radicals.

 

VIT.3. Functional and Structural Properties of Radical Enzymes

Figure VII-1 summarizes some generalizations of the properties of radical
enzymes (1). Firstly, these enzymes generally contain a metal - typically Fe,
Cu, Co or Mn (14). When the metal does not occur it is replaced by redox

active cofactors such as S-adenosyl methionine, whose function is analogous

 

0 that of the metal. In spite of these exceptions, radical proteins are also
eferred to as ”metallo-radical enzymes”. Secondly, these enzymes can be
ivided into three different families, namely the glycyl/thiyl group, the B12-
ependent group and the 02-dependent group. Thirdly, these enzymes have
ommon structural and functional principles across these families.
unctionally the metal center acts to generate the amino-acid radical, which
turn initiates catalysis by abstracting a hydrogen atom from the substrate.

ough there are modifications to this principle, the general scheme ”metal

   

 

generat
valid. '.
scheme

channe

metallc

 

and ge:
thermo
radical.

in initi;

 

 

 

161

generates radical and radical initiates catalysis by H-atom abstraction” remains
valid. This functional generalization is accompanied by a related structural
scheme: the metal, the redox-active side chain and the substrate-binding
channel all are physically close in the enzyme structure.

Figure VII—1 gives one example of the operation of the Oz-dependent
metallo-radical enzymes. In this family the role of the metal site is to bind O2
and generate activated metal-bound oxygen species, which serve as a
thermodynamic sink to provide oxidizing power to generate the amino acid
radical. Once generated, the radical then uses the substrate as a H-atom source

in initiating the catalysis.

V114. Example: Galactose Oxidase (GOase)

Galactose Oxidase (GOase) is an extracellular, type 11 copper protein (68 kDa) of
fungal origin (15). It catalyzes the oxidation of several primary alcohols to
aldehydes with the concomitant reduction of O2 to H202, via a two-electron
reaction. The crystal structure of GOase (16) reveals a unique mononuclear
c0pper site with two histidine imidazoles, two tyrosines, and an exogenous
water or acetate in a distorted square—pyramidal coordination. The equatorial
tyrosine is covalently linked to a cysteine residue by a CS bond at the ortho
position from the OH group (Figure VII-2).

The enzyme exists in three well defined and stable oxidation states: the
active, oxidized form is EPR silent, suggesting that the Cu2+ ion is

antzferromagnetically coupled to the free radical. The intermediate form of

 

 

 

 

the em

center:

The en
forms :
specific
molecu

removi

figure
radical
oxidati
mechar
Conditii
m
EPR ha
radical
thirties
Which 1
inhOInc
ENDOI

are USu

 

 

162

the enzyme shows a Cu2+ EPR signal, while the reduced form contains a Cu+

center:

Cu2+—Tyr'+ <62; Cu2"-Tyr g Cu”-Tyr
The enzyme can readily be interconverted between the active and the inactive
forms in a redox titration using ferri/ferrocyanide solution (5, 15). The
specificity of the enzyme for primary alcohols is low, ranging from small
molecules to polysaccharides (18). On the other hand, GOase is stereospecific
removing the pro-S methylene hydrogen of the C-6 alcohol in galactose.

The mechanistic scheme for the catalytic cycle of GOase is shown in
Figure VII-3. Two electrons are transferred from the substrate, one to the free
radical and one to the cupric ion. The O2 then restores the active form by
oxidation of the cuprous center and the modified tyrosine. In the proposed
mechanism, the aldehyde release precedes O2 reduction. Under anaerobic

conditions, galactose is oxidized into aldehyde and the Cu(I) state is stable.

V115. EPR Spectroscopic studies of tyrosyl radicals

EPR has proven to be the most informative spectroscopy to study the tyrosyl
radical in both biological and model systems. Since the tyrosyl radicals are
effectively immobilized in proteins, they give rise to powder EPR spectra, in
which small hyperfine couplings are not resolved, due to the
inhomogeneous line broadening. As a result, besides the customary CW-EPR,
ENDOR and pulsed EPR spectroscopies together with specific isotopic labeling

are usually necessary to obtain hyperfine information.

 

 

 

 

 

His 496

 

“It-Cu ' O 2-‘5

I
O l I '23., 1.94 O
(exogenous acetate (pH 45) Tyr 272

or water (pH 7)) S

C ys 228

Figure VII-2. The active site of Galactose Oxidase [from Ref. ( 17)].

Tyr 295 Tyr 295 Tyr 295
O RCHon
/ C + c / CU\/ /C“\ 0+
H20 o 7,, 272 H20 O‘Tyr 272 chzo: \OTyr 272
inactive active \H
' r r 295
H202 7
OH
H N\'::u / N
\ / \ o ‘-
H‘ R-C-O OTyr272
H/ J 1" transfer
ler 295 Tyr 295 Tyr 295
OH OH l
N H N N r N O”
> CU / \ CU / C (““5fo N \"C / N
O \o T 272 \ ‘ ’ u \ ‘
I Yr OTyr272 R.(l:.o PTyr272
O
/ 02 "\f O + H H
H t H
R

Figure VII-3. A proposed catalytic cycle of Galactose Oxidase

[from Ref. (17)].

 

 

 

 

ENDO
hyperti
and 3 5
CH3 pr
electror

bonds v

relatior

where
unpairi
dihedra

[19).

varioug
26 pro
Positio
all tyre

been ft

the phe

164

ENDOR has been found to be particularly useful in determining the

hyperfine tensor components for various hydrogens. The or-protons at the 2,6

and 3,5 ring positions are characterized by rhombic HFI tensors, while each [3-

CH2 proton is characterized by an axial tensor, which reports on the unpaired
electron spin density on Carbon-1 as well as on the dihedral angle of the CH
bonds with respect to the pZ orbital of G], in accordance with the well-known

relation,

Agzpc_1(B0+B2cosze),

where Al3 is the isotropic hyperfine coupling to the B-proton, p01 is the

 

unpaired electron density‘at Carbon-1, B05. 0 MHz, B2 = 162 MHz, 0 is the

dihedral angle of the C-H bond of the B-proton relative to the pZ orbital of C-1
(19).

TABLE VII-1 summarizes the hyperfine tensors for tyrosyl radicals in
various enzymes and in two model systems. In general, the coupling to the
2,6 protons is relatively weak, reflecting low unpaired spin density at these
positions. The 3,5 and the methylene protons are more strongly coupled. For
all tyrosyl radicals studied thus far, the same basic pattern of spin densities has
been found (Fig. VII-4).

ENDOR can also be valuable in characterizing the hydrogen bonding to

the phenol oxygen (20, 27). It has been shown that in the YD tyrosine of PS II

TABLE 1.

112.6 y
H16 2
H math l
H meth I
“met ll
Hmeth 1
llbond |]
Hbond I
0ll (2)

W

a) Hlpcr

hl Shi. u
0‘! TS}

165

ABLE VII-1. Hyperfine tensors of tyrosyl radicals.

 

Aqueous
tLtom/Position RNR GO YD (PSII) YZ (PSII) PGHS Glass
Fl 3,5 x —26.7 -26.4 -26.8 —25.7 —25.4
:1 3,5 y —8.4 —8.4 —7.9 —8.4 -7.2
91 3,5 2 —l9.6 —21.6 —l9.5 —19.5 -l9.5 -l9.5
'l 2,6 x 5.0 4.8 4.2 5.0 4.9
1 2,6 y 7.6 6.8 7.1 7.5 7.1 6.5
1 2,6 2 2 1 2.4 1.3
{math 11 61.2 43.4 31.5 35.4 69 3.0
‘lmcth .1. 53.7 39.8 27.2 29 58.2 2.3
{mcth n 2 1 11.4 9.8 7.9
1 meth J. -4.5 4.4 1.4
1 bond II 11.3
1 bond .1. no 3.1 3.0
'll (2) -125 ”l “
Lf‘crcnccs 20 , 21 22 23 ‘26 27 b) 2’ ”28' 2'13,

 

l Hchrfine tensor components given in MHz.

l 3171. W., HogansOn. C. W., Espe, M.. Bende

0., Tsai, A.-L.. unpublished"

r, C. 1., Babcock, G. T., Kulmacz, R. 1., Palmer,

 

 

 

 

and 11
bonde
(PGHS
confir
interat
conch
spectn

dilzedr

distrii

as we
invest
about
YOlatlt
aCCOl‘t
dispe'
quam

the re

Wam

PTOto

166

and in the tyrosyl of apo-galactose oxidase the phenol oxygen is hydrogen
bonded, while in the case of Y122 in RNR and Y385 in prostaglandin H synthase
(PGHS) it is not. For the YD and Y122 radicals these findings have been
confirmed by high-field EPR (30). It has also been shown that H-bonding
interactions perturb the basic spin-density pattern only slightly, leading to the
conclusion: the striking variation of lineshapes observed in the CW—EPR

spectra of tyrosyl radical enzymes and models is attributed to variations in the
dihedral angles of the Ii-CH2 bonds, not to variations in the spin-density

distribution.

 

B-CDZ- labeled tyrosyl radicals show unexpected zH-ESEEM lineshapes

as well as anomalous CW-EPR and ENDOR linewidths (23, 27, 28). A detailed

investigation has shown that the rotation mobility of the phenol head group
about the C1-C5 bond can account for this phenomenon. As the barrier to this
rotation decreases, the heterogeneity in the dihedral angles increases, and

accordingly, the magnitude of the isotropic coupling, Ag, will show increased

dispersion. An analysis of 2H—ESEEM spectra has provided a detailed

quantitative picture: both the rotational barrier and relative populations of

the rotamers can be deduced (28, 29)-

Utilizing the advantages of the 2H--ESEEM technique over 1H-ENDOR,

Warncke and McCracken have characterized the coupling of the 3,5 or—

PrOtOIIS, using tyrosine specifically 2H-labeled at these positions (31).

 

 

167

 

 

 

 

Figure VII-4. Spin—density distribution :of tyrosyl radicals.

 

E“
O\

l:

3

ml. be
mater
NaOl
and n
obtair
0.1M

in 10 :
12 M

ml w
P 11me

soluti

magn
3,5-le
then 1
aliqw
rapid
5mM
descr

the U

 

168

VII.6. Sample Preparation
0.1 M NaZHZEDTA stock solution: Weigh 5.84 g (20 mmol) H4EDTA into a 250

mL beaker, then add 100 mL water. Solubility of H4EDTA is small, most of the
material will not dissolve. Prepare 0.4 M NaOH solution by dissolving 1.6 g of
NaOH pellets in 100 mL of H20. Combine this solution with the H4EDTA; stir
and mildly heat the mixture until a clear solution (0.1 M NaszEDTA) is
obtained.

0.1M GdCl3 stock solution: Dissolve 372 mg of GdC13.6HzO (FW=371.70 gmol'l)
in 10 mL of H20. The pH of the solution will be ca. 5.8.

12 M LiCl stock solution: Weigh out 5.09 g of LiCl (FW=42.4 gmol"). Add 5-6

mL water. Stir until the crystals dissolve. Dilute with water to 10 mL.

 

Preparation of tyrosyl radicals by UV irradiation: Mix 250 uL GdC13 stock
solution with 260 11L N aZHZEDTA stock solution in a vial equipped with a
magnetic stirrer. Add 500 uL H20 and 3.4 mL 12 M LiCl solution. Add 5 mg

3,5-2D tyrosine (Cambridge Isotopes, FW=183 gmol‘l). 201,1L of 40 We NaOH is
then added to increase the pH and to cause the tyrosine to dissolve. An
aliquot of this solution is introduced into an EPR sample tube, then frozen
rapidly with liquid N2, which results in glass formation. This sample contains
5mM Gd(EDTA) and 5mM tyrosine. Tyrosine radicals were generated as
described in Ref. (31), using a 950 W Hg lamp with no filtering. The time of

the UV irradiation was 90 5, while the sample was immersed in liquid N2.

Figur
above
limite
parar

very

pout
dean
EXpei

twoe
1vhic
VVher
1111:
Posh

Prob

 

169

VII.7. Results and Discussion

Figure VII-5 shows the CW-EPR spectrum of the sample prepared as described
above. It is identical with those published in the literature (32) and provides
limited information due to the inhomogeneous broadening. Gd3+ is a strongly
paramagnetic ion, containing 7 unpaired electrons (S=7/2). It gives rise to a
very broad EPR signal, which is hardly observable in derivative EPR spectra. ‘

The echo-detected EPR spectrum of the same sample is shown in
Figure VII-6. The data for this spectrum were collected with a repetition rate
of 100 Hz, at 4.2 K temperature. A strong radical signal emerges owing to rapid
relaxation made possible by Gd3+ (in contrast, in the absence of Gd3", echoes
cannot be observed at repetition rates higher than 10 Hz).

Figure VII-7 shows the HYSCORE time-domain data set (128x128
points) (a), and the corresponding contour plot (b). The contour pattern is
clearly different from the axial case (cf. Chapter VI). Rhombic HFI tensors are
expected to give rise to horn-shaped contour lines, so we can assume that the

two clearly resolved arcs (designated by 1 and 2) are the two limiting arcs,

which form the boundary of the horn. They correspond to <1>=0° and (13:90”,
where (I) is the azimuthal angle specifying the position of B0 in the PAS of the

HFI tensor (Figure VII-8). In a randomly oriented sample B0 spans all possible

positions ((1) and 8 take all values between 0 and 90 degrees) with equal

probability. If we fix (I) and let B0 sweep G) from 0 to 90 degrees, we have a

 

 

 

 

 

 

.835 i5 25 m £3 .8228 :68: U: 2 S a 5 causes:

>3 .3 @368me 36:68 38.53%? md m5... m0 85.56QO Maw-730 .m-=> «gamma

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

        

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

- A t ~14. m.- d.d ...-1.
1-111. J1... .— u H.FL|«1....H1F. 1-H.. firithm .u.«.‘.. q “in. 1.. ... OM12 ..
.11 . Td1<fiqam «1:? 1fl1m1m. m1ml.d.1111m1_ .uh 11 U1 F1hl-51.D.111 11.11 1.91131 ..n1.11. I... 0.»! 1 1.1.“. m.“ .. . 1111 . a . ...m
I 1 1 1| 1| .0. .I .. I.1 .01.. 11-1.-.- 0 0|. .- .‘IO 1 1..\.”0. 11. .1 1 1 11. I 1 0' I ill.“- . '1 .1.
nnfa. -11.-12. .:...H ....1. 1.1-- .--.-H 1 m.-. .1 . ..m..1.”-...- ...... . ~1: _ . . .F.. .r: . H: .m .. ...
. . 11111.... n. 11. ... . . .11...m1 1 .lw- . .1”- . . . 1 .-.“u ... . .. 1.... .. ..
.39 lull-Ill .' 1.1 a“..Hov-|111111H6m*w.u1 11.1“1 c .1.Hn.m”.11h 1| 0. ..... 1 1 --.I.. 1....n... 9.! _1-.u|11|. . .. .
.Io . -I|‘O'oc|.. . 01 In” HALT'OIO.1O‘.I‘11‘.HHHH p ”HH‘” 1 N: .I. . I. 6" 0' .01" i0 1. H ........ ns. 1 I‘Jl 1 J .1 .III
.1 1.1.3 . . ..p 1 -1 .1 ......1 ...: - .L. 1.. .11 . . .. u
H‘ 11 -I 1 Ti 9110 .1 1 o. ...... 4 4 . 1!] 1| 151-1 In I 0 .1 . 111cm . .
II n-i. oeokbv ..... '0- II-. “molt. 1“ . a b l 1 1' u 11% m . m
M 1 .. ..._.r. .u UH”.- .H...11.1.l.4.. . _ . . ...-m 1.1. .. . . .. ..
.1 1. . 1 .1 1 .111 ..... . .1111 .. 1 1.. .
.. 11... 1- 1. . 1..-Ill . 1 . 111' 1 .. . ..
_ . .. 1‘ . .1111. .11... 1'; l” 1 v .
. _ . . . 111...._..-. u ...-.11. . .h 11..
1 1 . . 1.4-. 141111 ... . . _
e .. - 1 .. 11.111.1m1 1 11111.11.- ..l 111' 11.1w . 111 . . _ ~. .
I‘ll. 10.1.!ilrllll '1' ‘III. p o _ III; -
1 . . . II. . . . . —. .IIA — .. .
1 . . II 1 . .1 . Illll M . .— .
. . . 1 .-.1L .. _. r .111 . -H u ..
. : ._ . -1...L. 1. . . -... _. .
- . 1 .. 1 . .11111 . ..~ r. ..-. 1.1111 .... .. 1. .
H- . . . .... - u “1.1.1.... “2.... .. .. .11 ..... _. - .. __..
11111 1 _ D . .1. . 11 .1 b_.w1. w..1l 1 A L \II 1 .. .1. .. . 1. . 1—
-11.1 1.1-1.1..114: 1-111 1111.1 .171.-.1.I.111|.1.1« 11-. ...; .. .:.... ”H1 H- -. .H. . ...—...: ..._
.. 7H-.. .- . . u. .1 . Hid-11.13.32... _._... .-1F1-.:..L - .. -u. ._ -1 . w. ..w ._
I.|l . 1 . 1- ..~1— o .0 1 0.111 I 1‘.11l1l II. a. . — 11' . ‘4 .
I. II . 11. 11 ..lv. I'll “....H o. . ..- . I. 1 104 .1 11!... \1 . all .1 o .—
11.-. . u. an” 14“.”. ... . “11w ...- . ... .....H11-Iim .. H- .--.,.1..1.. . -....N . ..T H” T. n m
- ..l.u..:- -.-. - 1.1 1 “...”. ...1. 11- 2.11.... 1M. .1... ._ .. ... .
111.11 .I .- 1 1 1mm . 1111‘ 1111. . «1“:— .1*' ..ovllL. Illll 0.1 I. I N ” ...
- 1-1. .. - s . .- . .--- . .. .1. - -.1 I .. . ,. ..-.. .... ... .
110 . .rrlul l ..u n.“ a 1 .. 1 II 1mm.“ .1- . . . “fen _ ..d— '1 .m H .. 1111. . “”171 m
.1. ll L -..r 3% . , -111. .. 1 :11 ...h . _ . .. .1. ...w. .
1.111 1 ... T 11 . cll‘ll -D . 1 1.. 11"; 11 . I .N 111..- .—.. _
1'1 1. o. 1 1 1 o . . 1 I. I. . I . V 1 . .l ...c—
. I! II.- .. - .1 . y 1 It... . 1 .. 1 . . .l ..-
IHI 11”.” 1n n ....... .1 .1 bu . 1. .ifiHlH ........... _. ...1 . — .‘ H1s.~..1111 .......—.
a... 1.1 07. ‘11: .710 .m. . 1 v A..H1_11n. . 1llulllr|I11 I111 .111. h. d .11
.FM 1. . ~11. . 11.111. ..1..1- . 1 A. . .1 ... - .- .. .... .n .m
.m. 1.1.1 . 1+1. .m.... .m__ ..... l ..... ... 11.... ” ... .m .11. _ ...
. 9.. 1 . 1 t1. .m.1. . .11lbl .. . .— 1 111 ... .1. . ._ .- _.._
1 1 1 1—1 0 ll 1 .l1 1 ..-lb Jmu . . .1 V- c 1.. 1' s 4 1 ...-. .
.0 ... .c ...... 1| Oil-l .... .1. 1 -1 1. 1.. ol1 m I . . .
..H L _ . r. . .11 i .. .1. ...... ._ 1. _ _ .
7 H.... -.H. “.1 1 1. .11 usual q. 1. .11 . -... 1.. . . .1- .:.... ..
1 Hnmu.wnumu.n1.11.u . ”I1 ._ .. H... "11.11 “m. .- — m. ... .
. n; ..r.. mm.— 44 1 .1. 1 . t. 1 ml . u. ..
.. n I. 1 ...d m 1 .”H\mo 1._ 1 u m.. .
..:.... - -. u: .l- m. - H . . _
”a... .1.- 1.- .. ”1.1- .. . . _ _
.m. .o—O . c 001- m. . H— — .mp—
. . ..I 1.111 1111... . .
- m. 1... .111. 1 m . . . _ 1 .‘—.~ .
.11 1 ... 1 . 11" — 01. o 1 w p .1! I
. . .. 111111 11”]..Il1qfi.1llJ 1111.1 . .I 1.uJ. 1H . n H 1 m.” .
. 1 . .ill11 . 1. ... 1 . 1 ... .
m. . u - .111 m.“ 1 W“ . . ..n 1111 . I“ ._“” . ”
_.-. . ...llu . .. 1. - 1.. 111 1 .o . 111 . .. .
.. 1 n m . ... . l .o .
.. . 1 . 1|“. . 1'11... A” . .. 1111.. 1.11 . .m . l1 - — .. .
. 1 III . ‘Ilit. .1“$ 6.. 1.1.4 - II.1 ...b 1. 1"1 n .m. —
— w. I. _ IHlI “m1. 1... 4 I- .1. 911'” 1|§1H1-&.. 1 1 ll . 1*. hr 4
4] ml. 1 I ‘0 . . _n 1
.. ... . _-_ 11;... __. “tin”. 7H. r. ...1” 3.... ..n .:.z ..u. . . .. .
..L 1 n 1— o . coll-ls . .m 1. . u m 1 .1 '0 l1 . 1.‘. 1 . l1 ”m - ... --.. . 1. . .
- .v . u .1 m 6 .0 u... . .... . 1 1 1 III”.- —.#1.3 .111 1 5 w.. u.- 1. I... 1 . ..
..pfi . _. .. 1 .l .... . ...—m. . . . 1 .1... _. .1»; .. .11 . 11 ._. 1’. ... 1.111. g. n .-
1v~ 1 m. 1 1... 11 Ill u... . 1.. . . ...Hu 1 . . 11.1 .. .11.“ .lnloo . V . __. I11. . m ...M
.... 1 . . .. .. II It; .... . ...... . . .ll . c ’11—. .r . 11'“. .1 1 a. .
..... .. . a : H: U 3““. ”H“... w x “3:“. ." ... ..11 1 Hum 1.. Sit. lb _3 .-HH .1” H. _H h .n
H; “3.. L. .11. u... .3. .. . 18 . J n. mu c. .11 a 1. . 4 .. ...
. P 1 1 u 1.. .111 0 1 . l . .1 ll . . .. ..
..J H .. ...1. 1 . .....1: ..l..m.:. .. .. .._ .1. _. fill... $an +. .11..”. .-1. T. t. m: .1”. . _. .. ...
. .h 1.. . 1-11.. . I 1.1 . . 1 II . .0.1 1 ... . 1 .1 . ..
...... m_. m|_u 4—. —f1h Julio. .541. .o {my m.” .1‘1 uo~.‘..v.H 1‘ 1.9.1111... 0-.. .‘h.. ..... ”L“ A.#.— .II I in! 1 1 u _c
T. .... .n: .11 I: .. n. 3; ...: .3: 11111191-. Ti.-. ..fi ...-.- . -1. . _m r ..
._ .. 1. ".111... . .nHu ...H1. .1.“ 1 1111..-. 1- 11.. .:.m . 1--. ... u. _
3 manila” ..:.-hr--. v.11 .2. ...-1 _.Z. ... L ..w 11.13%.“ ....h .- ”nun.” HR 2 E T. .111. _ . T . ”.1.
1 I151 .1 In m... 1 .31100164105 . 1 1 ' . .1. 114 1 . 111-II 1..1 ‘A UIL. 111. .1 A 1 . ... q
. .+..L1~.1.. ...-m .-...- . 1 .111 _ 1 i. a-.. :- ...Hb ..n . .00 1. EH 4 H1 1.. 1 U ”MW. . .111.“ . m" . .
1| nip 4 .1 ‘ .m. 1 A11 I! .1III .1 .D A u 0.04 .0 m I 1.. T 1 .1 II. O If Ill ‘51! “I‘d a. . 1.“ 11ch . “H .... .
... ..ULL. .3. :n ...-.11. 11... .11.? . :1-.." .-n. u; “.71.- ...Ulll...: ..fi .1... u . ”Hutu ..-- :.... “ ....- ...111 91:. ..:....H
,. ..-: .1.- :. T e.-. ”Hut... .. T. H... .-:U H ..:.. H.“- . 1.- - 1 - 111-. -L . . -.. ...1 .... ....
.. .A.’_1.m _. .1 1. 1111” 01.— 1 1 ..1. 1 . 1». 1 .h . 11.1 1.1 1 I l'll| I1 -11.. 11.. .411lloa. .I 1:1 u. 11‘. 15.111 . v.1. . . ~m.
.. 7 cf... .o~ o_.9 o I. III ..-] . 1. . . ~mm. . m 111“ 1 .. Lfil AHMII III ‘17 o .1 00.1..) ..11—0- “A”. v” 1 AL I4 .140 ... .
.. d 1.1.. .l. 1. M1 .17 - 1 111111 .11 .T..w1.~..l fl...“ 0». . . .Ilvt [rill 11 .Illr.~. . 1. 11‘ 1..; u 7H ..llo _-.1 . u.. .m.
.u .11. o no 31H.1 1. 1 Jill- ! H4. 0 .- 1—1. I” I . ~ .fiv 1. .l .Ilrllll A .1 s”!- ol‘- _- O a .111 OaIIIIOL ..l u A... u
14 .I.. or» a... 1 . . (11.110 1. . . .. a m A a 11 10.1.1.1 111111 .' try “11.11 111.1IIIL ... ..r _v. _ .
...» II..- ‘41ol .HIOI .¢- .IIIII 1"ljll4‘v. to m. l «.111 c! .1 _I '119111 AT'I I‘ll'lpn-Jw 0.14.! 41 .r r . P t 1
1M m m.* W11 0 1! I... I'll.» ...w-l no.“ 11 ....vo ..13. mo 2" [In I 4". 0 I" H1. ”H4 H1.- 1 111- .-.1~L-__ A“ o .
11.. m". 9" 1E1. 1TH Izilo ill... —.. ...1- on 011. 1* .4~1d... . [I . 1 1 1.1. .9 ””1. 1“.”11 .— ...«b .. u.
null. 1 .1 max abnh1 .Plu . vulva-I. 9.51m av . .1”... a 1”. 1.“ .4. H“ . 1. «u no 1. m . onm—H 1v.
n....xx...-..n1...n h d” "11-..... L3... .-- .. . . .. .1 .H . 111-H- . .3 _1... _.
...I T... l. .1.|.O .101. 10' .51.. um!.... c 11101.... ..— . 1U WU“. “It- 1.‘1“ . w.— L1 m..1 . u
..p.f. 0.4” ul. .0.‘ .101 .1 1 l1...‘ 3.1..» u T... v.-l. O a on u 111111 1 cc” . 1.
. . a . 1 1 . c n .. h 0 10. 0 III.‘ Filling. .1 . me. I ..... .l‘ c.” a L .1 £11 a Q!- 1 . n m H m
.11—1—h.- on_ 3!! min b .1 oHI-1m .1 >13 .1..— J.- .anIH Ha.p.h.b no .... n H b h - ”Nil-1 r» P._M - m.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E
‘-
j.
...

 

. .:....
.21... .:......”e 37.1.? 13429633. bzuvéfi

1..: can . 2.123- .. - .2.
80¢ .6 o... S 2.6a .a.)
at. 18.-3° 9:5! . .\ 13‘

.. <>¢<sofu$uttli .- .- .1 1 . 11-- ...11 -, . - . 1 1 1

II

 

 

 

 

 

emu O.w-V “LGIOQ umUmu ZED \Im H OHfiE-UU t0 LQQIDUO ..UCUEEUU
N10 OOnm.D “.001..— NLU11LxuNQ "UH-«1.

171

 

:5 26

m wEEmEOU .mEEam 3m0ub-$~ Rm 9: *0 ~75 UBuBmv ozumémmw .o-=> 95me

Am:

cumcmgpm ufimfi$

0.000“

 

 

b p

— - — p — — J

 

 

 

 

 

can: oomuh omwuaou
x om.vu.aemu

emu o.mvngmzoa

Nzw oo«m.mu.aoam

mmu.oaoc§ n: ocunouoa .amc
mmmulmutvuouou

”.uw zen \2. H anneom we Laoaommuucoeeou
mcoo.c>umouon«u

ma

apnqttdwe DUDe

tha

 

Figu]

conu

172

 

 

 

 

 

 

 

(a) \
- 4000
— 3500
- 3000
— 2500
— 20003
2:.

— 1500

5

g --1000

E

a — 500

O

C

3

o
3500 4000

 

 

 

 

 

 

 

 

25 4.5
t1(MHz)

Figure VII-7. (a) Time-domain HYSCORE data (128x128 data points

2

) from 3,5

2H-tyrosyl, with 1:200 ns; (b) the corresponding HYSCORE frequency-domam

contour plot. 30:3129 G, v=8.910 GHz.

 

173

 

 

 

 

Figure VII-8. Definition of the position of the external magnetic field vector

(30) in the Principal Axis System of the (rhombic) hyperfine tensor.

 

situa

such

Whil

Com

HYS
coup
Furt‘
nega

t0 tl

 

174

situation same as the axial tensor, which results in a pair of arcs. There is one

such pair for each (I) and the ensemble of these arcs creates the horn.
When (13:00, the two principal values of this hyperfine tensor will be

A i=Axx and Al |=Azz. When <I>=9O0 these parameters will be A i=Ayy and
Al l=Azz. Thus, evaluating these boundary arcs can, in principle, yield all three
principal values of the rhombic hyperfine tensor. Analysis of arc 1 ((13:90))

following the method of Dikanov and Bowman (33) leads to the axial

hyperfine matrix

:08 0 0
A(90°) = 0 i0.8 0 (MHz), [VI-1]
0 0 :35

while for arc 2 ((19:00) we obtain

$3.1 0 0
A(OO) = 0 i3.1 0 (MHz). [VI-2]
0 0 :35

\

Combination of these leads to the rhombic hyperfine matrix,

i0.8 O O
A = O i3.1 0 (MHz). [VI-3]
0 0 :t3.6

HYSCORE cannot provide direct information on the sign of the hyperfine
coupling; however, it reports on the relative sign of Aiso and T (33).

Furthermore, it is known that Aiso<0 for oc-protons, which is expressed in the
negative value of the McConnell constant (Q(2H)=-10.7 MHz) (19). This leads

to the hyperfine tensor

 

 

 

gvrm

perfc
been
valm

expe

Wan
mat
the s
also

map

agre
Cert;
C0n:

Gen

 

175
—0.8 0 0
A = 0 —3.1 0 (MHz). [VI-4]
0 0 —3.6

These principal values fall in the range of other tyrosyl 3,5, a-proton values

(cf. TABLE VII-1; note that 1H coupling parameters are obtained by

multiplying the corresponding 2H values by 6.5, i.e. by the ratio of the
gyromagnetic ratios of proton and deuteron, 7(1H)/ 7(2H)).

The validity of the above hyperfine parameters has been checked by
performing 1D 3-pulse ESEEM at various T values. Numerical simulation has

been done with the FORTRAN program ”sedeut", using the above principal

 

values as input parameters. The simulations show good agreement with the
experiment (Figure VII—9).

The result of our study is close to the hyperfine tensor determined by
Warncke and McCracken (31) for a tyrosyl model trapped in 40% NaOH glass
matrix. The difference between the hyperfine values is expected considering
the significantly different media in the two experiments. The assignment has
also been confirmed by a numerical simulation of the HYSCORE contour
map performed with the MATLAB routine ”hyslineRom.m” (Appendix).
The simulated and experimental spectra are shown in Figure VII-10. The
agreement is satisfactory with significant discrepancy in the intensities at
certain parts of the contour lines (e.g. the tail of arc 1). This can be understood
considering that the intensities were calculated using the formulae of

Gemperle et a1. (34) derived for S=1/2, 1:1/2. While the frequencies can safely

 

 

176

Figure VII-9. 3-pulse ESEEM spectra of 3,5 2H—tyrosyl radical with the

corresponding simulations; (a) experimental, 1:293 ns. (b) simulated; (c)

 

experimental, 1:438 ns. (d) simulated. Simulation parameters: the principal
values of the HFI tensor (MHz), Axx=0.8, Ayy=3.1, AZZ=3.6; nuclear Larmor

frequency, VL=2.1 MHz.

 

 

 

 

 

 

 

 

(a) ‘ cement: 30.. tyrosyl. tau-293: from: 8.0100 6H2
f11c:0tyn3p.02 f191023155.00 s
tau: 293 no
96 {-m" "l’ 1 F “"’ T j 1"“ TI 1 I ‘l ‘1— !
_ a
, Lt
4’ -+
.J
L J l L l _ _ l —
6 8 10 12

 

frequency (MHz)

 

amplitude

 

 

 

 

frequency (MHz)

 

 

 

 

 

178

C cement: 30.. tyrosyl, tau-438 freq; 8.8100 8H:
file: Bryan .04 ”0102315500 9
tau: 438 ns

 

 

 

 

   

frequency (MHz)

 

amplitude

   

 

 

45M

2 4 6
0 frequency (MHz)

 

 

Fig

Ml

La

pa

179

Figure VII-10. (a) HYSCORE contour plot of 3,5 2H—tyrosyl (same as Figure VI-
7b); (b) the corresponding frequency-domain simulation performed with the

MATLAB program ”hyslineRom.m” using the following parameters: nuclear

Larmor frequency, vL=2.1 MHz; A3526 MHz, T=0.80 MHz, asymmetry

parameter of HFI tensor, 5:0.53; tau-value, 1:200 ns.

 

 

180

 

. (a)

 

 

 

 

 

. a»)

 

 

 

 

 

 

 

 

bec

inle

qua

COI

ter

58]

3P

51]

EE

Ci

181

be calculated with this formula (the small nuclear quadrupole interaction of
the deuterium does not alter the ENDOR frequencies substantially), the
intensities of the cross peaks may be more significantly affected by the

quadrupole interaction.

The “Ir-suppression effect in HY SCORE has been discussed in Chapter 11.

One can easily infer from Eq. [II-11c] that at a certain ’r-value the intensity of

certain cross-peaks becomes zero (blind spots). In randomly oriented samples

this leads to a distortion of the lineshape, which makes the lineshape analysis
of the 1D ESEEM very involved. In HYSCORE, however, partially suppressed
contour lines still provide sufficient information to deduce the hyperfine

tensor parameters without spectral simulation (33). Figure VII-11 shows a

series of HYSCORE contour plots obtained at different t-values. Suppression
of various parts of the lineshape depending on 1: can be observed. These

spectra also demonstrate that the shortest t—value (200 ns) has yielded the least

suppressed features, in agreement with Hofer’s analysis (35).

In conclusion: this work presents an example of the application of 2Iri—
ESEEM, which is an alternative method to 1H-ENDOR to obtain detailed
structural information on organic radicals. It shows that a low-temperature
study of slowly-relaxing radicals is possible if rare earth ions (such as Gds",
Eu“) are added to the sample in sufficient concentration. This study
demonstrates that the contour lineshape analysis of Dikanov and Bowman

can be applied to the tyrosyl radical as well; it yields the parameters of the

 

 

182

rhombic hyperfine matrix Without lengthy spectral simulation.

 

 

 

183

 

Figure VII-11. The I-suppression effect in our experiments. HYSCORE spectra

of the 3,5 2H-tyrosyl radical taken at various values of t: (a) 200 ns; (b) 300 ns;

(C) 400 ns; (d) 500 ns; (e) 600 ns.

 

184

 

 

 

 

 

 

 

 

 

 

 

 

 

l-lz)

185

 

 

 

 

 

 

 

 

 

 

      

    

2:5
mum)

1.5 2

 

0.5 1

186

 

 

_ Le)

 

 

 

 

 

 

 

 

 

 

10.

11.

12.

13.

14.

15.

16.

17.

187

References

Sigel, H.; Sigel, A., Eds. Metal Ions in Biological Systems, Vol 30.;
Marcel Dekker: New York, 1994.

Sjc’iberg, B.-M.; Graslund, A. Adv. Inorg. Biochem. 1983, 5, 87.
Barry, B. A.; Babcock, G. T. Proc. Natl. Acad. Sci. USA. 1987, 84, 7099.

Smith, W. L.;E1ing, T. E. Kulmacz, R. J.; Marnett, L J.; Tsai, A. L.
Biochemistry 1992, 31, 3.

Whittaker, M. M.; Whittaker, I. W. I. Biol. Chem. 1990, 265, 9610.

Jones, S. M.;.Mu, D.; Wemmes, D.; Smith, A. J.; Kaus, S.; Maltby, D
Burlingname, A. L.; Klinman, I. P. Science 1990, 248, 981.

DeFillipis, M. R.; Murthy, C. P.; Faraggi, M.; Klapper, M. H.
Biochemistry 1989, 28, 4847.

Priitz, W. A.; Butler, J.; Land, E. J.; Swallow, A. J. Free Radical Res.
Commun. 1986, 2, 69.

Sivaraja, M.; Goodin, D. B.; Smith, M.; Hoffman, B. M. Science 1989,
245, 738.

Wagner, A. F. V.; Frey, M.; Neugebauer, F. A.; Schafer, W.; Knappe, J.
Proc. Natl. Acad. Sci. USA. 1992, 89, 996.

Weindruch, R. Scientific American 1996, Ianuary, 46.

Shinkarev, V. P.; Wraight, C. in The Photosynthetic Reaction Center,
Vol I.; Deisenhofer, J.; Norris, J. R., Eds; Academic Press: San Diego,
1993.

Reichard, P.; Ehrenberg, A. Science 1983, 221, 514.

Frey, P. A. Chem. Rev. 1990, 90, 1343.

Whittaker in Ref (1), pp. 315-360.

Ito, N.; Phillips, S. E. V.; Yad, K. D. S.; Knowles, P. F. ]. Mol. Biol. 1994
238, 794.

I

Fontecave, M.; Pierre, ].-L. Bull. Soc. Chim. Fr. 1996, 133, 653.

 

 

 

 

 

 

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

188

Johnson, J. M.; Halsall, H. B.; Heineman, W. R. Biochemistry 1985, 24,
1579.

Gordy, W. in Techniques in Chemistry, Vol. XV: Theory and
Applications of Electron Spin Resonance; West, W., Ed.; Wiley: New
York, 1980. g

Bender, C. J.; Sahlin, M.; Babcock, G. T.; Barry, B. A.; Chandrashekar, T.

K.; Salowe, S. P.; Stubbe, J. ; Lindstrom, B. ; Petersson, L.; Ehrenberg, A. ;
Sjoberg, B.-M. I. Am. Chem. Soc. 1989, 111, 8076.

Hoganson, C. W.; Sahlin, M.; Sjoberg, B.-M.; Babcock, G. T. J. Am.
Chem. Soc. 1996, 118, 4672.

Babcock, G. T. ; El-Deeb, M. K.; Sandusky, P. 0.; Whittaker, M. M.;
Whittaker, J. W. J. Am. Chem. Soc. 1992, 114, 3727.

Warncke, K.; McCracken, J.; Babcock, G. T. J. Am. Chem. Soc. 1994, 114,

7332.
Rigby, S. E.; Nugent, J. H. A.; O’Malley, P. J. Biochem. 1994, 33, 1734.
Hoganson, C. W.; Babcock, G. T.; Biochemistry 1992, 31, 11874.

Tommos, C.; Madsen, C.; Styring, S.; Vermaas, W. Biochemistry 1994,
33, 11805.

Tommos, C.; Tang, X.-S.; Warncke, K.; Hoganson, C. W.; Styring, S.;
McCracken, J.; Diner, B. A.; Babcock, G. T. J. Am. Chem. Soc. 1995, 117,
10325.

Warncke, K.; McCracken, J. ]. Chem. Phys. 1995, 103, 6829.

Warncke, K.; Babcock, G. T. McCracken, J. J. Phys. Chem. 1996, 100, 4654.

Un, S.; Atta, M.; Fontecave, M.; Rutherford, A. W. ]. Am. Chem. Soc.
1995, 117, 10713.

Warncke, K.; McCracken, J. J. Chem. Phys. 1994,101, 1832.

Barry, B. A.; El-Deeb, M. K.; Sandusky, P. 0.; Babcock, G. T. J. Biol.
Chem. 1990, 265, 20139.

Dikanov, S. A.; Bowman, M. K. J. Magn. Reson. 1995, Series A 116, 125.

 

 

j

 

 

 

189

34. Gemperle, C.; Aebli, G.; Schweiger, A.; Ernst, R. R. J. Magn. Reson.1990,
88, 241.

35. Hc'ifer, P. J. Magn. Reson., 1994, Series A 111, 77.

 

 

 

 

 

 

 

 

Appendix

This Appendix contains the printouts of MATLAB programs that are
necessary to run HYSCORE experiments. It also has two MATLAB scripts
which are capable of simulating HYSCORE contour plots (hyslinel and
hyslineRom). The function of these programs is briefly described at the
beginning of each script file. The operation of the program can easily be

followed by reading the notes embedded in the script.

thyscore_read.m
This m-file reads data output by the 4-pulse hyscore

% data collection program

%

t The data will be placed in a square matrix of dimensron
% cval(8), npts by npts of the format data(tl,t2)

%

fname=input('enter hyscore data file name:','s’);
fp=fopen<fname);

%

% set up floating point parameters

%

freq=fscanf(fp,’%lf',l);
field=fscanf(fp,'%lf',1);
power=fscanf(fp,'%lf’,1);
temp=fecanf(fp,’%lf’,1);
%

% get integer control parameters
%

cval=fscanf(fp,'%d',8);
tau=cval(1);
npts=cval(8);
tlstart=cval(2);
tincr=cval(3);

%

t read in the date of expt
%

date=fscanf(fp,'%d’,3);

%

% read in the data array

%

datazfscanf(fp,'%d');
%

190

 

 

 

 

 

 

191

thyscore_plot2.m
% This m—file uses the time domain data array created by the

% program 'hyscore_read.m' and transforms it into a matrix
% of dimension npts x npts; then it creates a 3—D plot

thyscore_read;

tcreate the data—matrix (D) using the data array 'data' created by
%- 'hyscore_read;
for i=1:npts
for j=lznpts
D(i,j)=data((i-1)*npts+j);
end

end
trun 'baseline' to get the baseline corrected data matrix DBC

%

baseline;

%

tto save memory space delete D and B and b
%

clear D;

clear B;

clear b;

%

k=1znpts;
T(k)=tincr*(k-1)+tlstart;
mesh(T,T,DBC)

xlabel(’t2 (ns)’);
ylabel('t1 (ns)');
zlabel('echo amp. (a.u.)');

 

%
’ call tdplot to plot the time—domain in 3—D (using the mesh routine);

9

ttdplot;

 

 

192

ttdplot.m plot the time-domain in 3—D usrng the plot3 routine
t

%
N=input(’enter the # of slices you wish to see : ’);
num=round(npts/N);
E=eye(npts);
for i=2:npts
if rem(i,num)==
E(i,i)=1i

SH=DBC’*E;
% plot using the plot3 function .
% first generate the matrices needed for the plot3 routine

for i=1:npts

for j=lznpts
X(1.j)=T(1);
end
end
Xm=X*E;
Y=X’;
Ym=Y*E;
plot3(Xm,Ym,SH)
xlabel('t2 (ns)’)
ylabel('tl (ns)’)
zlabel(’echo amp. (a.u.)’)

I I .
ans=input('do you want to change the i of slices ? (y/n) ', 5 )1
1f ans=='y',
tdplot;
end
clear X;
clear Xm;
clear Y;
clear Ym;
clear SH;
clear T;
clear E;

return;

 

 

193

' tfrplot512.m
% this script creates a 3—D plot of the (++) quadrant of the hyscore
% the matrix obtained by the 2—D FT of the time domain data matrix, DBC.

%

t

np=512;
YY=fft2(DBC',np,np);
SPEC=abs(YY);

clear YY;

disp512

end

% disp512.m this file displays the desired portion of the 2—D spectrum,
by the file frplot512;

frinc=1000./(tincr*(hp—1));
f=(0:(np-1))*frinc;

frmin=input(’enter fmin: ’);
frmax=input('enter fmax: ’);

lmin=round<frmin/frinc)+1;
lmax=round(frmax/frinc)+1;

x=lminzlmax;
meSh(f(x),f(x),SPEC(lmin:lmax,lmin:lmax));
xlabel(’f1(MHz)');

ylabel('f2(MHz)');
zlabel(’ampl.(a.u.)’);
PL=SPEC(lmin:lmax,1minzlmax);
ans=input(’would you like to see anot
if ans==’y’,

disp512

end

I’IBI)’.

her frequency range ? (y/n):

 

194

DQZ.m .
this script creates a 3—D plot of the (-+) quadrant of the hyscore
% the matrix obtained by the 2-D FT of the time domain data matrix, DEC.

0P 0P

YY=fft2<DBC',np,np);
SPEC=abs(YY)i

clear YY;
frinc=1000./(tincr*(hp-1));
f=(0:(np-1))*frinc;

frmin=input('enter fmin: ');
frmax=input('enter fmax: ’);

 

lmin=round(frmin/frinc)+1;
lmax=round<frmax/frinc)+1;

x=lmin+1zlmax+l;

Lma=np-lmin+1;
Lmi=np-1max+l;

mesh(f(x),f(x),SPEC(lmin:lmax,Lma:—1:Lmi));
PL=SPEC(lmin:lmax,Lma:—1:Lmi);

xlabel(’-fl(MHz)');
ylabel(’f2(MHz)');
zlabel ’am l.(a.u.)’)' , . I .
ans=input(?would you like to see another frequency range ? (Y/n)‘ ' S )I
if ans==’y',
002;

end

return;

 

 

195

% hysline.m this script generates 2—D frequency correlation maps for dis-
tordered systems with axial hfi tensor;
9

vi=1nput('enter the nuclear Larmor frequency (MHz): ');
ia=input('enter the isotropic hyperfine constant (MHz): ');
T=input('enter the anisotropic HFI constant (T), in MHz ')
tau=input('enter the tau value of the experiment, in ns : ')
del =input('enter the time increment of the exp., in ns: '
wd= input(’enter the gaussian linewidth (MHz): ’);
np= input('enter the # of "data-points”: ');

fmax=1000/del;

define the limiting (principal) values of the HyperFine Interaction
along the two principal axes of the HFI tensor (i.e. along and perpen-
dicular to the external magnetic field

o\° o\° 0P o\° 01°

VAP=—vi+(a+2*T)/2;
VBP=—vi—(a+2*T)/2;
VAM=-vi+(a-T)/2;
VBM=~vi-(a-T)/2;

% generate nmr frequency array;

dt=pi/120;
%t=(pi/2-dt*20):dt:pi/2;
t=0:dt:pi/2;

c=cos(t);

csq=c.*c;

s=sin(t);

ssq=s.*s;

VAN=VAM*VAM;
VAA=VAP*VAP-VAN;,
va=sqrt((VAA*csq)+VAN);
VBN=VBM*VBM;
VBA=VBP*VBP—VBN;
vb=sqrt((VBA*csq)+VBN);
SS=abs(vi‘2—((va+vb).“2)/4);
SS=SS./(va.*vb);
CS=abs(vi“2—((va-vb).“2)/4);
CS=CS./(va.*vb);

k=4*SS.*CS;

9.

o
9-

tau=tau/1000;

sa=sin(va*2*pi*taU/2)i

sb=sin(vb*2*pi*taU/2)i

amP=(k/4) .*ab3 (8a.*sb) . *Sqrt (2* (CS. A2'l'SS-A2) );
9..

generate the 2-D frequency-domain data matrix, SPEC

o\° 0P 0

%r= [VAM,VBM,VAP,VBPli
tfmax=max(abs(r))*l.li

 

 

196

frinc=fmax/(np—l);
f=(0:(np—1))*frinc;
SPEC=zeros(np,np);
for i=1:61
nchana=round(va(i)/frinc)+l;
nchanb=round(vb(i)/frinc)+1;
SPEC(nchana,nchanb)=amp(i);
SPEC(nchanb,nchana)=amp(i);
end

perform the convolution in 2 steps along the matrix SPEC;

o\° o\“

% set up Gaussian line—shape function for convolution

%
Gauss;

for i=1:np
L(i,:)=conv(SPEC(i,:),fg);
end

SP=zeros(np+40,np+40)i
for i=1:(np+40)
SP(:,i)=conv(L(:,i),fg)i
end

we have to account for the size differnce between the matrices before
and after the convolution

o\" o\“

Ds=zeros(np,np);
I P

for 1=1:n
for j=lznp
DS(i,j)=SP(i+20,j+20);
nd
end

hysplotl

hysplotl.m this routine displays the desired frequency range of the matrix
generated by 'hyslinel.m

0P 0'

frmin=input('enter fmin: ’);
frmax=input(’enter fmax: ’);

lmin=round(frmin/frinc)+1;
lmax=round(frmax/frinc)+1;

X=lmin:lmax;
mesh(f(x),f(x),DS(lmin:lmax,lmin:lmax));

xlabel(’f1(MHz)’);

ylabel('f2(MHz)');

zlabel(’ampl.(a.u.)’);

PL=DS(lmin:lmax,lmin:lmax);

ans: input(’would you like to see another frequency range ? (y/n)- . S )r
if ans=='y',

hysplotl;

end

 

 

197

% hyslineRom.m this script generates 2—D frequency correlation maps for
% systems with rhombic hfi tensor;
%

vi=input('enter the nuclear Larmor frequency (MHz): ');
a=input('enter the isotropic hyperfine constant (MHz): ');
T=input('enter the anisotropic HFI constant (T), in MHz : ’);
D=input('enter the rhombicity parameter of the HFI tensor : ’);
tau=input(’enter the tau value of the experiment, in ns : ');
del =input('enter the time increment of the exp., in ns: ’);
wd= input('enter the gaussian linewidth (MHz): ');
np= input(’enter the 0 of 'data-points': ’);

fmax=1000/del;

define the limiting (prinCipal) values of the HyperFine Interaction
along the two principal axes of the HFI tensor (i.e. along and perpen—

dicular to the external magnetic field

oPoPoPoPoP

Vza=-vi+(a+2*T)/2;
Vzb=-vi-(a+2*T)/2;
%

 

an=-vi+(a—T*(1-D))/2;
be=~vi-(a-T*(1-D))/2;
%
Vya=-vi+(a-T*(1+D))/2;
Vyb=-vi-(a-T*(1+D))/2;
%

V22a=Vza*Vza;
V22b=Vzb*Vzb;
V2xa=an*an;,
V2xb=be*be;
V2ya=Vya*Vya;
V2yb=Vyb*Vyb;

% generate nmr frequency array;

dt=pi/120;

t=0zdtzpi/2;

c=cos(t);

csq=c.*c;

s=sin(t);

ssq=s.*s;
V2xya=V2ya-(V2ya-V2xa)*csq;
V2xyb=V2yb-(V2yb-V2xb)*csq;
tau=tau/1000;
SPEC=zeros(np,np);

 

for i=1:61 .
VA=sqrt((V22a-V2xya(i))*csq+V2xya(i))i
VB=sqrt((V22b—V2xyb(i))*csq+V2xyb(1));

SS=abs(vi“2-((VA+VB).“2)/4)i
ss=ss . ./ (VA. *vs);
CS=abs(Vi‘2-((VA—VB).‘2)/4)i
CS=CS./(VA.*VB);

 

 

k=4*SS.*CS;

sa=sin(VA*2*pi*tau/2);
sb=sin(VB*2*pi*tau/2);

amp=(k/4).‘abs(sa.*sb).*sqrt(2*(cs.‘2+ss.‘2));
’.

o

generate the 2—D frequency—domain data matrix, SPEC

Po?

frinc=fmax/(np—1);
f=(0:(np—1))*frinc;

for i=1:61
nchana=round(VA(1)/frinc)+1;
nchanb=round<VB(1)/frinc)+1;
SPEC(nchana,nchanb)=amp(i)+SPEC(nchana,nchanb);
SPEC(nchanb,nchana)=amp(i)+SPEC(nchanb,nchana);
end
end

perform the convolution in 2 steps along the matrix SPEC;

0? AP

‘ set up Gaussian line—shape function for convolution

O

o\°

Gauss;

for i=1:np
L(i,:)=conv(SPEC(i,:),fg);
end

SP=zeros(np+40,np+40);
for i=l:(np+40)
SP(:,i)=conv(L(:,i),fg);
end
‘ we have to account for the size differnce between the matrices

Ds=zeros(np,np);
for i=1:np

for j=1znp
DS(i,j)=SP(1+ZO,]+20);
end
end
hysplotl

 

 

 

HICHIGQN STQTE UNIV

”l! ll ll Buggy ' 3 *