LIBRARY
Michigan State
University

 

 

 

PLACE ll RETURN BOX to remove this Maui from your record.
TO AVOID FINES ratum on or before ddo duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU chn Afflrmdivo Mommy-l Opportunity Instituion
Wain-9.1

 

ADVANCED ELECTRON MAGNETIC RESONANCE STUDIES
OF NITROGEN LIGATION IN PHOTOSYNTHETIC SYSTEMS

By

Michelle Mac

A DISSERTATION
Submitted to
Michigan State University

in partial fulfillment of the requirements
‘ for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemistry

1 996

ABSTRACT

ADVANCED ELECTRON MAGNETIC RESONANCE STUDIES
OF NITROGEN LIGATION IN PHOTOSYNTHETIC SYSTEMS

By

Michelle Mac

Oxygenic photosynthesis in higher plants and cyanobacteria requires the interplay
of two membrane bound pigment-protein reaction centers, Photosystem I and 11 (PSI and
P811). In each of these, the primary photochemical event proceeds via the light induced
generation of a chlorophyll a cation radical Species. These specialized chlorOphylls
initiate a series of unidirectional electron transfer events that ultimately either oxidize
water (PSII) or reduce nicotinamide adenine dinucleotide (PSI). To understand better
both the initial charge separation and subsequent electron transfer events, knowledge
regarding the geometric and electronic structures of the primary electron donor is
desirable. Recent X-ray crystallographic studies have shown that the primary donor in
PSI, P700, consists of a dimer of chlorophyll a molecules; the electronic structure of the
cation radical, however, remains elusive. It has been suggested, on comparison to the
primary donors in the bacterial systems and recent magnetic resonance studies, that the
unpaired electron is delocalized either monomerically or highly asymmetrically over the
two halves of the dimer. By using a combination of histidine-tolerant mutants of

Synechocystis PCC 6803, isotopic enrichment, advanced electron magnetic resonance

spectroscopies and numerical simulations of experimental data, we were able to
determine the magnetic coupling parameters for the pyrrole nitrogens in Pm‘ and.
moreover, identify definitively an axial histidine ligand to P7001 The magnitude of the of
the spin density on the pyrrole nitrogens is supportive of a model for Pm‘ where the
excited state orbital makes a significant contribution to the spin density distribution of the
cation radical. The methodology utilized to determine this structure. a combination of
isotope enrichment, electron magnetic resonance spectroscopy and numerical simulations.
provides a means by which complicated couplings in biological systems can be

consistently and systematically evaluated.

ACKNOWLEDGMENTS

The journey through graduate school can be long and arduous, one rarely makes
the trip alone. The people I’ve had the pleasure of working with over the past six years
have made graduate achool a rewarding and memorable experience.

The histidine ligand work would not have been possible without the histidine
tolerant mutant samples developed by Dr. Bruce Diner and Dr. Xiao-song Tang at
Dupont. Dr. Neil Bowlby was also instrumental in the PSI work, teaching me the PSI
prep and walking me through it several times. Neil was also my source of biochemical
knowledge, and knew just enough spectroscopy to make him dangerous at group
meetings, where his questions always caught me off guard! I am also indebted to Pierre
Dorlet for helping me make oxygen evolving PSII samples and also for his willingness to
come in at the crack of dawn just to help me with sample illumination.

I never had the privilege of a direct collaboration with either Professor Charles
Yocum, or Professor Dave Britt, however the interactions I had with both of them will be
remembered fondly. In particular, the many lively discussions (both at group meetings
and at Dag’s) I had with Charlie about “information rich” spectroscopy and “sporting
methods”, and the bullboys, for whom Dave will always be remembered (even though he

got a graduate student to do the dirty work!).

Choosing to work jointly between Professors Jerry Babcock and John McCracken
was the best decision I made in graduate school. and I have to thank Dr. Costas Varotsis
for suggesting the collaboration during one of our many late night talks during my first
year.

I was also lucky enough to be assigned to work with Dr. Matt Espe. a senior grad
student at that time. Matt became both a mentor and friend and taught me everything he
knew about EPR and BBY preps. He also showed me that signal averaging can be fun!

The many interactions with the members of the Babcock and McCracken groups
made my years at MSU enjoyable. In particular, I would like to thank Dr. Kurt Wamcke,
Dr. Curt Hoganson, Dr. Nikos Lydakis-Simatris and Cecilia Tommos for our many
discussions/arguments about magnetic resonance, not to mention our “lively” EPR
meetings. Kurt was not only the best post—doc with whom I have worked, but a great
person and friend as well.

There are some people in our lives with whom we connect and about whom we
cannot say enough. Kerry Reidy-Cedergren is one of these people. We started grad
school together and have shared every milestone in our lives, both professionally and
personally. She was always willing to take time away from anything she was doing to
talk to me about my research, my life, or, on several occasions, to talk me out of quitting!
She is the most giving and caring person I have met and I cannot express how much I will
miss her when I leave.

I was also lucky enough to have a circle of friends that were supportive and could
always be counted on for a celebration. The closeness and camaraderie I felt with this

group of friends: Matt Gardner, Sara Helvoigt, Chris Powell, Per Askeland and Tom
ii

Halasinski, was the one constant in graduate school. They were always there to share the
good times or to commiserate over a bad day in lab; I’ll miss them tremendously and
hope to be fortuitous enough to find such a good group of friends. I’ll especially miss our
Friday “Happy Hours”. and our late night trips to Dag’s, where T., Brenda and Randy
always had a cold beer and a triangle fish waiting!

There were also other people in the department without whom the thesis writing
process would have been (even more) stressful. I’d like to thank the staff in the Main
Office and the Graduate Office, especially Lisa Dillingham and Beth Thomas. for their
assistance in preparing the final thesis document and also for their friendship. I am also
indebted to John and Jerry’s secretary, Vada O’Donnell, for moral support over my
graduate career. I also have to thank Erik Ruggles for keeping me focused on thesis
writing, which allowed me to finish an unheard of TWELVE hours before the deadline!

The two individuals, however, who have had the greatest influence on my life and
my development as a scientist over the last six years are my advisors, Jerry Babcock and
John McCracken. Through their teachings, guidance and inspiration I have learned to
think independently, to solve problems and, more importantly, to do control experiments!
The confidence that I gained in graduate school is the direct result of the interactions I
had with them and their respective research groups and I would not have done as well as I
have without them. I could not have asked for two better people as advisors and I have
thoroughly enjoyed my time in the lab.

'Lastly, I would like to thank my parents and brothers for their love and support

over the last six years, for believing in me and for teaching me that I could do anything

that I chose to.

iii

To my parents.

iv

TABLE OF CONTENTS

CHAPTER 1:

INTRODUCTION TO PHOTOSYNTHESIS

IN HIGHER PLANTS AND CYANOBACTERIA ............................................................ 1

Photosystem I .............................................................................. 4
P700 .................................................................................. 6
Electron Transfer in PSII ............................................................... 15
Structure ............................................................................ 19

The Oxygen Evolving Complex ................................................. 23
Manganese ......................................................................... 26

EPR ............................................................................................................ 27
EXAFS ....................................................................................................... 30
Calcium Effects .................................................................... 32
Nitrogen Ligation ................................................................... 34

List of References .......................................................................... 35

CHAPTER 2:

ELECTRON MAGNETIC RESONANCE THEORY .......................................... 41
EPR ............................................................................................ 41
Electron Nuclear Double Resonance ..................................................... 55
Electron Spin Echo Envelope Modulation (ESEEM) ................................. 60

The Two-Pulse ESEEM Experiment ........................................... 61
Density Matrix Treatment of ESEEM ......................................... 68
Nuclear Quadrupole Effects ............................................................... 76
List of References .......................................................................... 82

CHAPTER 3:

IDENTIFICATION OF HISTIDINE AS AN AXIAL LIGAND TO Pm“ ................... 83
Background ................................................................................. 83

Chlorophyll ........................................................................... 86
Axial Ligation of Pm’ in PSI92
Materials and Methods ...................................................................... 94
Results ........................................................................................ 96
Discussion .................................................................................... 99
List of References ......................................................................... 108

CHAPTER 4:

MULTIFREQUENCY ESEEM STUDIES OF THE PRIMARY ELECTRON

DONOR TN PSI: THE ELECTRONIC STRUCTURE OF P70; ............................ 112
Background ................................................................................ 112
Materials and Methods .................................................................... 117

V

Growth of Cells ..................................................................... 117

Preparation of Photosystem I Particles ......................................... 117
Spectrosc0py and Spectral Simulations.........................................ll8
Results ....................................................................................... 121
Discussion ................................................................................... 141
Analysis and Interpretation of
Nuclear Quadrupole Interactions in P700’ ....................................... 144
Analysis and Interpretation of
Hyperfine Interactions in P700+ ................................................. 147
Conclusions .................................................................................. l 5 5
List of References ........................................................................ 156
CHAPTER 5:
THE EFFECT OF SR2+ SUBSTITUTION ON THE
OXYGEN EVOLVING COMPLEX OF PHOTOSYSTEM II ............................... 159
Introduction ................................................................................. 1 59
Materials and Methods ................................................................... 161
Results ....................................................................................... 164
Discussion ................................................................................. 172
List of References ........................................................................ 174

vi

LIST OF TABLES

Table 3-1 - Calculated 1: Spin Densities for Metalloporphyrin Cation Radicals ............ 91
Table 4-1- l’N ESEEM Simulation Parameters ................................................ 136
Table 4-2- ”N ESEEM Simulation Parameters ................................................. 139

Table 4-3- Quadrupole parameters for chlorophyll and bacteriochlorophyll
species in vivo and in vrtrol46

Table 44- Experimental and theoretical hyperfine coupling tensor

components for Pm" and Chl a+ ....................................................... 151
Table 4-5- Dipolar couplings and spin densitites for pyrrole nitrogens in Pm’ ............ 152
Table 4-6— Experimental and calculated reduction factors for P700‘ ......................... 154

vii

LIST OF FIGURES

Figure 1-1 - The Z-scheme of photosynthesis representing

non-cyclic electron flow in higher plants ........................................... 3
Figure 1-2 - Proposed architecture for the PSI reaction center .................................. 7
Figure 1-3 - EPR spectrum from P70; from Synechocystis PC C 6803 .......................... 9
Figure 1-4 - Proton “rocking” of YD .............................................................. 17
Figure 1-5 - Proton “sloughing” of Y2 ............................................................ 19

Figure 1-6 - Proposed structure of PSII/OEC including polypeptides and cofactors. . .....21

Figure 1-7 - The Kok cycle ......................................................................... 23
Figure 1-8 - Proposed mechanism of water oxidation in PSII ................................. 25
Figure 1-9 - Multiline EPR spectrum from reaction center core complexes ................. 29

Figure 2-1 - The interaction of the applied magnetic field H,
with the magnetic moment, it ........................................................ 42

Figure 2-2 - Energy level diagram for an S = 1/z , I = ‘/2 spin system ......................... 48
Figure 2-3 - The effect of spin polarization on the spin density at the hydrogen atom ..... 51
Figure 2-4 - The dipolar interaction in an applied magnetic field

between the magnetic moments of the electron spin

and the nuclear spin for a nuclear spin = V2 . .............53

Figure 2-5 - Energy level diagram for S = V2 , I = ‘/2 system
showing the allowed EPR and NMR transitions ................................. 57

Figure 2-6 - ENDOR spectrum showing the first derivative lineshapes
expected for an axial dipolar coupling (top) and the
corresponding powder pattern absorption (bottom) ............................... 59

Figure 2-7 - An inhomogeneously broadened EPR absorbance .............................. 61

viii

Figure 2-8 - Classical description of the formation of a spin echo ............................ 62

Figure 2-9 - Representative time domain trace from twopulse ESEEM experiment ....... 65

Figure 2-10 - Nuclear modulation effect ......................................................... 67

Figure 2-11 - a) Inhomogeneously broadened line indicating
the distribution of values of Ho. b) Energy levels
of a quantized system. Microwave radiation of
frequency (0 is able to induce transitions from any
a state to more than one [3 state

Figure 2-12 - Pulse sequences for the Hahn (top) and
stimulated (bottom) echo experiments indicating
periods ofnutation and precession......................

Figure 2-13 - Energy level diagram for S = V2 , I = ‘/2 system.
The vectors indicate the transition matrix elements
connecting the eigenstates ........................................................................

Figure 2-14 - The interaction of a quadrupolar nucleus with four
pointcharges showing the lowest evergy state for
the nucleus (nucleus B)

Figure 2-15 - Energy level diagram for an S = ‘/2 , = 1 system at exact cancellation...

Figure 3-1 - Comparison of A) porphyrin, B) chlorin and

C) bacteriochlorin compounds .....................................................

Figure 3-2 - Energies of the highest occupied and lowest unoccupied
molecular orbitals for Zn porphyrin, chlorin and
bacteriochlorin compounds ................. 89

Figure 3-3 - ENDOR spectra collected at 6K for Pm‘ globally

labeled with 1’N (top), specifically labeled with I’N

histidine (middle) and "reverse" labeled, I"N histidine

in the presence of IN nitrate (bottom). Spectrometer

conditions for all spectra, unless otherwise indicated:

microwave power, 1.99 mW; magnetic field strength,

3375 G (top), 3356 G (middle), 3367 G (bottom); RF

power, 200 W; RF frequency modulation, 100 kHz ................................ 98

ix

.70

.71

74

.77

.79

87

Figure 3-4- Fourier transformations of three-pulse ESEEM
data from P700 containing natural abundance l"N
(top) and specifically labeled with l’N histidine
(bottom). Spectrometerconditions: magnetic field
strength, 3210 (top), 3195 (bottom); microwave
pulse power, 45 dBm; microwave pulse length
(F WHM), 15 ns; pulse repetition rate, 20 Hz; tau

value, 175 ns and sample temperature, 4K ...............................

Figure 3-5: Three-pulse ESEEM frequency spectra from Pm‘
globally enriched with 1’N (top) and "reverse" labeled
(bottom). Spectrometer conditions: magnetic field
strength, 3195 G; microwave pulse power, 45 dBm;
microwave pulse length (FWHM), 15 ns; pulse
repetition rate, 90 Hz; tau value, 250 ns; sample

temperature, 4K .................................................................

Figure 3-6 - Fourier transformations of three-pulse experimental (top)
and simulated (bottom) ESEEM spectra from Pm’ globally
labeled with I’N. Experimental conditions were identical
to those of Figure 3-5. Simulation parameters: Ar = -0.64

MHz, A" = 1.28 MHz, all other parameters as in experiment ..........

Figure 4-1 - EPR spectra of P700’ containing natural abundance
l"N (top) and istopically enriched with 15N. Spectrometer
conditions: microwave power, 20 dB; center field, 3380 G;
modulation amplitude. 20 Gpp; modulation frequency,
100 kHz; time constant, 200 ms; sweep time, 200 3;

sample temperature, 8 K..........

Figure 4-2 - Multifrequency stimulated echo ESEEM time domain
decay traces from Pm+ containing natural abundance l"N.
Spectrometer conditions: magnetic field strength, as noted;
tau value, 250 ns, B) 162 ns, C) 300 ns, D) 400 ns; microwave

pulse power, A) 45 dBm, B) 49 dBm, C) 55 dBm, D) 60 dBm;

pulse repetition rate, A) 20 Hz. B) 60 Hz, C) 30 Hz,
D) 60 Hz; pulse width (F WHM) 22 ns; sample temperature,

Figure 4-3 - Cosine Fourier transformations of three-pulse ESEEM
time domain traces at multiple microwave frequencies.

Spectrometer conditions as in Figure 4-2 ..................................

...... 100

...... 101

...... 105

.120

...122

...... 123

Figure 4-4 - Multifrequency stimulated echo ESEEM time domain
decay traces from Pm’ isotopically enriched
with 15N. Spectrometer conditions: magnetic
field strength, as noted; tau value. A) 222 ns.

318 ns C) 400 ns; microwave pulse power,

A) 45 dBm, B) 45 dBm, C) 57 dBm; pulse
repetition rate, A) 90 Hz B) 30 Hz C) 30 Hz;
pulse width (F WHM) 22 ns; sample temperature,

Figure 4-5 - Cosine Fourier transformations of three pulse
ESEEM time domain traces at multiple microwave
fi‘equencies. Spectrometer conditions: as in
Figure 4-4 ............................................................................ 127

Figure 4-6 - ENDOR spectrum collected at 6 K for P700+ globally labeled
with 1’N. Spectrometer conditions: microwave power, 1.99
mW; magnetic field strength, 3375 G; RF power, 200 W; RF
fiequency modulation, 100 kHz. ............................................................. 129

Figure 4-7 - The effect of hyperfine anisotropy on the ESEEM
spectrum. Simulation conditions for all spectra unless
otherwise noted: A l: 1.84 MHz (top), 1.64 MHz (middle),
1.44 MHz (bottom); AH : 2.17 MHz (top), 2.56 MHz(midd1e),
1.45 2.97 MHz (bottom); Am: 1.95 MHz; eZQq: 2.69 MHz;
1.46 n: 0.87 MHz; magnetic field strength: 3480 G; tau
value: 162 ns; Euler angles (a,[3,y): 0,0,0 ......................................... 133

Figure 4-8 - The effect of the asymmetry parameter on the ESEEM
spectrum. Simulation conditions for all spectra unless
otherwise noted. A i: 1.64 MHz ; AH : 2.56 MHz; Aim: 1.95
MHz; equ: 2.69 MHz; n: 0.2 (top), 0.4 (middle), 0.8
(bottom); magnetic field strength: 3480 G; tau value: 162 ns;
.Euler angles (a,B,y): 0,0,0134

Figure 4-9 - Experimental (top) stimulated echo ESEEM spectrum
from Pm’ containing natural abundance l"N and
corresponding numerical simulation (bottom).
Experimental conditions: as in Figure 4-3B. Simulation
parameters: as in Table 4-1; additional parameters as in
experimental .......................................................................... 136

xi

Figure 4-10 - Experimental (top) stimulated echo ESEEM spectrum
from Pm+ containing natural abundance "N and
corresponding numerical simulation (bottom).
Experimental conditions: as in Figure 4-SB. Simulation
parameters: as in Table 4-2; additional parameters

as in experimental ............................................................

Figure 4-11 - F AB-MS data on chlorophyll a extracted from PSI isotopically
enriched with 15N. Peaks at 896.65 and 873.68 indicate the

presence of residual l"N ......................................................

Figure 4-12 - Numerical simulations of multifrequency ESEEM spectra
fi'om Pm‘ containing natural abundance l‘N. Simulation
parameters: as in Table 4-1; all other parameters as in

experiment (see Figure 4-3 A-D). . . .. ..................................................

Figure 4-13 - Quadrupole tensor axis systems for a) irnino nitrogens

in imidazole as compared to pyrrole (b) nitrogens ......................

Figure 5-1 - Light minus dark multiline EPR spectra from untreated
(top) and Sr2+ substituted (bottom) RCCs. Spectrometer
conditions: microwave power, 20 mW; modulation amplitude,
20 Gpp; modulationfrequency, 100 kHz; sweep time, 100 s;

time constant, 200 ms; sample temperature, 8 K ........................

Figure 5-2 - Fourier transformation of light minus dark two-pulse
ESEEM data from untreated RCCs. Spectrometer
conditions: magnetic field strength, 3440 G; microwave
pulse power, 44 dBm; microwave pulse width, 22 ns;
pulse repetition rate, 90 Hz; tau value, 200 ns; sample

temperature, 4.2 K .

Figure 5-3 - Fourier transformation of light minus dark two-pulse
ESEEM data from Srz’ substituted RCCs.
Spectrometer conditions: magnetic field strength,

33250 G; all other conditions as in Figure 5-2 . .

Figure 5-4 - Fourier transformation of light minus dark three-pulse
ESEEM data from untreated RCCs. Spectrometer

...... 137

...... 139

..... .140

...... 146

...... 160

...165

..166

conditions as in Figure 5-2 ........................................................ 168

xii

Figure 5-5 - Fourier transformation of light minus dark three-pulse
ESEEM data from Sr” substituted RCCs. Spectrometer
conditions as in Figure 5-3 ......................................................... 169

Figure 5-6 - Three-pulse ESEEM spectra from untreated RCCs after ,
illumination (top), prior to illumination (middle) and the
light minus dark difference spectrum (bottom) collected
at g=l .95. Spectrometer conditions: microwave frequency.
8.955 GHz; magnetic field strength, 3290 G; microwave
pulse power, 20 W; tau value, 213 ns; pulse repetition rate.
90 Hz; sample temperature, 1.8 K170

xiii

Chapter 1

Introduction to Photosynthesis in Higher Plants and Cyanobacteria

Although photosynthesis has been studied extensively since the discovery of
oxygen evolution from plants in 1780 by Joseph Priestley, many of the molecular
intricacies of the process remain an enigma. The mechanism of photosynthesis utilizes
photochemical energy from sunlight to disrupt chemical bonds in stable substrates. The
free energy inherent in this process can be trapped by the organism to synthesize vital
amino acids, proteins, nucleic acids and other essential biomolecules. Two classes of
organisms utilize this photosynthetic process; these classes can be distinguished from one
another by the presence of molecular oxygen as a catalytic product. Because
photosynthetic bacteria do not evolve oxygen, the architecture that supports catalysis and
electron transfer in these systems is simpler than that of higher plants and cyanobacteria,
organisms that produce oxygen. The photosynthetic process of these oxygen-producing

organisms can be described with a deceptively simple reaction
C02 + H20 ‘——_" (CI—120)" + 02

The mechanism behind this reaction is complex and requires the interplay of many
proteins and cofactors, many of which have not been characterized completely. Whereas
bacteria utilize only one pigment protein reaction center, or photosystem, for conversion
of light energy into chemically useful energy, the more complex plant systems contain

two interacting photosystems. The first of these, Photosystem I (PSI) mediates the
l

2
production of oxidized plastocyanin and reduced nicotinamide adenine dinucleotide

phosphate (NADPH), which is used by the plant to fuel the enzymatic cycles associated
with the assimilation of carbon dioxide in the Calvin cycle.‘ Photosystem II (PSII) is
associated with water oxidation and is the site of oxygen evolution.2 Both of these
systems and their respective electron transfer cofactors are represented schematically in
Figure 1-1. This model. known as the Z-scheme,3 indicates the flow of electrons in
higher plants and cyanobacteria; the vertical position of each electron carrier corresponds
to its reduction potential at pH = 7.0.

An interesting aspect of the electron transfer process in photosynthesis is its
vectorial nature. The photosynthetic machinery is located in the plant cell in an organelle
known as the 'chloroplast. The presence of an inner and outer membrane make
chloroplasts analogous to the mitochondria found in respiring eukaryotic systems. The
inner membrane of the chloroplast surrounds an aqueous protein matrix which contains
closed vesicles, or thylakoid membranes. The thylakoid membranes separate two
chemically different environments: the stroma, located on the exterior of the membrane,
and the lumen, located on the interior. The membranes appear on electron micrographs
as flattened sac-like structures and are “stacked” in certain regions of the chloroplast.‘ It
has been estimated that as much as 85% of the P811 reaction centers are concentrated in
these stacked regions, known as grana. PS1 is found to a large extent, in the unstacked
stromal regions of the thylakoid membrane. This asymmetric arrangement assists in
providing directionality to the electron flow; the ultimate consequence of this
directionality is the establishment of both charge and proton gradients across the

photosynthetic membrane that provide the free energy necessary for ATP synthesis.

..ED<Z

.._D<Z

«.EmE .23... E Bo: .5583 0:93-55 wEEomen—B flmoficxmoaozq Co 050:3. N 2:. "T— 9...»;

 

v

 

 

 

_ £89382..—

.‘3 oEeSueiU

 

 

/

Ion.

 

 

 

 

mum
So

 

 

Om

: 529.333.—

 

 

 

cach—

E

owcl

 

 

 

onEou
.23:

 

 

 

 

.5. + No

ofm

4
Because PSI and P811 are linked by electron transfer reactions, their physical

separation necessitates the presence of an additional membrane bound complex utilized
as an electron transfer mediator. Cytochrome béf, distributed evenly between the grana
(PSII) and stromal (PSI) regions of the thylakoid, is the ideal charge transfer interface

between PSI] and PSI.’

Photosystem I

The primary charge separation event in photosynthesis proceeds via the light
induced generation of a chlorophyll (in higher plant and cyanobacteria] systems) or
bacteriochlorophyll (in bacterial systems) cation radical species. This species, PM”,
where hmax indicates the wavelength at which the pigment absorbs maximally (kmax =
700 nm for PSI, 680 nm for P811 and either 960 or 870 nm in the bacterial systems), is
distinguished from the light harvesting and energy transfer chlorophylls by its spectral
and redox properties. Subsequent electron transfer occurs unidirectionally through a
series of redox active cofactors whose stabilization of the transient charge separated states
prevents back reactions and facilitates forward electron transfer. The plant is amazingly
efficient at charge stabilization as the forward reactions occur with a quantum yield
approaching unity.6 The photoactive components responsible for this unidirectional
electron transfer in PSI are located on three polypeptides genetically encoded by the
psaA, psaB and psaC genes (the protein names are derived from their gene loci). These

cofactors include the aforementioned primary electron donor P700. a specialized

5
chlorophyll a species, a primary electron acceptor (A0, also a chlorophyll a species), an

intermediate quinone acceptor (A,. a phylloquinone derivative) and three terminal iron-
sulfur clusters (F,, F A and F3). Reduction of the NADP’ is accomplished by a ferredoxin
protein. The reaction center heterodimer consists of two polypeptides, psaA and psaB.
containing 750 and 734 amino acids residues and with molecular weights of 82 and 83
kDa, respectively. The reaction center complex in plants also includes approximately 100
antenna chlorophst responsible for light harvesting and energy transfer, as well as a
series of intrinsic, which can only be removed by detergent washing, and extrinsic, which
can be removed by salt washing or treatment with chaotropic agents, polypeptides, many
of whose functions are as yet undetermined.

The amino acid sequences of psaA and psaB have been determined and show
sequence homologies of approximately 95% amongst subclasses of higher plants.7 This
high degree of homology remains even upon comparison to the amino acid sequences of
more distantly related organisms. for example the cyanobacterium Synechococcus PCC
7002, especially when conservative replacements are considered. It is interesting to note
that the psaB polypeptide contains a sequence of 100 amino acids that are virtually
invariant amongst species. These amino acids include the cysteine residues ligating the
Fe-S clusters“ as well as the leucine zipper motif’ thought to bind the two hydrophobic
psaA and psaB proteins together.

The arrangement of the polypeptides and the presumed binding sites of the
electron transfer cofactors of PS1 are shown in Figure 1-2. Although the binding sites for
many of the redox active cofactors are not known, it is expected that P700, A0 and A1 are

bound to either psaA or psaB of the heterodimer. The presence of a series of conserved

6
histidine residues in helix XIII of psaA and psaB indicates that these residues may be the

putative axial ligands to the Mgz' atoms of P700 and. or. A0.9 As mentioned. the cysteine
ligands of F, have been identified and show F, as an interpolypeptide bridge between
psaA and psaB. The terminal Fe-S centers (FA and F3) reside on a separate 8.9 kDa
polypeptide known as psaC. These positions were recently confirmed by the 4.5 A X-ray
crystal structure of PSI,'°" however, the positions of the other cofactors could not be

determined unambiguously.

Because the highly unidirectional and efficient forward electron transfer events in
PSI are controlled by the spatial and electronic arrangement of the cofactors involved,
knowledge regarding the structure of these radicals is necessary to understand better the
electron transfer processes in photosynthetic systems. The primary charge separation
event in PSI proceeds via the light induced generation of a specialized chlorophyll a
cation radical, Pm‘. The electronic and geometric structure of this radical can provide
insights into the primary electron transfer events of PS1 and ultimately shed light on the
overall mechanism of electron transfer in PSI. Until recently, the lack of a high

resolution crystal structure has precluded determination of the nuclearity of P700.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

F errodoxin
PsaD
F errodoxin P3 0 E
Docking
FA Fa
PsaC
STROMA
Fl‘
40 A
A1 A1
A0
P
700
LUMEN PSdA PsaB
PsaF
Plastocyanin
Plaslocyanin Docking

 

 

 

Figure 1-2: Proposed architecture for the PSI reaction center.

8
Although resonance Raman" and optical hole-buming experiments12 have

indicated a dimeric structure for P700 in its ground state, no evidence confirming these
findings has been presented in the literature. Optical absorption data, however. provide
evidence for a monomeric structure: deconvolution of the P700 chlorophyll envelope from
these studies shows that P700 has a bandwidth and area consistent with a ratio of 1:2 for
the oxidized species to reduced species.'3 These findings have led to the hypothesis that

.9

P700 in its ground state is structurally a dimeric species while P700 is electronically
monomeric. Although this hypothesis has been investigated quite thoroughly by several
research groups with multiple magnetic resonance techniques (vide infia)," no definite
conclusions regarding the electronic structure of the cation radical have been made and
the ambiguity surrounding its nuclearity and electronic structure remains.

Some insight into the structural arrangement of the PSI reaction center can be
gained from comparison to the simpler bacterial reaction centers. Since the three
dimensional structures of the bacterial reaction centers from Rhodopseudomonas viridis '5
and Rhodobacter sphaeroides" have been determined to high resolution with X-ray
crystallography, these systems can been used as a template for the more complex oxygen
evolving systems of higher plants and cyanobacteria. The bacterial reaction centers
consist of a single photosystem with approximate C2 symmetry. Electron transfer
proceeds unilaterally down the L branch of the protein. The analogies to PSI extend to
the structural arrangement of the primary donor, which has been shown to be a
bacteriochlorophyll dimer with histidine residues as axial ligands to the Mg".”"" These

analogies, although a starting point for hypotheses regarding the structure of P700, must be

approached carefully as the presence of quinones as terminal electron acceptors in the

9
bacterial reaction centers indicates a similarity to PSII rather than PSI, which uses Fe-S

centers in addition to quinones as terminal electron acceptors. Even with these

differences the basic motif for photochemical charge separation is preserved in the

bacterial reaction centers, P811 and PSI.

2200 _
2000 -

1800

 

1600 -

EPR amplitude (a.u.)

1400

W

1200 .

 

l I l A l A I

1 A
3300 3350 3400 3450 3500
Magnetic field strength (Gauss)

Figure 1-3: EPR spectrum from Pm” from Synechocystis PCC 6803.

The presence of an easily trapped light-induced radical species makes each of
these systems (eukaryotic and prokaryotic) especially amenable to the application of
electron magnetic resonance techniques. Oxidation of P700 can be achieved either

photochemically or chemically and leads to changes in the absorbance spectrum at 700,

10
685 and 435 nm. The EPR signal (known as Signal 1. see Figure 1-3) from this cation

radical was initially observed by Commoner e! a].’7 in the 19505 and appears as a
structureless Gaussian centered at g=2.0026 with a peak to peak linewidth of 7-8 Gauss.
Because the EPR linewidth of P700’ is narrowed by a factor of 1N2 with respect to
the chlorophyll cation radical monomer in vitro," it was thought that the in vivo species
consisted of a dimer of chlorophyll a with unpaired electron spin density distributed
equally over the two chlorophyll a macrocycles. Subsequent electron nuclear double
resonance (ENDOR) experiments supported this idea; proton hyperfine couplings in P700
decreased with respect to the chlorophyll a cation radical monomer models.‘9 The
dimeric model was challenged, however, by Wasielewski er 01 20whose second moment
analysis of "C and 2H enriched PSI reaction centers suggested that Pm’ consists of a
single chlorophyll a macrocycle. Moreover, this analysis also predicted the dimeric
nature of Pm’, the primary electron donor in bacterial systems, a structure later confirmed
by the high resolution crystal structure. The evidence for a monomer has also been
corroborated by molecular orbital calculations which have shown that the linewidth of
chlorophyll a cation radicals in vitro is extremely sensitive to environmental factors, most
notably to the strength of ligands bound to the central Mgz’ atom.2| Molecular orbital
calculations on chlorophyll a monomers have also shown that the availability of a low
lying excited state with decidedly different spin distribution can influence the electronic
structure of the radical.22 Because the proton hyperfine couplings measured by ENDOR
are a reflection of the spin density distribution in the radical, they can be affected

dramatically by changes in the electronic state of the cation. O'Malley and Babcock23

11
suggested that the decreased proton hyperfine couplings observed for P700’ in viva did not

occur as a result of electron delocalization over two macrocycles. but instead were the
direct result of differences in the composition of the ground-state electronic configuration
compared to that of the chlorophyll a in vitra. The presence of an energetically accessible
excited state led to the idea that interactions of P700’ with its protein environment. either
through axial ligation of the Mg” atom or by electrostatic interactions of the radical with
nearby charged amino acid residues, could provide the perturbation necessary to mix the
ground and excited state molecular orbitals. By mixing 25% of the first excited state
orbital with the ground state electronic configuration, O'Malley and Babcock were able to
account for the reduced proton hyperfine couplings of Pm’ observed with ENDOR.
Because the EPR signals from chlorophyll species both in vivo and in vitra consist
of a single resonance line inhomogeneously broadened by the presence of multiple small
overlapping hyperfine lines, internal spin redistributions resulting from environmental
influences may not be detected if the perturbation occurs at sites that are either invisible
to magnetic resonance or whose hyperfine coupling constant is small with respect to the
overall linewidth. To overcome this obstacle, many laboratories have used a combination
of isotopic enrichment and advanced magnetic resonance techniques to attempt to
elucidate the electronic structure of chlorOphyll radicals in vivo and in vitro. ENDOR”
and electron spin echo envelope modulation (ESEEM)” spectroscopies have been used
extensively to measure the nitrogen hyperfine and quadrupole interactions in both
chlorophyll model compounds and in the primary donors. It was clear from early
ESEEM experiments on PSI that the nitrogen hyperfine couplings could only be obtained

by using "N isotope enrichment since modulations observed from PS1 reaction centers

12
containing natural abundance I‘N (I=1) were dominated by nitrogen quadrupole

interactions whereas the ESEEM spectrum from PSI enriched with "N (I=l/2) exhibited
only nitrogen hyperfine couplings. In ESEEM experiments performed at 77K. Bowman et
a1. 2 6 used the experimental spectra obtained in conjunction with molecular orbital
calculations to estimate the magnitude of the nitrogen hyperfine and quadrupole
couplings in chlorophyll a model compounds. These studies were extended to the radical
in viva by Astashkin et al”. and the limits for the isotropic and anisotropic parts of the
nitrogen hyperfine coupling originally determined by Norris e101." were confirmed.

With the advent of the high resolution crystal structure from bacteria and
subsequent single crystal EPR and ENDOR studies, elucidation of the electronic structure
of P’ in the bacterial systems was possible.” The unpaired electron was found to be
delocalized asymmetrically over the two halves of the bacteriochlorophyll dimer,
favoring the L branch by approximately a 2:1 ratio. This asymmetry has been attributed
to structural differences in the protein environments of the two bacteriochlorophylls
which manifest themselves as shifts in the energy of the highest filled molecular 1t
orbitals of the bacteriochlorophyll monomers and may contribute to the high efficiency of
forward electron transfer by moving the hole away from the primary electron acceptor.

Recent "H and "N ENDOR studies on frozen solutions of PSI have indicated
either an analogous asymmetric, or completely monomeric, unpaired electron spin density
distribution over the chlorophyll a dimer for P700’.2"2’ However, in systems with multiple
anisotropic hyperfine couplings, unequivocal structural assignments from powder pattern

ENDOR spectra are difficult to make, as often only the turning points of the hyperfine

l3
coupling tensor are observed and correlation of these components to a specific "N

nucleus is difficult without using specific isotope labeling. Also. the relative intensities
of the ENDOR lines are not proportional to the number of coupled nuclei contributing.
The modulations observed in an ESEEM experiment, though. are related directly to the
relative populations of spin systems in the sample (see Chapter 2 for a complete
description of these magnetic resonance techniques). Recent one and two-dimensional
ESEEM experiments have been used to measure the IN hyperfine couplings and nuclear
quadrupole interactions in Pm’. These parameters were not determined conclusively,
since spectral simulations of the experimental data were not successful.

Additionally, ESEEM and ENDOR spectroscopies have been used with site-
directed mutagenesis in an attempt to identify the axial ligands of P700+ in the protein

system.30

Since the Mg” atom in the chlorophyll a macrocycle is coordinatively
unsaturated, amino acid residues with electron donating capabilities have been suggested
as possible ligands; of these residues, histidine is the most likely candidate. The presence
of histidine ligands to P’ in the bacterial reaction centers, as well as conserved histidine
residues in the amino acid sequence of psaB, strengthens this argument. Recently, Cui et
al.’° have performed proton ENDOR and ESEEM experiments on mutants of
Chlamydomas reinhardtii that indicate that the spectroscopic properties of the primary
donor in this system are not affected by mutations of hi3523, a possible ligand to P700.
Three mutants were studied: H523Y, substitution of histine with tyrosine; H523Q, a
glutamine substitution; and H523L, a leucine substitution. Of these, only the glutamine

and leucine mutants were able to grow photoautotrophically, the tyrosine mutant required

exogenous acetate to grow photoautotrophically. However, the growth rate of the H523Q

14
and H523L mutants was decreased with respect to the wild type. All mutants displayed

decreased activity when compared to the wild type. in the case of the H523Y mutant, no
activity (measured as oxygen uptake) was observed. Additionally. this mutant lacked
both an EPR signal and an optical absorption spectrum attributable to Pm’, thus
confirming electron transport and Western blot assays that indicated a lack of PSI.

Spectroscopic investigations of H523Q and H523L by ENDOR and ESEEM
showed no differences in the P700’ signals when compared to wild type. Because
mutations of the axial ligand to P’ in bacterial systems result in the loss of the Mgz’ and
the generation of a bacteriochlorophyll-bacteriopheophytin heterodimer;31 it was expected
that the mutations of the putative his ligand to P700 would result in the generation of a
pheophytin. The spin density distribution in a pheophytin molecule differs considerably
from that of chlorophyll and this difference should be manifested in the ENDOR and/or
ESEEM spectra of P700122 Since no changes in the lineshapes or splittings in the ENDOR
and ESEEM spectra were observed, it was concluded that H523 is not a ligand to P700.
However, since water is ubiquitous in its role as a chlorophyll ligand; the induced
mutations may have allowed water access to the binding site as a bridging ligand,
therefore preserving the integrity of the P700 structure.32 Moreover, changes in the
ENDOR and/or ESEEM spectra would not be expected unless the coupling of the axial
ligand was considerable in magnitude; this idea will be explored in greater detail in
Chapter 3.

More recent site-directed mutagenesis experiments and subsequent spectroscopic
analysis of a PSII lacking strain of Chlamydomonas reinhardtii have indicated that

histidine 656 of the psaB polypeptide is the axial ligand to P700.33 Although no direct

15
spectroscopic evidence of the ligand was presented. mutations of H656, located in the

helix X region of psaB and conserved in all sequences. to Asn or Ser result in changes in
the oxidation midpoint potential. electronic structure and optical properties of P700’ .
Analogous investigations by Lin et al.” on the bacterial reaction center revealed that
changes resulting from mutations altering the ligand environment of the special pair have
a drastic impact upon these properties. This led Webber et al. to conclude that H656 is
close to and most likely a ligand to the central Mg of one chlorophyll of the P700 dimer.
The conclusion must be considered as tentative and rather weak in the absence of direct
spectroscopic support for it.

Although evidence supporting both monomeric and dimeric structures for P70;
exists in the literature, the actual geometric and electronic structure of the radical species
in viva, until now, has yet to be determined. A complete spectroscopic study of P700‘,its

axial ligands and electronic structure can be found in Chapters 3 and 4.

Electron Transfer in PSII

The initial step in the electron transfer reactions of PSII occurs when a photon of
light is absorbed by a light harvesting protein antenna complex (called the LHC)
composed of non-covalently bound chlorophyll a, chlorophyll b and carotenoid molecules
in higher plants and green algae. These antenna pigments then transfer energy via exciton
interaction and Forster transfer to the reaction centers. All of the energy absorbed by the
LHC is transferred to Pm, the primary electron donor, promoting it to an excited singlet

state, which, within 3 picosecondsl’S reduces a nearby pheophytin molecule (Pheo). To

16
prevent recombination. the Pheo' quickly (within 300-600 ps)“ reduces a nearby

plastoquinone, Q A.

The electrons are shuttled out of PS 11 by electron transfer from QA' to another
plastoquinone, QB in about 200 us.37 An interesting change occurs at this point; the
photosynthetic process, previously composed of single electron events. becomes a
multielectron process. QB' remains tightly bound in its binding site until a second
photochemical event reduces it to QBZ’ whereupon it is protonated and released from its

binding site as plastoquinol. The plastoquinol is oxidized by cytochrome b6f and the

electrons are transported to PS1 by a plastocyanin molecule. The QBH2 is replaced in its
binding site by an oxidized quinone from the membrane associated quinone pool.”

Recombination between Pm’ and Q A‘ (which occurs with a half time of
approximately 100 us) would be an energetically "wasteful" process. To prevent this
recombination, oxidation of QA‘ or reduction of P680’ must occur on a timescale
competitive with recombination. The electron transfer between QA‘ and Q8 has been

measured to be in the 100-500 us time range.38

The reduction of P630’, therefore must
occur in the submicrosecond time range. This reduction, found to be dependent on the
oxidation state of the OEC,’9 indicates the presence of a redox active donor that operates
as a charge transfer interface between the CBC and Pm’ in an equilibrium that depends
on the net charge of the CBC.39 Barry and Babcock‘0 identified this donor as tyrosine 161
of the D1 polypeptide (Y1), although under non-physiological conditions other complexes

may donate to P6805 The oxidizing power generated in producing Pm,“ must be directed

17
eventually at water. Tyrosine oxidation, producing the highly oxidizing (Em°=1.0 V) Y;

radical is a means by which to achieve this.

 

 

[B B
x /
/H [H
a o’
-e' A
‘ +6.
/CH2 /CH2

Figure 1-4: Proton "rocking" of YD.

There is a second redox active tyrosine in PS 11 known as YD.‘I It has been
identified as Tyr-160 of the D2 polypeptide by site directed mutagenesis experiments,
and in its oxidized form, is responsible for the dark stable EPR signal known as Signal 11.
The fact that Y; is not involved in the electron transfer reactions that lead to water
oxidation was confirmed with site-directed mutagenesis experiments by Debus et al."
where phenylalanine was substituted for both Y2 and YD. In the YZ case, photosynthetic
growth was absent after the deletion whereas in the YD case photosynthetic growth

continued. The two tyrosine moieties are related by an apparent C2 symmetry. These

18
symmetry related branches. however. are not related functionally. which has interesting

fundamental implications.

Insight into the function of these two apparently related moieties has recently been
garnered by a series of magnetic resonance experiments on both specifically labeled
model tyrosine compounds and on isotopically labeled radicals in vivo.42 The suggestion
that YD acts purely as an electron transfer cofactor is supported by ENDOR experiments
indicating the presence of a well-ordered hydrogen bond.’3 A proton “rocking” motion is
therefore proposed for YD (Figure 1-4). YD is most likely hydrogen bonded to H189(D2)"
and upon oxidation, the “sense” of the hydrogen bond is reversed. allowing the proton to
remain in the site. The hydrophobic environment","3b of YD facilitates this and nuclear
motion along the reaction coordinate is minimized, keeping the reorganization energy
small and, therefore, maximizing electron transfer rates.

In contrast, the close environment of YZ is hydrophilic,“5 facilitating proton
movement. Additionally, the lack of a well ordered hydrogen bond as well as increased
rotational mobility about the CB-C. bond in Y2 has provided support for a model where

Y2 acts as a hydrogen atom abstractor in a purely catalytic function.“7

This proton
“sloughing” model where oxidation is coupled to deprotonation, is shown in Figure 1.5.
Studies of proton release patterns of PSII support this hydrogen atom abstraction model,
each S state transition is accompanied by a concomitant proton release and the release is
coupled to YZ oxidation.” Site directed mutagenesis experiments have supported this

model, as mutations of H190( D1), which facilitates the concerted hydrogen atom transfer

upon oxidation of Y2 by acting as a transient proton acceptor species, have decreased the

19
rate of P680’ reduction by as much as 200 times.’9 The tyrosyl radical produced by this

“sloughing” has an essential function in water oxidation (vide infi'a).

,3 ,B B
/H O/H O
H’
-e' -e'
fast
H20\ H20\ H2C\

Figure 1-5: Proton "sloughing" of Y2.

Structure

A better understanding of the functional aspects of the PS II/OEC can be obtained
with an in depth discussion of the polypeptides and cofactors associated with this
membrane bound complex. There are about twenty polypeptides associated with the PS
II/OEC, making it a moderately complex system.’0 This complexity, however, is
remarkably easy to resolve biochemically. Detergent solubilization methods have

allowed isolation of a number of smaller assemblies which have led to the construction of

20
a reasonable working model of the PS Il/OEC. The proposed architecture of PSII

showing relevant polypeptides and cofactors is depicted schematically in Figure 1-6.’0

The PS II/OEC contains both extrinsic and intrinsic polypeptides as well as a
number of enzymatically active cofactors. There are three extrinsic polypeptides. which
serve to stabilize the CBC and preserve oxygen evolution capabilities. The 33 kDa
protein protects the manganese complex associated with water oxidation by maintaining
the structural integrity of the Mn ensemble. Removal of this polypeptide results in the
loss of Mn and, consequently, a loss in oxygen evolution capabilities.SI This polypeptide
also binds directly to the intrinsic polypeptides, an interaction further stabilizing the Mn
complex. Recent crosslinking studies have determined the stoichiometry of this
polypeptide to be 1:1 with an intrinsic polypeptide (the 47 kDa).52 It also serves an
additional function; it isolates the redox active species involved in water oxidation from
the aqueous environment. This is an important function in that the water oxidizing
system is susceptible to spurious reductants and this biochemical barrier provides a redox
shield that allows it to perform its water splitting function successfully. Unfortunately, it
is difficult to study quantitatively due to rearrangements of the intrinsic polypeptide core
that occur upon removal of these soluble subunits.53

The remaining extrinsic polypeptides, the 17 and 23 kDa proteins, also function as
binding sites for cofactors necessary for water splitting chemistry, in particular the
calcium and chloride ions. These two polypeptides also protect the manganese ensemble
from exogenous reductants. Upon removal of these polypeptides, oxygen evolution can

only be maintained with increased concentrations of both calcium and chloride.“

21

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

47 kDa
43 kDa
9.9!... —— 1
D1 Q D2 *-
B :
F2:
b s b
‘1’ ’1’
Bheo
Pm? ._.
......... Yz _ Yo
“‘ (Mn). j
Ca2+ Cl'
33 kDa
23 kDa
l7kEm

 

 

 

Figure 1-6: Proposed structure of PSIl/OEC including polypeptides and cofactors.

22
The photochemical core of PS 11 is formed by two membrane spanning intrinsic

polypeptides. These heterodimers (MW approximately 32 kDa) . commonly known as
D1 and D2, bind the chlorophylls, pheophytins and quinones that mediate the light driven
charge separation reactions.” Sequence and functional homologies between the L and M
subunits of the bacterial reaction center from Rhodopseudomanas viridis and the DVD?
heterodimer have led to the suggestion that D1 and D2 each contain five membrane

56.57

spanning helices. This structural arrangement leads to a pseudo-C2 symmetry for the

PS II/OEC which is broken by the incorporation of the manganese ensemble, proposed to
lie off of the C2 symmetry axis."

The remaining intrinsic polypeptides have a wide variety of functions. By
providing binding sites for accessory chlorophylls the 43 and 47 kDa proteins have
mainly a light harvesting function. The 23 kDa polypeptide can be isolated in its pure

form, and in addition to binding chlorophyll a. has been implicated in maintaining the QB

site on D1."’,"0 Conformational changes to D1 are modulated depending on whether this
quinone-binding site is occupied. The smaller (22 and 10 kDa) polypeptides do not have
binding sites for any of the cofactors. However, upon removal of these two proteins an
increased accessibility of exogenous acceptors like DCBQ to Q A' has been observed; this
leads to the conclusion that thesepolypeptides influence the environment around these

quinones. The final two intrinsic polypeptides (4 and 9 kDa MW) each provide a single

histidine ligand to the heme of cytochrome bmf’I

23
The Oxygen Evolving Complex

The OEC is the site of water oxidation in the photosynthetic process. The half

reaction corresponding to this process.
214,0 __+ 02 + 4H- + 4c-
has an average reduction potential of 0.93 V at pH=5.0.62 The energy required to drive
this reaction comes from the oxidized P630. The midpoint potential for this reaction
P680 ——0 P680’ + c'

has been estimated to be 1.17 V. The stoichiometries of these two reactions bring up an
interesting mechanistic question; a single absorbed photon generates only one oxidizing
equivalent, however the water splitting process is a four electron process. The oxidizing

power of four photons needs to be combined in order to facilitate water oxidation.

e' e' e' e'

K

02

Figure 1-7: The Kok cycle.

Insight into the resolution of this paradox began with the now classic 02 flash

yield measurements of Joliot et aI."3 These experiments showed that the 02 produced by

24
a sequence of brief (about 10 us) saturating light flashes showed a damped oscillatory

pattern with a periodicity of four following the third flash. These data were soon
reproduced by Kok er al. and a working kinetic model was established.“ This model
showed that the PS 11 units function independently in accumulating the four oxidizing
equivalents necessary to split water. This so-called S-state cycle is driven by successive
photoactivations of PS 11. A schematic representation of this cycle is shown in Figure 1-
7.

The S represents the water splitting center and the subscripts indicate the number

of oxidizing equivalents accumulated. Oxygen evolution occurs only after the S4 state

has been achieved. This model accurately explains the details of the flash experiments.
Maximal oxygen evolution occurs after the third flash with repeated maxima after
successive four flash intervals which can be explained if both S0 and SI are dark stable
states.“ The oxygen evolution vs. flash pattern eventually dephases and reaches a steady
state value. The slow dephasing seen is due to the non-zero probability of "misses" and
"double hits" resulting in the loss of S-state coherence.

The mechanism of water oxidation has been long debated. A recent model,"
based on experimental evidence as well as comparison to other enzyme systems,"5 is
presented in Figure 1-8. In this model, the CBC comprises Yz , the Mn cluster, Ca“ and
Cl’ and involves the repetitive abstraction on each S state transition of protons and
electrons fiom water and OH“ bound as substrates. Each of the components maintains an
integral firnction in the water oxidation mechanism: Yz acts to abstract the protons and

electrons from substrate water bound to the Mn tetramer; the Mn cluster provides a means

 

o and- g .
Mn T1 "Q Mn
50% E) S] or... no

 

53 0

01— CF

Figure 1-8: Proposed mechanism of water oxidation in PSII.

26
by which water binding sites are provided as well a site for the delocalization of oxidizing

equivalents; and Ca2’/Cl' prevent the production of potentially toxic or reactive
intermediates like peroxides. This model is also attractive because it retains
electroneutrality throughout the S state transitions. Water oxidation occurs as a concerted
process during the 83-) S,--)S0 transition.

This model is is consistent with recent pulsed ENDOR experiments by Gilchrist et
a]. where the source of the split S3 EPR signal in Cal” depleted PSII samples was found to
be Yz in magnetic contact to the Mn cluster.“ By measuring the hyperfine coupling of
the B methylene protons, and assuming a purely dipolar interaction, the distance between
the Mn cluster and Y2 was estimated to be 4.5 A This physical proximity is necessary, as
studies of other metalloradical enzyme systems have shown that the oxygen activating

metal center is usually less than 10 A from the redox active amino acid.“7

Manganese

The first row transition metal manganese has long been associated with the
oxygen evolving complex of photosynthesis. Its many stable oxidation states (+2 to +7)
make it an ideal metal center for the multielectron chemistry that occurs in the CBC. The
earliest experiments linking Mn to oxygen evolution were performed in the 1930's and
showed that algae grown on low levels of manganese exhibited a decreased ability to
evolve oxygen.“ Later experiments placed the location of manganese directly in PS 11. 1

By using heat shock techniques, Cheniae and Martin"9 demonstrated a loss of oxygen

27
evolution concomitant with the release of Mn. It was later shown that the major fraction

of Mn2+ in highly oxygen evolving thylakoid membranes was in an EPR silent form.70

This amount, calculated to be 4 per 400 chlorophyll. was resistant to release by Caz- and
was released by using acidification processes and subsequently quantified with EPR
measurements. A stoichiometry of four Mn per PS 11 was determined for optimum
functioning of photosynthetic oxygen evolution. Dimers, trimers and tetramers have all
been suggested as structures for this multinuclear cluster. A variety of techniques have
been employed to determine the oxidation state changes of this functional Mn ensemble

that correspond to S-state transitions.

EPR

Perhaps the greatest amount of information regarding the structure of the Mn
ensemble has come from electron magnetic resonance techniques, particularly electron
paramagnetic resonance (EPR). EPR is used as a sensitive probe into the electronic and
geometric structures of paramagnetic substances; it is particularly useful in protein
systems where spectral congestion, the result of multiple interacting nuclei, precludes the
application of other structural determination methods. The Mn ensemble in the oxygen
evolving complex, as it cycles through the S states, provides a light induced, easily
trapped paramagnet (82 state) that makes study of the complex especially amenable to
EPR. Both continuous wave (cw) and pulsed EPR techniques have been employed to

characterize the structure of and ligands to the Mn ensemble.

28
The first EPR signal associated with the Mn ensemble was reported by Dismukes

and Siderer in 1980." This so-called multiline signal was initially observed in samples
that had been rapidly frozen following a flash of light at room temperature. The EPR
signal consists of 18-20 hyperfine lines separated by 85-90 gauss. is centered at g=1.982
: 0.002 and has a linewidth of 1500-1800 gauss. see Figure 1-9, and can only be
observed at temperatures less than 35 K. The signal resembles that of synthetic
antiferromagnetically coupled Mn(III)Mn(IV) dinuclear clusters.72 These mixed valence
compounds show approximately 16 hyperfine lines which have been attributed to EPR
transitions of the two chemically inequivalent ”Mn nuclei. Although the multiline EPR
spectrum is broader and has several more resolved transitions. its resemblance to the
spectra from this and other mixed valence Mn dinuclear clusters have suggested that the
OEC consists minimally of a dinuclear Mn cluster.

The multiline signal amplitude oscillates with flash number, having maximal
intensity after the first and fifth flashes. This flash induced periodicity is analogous to the
oscillations seen by Joliot and Kok in the S-state experiments and the multiline signal has
been attributed to the S2 state of the Kok cycle.

In addition to the multiline signal. a second EPR signal has been associated with
the S2 state.’3 This signal, in non-oriented samples, appears at g=4.1 and is approximately
1000 G wide with no discernible 55Mn hyperfine structure. The signal appears
subsequent to low (140 K vs. 195 K for multiline) temperature illumination and is
converted to the multiline signal when PSII samples are warmed to 195 K. This signal is
stabilized by increased sucrose concentration and by various 02 evolution inhibitors, e.g.

F‘ and small amines?”

29

200 _

-200 -

EPR amplitude (a.u.)

 

 

I A l . I A J . I
2500 3000 3500 4000 4500

Magnetic field strength (Gauss)

Figure 1-9: Multiline EPR spectrum from reaction center core complexes.

The advancement of the S states was also found to be markedly temperature dependent;
S.---->S2 can occur at temperatures as low as 130 K and Sz--->S3 occurs readily only at
temperatures above 210 K. Brudvig and coworkers found that a single intense flash of
light became progressively less effective in generating the multiline signal at
temperatures below 240 K and. at lower temperatures (160 K), continuous illumination

was needed."

30
EXAF S

A number of possible Mn cluster structures have been proposed based on the
experimental constraints provided by Extended X-ray Absorption Fine Structure
(EXAF S) studies. In EXAFS, backseattering of X-ray generated photoelectrons from the
electron charge density of nearby atoms causes modulations of the X-ray absorption cross
section as a function of the electron de Broglie wavelength. These modulations can be
utilized to determine the distance of the electron density from the parent atom. Since
each element has a uniquely defined ionization energy, the results are specific for each
element and provide a probe into the protein environment that focuses exclusively on the
active site. The only drawback to this technique is that EXAFS cannot distinguish the
individual environment of each Mn ion in the cluster, but instead reveals the summation
of electron distributions surrounding all four Mn ions.

Sl has been characterized best by the application of this technique. A strong
backseatterer observed at 2.7 A is consistent with the presence of a first row transition
metal like Mn.77 It is also consistent with Mn-Mn distances measured in di-u-oxo
bridged Mn(III)Mn(IV) model compounds.78 The presence of a single backscatterer at
2.7 A lends support for a dimer of dimer model for the OEC.

An additional backseatterer at 3.3 A is present in the EXAFS data, however
definitive assignment of this to a distance in vivo has not been made. Evidence
identifying this scatterer as arising from either Mn77 or Ca2+” has been presented in the

literature. Peaks in the 1.8-2.0 A range arising from ligands, either nitrogen or oxygen, to

31

the OEC are also apparent.” However, it is not possible to distinguish between these two
atoms with EXAFS. Moreover, since this peak contains a sum of all ligands to all Mn
ions, definitive assignment of these features to a specific ligand is impossible.

These data have not resulted in a unique structure for the OEC, however they have
provided useful constraints for the construction of a likely model. The “dimer of dimer”
model described by Klein et a1”. is perhaps the best representation of the geometrical
structure of the OEC to date. In this model the dimers are linked via u-oxo di-u-
carboxlato bridges and the planes of the two dinuclear Mn cores are tilted with respect to
the membrane normal, presenting a “C” like structure.

EXAFS has also been used to examine the structural changes in the cluster as a
function of the S states. Small changes were observed upon oxidation of the cluster from
$1 to 82, indicative of little or no structural changes."1 Interestingly, though, low
temperature illumination resulting in the g=4.1 signal resulted in a change in
backseattering from 2.7 A to 2.8 A.”2 This is consistent with the structural changes
expected with a change in ground spin state (from S=1/2 for the multiline to an S=3/2 or
5/2 for the g=4.l).

X-ray absorption Near Edge Structure (XANES) spectroscopy has also been used
to examine the oxidation state changes in the OEC. Shifts in the Mn K edge absorption
as a result photoexcitation was used initially as evidence that Mn was being oxidized
during the S, to S2 transition.” This shift is a manifestation of the additional photon
energy necessary to achieve photoionization of the core electrons upon oxidation state

changes. Based on XANES data, the current consensus on the oxidation states in the S,

32
state is Mn(III)Mn(III)Mn(IV)Mn(IV).” Photooxidation to the S, state, then, produces a

paramagnetic state: Mn(III)Mn(IV)Mn(IV)Mn(lV), which yields an S=1/2 state when
antiferromagnetically coupled. The lack of evidence for Mn oxidation during the S, to S3
transition has been interpreted as the oxidation of a nearby amino acid residue. However.
it must be pointed out that the EXAFS data indicates a large structural rearrangement
upon this transition, and edge shifts are only accurate if the geometry and ligation sphere

of the core remain unchanged.

Calcium effects

The role of calcium in PS 11 has been well documented. Initially observed in
photosynthetic membranes from cyanobacteria, calcium requirements have played roles
in higher plant photosynthesis as well.” Reconstitution of calcium depleted membranes
was found to reverse the inhibition of oxygen evolution.“ Depletion of calcitun can be
accomplished by salt washing either in the light or dark with 1-2 M NaCl. This treatment
removes the 17 and 23 kDa polypeptides and allows access to the proposed calcium site.
Salt washing in the light is facilitated by the addition of a chelating agent such as
ethylene diarnine tetraacetic acid (EDTA). Because illumination was found to have an
effect upon the ease of calcium extraction, the role of calcium in PS 11 was associated
with the S-states of the OEC. The ease of extraction progresses in the order S3 > S, .. SD

> S'-

33
The multiline spectrum has been studied to determine the effect of calcium

depletion upon S-state turnover. The S-state transition inhibited by calcium depletion has
been-controversial. Boussac and Rutherford provide evidence for blockage of the S, --->
S0 transition in NaCl washed samples.”7 Continuous illumination of the calcium depleted
samples at 200 K results in a nonnal S, multiline signal whose intensity is comparable in
amplitude to the signal from calcium reconstituted samples. Flash experiments showed
normal multiline formation upon the first flash with a typical decrease in amplitude after
the second flash. However, no increase in intensity was observed upon the fifth flash.
The typical S-state oscillation pattern was present upon calcium reconstitution of these
samples.

Reconstitution of the Caz’ depleted samples with Srz’ restores approximately 60%
of the oxygen evolution capabilities of the unperturbed sample. This decrease in catalytic
activity has been attributed to a decrease in the rate of the S, -) (8,) 9 S0 transition. The
magnetic properties of the cluster are affected by the Sr” substitution and the multiline
spectrum is modified.” These modifications are manifested as losses, splits. shifts and
redistributions of the ”Mn hyperfine lines indicative of small perturbations in the OEC
structure. Sr” substitution also stabilizes the g=4.1 signal. ’9 I will return in Chapter 5 to
a discussion of Ca” effects in PSII and a study of the multiline signal in samples

reconstituted with Sr”.

34
Nitrogen ligation

Electron spin echo envelope modulation (ESEEM) techniques were used to
observe nitrogen interactions between the Mn complex and a proposed histidine ligand in
both cyanobacteria and spinach thylakoid preparations.”o By studying ammonia inhibition
mechanisms in PS 11. Britt and coworkers90b were able to identify hyperfine and
quadrupole frequencies associated with the ”N (I=l) of ammonia. Subsequent isotopic
exchange experiments with "N (I=1/2) were used to confirm the nitrogen from ammonia
as the source of these frequencies. Analogous experiments were performed on oxygen
evolving cyanobacteria preparations.°°’ Cultures of the cyanobacteria Synechococcus
were grown with KNO, as their only source of nitrogen. Again, hyperfine and
quadrupole frequencies were determined by using ESEEM spectroscopy. Isotope

labeling experiments confirmed nitrogen as the origin of these frequencies.

LIST OF REFERENCES

LIST OF REFERENCES

 

‘ For reviews see: a) Golbeck, J. (1992) Annu. Rev. Plant Physiol. Plant Mol. Biol. 43,
293 b) Malkin, R (1996) in Advances in Photosynthesis (Ort, D. & Yocum, C.F., eds.)
Kluwer, Dordrecht, in press.

2F or reviews see: Diner, B.A. & Babcock, G.T. (1996) in Advances in Photosynthesis
(Ort, D. & Yocum, C.F., eds.) Kluwer, Dordrecht, in press. b) Britt, RD. (1996) in
Advances in Photosynthesis (Ort, D. & Yocum, C.F., eds.) Kluwer, Dordrecht, in press.
3 Hill, R. & Bendell, F. (1960) Nature 186, 136.

‘ Staehelin, LA. (1986) in Encyclopedia of Plant Physiology, New Series, Vol. 19,
Photosynthesis III (Staehlin, L.A. & Amtzen, C.J., eds.) Springer Verlag, Berlin, 476.

5 For review, see Malkin, R. (1992) Photosynth. Res. 33, 121.

6 Setif, P. (1992) in The Photosystems: Structure, Function and Molecular Biology
(Barber, J ., ed.) Elsevier, B.V., 471.

7 Andersen, B., Koch, B., Scheller, H. V., Okkels, J.S., Moller, BL. (1990) in Current
Research in Photosynthesis (Batscheffsky, M., ed.) Kluwer, Dordrecht, 671.

8 Golbeck, J.H.& Cornelius, J .M. (1986) Biachim. Biophys. Acta 849, 16.

9 Kossel, H., Dory, 1., Istvan, G. & Maier, R. (1990) Plant Mol. Biol. 15, 497.

‘° a) Krauss, N., Hinrichs, W., Witt, 1., Fromme, P., Pritzkow, W., Dauter, Z., Betzel, C.,
Wilson, K.S., Witt, H.T. & Saenger, W. (1993) Nature 361, 326. b) Schubert, W.D.,
Klukas, 0., Kraub, N., Saenger, W., Fromme,P. & Witt, HT. (1995) in Photosynthesis:
fiom Light to Biosphere, Vol. III (Mathis, P., ed.) Kluwer, Dordrecht, 3.

” Moenne-Loccoz, P., Robert, B., lkegami, I. & Lutz, M. (1990) Biochemistry 29, 4740.

'2 Ratajczak, R. Mitchell, R. & Haehnel, W. (1988) Biachim. Biophys. Acta 933, 306.

35

36

 

'3 Ikegarni, I. & Itoh, s. (1989) Biochim. Biophys. Acta 934. 39.

“ See, for example, a) Astashkin, A.V., Dikanov, S.A. & Tsvetkov. Yu. D. (1988) Chem.
Phys Lett. 152, 258 b) Kab, H., Fromme, P. & Lubitz. W. (1996) Chem. Phys. Lett. 25 7.
197.

" Deisenhofer, J., Epp, 0., Miki, K., Huber, R. & Michel, H. (1985) Nature 318. 618.

'6 Allen, J.P., Feher, G., Yeates, T.O., Komiya, H. & Rees, D. (1987) Proc. Natl. Acad.
Sci. USA 84, 5730.

‘7 Commoner, B.,.Heise, J.J. & Townsend, J. (1956) Proc. Natl. Acad. Sci. USA 42, 710.

" Norris, J .R., Uphaus, R.A., Crespi, H.L. & Katz, J .J . (1971) Proc. Natl. Acad. Sci. USA
68, 625.

'9 Davis, M.S., Forman, A. & Fajer, J. (1979) Proc. Natl. Acad. Sci. USA 76, 4170..

2° Wasielewski, M.R., Norris, J .R., Crespi, H.L. & Harper, J. (1981) J. Am. Chem. Soc.
103, 7664.

2‘ Fajer, J ., Davis, M.S., Brune, D.C., Spaulding, L.D., Borg, D.C. & Forman, A. (1977)
Braokhaven Symp. Biol. 28, 74.

22 Petke, J.D., Maggiora, G.M., Shipman, L.L. & Christoffersen, RE. (1980)
Photochemistry and Photobiolag 31, 243.

2’ O’Malley, P.J. & Babcock, G.T. (1984) Proc. Natl. Acad. Sci. USA 81, 1098.

2‘ Kab, H., Bittersmann-Weidlich, E., Andreasson, L.-E., Bonigk, B. & Lubitz, W. (1995)
Chem. Phys. 194, 419.

2’ a) Davis. I.H., Heathcote, P., MacLachlan, D.J. & Evans, M.C.W. (1993) Biochim.
Biophys. Acta 1143, 183 b) Kab, H. & Lubitz, W. (1996) Chem. Phys. Lett. 251, 193.

2‘ Bowman, M.K., Norris, J .R., Thumauer, M.C., Warden, J., Dikanov, S.A. & Tsvetkov,
Yu. D. (1978) Chem. Phys. Lett. 55, 570.

2’. Astashkin, A.V., Dikanov, S.A., Tsvetkov, Yu. D. & Goldfeld, MG. (1987) Chem.
Phys. Lett. 134, 438.

1’ Norris, J.R., Uphaus, R.A., Crespi, H.L. & Katz, J.J. (1971) Proc. Natl Acad. Sci. USA
68, 625.

37

 

2" a) Lendzian, F., Lubitz, W., Scheer, H., Hoff, A.J.. Plato. M., Trankle. E. & Mobius. K.
(1988) Chem. Phys. Lett. 148, 377. b) Lendzian, F., Huber. M.. Isaacson. R.A.,
Endeward, B., Plato, M.. Bonigk, B., Mobius, K.. Lubitz, W. & Feher. G. (1995)
Biachim. Biophys. Acta 1183, 139.

3° Cui, L., Bingham, S.E., Kuhn, M., Kass, H., Lubitz, W. & Webber, AN. (1995)
Biochemistry 34, 1549.

3' Schenck, C. C., Gaul, D. S,teffen, M, Boxer, S. G. McDowell, L.. Kirrnaier, C. &
Holten, D. (1990) 1n Reaction Centers of Photosynthetic Bacteria (Michel- Beyerle. M. E..
ed. )Springer-Verlag, Berlin, 229.

’2 Goldsmith, J.O., King, B. & Boxer, 8.0. (1996) Biochemistry 35, 2421.

3’ Webber, A.N., Su, H., Bingham, S.E., Kass, H., Krabben, L., Khun. M.. Jordan. R..
Schlodder, E. & Lubitz, W. (1996) Biochemistry, in press.

3‘ Lin, 8., Murchison, H.A., Nagarajan, V., Parson, W.W., Williams, J .C. & Allen J .P.
(1994) Proc. Natl. Acad. Sci USA 91, 10265.

3’ Schatz, G.H., Brock, H. & Holzwarth, AR. (1988) Biophys. J. 54, 397.

3" Nuijs, A.M., van Gorkom, H.J., Plijter, J .J . & Duysens, L.N.M. (1986) Biochim.
Biophys. Acta 848, 167.

’7 Hansson, O. & Wydrzynski, T. (1990) Photosynthesis Res. 23, 131.
’8 Crofts, A.R. & Wraight, CA. (1983) Biochim. Biophys. Acta 726, 149.

’9 Witt, H.T.,Schlodder, E., Brettel, K., and Saygin, O. (1986) Phatosynth. Res,
10, 453.

“Barry, BA. and Babcock, G.T. (1987) Proc. Natl. Acad. Sci. USA, 84, 7099.

" Debus, R.J., Barry, B.A., Babcock, G.T., and McIntosh, L. (1988) Proc. Natl. Acad.
Sci. USA, 85, 427.

‘2 Hoganson, C.W., Lydakis-Simantiris, N., Tang, X.-S., Tomrnos, C., Warncke, K.,
Babcock, G.T., Diner, B.A., McCracken, J. & Styring, S. (1995) Photosynthesis Research
46, 177.

38

 

‘3 Reviewed in a) Barry, BA. (1993) Photochem. Photobiol. 57. 179. b) Hoganson, CW.
and Babcock, G.T. (1994) Photosystem II: in Metal Ions in Biological Systems. Vol. 30
(Sigel, H. and Sigel, A., eds.) Marcel-Dekker, New York, 77.

“ a) Tang, X.-S., Zheng, M., Chisholm, D.A., Dismukes, G.C. & Diner. BA. (1995)
Biochemistry, submitted. b) Tommos, C., Tang, X.-S., Warncke, K., Hoganson, C.W..
Styring, S., McCracken, J., Diner, B.A. & Babcock, G.T. (1995) J. Am. Chem. Soc.

1 I 7,10325.

’5 Svensson, B., Vass, I. & Styring, S. (1991) Z. Naturforsch 46c, 765.

‘6 Svensson, B., Vass, 1., Cedergren, E. & Styring, S. (1990) EMBO J 9. 2051.

‘7 Warncke, K. & McCracken, J. (1995) J. Chem. Phys. 103, 6829.

‘8 Lubbers, K., Haumann, M. & Junge, W. (1993) Biachim. Biophys. Acta 1183, 210.

‘9 a) Diner, B.A., Nixon, P.J. & Farchaus, J .W. (1991) Curr. 0p. Struct. Biol. 1, 546. b)
Tang, X.-S., Chisholm, D.A., Dismukes, G.C., Brudvig, G.W. & Diner, BA. (1993)
Biochemistry 32, 13742. c) Kramer, D.M., Roffey, R.A., Govindgee & Sayre, RT. (1994)
Biachim. Biophys. Acta 1185, 228. d) Roifey, R.A., Van Wijk, K.J., Sayre, R.T. &
Styring, S. (1994).]. Biol. Chem. 269, 5115.

5° Babcock, G.T., Barry, B.A., Debus, R.J., Hoganson, C.W., Atamian, M.. McIntosh,L.,
Sithole, I., and Yocum, CY. (1989) Biochemistry, 28, 9557

5’ Ghanotakis, D.F., Topper, J .N. and Yocum, CF. (1984) Biochim Biophys Acta 767,
524.

’2 Enarni, I., Kaneko, M., Kitamua. N., Koike, H., Sonoike, K., Inoue, Y., and Katoh, S.
(1991) Biachim Biophys Acta, 1060. 224.

’3 Hoganson, C.W., Babcock, G.T., and Yocum, C.F. (1989) Phatasynth. Res, 22, 285.
5‘ Nakatani, H.Y. (1984) Biochim. Biophy. Res. Commun, 120, 299.

5’ Nanba, O. and Satoh, K. (1987) Proc. Natl. Acad. Sci. USA, 84, 109.

’6 Diesenhofer, J., Epp, 0., Miki, K., Huber, R., and Michel, H. (1985) Nature 318, 618.
’7 Sayre, R.T., Andersson, B., and Bogorad, L. (1986) Cell, 47, 601.

5’ Rutherford, A.W. (1989) Trends. Biochem. Sci, 14 (6), 227.

39

 

’9 Ghanotakis, D.F., Demetriou, D.M., and Yocum, CF. (1987) Biachim. Biophys. Acta.
891, 15.

6° Bowlby, N.R., Ghanotakis, D.F., Yocum, C.F., Petersen, J .. and Babcock. G.T..(1988)
Light Energy Transduction in Photosyntheisi: Higher Plant and Bacterial Models
(Stevens, S.E. Jr., and Bryant, D.A., Eds.) p. 215, The American Society of Plant
Physiology Publishers, Baltimore, MD.

6' Babcock, G.T., Widger, W.R., Cramer, W.A., Oertling, W.A., and Metz. J .G. (1985)
Biochemistry, 24, 3638.

‘2 Klimov, V.V., Allakhverdiev, S.l., Demeter, S., Krasnovskii, AA. (1980) Dokl.Akad.
Nauk SSSR, 249, 227.

‘3 Joliot, P., Barbieri, G., and Chabaud, R. (1969) Photochem. Photobiol.. 10. 309.
6‘ Kok, B., Forbush, B., and McGloin, M. (1970) Photochem. Photobiol., 11.457.

‘5 Sigel, H. & Sigel, A. (1994) (eds.) in Metal Ions in Biological Systems. Vol. 3 0.
Marcel-Dekker, New York.

66 Gilchrist Jr., M.L., Ball, J.A., Randall, D.W. & Britt, R.D. (1995) Science, submitted.
67 PiCOt, 1)., L011, P.J. & Garavito, R.M. (1994) Nature 367, 243.

68 Pirson, A. (1937) 2. Bot. 31. 193.

6’ Cheniae, GM. and Martin, LP. (1966) Brookhaven Symp. Biol. 19, 406.

7° Yocum, C.F., Yerkes, C.T., Blankenship, R.E., Sharp, RR, and Babcock, G.T. (1981)
Proc. Natl. Acad. Sci. USA, 78, 7507.

7‘ a) Dismukes, GO and Siderer, Y. (1980) FEBS Lett, 121, 78. b) Dismukes, GO and
Siderer, Y. (1980) Proc. Natl. Acad. Sci. USA, 78, 274.

72 Cooper, S.R., Dismukes, G.C., Klein, M.P. & Calvin, M. (1978) J. Am. Chem. Soc.
100, 7248.

7’ Zimmerman, J .L. & Rutherford, A.W. (1986) Biochemistry 25, 4609.
7‘ Casey, J. & Sauer, K. (1984) Biochim. Biophys. Acta 767, 21..

7’ 3) Beck, W.F. & Brudvig, G.W. (1986) Biochemistry 25, 6747 b) Beck, W.F. &
Brudvig, G.W. (1988) Chemica Scripta 28A, 93.

40

 

7" Brudvig, G.W., Casey, J .L., and Sauer. K. (1983) Biochim. Biophys. Acta. 723.366.

’7 Penner-Hahn , J .E., Fronko, R.M., Pecoraro. V.L., Yocum. C .F.. Betts. SD. 8: Bowlby.
N.R. (1990) J. Am. Chem. Soc. 112, 2549.

7’ Sauer, K., Yachandra, V.K., Britt, R.D. & Klein, MP. (1992) in Manganese Redox
Enzymes (Pecoraro, V.L., ed.) VCH Publishers, New York, 141.

7’ Latimer, M.J., DeRose, V.J., Mukerji, 1., Yachandra, V.K., Sauer, K. & Klein. MP.
(1995) Biochemistry 34, 10898.

’° Yachandra, V.K., DeRose, V.J., Latimer, M.J., Mukerji, I., Sauer, K. & Klein. MP.
(1993) Science 260, 675.

" A) DeRose, V.J., Mukerji, I., Latimer, M.J., Yachandra, V.K., Sauer. K. & Klein. MP.
(1994).]. Am. Chem. Soc. 116, 5239 b) Mukerji,1., Andrews, J.C., DeRose. V.J., Latimer.
M.J., Yachandra, V.K., Sauer, K. & Klein, MP. (1994) Biochemistry 33, 9712.

‘2 Liang, W., Latimer, M.J., Dau, H., Roelofs, T.A.. Yachandra, V.K., Sauer. K. & Klein.
M.P. (1994) Biochemistry 33 , 4293.

’3 Goodin, D.B., Yachandra, V.K., Britt, R.D., Sauer, K. & Klein, MP. (1984) Biochim.
Biophys. Acta 76 7, 209.

8" Riggs, P.J., Mei, R., Yocum, C .F. & Penner-Hahn, J.E. (1992) J. Am. Chem. Soc. 114.
10650.

85 England, R.R. and Evans, EH. (1981) FEBS Lett., 143, 175.
3" Ghanotakis, D.F., Babcock, G.T., and Yocum, CF. (1984) FEBS Lett, 16 7, 127.
’7 Boussac, A. and Rutherford, A.W. (1990) FEBS Lett., 277, 69.

’8 Ghanotakis, D.F., Babcock, G.T.. and Yocum, CF. (1984) .Biochim. Biophys.
Acta” 765, 388.

’9 Boussac, A. & Rutherford, A.W. (1988) Biochemistry 27, 3476.
9° a) DeRose, V.J., Yachandra, V.K., McDermott, A.E., Britt, R.D., Sauer, K., and Klein,

M. (1990) Biochemistry, 30, 1335. b) Britt, R.D., Zimmerman, J .L., Sauer,K., and Klein,
M. (1989).]. Am. Chem. Soc.,111,3522

Chapter 2
Electron Magnetic Resonance Theory

Although the intellectual environment prior to the first reported electron
paramagnetic resonance (EPR) spectrum in 1945 by Zavoisky was favorable for the
development of EPR methodology. significant advancements in the technique did not
occur until the advent of available microwave technology following World War 11.
Subsequent improvements in pulse generators, stable microwave sources, amplifiers and
the recent development of a commercial instrument have similarly catalyzed the growth
of pulsed-EPR as an effective and applicable spectroscopic tool.

Electron magnetic resonance spectroscopy has been used successfully in
chemistry and physics to understand electronic and geometric structures in a plethora of
systems encompassing fields from biophysics to material science. The presence of stable,
easily trapped paramagnetic species in photosynthetic systems makes them especially
amenable to the application of these techniques. This chapter will provide a basic
introduction to the theory behind these EPR techniques and is not meant to be a
comprehensive review; a number of excellent treatises' and textbooks2 have been

dedicated to the complete theoretical description of magnetic resonance spectroscOpy.

EPR

Electromagnetic radiation is composed of mutually perpendicular oscillating
electric and magnetic fields. Interaction of these fields with molecular species initiates

transitions between energy levels, which provides information regarding both geometric
41

42
and electronic structure. These transitions are observed only if the incident photon has

sufficient energy to bridge the gap between energy levels and the interaction perturbs
either an electric or magnetic dipole present in the sample. In EPR. the interaction of the
electron magnetic moment with the magnetic field component of the electromagnetic
radiation is of primary interest. The energy levels corresponding to different orientations
of these magnetic moments are normally degenerate. however. the presence of an
externally applied static magnetic field effectively splits the degeneracy of these spin
states and, upon introduction of oscillating electromagnetic radiation, allows observation
of an absorption attributable to a magnetic interaction. This interaction is the basic tenet

of both EPR and nuclear magnetic resonance (NMR) spectroscopy.

 

Figure 2-1: The interaction of the applied magnetic field H, with the magnetic moment,
a.

The macroscopic magnetic moment present in a paramagnetic material is the
physical manifestation of the presence of electron spin. The classical energy of this

magnetic dipole moment in a magnetic field is given by

43
E = ~11 0 Fl
where u is a vector describing the macrosc0pic magnetic moment and has contributions

from both spin and orbital angular momentum and H is the applied magnetic field

strength (see Figure 2-1). Because electrons follow quantum rather than classical

mechanics, u is replaced in this expression with the appropriate spin operator to yield the
Hamiltonian for an electron in a field
21 = geBS 0 H

where B, the electron Bohr magneton, is equal to the magnetic moment for a single unit of
quantum mechanical angular momentum, and g, the electronic g factor, represents, from a
classical viewpoint, the correction factor for the anomalous magnetic moment of the
electron (vide infia). Application of a static magnetic field along 2 with magnitude H0
will quantize the spins along this direction and allow substitution of S, for S. The scalar
product, therefore, simplifies to

74‘! = geBSZl-l 0.
The eigenvalues of this Hamiltonian are the eigenvalues of the operator 3,; these
eigenvalues, M,, for an electron are + 14 (at state) and - '/2 ([1 state). This so-called
electronic Zeeman splitting results in the formation of two energy states that are
degenerate at zero field, and whose separation increases linearly with increasing magnetic

field.

The g value accounts for the coupling of spin and orbital angular momentum and

is given by

44
_ 3J(J+l)+S(S+1)—L(L+1)
gJ ' 2J(J+ 1)

 

where J represents the total angular momentum of the system. S the spin angular
momentum and L the orbital angular momentum. For the free electron. gc = 2.00232. In
the photosynthetic systems studied in this thesis. the electronic g factor is near gc and very
nearly isotropic. This assumption of an isotropic g value is upheld throughout this
discussion and the data analysis in Chapters 3-4.

Although the quantum mechanical representation of a magnetic moment

interacting with a field gives a mathematically complete description of the interaction, a
more “classical” approach can assist with visualization of the physical motion of the
magnetic moment. Classically. Newton’s third law for rotational motion states that the
rate of change of angular momentum in a system is equal to the torque. For the electron
with magnetic moment 11 this relationship can be written
fighgefilfiflll-
If H is applied parallel to the z direction with magnitude H0, then the solutions of this
differential equation describe precession about the z axis with a characteristic Larrnor
frequency, (Do. This precession provides a physical picture of the motion and is also
consistent with the properties of quantum mechanical angular momentum in that the 2
component of the motion is sharply defined. while the x and y components oscillate about
some average value; this behavior is reflected in the commutator relationships found for
the spin angular momentum operators.

Transitions between energy levels can be effected provided that the photon

energy, hv, is equal to the energy splitting between levels, AE,

45
AE = hv = geBl-I.

This is the basic resonance condition for a free electron. Because of the narrow
bandedness of available low noise microwave sources. the typical EPR experiment is
performed by holding the microwave frequency constant and sweeping the magnetic field
until the resonance condition is met.

Microwave radiation is used to provide the energy necessary to effect these
transitions. The magnetic field component of this radiation interacts with the electron’s
magnetic moment much in the same way as the static field does. This field, H, (H,<<H,)
rotates about 2 with precessional frequency a). If this precessional frequency matches
exactly the Larmor frequency of the magnetic moment. then the magnetic moment will
experience a constant field in the xy plane and will precess about this with frequency

(1)1 = geBH].
Since (D,<<(.00, the magnetic moment will spiral down until its projection on -2 has the
same value that it had on +2. This oscillatory motion between +2 and -z will continue as
long as H, remains on; the frequency of these oscillations is (Do.

Although this classical description of the behavior of a magnetic moment in an
applied magnetic field does not rigorously hold for quantum mechanical particles, a
requirement of which is the discrete exchange of energy and matter, it daes lend insight
into the physical manifestations of this behavior. For example, the application of an
oscillating magnetic field applied perpendicularly to an applied field containing an

oriented magnetic moment can cause changes in the value of the projection of that

46
magnetic moment and this reorientation occurs only if the oscillating radiation is in

resonance with the natural frequency of the system.

The relative populations in each of the spin states at thermal equilibrium lends
insight into the form of the EPR spectrum obtained experimentally and can be calculated
by using the Boltzmann expression

=expr-A%.)=exp1-1gm)).

lower

 

At room temperature and 3400 G this ratio is 7.6 x 10". This population difference is
essential in observing the resonance phenomenon since transitions are effected with equal
probability upward and downward. If the populations of the two states were equal, as
much energy would be emitted as absorbed and no net effect would be observed. Since
there are more spins in the lower spin state. a net absorption is measured in EPR.

The power of the incident microwave radiation can alter the intensity of the EPR
absorption. To understand this qualitatively, the differences in the relative populations of
the spin levels as a function of microwave power will need to be examined. Transitions
between spin states are driven more rapidly as the power of the incident radiation is
increased. This, initially, will result in an increase in absorption as more transitions are
effected. However, for inhomogeneously broadened lines, this increase in intensity will
ultimately level off and eventually decrease as the transition rate becomes competitive
with relaxation pathways. As a result of this, the relative populations in the two spin

states becomes equivalent and no net absorption is observed. This phenomenon, known

47
as “saturation.” can be avoided by performing intensity vs. power studies and choosing a

non-saturating power at which to perform the experiments.

The expressions derived above describe the energies of electron spin states in an
applied magnetic field. If this was the only interaction present. EPR spectroscopy could
be used only to detect the presence. and perhaps the kinetics. of paramagnetic species.
However, there are always magnetic fields within the paramagnetic substances that
interact with the electron spin moment to perturb the resonance. The ensuing complexity
of the energy levels as a result of these interactions is manifested in the experimental EPR
spectrum as splittings or alterations in the frequency or linewidth of the absorption. The
EPR spectrum is therefore interpreted as the allowed transitions between the eigenstates
of a spin Hamiltonian that consists of terms indicative of these interactions. For the
photosynthetic systems studied, this spin Hamiltonian, operating on the spin only part of
the electronic wavefunction, is of the form

74:71 +71 +74 +7.
ez nz hf nqr

where W., is the Hamiltonian for the electronic Zeeman interaction, 7,, the nuclear
Zeeman Hamiltonian, 1,, the Hamiltonian for the electron-nuclear hyperfine interaction
and 2!", represents the Hamiltonian for the nuclear quadrupole interaction. Although
there are other interactions that can contribute to this spin Hamiltonian, only those terms
directly relevant to the systems studied in this thesis have been included. The energies of
the states can be evaluated by applying the spin Hamiltonian composed of the electronic
Zeeman, nuclear Zeeman and hyperfine terms. These energies and the corresponding

energy level diagram for an S= ‘/2 . 1= 1/2 system are shown in Figure 2-2. This figure is

48

 

 

 

laa>
.1293 1
11
>
"PET—<1 ganH 119.0
"0 ‘4 ‘
|aB> ‘~._JL_
|a13>
gBH

 

 

\
\

Electronic Zeeman Nuclear. Zeeman Hyperfine
Interaction Interaction

Interaction
Figure 2-2: Energy level diagram for an S=1/2, I=1/2 spin system

49
representative of the chlorophyll species studied in Chapters 3 and 4 of this thesis. In

these systems, the unpaired electron is delocalized in a 1: molecular orbital. and is coupled
to an "N (1= ‘/2 ) nucleus on a ligated histidine residue through the electron-nuclear
hyperfine interaction. The interaction is nearly isotropic and represents coupling through
a covalent band. This energy level diagram will also be utilized later in this chapter to
understand the ENDOR spectrum arising from this chlorophyll species.

The electronic Zeeman term of the Hamiltonian acts on the electron spin and
represents the interaction of the electron magnetic moment with the field. The resulting
energy levels representing the M,=+ ‘/2 (or | a > state) and M5 =- V: (or | 13> state) spin
manifolds and transitions were previously described and are shown in Figure 2-2.

The magnetic field also has an effect on the magnetic moments from nuclei with
nonzero spin (1 ¢ 0) present in a sample. The second term in the spin Hamiltonian

represents the nuclear Zeeman interaction and is given explicitly by
’an = -gn0nH o I.
This Hamiltonian is analogous to the electronic Zeeman interaction described above. The
nuclear spin quantum numbers M,, are valid if the static field is applied parallel to the z
direction, the corresponding energies are given by
E = —gnl}nHMI

For an I = '/z nucleus, this yields two energy levels for M, = + ‘/2 (I01?) and M, = - V2 (IB,>,
see Figure 2-2). This interaction is approximately three orders of magnitude smaller than

the analogous electronic Zeeman interaction. This is due to the difference in relative

masses of the proton and electron. Because of this difference, transitions between nuclear

 

51318

else
1111!
free

11101

1111
is i:
orb
110)

dis

Fig

filer

50
states can be effected by photons in the radiofrequency range. These transitions can be

probed by using NMR techniques.

The third term in the Hamiltonian describes the interaction between the unpaired
electron and magnetic (I at 0) nuclei in the system. The magnitude and form of this
hyperfine interaction is a direct reflection of the electronic and geometric structure of the
free radical and originates from the fact that the interaction of the electron’s magnetic
moment with the external magnetic field is modulated by the presence of fluctuating
magnetic fields induced by nuclear magnetic moments. These interactions are of two
types, a contact interaction and a dipole-dipole interaction. The Fermi contact interaction
is isotropic and arises from electron spin localized at the nucleus. Only the spherical s-

orbital has a nonzero probability. ‘1’*(0)‘I’(O), of the electron being located at the nucleus,

however, for orbitals with (at 0. ‘1’*(O)‘1’(0) is non-zero and this contact interaction

disappears. The energy for the isotrOpic interaction between the electron and nuclear

magnetic moments can be obtained from solving the Hamiltonian for the interaction

A

7111050) = A0 5 '1
where Ao=(8Tt/3) g,g,,B,B, |‘P(0)13. The 2 components of the electronic and nuclear spin
angular momentum operators can be substituted in place of So I , providing that the

hyperfine interaction is less than the electronic Zeeman interaction. the ensuing energies

are then given by
E 150 = AOMSMI
Figure 2-2 shows the energy level splittings incorporating this term in addition to the

electronic and nuclear Zeeman terms.

51
The isotropic coupling corresponding to this energy can be related to the unpaired

electron spin density distribution in the molecule and arises. in part. from the
phenomenon of spin polarization that is induced by the presence of an unpaired electron
in a p orbital. For example, in the CaH radical fragment. Hund's Rule predicts that the
state with the highest spin multiplicity will be the ground state. This is due. to some
extent, to the favorable exchange energy between two electrons with parallel spins
located in different orbitals. This exchange interaction. between the unpaired electron on
the C,l and the electrons in the 6 bond of the CQH fragment. can influence the sign of the
hyperfine coupling between the unpaired electron and the hydrogen nucleus (see Figure
2-3). For example, an unpaired electron with an or spin. located in the pz orbital on C,
will constrain the electron from C participating in the 0 bond to an or spin state and. since
electrons participating in a covalent chemical bond must be of opposite spin. this requires
that the electron spin on the proton have a [3 spin (see Figure 2-3). The presence of this
negative spin density on the proton will yield a negative hyperfine coupling constant for

the hydrogen atom.

292

0.

or
sp2 hybrid .fiw B

Figure 2-3: The effect of spin polarization on the spin density at the hydrogen atom.

52
In 1: radicals, like the chlorophyll cation radicals in PSI, the unpaired electron is

delocalized in a 1: molecular orbital. 1n symmetric molecules like benzene, this unpaired
electron is distributed equally amongst the six carbon atoms that contribute to the rt
molecular orbital. The proton hyperfine coupling. originating from the interaction of the
unpaired spin with the ring protons, in this case. is equivalent for each atom. The
magnitude of the hyperfine coupling, therefore, reflects the distribution of the unpaired
electron. Moreover, this distribution is a direct probe into the symmetry of the molecule
and the molecular orbitals containing the unpaired electron. The McConnell expression3
describes this relationship
A 150 = P1! Q
where p" is the spin density in the p" orbital and Q is an experimentally determined

proportionality constant specific to the nucleus of interest.
For orbitals with [:0 (e.g. p or d orbitals). the hyperfine coupling is anisotropic

and arises from a dipolar, through space interaction between the unpaired electron and

nearby magnetic nuclei (see Figure 2-4). The classical expression for this interaction is

given by

E J.M. 3(11e-r)(un°r)
dipolar“ 1‘3 T 1'5

 

 

The corresponding quantum mechanical expression for this interaction is obtained by

substituting the appropriate operators. The Hamiltonian is, therefore,

s i as i
fidipolar=_gegnfiel3n[ :3 ' ( .:Z( .r)]

Ur
OJ

 

Figure 2-4: The dipolar interaction in an applied magnetic field between the magnetic
moments of the electron spin and the nuclear spin for a nuclear spin = 1/2.

To remove the explicit spatial dependence of this Hamiltonian. one must integrate
ayer the electron wavefunction to obtain a Hamiltonian containing only spin operators.
This can be accomplished by expanding the Hamiltonian in the Cartesian coordinate
system, then integrating. This integration gives

“Bipolar = rxxrxs, + TyylySy + TZZIZSz +
T,,,,(I,,Sy +1ny)+ Ty,(lyS, +ley)+ sz(IzSx +IXSZ)

with coefficients

30::2 —r2
Tact =gegnBeBn< 5 >

3010
T010 =gegnBeBn< r5 >

where the brackets indicate integration. A more convenient, and succinct, way to write

this Hamiltonian

a = i . 1 - s
where Tis a matrix with elements T0,. This dipolar coupling matrix is traceless and

symmetrical and vanishes when averaged over all orientations. It. therefore. does not
contribute to the peak positions and splittings observed in solution EPR. Because this
matrix is symmetrical, an axis system can always be chosen in which it is diagonal.
These diagonal elements are the principal values of the tensor and characterize
completely the anisotrOpy of the hyperfine interaction. For systems with axial symmetry
(x = y at z), the anisotropy can be described by two terms. A , and A“.

The EPR spectrum observed experimentally is a composite of the EPR spectrum
for each orientation, with buildups of intensity at the principal values of the A tensor (AL
and A,I for a system with axial symmetry). The primary electron donor of PSI studied in
Chapters 3 and 4 is a chlorophyll a species, and because this molecule is planar, an axial
hyperfine tensor for each of the pyrrole nitrogens is expected. Additionally, because the
dipolar coupling term represents a through-space interaction, the magnitude of this
coupling, dependent on r", can yield distances between the paramagnet and interacting
nuclei.

The full hyperfine interaction requires the addition of the isotropic portion of the
coupling.

Atotal = A 150 + Adipolar
The hyperfine Hamiltonian can therefore be represented as

th =h§0§0l

where 1: A 0

lit—I

+1

55
Here A, represents the isotropic portion of the hyperfine coupling and l is the unit tensor.

The final term in the spin Hamiltonian. the nuclear quadrupole coupling term does
not contribute to the EPR spectrum from P7,, . however the effects on the ESEEM
spectrum are substantial. These effects. and the functional form of the Hamiltonian, will
be explained in the ESEEM section of this chapter.

The presence of multiple, small hyperfine couplings in conjunction with the
complexities introduced by the powder pattern lineshapes expected for EPR spectra of
frozen solutions lead to an inhomogeneously broadened EPR line. Because of this
spectral congestion, the components of the dipolar tensor are obscured. To resolve these
small hyperfine couplings in the photosynthetic systems, pulsed and double resonance
techniques have been applied. These techniques yield the weak hyperfine couplings
desired, but have the advantage of reducing the number of peaks observed in the spectrum

as well as decreasing the linewidths associated with the transitions.

Electron Nuclear Double Resonance (ENDOR)

The spin Hamiltonian describing the ENDOR experiment is identical to the one
used for EPR (see previous section) and consequently the energy level diagram
corresponding to the observed ENDOR transitions is also the same (see Figure 2-2). The
ENDOR experiment measures the change in intensity of an EPR absorption as a function
simultaneously applied RF radiation. To illustrate the ENDOR effect, an S = V2, 1 = ‘/2

system will be utilized. Again. this system mirrors the interaction of the unpaired

56
electron on Pm’ with the "N nucleus on the directly ligating histidine that is discussed in

Chapter 3. The Hamiltonian for this interaction is
¢= geBeHSz - ganle + AisoSzIz

with gn = -O.566378 for the "N nucleus. Am, = 0.64 MHz for the interaction (see Chapter
3) and v" = 1.45 MHz for "N in a 3200 G static field. The selection rules for EPR are
given as AM, = i 1, AM, = 0: the allowed transitions for EPR are shown in Figure 2-5.
The analogous allowed nuclear transitions representing a nuclear spin flip. with selection
rules AM, = :1, AM, = 0, are also indicated. As discussed earlier, these transitions fall in
the radiofrequency (RF) range.

The ENDOR experiment is performed by fixing the externally applied Zeeman
field so that the resonance condition for a single EPR transition (eprl, for example) is
met. The microwave power is increased to equalize the populations of levels 2 and 3.
This results in partial saturation of this transition subsequently decreasing the intensity of
the EPR absorption. Irradiation with swept RF radiation will effect nuclear transitions
when the frequency of the radiation meets the resonance condition for the nuclei (nmrl or
nmr2). Focusing on nmrl, transitions between nuclear levels 1 and 2 will reinstate the
population difference between 2 and 3 and therefore desaturate the EPR transition, the
absorption due to this transition will therefore increase. This change in the EPR
absorption is detected and plotted as a function of the RF frequency.

While the peaks observed in an EPR spectrum reflect the interaction of the

unpaired electron with all nearby magnetic nuclei in a multiplicative fashion

57

X
peaks observed = H 2nil + 1

i=1
where n is the number of chemically equivalent nuclei and 1 represents their respective

nuclear spins, the ENDOR effect provides the same hyperfine information in an additive

 

 

 

 

 

laean> 11 A |1>
nmrl

Iae n> A [2 >

eprl epr2
r

113.13,,>‘ ,, |3>
nmr2

19.9.9 " 14 >

 

 

 

Figure 2-5: Energy level diagram for S~= 1/2, 1 = 1/2 system showing the allowed
EPR and NMR transitions.

manner. ,One pair of peaks appears for each chemically equivalent set of interacting
nuclei. This decreases the spectral congestion and moreover, simplifies assignment of the

salient features. These peaks. for an isotropic hyperfine coupling, Am, appear at
frequencies

 

58

where vn is the free precessional frequency of the nucleus of interest. For the "N studies
reported here, the ENDOR experiments were performed at a static field strength of 3375
G, where the Larmor frequency of "N is 1.45 MHz.

The presence of hyperfine anisotropy, discussed above in the EPR section. also
contributes significantly to the lineshapes observed in frozen solution ENDOR. The
lineshape characteristic of an axial hyperfine coupling tensor is shown in Figure 2-6a.
ENDOR data are collected in the first derivative mode, the corresponding absorptive
powder patterns are shown in Figure 2-6b. The turning points in this ENDOR spectrum
represent the most probable orientations and correspond to A, and A1,. The peaks are split

about the Larmor frequency according to

2
Adipolar=gegn13el3n(3cos 9 /r3)

where 0=90° for A, and 0° for A“. This equation predicts that A“ is twice as large as A ,.
The absorptive peaks in Figure 2-6a are then split by an amount equal to AH and the
derivative shaped peaks are split by A ,. In the limit of small dipolar coupling, only a
single pair of derivative shaped peaks is observed. By recognizing the lineshapes
associated with different tensor symmetries, the geometric and electronic structure of the
paramagnetic system can be deduced. The absorptive peaks associated with All are often
difficult to resolve experimentally since the number of orientations with parallel
symmetry contributing to the overall lineshape is small. This precludes the use of
ENDOR in systems with large dipolar couplings, however the applicability of the
technique to systems containing large, isotropic hyperfine couplings has been shown in

numerous studies of biological systems.‘ For systems containing small, anisotropic

59

 

 

 

 

 

 

 

 

 

 

J
1 A11 .
l A .L #1

 

 

 

4 6 8 10 12
frequency (MHz)

Figure 2-6: ENDOR spectrum showing the first derivative lineshapes expected for an
axial dipolar coupling (top) and the corresponding powder pattern absorption (bottom)

60
hyperfine couplings, the technique of electron spin echo envelope modulation

spectroscopy can be applied to extract hyperfine and nuclear quadrupole coupling
constants and therefore information regarding the structure and funcfion of the
paramagnetic species. The complementarity of these two spectroscopies is exhibited in

the study of axial ligation to Pm,+ presented in Chapter 3.

Electron Spin Echo Envelope Modulation (ESEEM)

Electron spin echo envelope modulation (ESEEM) is a sensitive technique for
measuring the small superhyperfme couplings in systems with large inhomogeneously
broadened EPR lines. lnhomogeneity in an EPR line can result from the presence of
anisotropy in the hyperfine term of the Hamiltonian. The electron is subjected to slightly
different local fields such that only a small fraction of the spins are in resonance at a
given time. The EPR spectrum is then composed of a superposition of these “spin
packets” (see Figure 2-7). Inhomogeneous broadening can also be the result of numerous
hyperfine couplings, which can be resolved effectively with the use of ENDOR
spectroscopy, or fluctuations in the homogeneity of the applied magnetic field, although
the latter is only observed at large field values. ESEEM spectroscopy circumvents the
inhomogeneity problem by causing rotations of net electron spin magnetization that allow

observation of the individual spin packets.

61
The Two-Pulse ESEEM Experiment

To explain the theory behind a two-pulse ESEEM experiment. a semi-classical
model will be utilized. The choice of coordinate systems for this model is crucial. and for

the ease of explanation, a rotating coordinate system will be used. Figure 2-8 depicts this

 

Figure 2-7: An inhomogeneously broadened EPR absorbance.

coordinate system, with rotational frequency (1),, and with axes X’,Y’,Z’. As in previous
discussions, a static magnetic field with magnitude H0 is applied along the 2’ direction.
Detection of the build up and decay of magnetization occurs along Y’.

The description of the two pulse experiment will focus on two spin packets (see
Figure 2-7) with characteristic Larmor frequencies that precess slower ((1),) or faster ((1),)

than the rotating frame. At thermal equilibrium, a net macroscopic magnetization will lie

62

 

 

 

C) D)

 

 

 

 

 

E)

 

Figure 2-8: Classical description of the formation of a spin echo.

63
along the z axis. This magnetization, which arises from the population difference of the

a and 0 spin states, is due to the vector sum of the individual magnetic moments.
Application of a short, high field microwave pulse will rotate this magnetization. It can
be described quantitatively by the fundamental Bloch equation

dM
dt -Y(H><M) .

where H is the effective field in the rotating frame. The effective field that is interacting
with the magnetization is composed of a 2 component (Ho) and an x component (H,). The
physical manifestation of this is, as discussed earlier, a precession about H with

magnitude

 

1
2 030‘“) 2 2
Heffective = Hl T H0 (D

To determine the magnetization before and after the H, field is turned on, the Bloch
equations describing the time dependent behavior of the magnetization will need to be

solved. For t=0, when the H, field is turned on, the solutions are

M2 = M0 cosm1t
v = M0 sintn II

where M0 is the magnitude of the magnetization at thermal equilibrium and v is the out-
of-phase component of the magnetization in the rotating coordinate system; these
solutions describe a circle in the Y’Z’ plane of the rotating flame, comparable to the

oscillatory motion of the magnetic moment as continuous microwave radiation is applied

(see EPR section).

64

Subsequent to the application of a microwave pulse of duration Tp. the
components of magnetization are

Mz(tp)= M0 cosmltp

v rp)= Mo since 11,,

If the microwave pulse is applied with (0,1,, = rt/2 then

M z (t p ) = 0
v (r p) = —M0
indicating that the magnetization, with its equilibrium value, now lies along the Y’ axis.
It is interesting to examine the magnitude of H, needed to effect such a rotation.
For NMR, n/2 pulses are regularly applied to the spin system to bring about rotations in
an analogous manner. However, because nuclear relaxation occurs on a much slower
timescale than the analogous electron relaxation, pulses can be applied for a longer time
period to create the H, necessary for a 1t/2 rotation of the magnetization. With pulses in
the microsecond range, H, magnitudes on the order of 2-3 G are needed to produce a 1r/2
rotation. In pulsed-EPR, pulses of 10-20 ns are necessary to ensure a reasonable
frequency spread and allow neglect of the effects of electronic relaxation. This requires
the generation of a microwave field with an H, in the 25-30 G range, a substantial
increase as compared to NMR. Design and implementation of hardware capable of
generating nanosecond pulses of this power was one of the primary impediments to the
development of pulsed EPR as an effective spectroscopic tool.
Following the first rt/2 pulse, the magnetization lies along the positive Y’ axis

(see Figure 2-8b). If the frequency of the rotating frame equals the precessional

65

 

1 200 y r
1000 ' I t
800 '
600 r

400 *

Echo Amplitude (a.u.)

E

 

 

o _L a 1 1
0 500 1 000 1 500 2000 2500

 

Figure 2-9: Representative time domain trace from two-pulse ESEEM experiment.

frequency of the spin packet, it will remain along the Y’ axis during the interpulse time.
1. However, if the spin packets have different offset frequencies (for example. a), and (1) ,-
in Figure 2-7), then they will start to dephase immediately (see Figure 2-8c). A second
microwave pulse of sufficient duration to rotate the spin packets 180° about the X’ axis is
then applied. The precessional frequency and direction of these spin packets remains
unaltered after the rotation. After time t, the spin packets refocus along -Y’, which
creates the build up of magnetization that is known as a spin echo. The amplitude of this
spin echo is measured as a function of the interpulse time, t.

The decay in the echo amplitude as a function of t is not monotonic (see Figure 2-
9), but instead modulated by interactions with nearby magnetic nuclei. The frequency of
these modulations can be related directly to the NMR frequencies in Figure 2-5. The

modulations of the spin echo decay are the direct result of the fact that microwaves can

66
induce transitions that initially involve one pair of energy levels, but then branch to a

second pair upon application of additional pulse(s). The resulting interference effects
then rely on the ability of the microwave pulses to excite coherently all four EPR
transitions shown in Figure 2-10. The branching of transitions is allowed quantum
mechanically because the anisotropy of the hyperfine interaction effectively mixes the
nuclear states contributing to the energy level diagram, consequently the transition
probability for each EPR transitions becomes non-zero.

To illustrate this nuclear modulation effect, a semi-classical description will be
used (see Figure 2-10). The rotating coordinate system of Figure 2-8 will again be
utilized. The energy level system for an S = V2 , l = V2 system is shown in the top of
Figure 2-10. The four transitions represent both the allowed and semi-forbidden
transitions. Because the states are mixed by the dipolar part of the hyperfine coupling,
the |a> and |B> labels used in Figure 2-2 are no longer valid, the states. therefore. have
been labeled numerically. This discussion will focus on the transition labeled “A” in this
energy level diagram. Transition “A” corresponds to a spin packet that is precessing
slower than the rotating frame, therefore after the rt/2 pulse, the spin packet has dephased
and its precessional frequency is given by (mo—toot (Figure 2-10a). Application of a 1:
pulse will excite transition “A” , and because of branching, the semi-forbidden transition
“C” will be excited as well; these transitions will not, however, be excited with equal
probabilities. Spin packet “C” has a characteristic precessional frequency that is faster
that the rotating frame, therefore after time t, spin packet “A” will have refocused along -

Y’ but spin packet “C” will not be aligned along the rotating frame.

67

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A
Cl’0
A & B: allowed transitions
C & D: forbidden transrtions
A) B)
X' X'
time t after
n/Z ulse
p I Y' 1! pulse Y‘ a)
C
(DA (DA
‘90
C) X'
03C
time 1: 0
——-——-> Y' _‘
(”A

 

Figure 2-10: Nuclear modulation effect.

68
The echo amplitude at time t is given by the projection of these spin packet “vectors"

onto -Y’. Spin packet “A” is along the -Y’ direction and therefore contributes fully to the
echo amplitude, however “C” does not contribute fully and may add to or subtract from
the echo amplitude. This depends on the projection of “C”, which is given by cos((1),,-
mc)r or cosmmt, where (0N, is the NMR frequency. The echo amplitude observed
experimentally is the sum of all of the spin packets in the excited region of the
inhomogeneously broadened line and this sum is the source of the modulation pattern
observed in the spin echo experiment. The fi'equencies of the modulations yield the
hyperfine frequencies that are the ultimate goal of the spectroscopist. To observe better
these frequencies, a cosine Fourier transform of the time domain data is performed and
the spectrum obtained subsequent to this yields an ENDOR-like spectrum: hyperfine
peaks can be expected to appear centered about the Larmor frequency of the nucleus from

which they arise and are split by Am.

Density Matrix Treatment of ESEEM

Because the generation of spin echoes and the nuclear modulation effect are both
quantum mechanical phenomena, they can be understood quantitatively by solving the
total spin Hamiltonian describing the interaction of the spin system with the static and

applied magnetic fields. This is most easily accomplished by utilizing the density matrix

formalism.

69
The density operator, p, describes the projection of one basis set onto another and

the elements of the density matrix characterize the quantum state of the spin system. It is
possible, therefore, to calculate all of the physical observables of |‘P(t)> from the density
matrix. The pertinent operator is given by

90) = |W(t)><‘1’(t)|

with matrix elements

p,..(t)=(u,, |p(t)lu..)

;

where ‘I’ is a state vector at time t that can be described in terms of an orthonormal basis

set |u,,> with coefficients c,(t) such that
119(1)) = 2 c, (1)] un).
11
The expectation value <A>(t) of an observable A can then be obtained by using

<A>(t) = Tr{p(t)A}.
The equation of motion for the density operator can be derived by using the Schroedinger

equation

d 1
211"“) = 5[W(t).p(t)]
where fit) is the time dependent system Hamiltonian.
Integration of this equation will yield the density matrix p, at the time of the
appearance of the echo in terms of the initial density p,,. The system Hamiltonian in this

equation of motion consists of a Stun of two components

“(:70 +11

70

 

 

 

 

 

 

 

 

 

a>
a) AHO, b) A
0)
V
v v IB>

 

 

 

 

 

T

H0,11vg

Figure 2-11: a) Inhomogeneously broadened line indicating the distribution of
values of H0. b) Energy levels of a quantized spin system. Microwave radiation
of frequency to is able to induce transitions from any a state to more than

one [3 state.

where 31,, is the Hamiltonian for precessional periods and W, the Hamiltonian for nutation.
The echo signal is proportional, then, to Tr(p ‘31,). To obtain this trace, two summations
must be performed: the first over the |a,> and |Bj> states of each quantized, this will yield
the envelope modulation function; and the second over the systems that make up the
inhomogeneously broadened line, this will produce the functional form of the spin echo
signal. The first of these summations can be factored out provided that the pulse width is
sufficient to excite all spin packets within :Afi’o, of an average value, 1.,, (see Figure 2-
11), the second, provided that A310, is large with respect to the energy spacing in each
spin manifold.

The general expression for the echo can be given as

E = nTr(pEWI)

71

precession

2 pulse
experiment / \

nutatron nutation \

spin echo

/ precession
3 pulse , _j \__ /
experiment I .

nutation nutation nutation

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2-12: Pulse sequences for the Hahn (top) and stimulated (bottom) echo experiments
indicating periods of nutation and precession.

where p,E is the density matrix of the system when the maximum echo appears and fl, is
the Hamiltonian describing the interaction of the spin system with the microwave pulse
and is given by

W, = cos,
Calculation of the initial and final density matrices is essential in the derivation of the
functional form of the echo expression for the two or three pulse experiment. These

expressions, as given by Mims (1972)’ are

E(t) = [2(21 + 1)]—1Tr[Q:M+P,+MQ,M+PTM + Hennitian conjugate]
QtMtpr:MQ,M+PTP,M

13(1 + r) =[4(2r + 1)]“1‘:
+QiQ¥M*P:MQTQ,M+P,M + 11c

72
The derivation of the echo expressions for the two and three pulse experiments is long

and, for the sake of brevity, only the summary of this derivation will be included here.
The details of the derivation can be found in the classic 1972 paper by W.B. Mims.5

Initially, the Hamiltonians corresponding to precession and nutation (see Figure 2-
12) need to be transformed into the rotating fiame. Then, the solutions for the equation of
motion for the transformed density matrix p, containing a time dependent Hamiltonian
need to be obtained. This can be accomplished by using general rotation operators of the
form

p(t) = exp[— 1m / h] p(0) exp[im / h]

The density matrix following the pulse sequence can be written as

pE(t)=R—lp(0) R where R=HR(“’p)(tp,t,T) and
up

Rm” (tp,t,T) = exp[(21ti / 11mmp (tp,t,T)]

Here 2!”, represents the various parts of the Hamiltonian in nutation and free precessional
periods and tp is the pulse duration. An appropriate basis set must be chosen such that 21,,
is block diagonal with 7‘, connecting elements amongst blocks but not within them. The
resulting basis set is built up by using the two electron spin manifolds (or and B) and
completed by including the nuclear spin eigenfunctions of 34,. This formalism allows the

71,, p5 and R“"’(tp,1:,T) matrices to be partitioned in submatrix form

or states Bstates

a States Aaa AaB

[3 states A130 A1313

73
Once this matrix has been set up and normalized, one finds that only A‘m and AB13 of

Rp(T,T) and the off diagonal block of 9’,” contribute to the echo signals; these matrices
are designated P,Q and M, respectively. The submatrices corresponding to P and Q are

diagonal with elements

Pi, = exp(ico‘,1t) and an = exp(itol,3,t)
where t represents either I or T depending on the experiment and a), is the energy of the
ith state belonging to the at or B manifold. The M submatrix is not diagonal and maps the

nuclear eigenvectors from one spin manifold to the other, it is given by

OM
M+0

11(1)]

“i=7

 

 

where M = MEMB. These unitary matrices describe the state mixing in the at and B ‘
manifolds with-elements given by
M,,, = a(i|n)0.
To obtain the terms involving the superhyperfine frequencies, the expressions for

the two and three pulse echoes need to be expanded by using the above definitions and

terms in the trace can then be rearranged to yield a frequency independent term and

a (I (I
(DU =(Di -0)j
mfin_mE—m§

The ultimate expressions for the envelope modulation functions can be written in the

following forms

Emod(r) =Ivl4 +|ul4 +lv|2|ulz[2coswabr + 2coswcdt —cos(a)ab —wcd)t —cos(wab +(ocd)t]

74

. a 1
cosmabr + cosmcdt + 25m” gored-r coswabh + T)
7 _
Emod(tiT)=lVl4 +lul4 +M‘lul2 7
+ 25m” Ewart)? coswcdir + T)

where the frequencies given by 0),, and a)“, indicate the intervals between states 1a>. lb>.
|c> and |d> (see Figure 2-13). The u and v terms are the transition probabilities for the
branching transitions and are the elements of the M matrices discussed above. It is easy
to see how the lack of branching (either u or v equal to 0) will result in no observed
modulations. These expressions also show the sum and difference frequencies that are

observed in the two-pulse experiment.

 

 

la
or

IL

It),

 

 

1 1

|V11V| I"! I“!

|C>r

 

 

 

 

1d-

Figure 2-13: Energy level diagram for S=l/2, I=1/2 system. The vectors indicate the transition
matrix elements connecting the eigenstates.

75
The transition probabilities for the branching transitions can be determined by

using transformations that can be described as rotations of the electron and nuclear

coordinate systems. The following Hamiltonians are obtained.

1' l
“(a =(DIIZ +-2'AIZ 3'5le

1 1 .
WB =wllz -§AIz -§le

Diagonalization of these Hamiltonians will yield eigenfrequencies given by

..-...=[(-;-A+wi’+iérs:lzl
.,-...=[(.;..-..)a(,32)21

these are the hyperfine frequencies where A and B represent the first and second order

101—-

[JI—

tenns of the hyperfine tensor (vide infra) and a), is the nuclear Larmor frequency.
Because the rotation operators utilized to diagonalize the above Hamiltonians
correspond to the M,, and M,,, the terms of the M matrix (the u and v components) can

easily be determined. The product |v|2|u|2 is then

2

and this can be used to give the final form of the envelope modulation functions

76

219,213?- . , 1 , ,1
Em0d(t)=1— ., 2 srn -wabtsm :11)ch

(”Eb“) cd

 

l
. 2_ _
20)]sz srn zwabrh coswcd(t+T)]
Emod(r+T)=l——— l
mabmcd + sin2 Ewcdtll — coswab(t + T)]

In these expressions the term (t1),B/tii),,,ti),,,)2 is called the modulation depth parameter and
it is clear from examination that if the nuclear Zeeman interaction, (0,, is equal to the
hyperfine frequency then the maximum modulation depth occurs. Furthermore, if the
assumption that the hyperfine coupling arises from a contact interaction, A,,,,, and a

classical magnetic dipole-dipole interaction, then the A and B parameters can be given as

ggnBBn

3 (3cos2 0 - l)

A=Aiso+ in

B = E’LBBEHRose sine)
hr

These assumptions have been made throughout the analysis of the ESEEM data presented

in Chapters 3-5.
Nuclear Quadrupole Effects

For nuclei with 131 interacting with the paramagnetic species, the ESEEM
experiment is substantially affected by the magnitude of the nuclear quadrupole
interaction.‘5 This interaction is present when the electric quadrupole moment of the
nucleus is surrounded by a non-spherical electron distribution. The quadrupolar nucleus

will interact with the electric field gradient from the non-symmetrical electron cloud

77
depending on the orientation of the charge distribution within the nucleus (see Figure 2-

14). Because the quadrupole moment is intimately tied to the nuclear spin through the
nuclear structure, the quadrupole interaction connects the spin directly to the molecular
network. The nuclear spin is, through its magnetic moment, also connected to the
magnetic fields produced by the spectrometer and unpaired electrons. Thus. the
quadrupolar interaction that is detected and measured by using ESEEM spectroscopy can
yield, when combined with the information extracted from the hyperfine coupling.

structural information regarding the paramagnetic species. The utility of measuring both

A B. "‘1
--q
0‘9

"‘1

Figure 2-14: The interaction of a quadrupolar nucleus with four point charges showing
the lowest energy state for the nucleus (nucleus B).

of these parameters is exhibited by the chlorophyll cation radical,P,,,,,+ , studied in
Chapter 4. By using multifrequency ESEEM and numerical simulations of experimental

data, the quadrupole coupling constants and asymmetry parameters were determined for

78
the pyrrole nitrogens in P,,,'. These results, and their implications for structure/function

relationships in PSI will be discussed in Chapter 4.

Quantitation of this quadrupolar interaction is through the quadrupole coupling
tensor, which is the product of the nuclear quadrupole moment and the gradient of the
electric field due to the surrounding electrons. This tensor is traceless and disappears in
the limit of a symmetrical distribution, e.g. cubic or octahedral, where the electric field
gradient is zero. However, a concentration of charge, for example from the electrons
participating in a covalent bond between a transition metal atom and a ligand, will result
in a non-zero gradient and the subsequent quadrupolar coupling is a measure of the
electron concentration in the bond. To understand the effects of this quadrupole coupling
on the energy levels, the Hamiltonian describing this interaction will be examined.
Because the quadrupole moment’s axis of quantization is collinear with the nuclear spin
angular momentum, the Hamiltonian for the quadrupole interaction is expressed in terms

of the operators for the nuclear spin. This operator is

2 2 2 2
anqi=x[312 —1 “1(111 —Iy)]

where x is the quadrupole coupling constant and reflects the extent to which the nucleus

is coupled to the electric field gradient and is given by

 

while 11 is the asymmetry parameter and measures the deviation from axial symmetry. It

is given by

79

 

 

 

 

1 1>
Electronic Nuclear Nuclear
Zeeman Zeeman Hyperfine Quadrupole

Figure 2-15: Energy level diagram for an S=1/2, I=1 system at exact cancellation.

80
qXX “‘1ny

1‘]:
922

 

 

where q”, qyy and qn are the principle values of the field gradient.

The effect of this Hamiltonian on the spin system will be investigated for the
special case where the electron nuclear hyperfine interaction is exactly twice the nuclear
Larmor frequency and nearly isotropic. This “exact cancellation” condition results in a
collapse of the energy levels in one spin manifold (see Figure 2-15).7 Complete
cancellation is only possible if the magnitude of the dipolar portion of the hyperfine
coupling is small compared to the isotropic contribution.8 The transitions in the
“cancellation” manifold arise from pure quadrupolar interactions and are easily resolved
with stimulated echo ESEEM spectroscopy. The peaks that arise from these “zero-field”
transitions appear at frequencies that correspond to pure quadrupolar interactions. The
energies of these transitions can be obtained by diagonalizing the 3 x 3 Hamiltonian

matrix. These frequencies are given by

v, =(3+n)x
v_ =(3-n)K
v0 =2nx

where the transitions v,,v, and v0 are shown in Figure 2-15.

In the other, non-cancellation, spin manifold, the effect of the hyperfine and
nuclear Zeeman fields must be taken into account. The external magnetic field due to
these interactions, with angles 0 and (p with respect to the quadrupolar axis system must

be incorporated into the Hamiltonian. Anisotropic hyperfine coupling with

81
\‘cff = iVi i‘ A/z‘

is assumed. The Hamiltonian matrix” corresponding to the non-cancellation manifold.

then, hasthe form
x(1+ n) veg sinGcoscp veg cosG
¢= veg sinOcoscp - 2x - ivcfi‘ sine sintp
veg cosO - IVefi‘ sinOsintp 1((1- n)

Trigonometric as well as graphical solutions to this Hamiltonian are available.'0
Experimentally, the exact cancellation condition (for l"N) is manifested by the
appearance of three sharp, low frequency peaks plus a broader “double quantum” peak at
higher fi'equency, which arises from transitions (AM, = -1 to AM, = +1) originating from
the non-cancellation manifold.” The appearance of this peak is indicative of the
contribution of the dipolar coupling to the total hyperfine coupling in the system. The
anisotropic hyperfine interaction acts to mix the eigenstates this alters the transition
probabilities between the eigenstates in the non-cancellation manifold. The probability of
the double quantum transition occurring is, therefore, decreased with a concomitant loss
of amplitude in the ESEEM spectrum. The anisotropic contribution can be estimated
based on the appearance of this peak and this information can be utilized to simulate
numerically the experimentally observed data This methodology has been used to
determine the magnetic properties of the chlorophyll cation radical involved in the

primary electron transfer events in PS1 (see Chapter 3-4).

LIST OF REFERENCES

(h s

LIST OF REFERENCES

 

' A) Mims, W.B. & Peisach, J (1981). in Biological Magnetic Resonance (Berliner, L.J .
& Reuben, J ., eds.) pp. 213-263, Plenum Press, New York. B) Kevan, L. (1990) in
Modern Pulsed and Continuous Wave Electron Spin Resonance (Kevan, L. & Bowman,
M.R., eds.) Wiley-Interscience. C) McCracken, J. In Handbook of Electron Spin
Resonance, Vol. II, (Poole, C.P., ed.) in press.

2 A) Atherton, NM. (1993) Principles of Electron Spin Resonance, Ellis-Horwood/PTR
Prentice Hall, New York. B) Wertz, J .E. & Bolton, J .R. (1986) Electron Spin Resonance:
Elementary T heary and Practical Applications, Chapman & Hall, New York. C) Gordy,
W. (1980) Theory and Applications of Electron Spin Resonance in Techniques of
Chemistry, Vol. XV (West, W., ed.) John Wiley & Sons, New York. D) Kevan, L. &
Kispert, L.D. (1976) Electron Spin Double Resonance Spectroscopy, John Wiley & Sons,
New York. E) Keijzers, C.D.-; Reijerse, E.J. & Schmidt, J. (1989) Pulsed EPR. A New
Field of Applications, North Holland Publ.

3 McConnell, H.M.; Heller, C.; Cole, T. & Fessenden, R.W. (1959) J. Chem. Phys. 82,
766.

" Mac, M.; Tang, X.-S.; Diner, B.A.; McCracken, J. & Babcock, G.T. (1996)
Biochemistry 35, 13288.

5 A) Mims, W.B. (1972) Phys. Rev. B. 5(7), 2409.

6 Lucken, E.A.C. (1969) Nuclear Quadrupole Coupling Constants, Academic Press, NY
7 Mims, W.B. & Peisach, J. (1978).]. Chem. Phys. 69, 4921.

3 Flanagan, H.L. & Singel, DJ. (1987).]. Chem. Phys. 87, 5606.

9 A) Casabella, P.A. & Bray, P.J. (1958).]. Chem. Phys, 28, 1182. B) Muha, GM. (1980)
J. Chem. Phys, 73, 4139.

'° Astashkin, A.V.; Dikanov, S.A. & Tsvetkov, Yu. D. (1984) J. Struct. Chem. 25, 45.
” Gerfen, G.J. & Singel, DJ. (1994).]. Chem. Phys. 100, 4127.

82

I."

Chapter 3

Identification of Histidine as an Axial Ligand to P,,,,,’

Background

Long range electron transfer and the chemiosmotic hypothesis developed in the
1960’s by Peter Mitchell‘ and others 2’3 proposed that the key electron transfer reactions
of respiration and photosynthesis were directed across the 35 A span of the membrane
dielectric. Although these ideas have gained support and general acceptance over the past
25 years, many of the factors governing long range electron transfer reactions have
remained unclear. The simple, non-adiabatic description of electron transfer given so
eloquently by Fermi’s golden rule indicates that the nuclear positions and environments
of reacting molecules play a central role in determining the electron transfer rate:

kravztrc
h

where V, is the electronic coupling of the reactant and product wavefunctions and PC is
the F ranck-Condon factor representing the integrated overlap of the reactant and product
nuclear wavefunctions. The classical view of nuclear motion described by Marcus‘ and
the analogous quantum mechanical explanations of Levich and Dogonadze’ and of
Jortner6 treat the nuclear reactant and product wavefunctions as harmonic oscillator
potential surfaces with the subsequent overlap of these wavefunctions representing the
Franck-Condon factors. This overlap will be maximal when the Gibbs free energy (-AG°)
is equal to the reorganization energy (it), the energy required to change the equilibrium

83

84
geometry of the reactant state to the product state without concomitant electron transfer.

The overall gaussian dependence of the electron transfer rate on this free energy predicted
by Marcus has been confirmed experimentally; the rate rises with increasing -AG° until -
AG° is equal to A. then falls as -AG° exceeds the reorganization energy. From a classical
viewpoint, the width of this gaussian, is proportional to the Boltzmann thermal energy.
However, if the effective frequency of the oscillator is large with respect to the thermal
energy, then a quantum mechanical approach is necessary. With the quantum
mechanically corrected view of the harmonic oscillator, nuclear tunneling becomes an
added concern and the breadth of the gaussian approximating the rate/Gibbs free energy
relationship becomes markedly temperature independent and proportional to thm .7

The application of these theories to the photosynthetic systems of bacteria has
provided a wealth of kinetic, thermodynamic and structural information on intermolecular
electron transfer.‘ These results indicate that the electron transfer rates in these systems
are modulated by the interdependence of -AG° and A, which are, in turn, dependent upon
the geometric and electronic structures of the cofactors involved in the electron transfer
reactions. The availability of a high resolution X-ray crystal structure of the reaction
centers from purple non-sulfur bacteria has been a tremendous catalyst to the study of
these electron transfer rates, providing an in viva geometric structure of the cofactors.’"°
Similar studies of oxygenic photosynthetic systems have been hindered by the lack of an .
analogous high resolution structure. Because the bacterial systems lack the oxidative

power necessary for oxygen evolution, a comparative study of the two systems, which

85
contain many structural and mechanistic similarities, is desirable from an evolutionary

perspective.

In PSI, the primary charge separation event proceeds via the light-induced
generation of a chlorophyll a cation radical species (P,,,,,*).‘1 This photochemically active
chlorophyll is distinguished from the light harvesting and energy transfer chlorophylls by
its spectral and redox properties. Initial charge separation occurs with a quantum yield
near unity, even though recombination back to the ground state is favored
thermodynamically. In spite of the fact that hypotheses rationalizing this unidirectional
electron transfer have been advanced,‘2 no definitive explanation has been found.
Elucidation of the geometric and electronic structures of the primary electron transfer
cofactors, specifically P700, is necessary to understand the directional specificity and
efficiency of forward electron transfer in PS1. Although significant progress has been
made in solving the crystal structure of PSI," the resolution is not yet sufficient to
provide a clear view of the geometric structure of P700; neither the nuclearity nor the
ligation spheres of the chromophore has been determined definitively. Because the
ligation and hydrogen bonding environments of chlorophyll species in viva can affect
profoundly the chemical and optical properties of the cofactor, and therefore, the forward
electron transfer rate, work exploring the ligation sphere of P7,,+ was undertaken and is
presented in this chapter. An analogous investigation of the electronic structure of the
cation radical utilizing isotope enrichment, ENDOR, multifrequency ESEEM and

numerical simulations of experimental data will be presented in Chapter 4.

86
Chlorophyll

The ability of a chlorophyll species in vivo to perform a variety of functions. from light
harvesting and energy transfer to primary electron donor and acceptor roles. reflects the
flexibility of the species and the sensitivity of its molecular orbitals (MO’s) to
modulation by the local protein environment. Characteristic optical. electronic and
magnetic properties of chlorophyll species in viva, which are photosystem specific.
emanate from the relative energies and electron configurations of the valence MO’s of the
neutral or ionic species."," These general features can be understood best by examining
the electronic structure of a porphyrin containing perturbed valence MO’s.

Porphyrins contain four fully unsaturated cyclic tetrapyrroles and have D4h
symmetry (see Fig. 3-1A). Saturation of ring IV results in the formation of a chlorin (Fig.
3-1B), additional saturation of ring 11 creates a bacteriochlorin (Fig. 3-1C). The
progressive saturation of the C,,-C,3 bonds changes the energies, symmetries and electron
distributions of the porphyrin MO’s, which destabilizes the valence orbitals and alters the

‘ The effect of these distortions on the

shape and symmetry of the cyclic rt system.l
chemical properties of the chlorin or bacteriochlorin can be observed by examining the
orbitals comprising the highest occupied and lowest unoccupied 1: molecular orbitals
(HOMO, HOMO-l, LUMO and LUMO+1) for the porphyrin derivatives relative to the
unperturbed porphyrin, see Figure 3-2; symmetry labels for the Zn porphyrin, chlorin and

bacteriochlorin in this figure are from the D,,, point group and have been utilized for

87

/| \ \
<//N\ ;.N\<

Fe

|\/

A. Porphyrin

 

B. Chlorin C. Bacteriochlorin

Figure 3-1 - Comparison of A) porphyrin, B) chlorin and C) bacteriochlorin compounds

88
clarity, in spite of the fact that the symmetry of chlorin and bacteriochlorin

compounds deviates from “pure” D,,, symmetry."

For the chlorin, ab initio molecular fragment and iterative extended Huckel (IEH)
calculations predict that the a“l MO lies in close proximity to another orbital of different
symmetry, the a,,,.'7 The relative energies of these two it orbitals can be affected by ring
substituents, the identity of the central metal and, or, the presence of axial ligands." In
the porphyrin, the rt“ LUMO and LUMO + 1, es, and e”, represent a “true” degenerate
pair, since the x and y axes are equivalent in a molecule with D,,, symmetry. Saturation
of the C3-C3 bonds in the chlorin, however, removes this degeneracy and destabilizes the
e,y as well as the HOMO (either an, or a,,). The destabilization of the HOMO with
respect to the LUMO is reflected experimentally in the increasing ease of oxidation
observed for bacteriochlorophyll a (E°",,,(CH,C1,) = 0.64 V vs. NHE)l9 compared to that
of chlorophyll a (E°",,, (CH,C1,) = 0.74 V vs. NHE).2° The ring saturation does not affect
the LUMO, the reduction potentials of the two moieties are similar.”21

The energy differences between the ground and excited states of the different
porphyrin derivatives directly affects their optical properties and is manifested in the
spectrum as an increasing red shift with concomitant changes in oscillator strength and
intensity as the symmetry of the molecule is decreased. Additional modulation of these
properties in viva as compared to in vitra has been observed and is attributed to
interactions of the chromophore with its protein environment through hydrogen bonding
or axial ligation of the central metal atom."m Semi-empirical and ab initio calculations

on bacteriochlorophyll analogues have correctly predicted the blue and red shifts

89

 

 

7t*
* ' . 0
"es" <1:>___.x’ -_---__ ......... _1t_*__

>
3
ES
0
C
D
'5 "alu" (T!)
e n r . v ' ' '
0 am. (it) . - ‘ '

"3211" (n) ,' ,’ "32,," (1t) "32“" (TI)

"31"" (n) "’o

Porphyrin Chlorin Bacteriochlorin

Figure 3-2: Energies of the highest occupied and lowest unoccupied molecular orbitals
for Zn porphyrin, chlorin and bacteriochlorin compounds."

90 ,
observed in bacteriochlorophyll compounds containing hydrogen bonded residues at the

13’-keto or 3-acetyl positions, respectively?3 Similar shifts are predicted with the
addition of water or imidazole ligands to the axial metal atom?" The experimental data.
though, show only minor perturbations, in ligating and hydrogen bonding solvents the
optical spectra of bacteriochlorophyll a and bacteriopheophytin a remain essentially
unaffected?4 However, because the chlorin and bacteriochlorin molecules are so large,
these calculations are performed with minimal basis sets and reflect only trends in
behavior and, therefore, are not quantitative. Nevertheless, the basic trends that have
emerged from these calculations combined with experimental results have provided a
foundation for understanding the general features of the chlorophyll optical spectra. the
redox properties and the unpaired spin density distributions in the anionic and cationic
chlorophyll and bacteriochlorophyll species.

The magnetic properties of the cation radicals are also affected by the decrease in
symmetry. The unpaired spin density distributions for the a,u and a,u MOs differ
markedly." The a,u orbital is characterized by large unpaired spin densities on the
carbons adjacent to the nitrogens while the unpaired spin in the a,u orbital resides mainly
on the meso carbons and the nitrogens. There is also unpaired spin density on the central
metal atom for the a,u MO. Calculations on Mgn and Co‘" metalloporphyrin cation
radicals where the unpaired electron is in either the a,u or a,u MO reflect this disparity in

unpaired spin density distributions, see Table 3-1?"

91
Table 3-1 Calculated 1: Spin Densities for Metalloporphyrin Cation Radicals

 

 

 

 

 

 

2A).. ZAzu
Mg“ Ca111 Mgn Com
C,1 0.098 0.102 0.0005 0.0005
C,, 0.027 0.023 0.019 0.014
Cm,o 0.000 0.000 0.151 0.135
nitrogen 0.000 0.000 0.043 0.069
metal 0.000 0.000 0.009 not measured

 

 

 

 

Magnetic resonance spectroscopy is an invaluable tool for studying these
differences because elucidation of the electron-nuclear hyperfine coupling constants for a
radical provides a direct measurement of its singly occupied molecular orbital (SOMO).
For example, in chlorophyll a cation radical monomeric model compounds, a marked
temperature dependence of the proton hyperfine couplings at positions 17 and 18 has
been observed.27 The magnitude of these couplings decreases as a function of increasing
temperature” and is evidence for thermal mixing of the ground and first excited states, as
the proximity of the a,“ MO allows thermal accessibility upon an increase in temperature.
A similar decrease in hyperfine couplings is observed when the chlorophyll a compound
is studied in ligating solvents." Axial ligation of the central Mg atom may induce mixing
between the ground a,u and first excited a,u states by lowering the energy of the a,u orbital.
This hypothesis is supported by IEH calculations.28 In the bacteriochlorophyll, however,
the singly occupied a,“ lies well above the an" and therefore the small perturbations in
electronic structure that result from axial ligation or interactions with the protein
environment do not perturb the energetics of the system enough to bridge the energy gap

between the orbitals and the probability of state mixing is smaller. This is evidenced

92
experimentally29 as axial ligation and hydrogen bonding have been found to have little

effect on the unpaired spin density distributions. These differences in unpaired spin
density distributions of the cation radicals of the primary donors in these systems and

their effect on function will be explored in greater detail in Chapter 4.

Axial Ligation of P7,,+ in PSI

X-ray crystallographic studies of the bacterial reaction centers have shown that the
primary donor in these systems consists of a dimer of BChl with each BChl having a
single histidine ligand.”'lo Recent chemically induced dynamic nuclear polarization
experiments have confirmed the presence of this nitrogen based.30 Specific mutation of
either of these ligands to leucine or phenylalanine” results in the loss of Mg?‘ from the
associated BChl and the generation of a BPheo.32 Interestingly, the replacement of
histidine with glycine does not result in the generation of a BPheo; a coordinating water
molecule is thought to stabilize the P+ structure.33

The identity of the axial ligand(s) to Pm,+ remains uncertain as previous attempts
at identification by using a combination of biochemical manipulation and subsequent
spectroscopic evaluation have been inconclusive." The presence of histidine ligands to
P’ in the bacterial reaction centers, as well as conserved histidine residues in the amino
acid sequence of psaB, the PSI protein known to bind P700, suggests this residue as the
most likely candidate'"” Recent proton ENDOR and ESEEM experiments on mutants of
Chlamydamas reinhardtii indicate that the observed spectroscopic properties of the

primary donor in this system are not affected by mutations of h13523, a highly conserved

93
residue that is a good candidate for the putative histidine ligand to P,o,°.‘” These

mutations were expected to generate a pheophytin. analogous to the bacterial system.
however the ENDOR and ESEEM spectra collected on these mutants (H523Q and
H523L) failed to detect any changes in the unpaired electron spin distribution for the
radical. Because the spin density distribution for a pheophytin molecule differs
considerably from that of chlorophyll“ , it was concluded that H523 is not a ligand to
P,,,,,‘. The possibility that the induced mutations could have allowed water access to the
binding site as a bridging ligand,33 preserving the integrity of the P,,,,,‘ structure. could not
be ruled out. Moreover, since the X-band ESEEM spectrum is dominated by
contributions from the pyrrole nitrogen atoms of the Chl a macrocycle. perturbations in
the hyperfine coupling resulting from changes to the axial ligand would be difficult to
detect.

However, because the EPR signal from P7,,+ consists of a single resonance line 3’
that is inhomogeneously broadened by the presence of multiple overlapping hyperfine
interactions, information regarding the magnitude of these hyperfine coupling constants is
obscured. These couplings can. be resolved by the application of advanced EPR
techniques. 'H and "N ENDOR and electron spin echo envelope modulation (ESEEM)
studies on frozen solutions of PS1 have indicated" that Pm,+ is a dimeric Chl a species
characterized by an asymmetrical spin density distribution similar to that found in the
’ bacterial reaction centers.39

The recent development of histidine-tolerant mutants‘0 of the cyanobacterium
Synechocystis PCC 6803 has allowed for specific isotopic labeling of the histidine

nitrogens. By using a combination of ENDOR and ESEEM spectroscopies and these

94
specifically labeled samples, we are able to present the first definitive spectroscopic

evidence for a histidine ligand in Pm‘.

Materials and Methods

Strains and Cell Growth. A histidine-tolerant strain was isolated from the
cyanobacterium Synechocystis PCC 6803 as described previously."0 Cells of this strain
were grown photoautotrophically at 30° C for 5 days in lO-L carboys by using cool-white
fluorescent lamps (7 W/m2). BG-ll medium"l was supplemented with 240 uM of either
DL-histidine containing only natural-abundance IN, or DL-histidine containing two "N
atoms in its imidazole group. For "N global labeling, 99.9% "N- nitrate was used as the
sole nitrogen source during cell growth. For I"N-histidine-reverse labeling, this "N
nitrate containing BG-ll medium was supplemented with 240 uM l"N histidine. The
growth medium was bubbled with 5% CO, in air. With these experimental conditions,
approximately 85% of the histidine molecules incorporated into thylakoid proteins were
from the histidine supplemented in the growth medium as shown previously by mass
analysis of histidine.‘0

Preparation of PS1 care complexes: PSI core complexes were purified according to the

42

procedure described earlier. The complexes were eluted from DEAE-Toyopearl 650$

column by using 50 mM MgSO, in 50 mM MES-NaOH buffer (pH=6.0) containing
0.03% dodecylrnaltoside, 25% (w/v) glycerol, 5 mM MgCl, and 20 mM CaCl,. The

samples were desalted by passage through a gel filtration column (Econo-pak 10DG, Bio-

95
Rad) equilibrated with 50 mM HEPES-NaOH (pH=7.5) buffer containing 10 mM NaCl

and 0.03% dodecylmaltoside, and then concentrated down to about 15 mg chl/mL by
using centricon 100 (Amicon). To generate P700’, 1 mM potassium ferricyanide was
added to the samples. Finally, the samples were loaded into 4 mm O.D. EPR tubes and
frozen in liquid nitrogen.
ENDOR spectroscopy and data analysis: The ENDOR data were obtained at X-band on
a Bruker ESP300e spectrometer with a Bruker ESP360 DS ENDOR accessory and home-
built demountable coils.’3 All ENDOR data were collected at a field position
corresponding to the center of the EPR absorption line. Constant temperature in the
cavity was maintained with an Oxford ESR900 continuous flow cryostat. Microwave
frequency was detemrined by using an EIP Microwave Model 25B frequency counter and
the static magnetic field strength was measured with a Bruker ER035M NMR gaussmeter
The peaks attributed to "N (I=1/2) electron nuclear hyperfine coupling are

expected to be split symmetrically about the nuclear Larmor frequency for "N (v ,S,, =
1.45 MHz at 3500 G). In solution ENDOR, the position of these peaks is indicative of
the strength of the isotropic hyperfine coupling according to the relationship“

v: = v": A,,°/2
where vN _>. Aha/2. However, in frozen solution, the anisotropy of the hyperfine coupling
gives rise to powder lineshapes described to first order by

v: = VN: A/2

where A = All c0520 + A, sin2 0

96
All and Ar in the above expression are the principal values of the axially symmetric

hyperfine tensor and 0 describes the orientation of the principal axis system of the
hyperfine tensor with respect to the laboratory magnetic field.45

ESEEM data collection and analysis: ESEEM data were collected on a home built
spectrometer“S by using a reflection cavity where either folded stripline‘7 or slotted-tube
structures" served as the resonant element. A three pulse or stimulated echo (90°-r-90°-
T-90°) pulse sequence was used. Dead time reconstruction was performed prior to
Fourier transformation as described.‘9 Computer simulations of the ESEEM data were
performed on a Sun Sparcstation 2 computer utilizing FORTRAN software based on the
density matrix formalism of Mims.so The analysis software for the treatment of
experimental and simulated ESEEM data was written with Matlab (Mathworks, Nantick,
MA). The experimental dead time was included in the simulations. An isotropic g-tensor
was assumediin all of the calculations. The success of the spectral simulations was based
on the modulation depth and duration observed in the time domain traces as well as the

lineshapes, peak positions and relative peak intensities in the frequency spectra.

Results

The use of ENDOR spectroscopy in conjunction with ESEEM in these studies
exhibits the complementarity of the two techniques. Couplings containing a large dipolar
component are usually difficult to detect with ENDOR, since the broadened powder

patttem lineshape precludes resolution of the turning points. In contrast, the dipolar

97
portion of the hyperfine coupling mixes the nuclear states within each electron spin

manifold and enables the "forbidden" EPR transitions that give rise to ESEEM.
Similarly, purely isotropic hyperfine couplings will result in discrete EPR transitions
from each hyperfine level and eliminate the interference effects that give rise to the
modulations.” ENDOR spectroscopy can, therefore, be used to resolve primarily
isotropic nitrogen couplings, while ESEEM is suited best for hyperfine couplings that
contain a considerable dipolar contribution. The degree of this contribution can be
quantitated by applying both techniques.

The ENDOR first derivative spectrum collected at 6 K for P7,; globally labeled
with "N is shown in Figure 3-3 (top). The analogous spectrum from Pm‘ specifically
labeled with "N histidine and “reverse” labeled so that all nitrogens are "N except those
of histidine, which contains natural abundance I"N, are shown in Figure 3-3, middle and
bottom, respectively. Features observed in the spectra of Figure 3-3 are expected to arise
from all "N nuclei coupled to the electron spin; these include nitrogens in the chlorophyll
macrocycle as well as in the putative histidine ligand. By using specific isotope labeling,
identification of thepeaks arising from this ligand is facilitated. The peaks at 1.13 and
1.78 MHz appear in the spectrum from the globally labeled sample (Fig. 3-3, top) as well
as in the spectrum from Pm,+ containing "N labeled histidine (Fig. 3-3, middle). Upon
reverse labeling, these peaks disappear (Fig. 3-3, bottom), thus confirming histidine as
their origin. These features are split symmetrically about the Larmor frequency for "N
and correspond to an isotropic hyperfine coupling of 0.64 MHz. Moreover, features in
the 2-3 MHz region can be assigned definitively to contributions fiom the nitrogens in the

chlorophyll a macrocycle. A comprehensive analysis of these chlorophyll pyrrole

98

 

 

 

All 15N
MW
’5 ‘15N histidine
5?.
OJ
‘0
2
‘5 M
E W
<
1:1:
CL
UJ
14N histidine
1.5 2 2.5 3 3.5

Frequency (MHz)

Figure 3-3: ENDOR spectra collected at 6K for P,,,,,’ globally labeled with "N (top),
specifically labeled with "N histidine (middle) and "reverse" labeled, “N histidine in the

presence of "N nitrate (bottom).

Spectrometer conditions for all spectra, unless

otherwise indicated: microwave power, 1.99 mW; magnetic field strength, 3375 G (tap),
3356 G (middle), 3367 G (bottom); RF power, 200 W; RF frequency modulation, 100

kHz.

 

99
nitrogen couplings by using multifrequency ESEEM and "N ENDOR is presented in

Chapter 4.

The presence of this nitrogen derived histidine coupling does not affect the
ESEEM spectrum as demonstrated in Figures 3-4 and 3-5. The cosine Fourier
transformation of three-pulse time domain data from P7,,+ containing natural abundance
I"N is identical (Figure 3-4, top) to the spectrum from P,,,,,‘ specifically labeled with "N
histidine (Figure 3-4, bottom). These spectra are dominated by 1"N modulations from the
Chl a pyrrole nitrogens and are typical for situations where the hyperfine interaction is
approximately equal to twice the nuclear Zeeman energy. Under these conditions the
energy level splittings in the spin manifold where the nuclear Zeeman and electron-
nuclear hyperfine interactions cancel is determined primarily by the l’N nuclear
quadrupole interaction. This exact cancellation condition is characterized by modulations
that are deep and long-lived. Because these modulations will dominate the ESEEM
spectrum, contributions from the nitrogen of the histidine ligand, which is away from
exact cancellation, are not expected to be observed. To eliminate this deep I"N ESEEM
from our measurements, analogous three-pulse ESEEM experiments were performed on
the "N enriched and “reverse” labeled samples (Figure 3-5, top and bottom, respectively).
No differences were resolved in the spectra, indicating a minimal contribution to the

ESEEM spectra from nitrogen hyperfine coupling originating from the histidine?

Discussion
The presence of a histidine ligand bound to the central Mg2+ atom of P,,,,,’ has

often been postulated as it has been shown to ligate the primary donor in the bacterial

100

 

all 14N

15N histidine
,ri
\ q. if WW

 

Amplilude (a.u.)

 

l "t

 

 

J I

8 10

 

4 6
Frequency (MHz)

Figure 3-4: Fourier transformations of three-pulse ESEEM data frOm P70; containing
natural abundance "N (top) and specifically labeled with "N histidine (bottom).
Spectrometer conditions: magnetic field strength, 3210 (top), 3195 (bottom); microwave
pulse power, 45 dBm; microwave pulse length (F WHM), 15 ns; pulse repetition rate, 20
Hz; tau value, 175 ns and sample temperature, 4K.

101

 

 

all 15N

 

 

 

 

 

 

 

3"

53.

CD

‘0

2

:3

E l

<

14N histidine
1 o ' l I J
0 1 2 3 4 5 6 7

Frequency (MHZ)

Figure 3-5: Three-pulse ESEEM frequency spectra from Pm,+ globally enriched with 1’N
(top) and ”reverse" labeled (bottom). Spectrometer conditions: magnetic field strength,
3195 G; microwave pulse power, 45 dBm; microwave pulse length (F WHM), 15 ns;
pulse repetition rate, 90 Hz; tau value, 250 ns; sample temperature 4K.

102
reaction center 9'" 5‘ but until now, not demonstrated unambiguously. Spectrosc0pic

characterization of this ligand is necessary to understand the optical and magnetic
properties of the cation radical. Modulation of these properties is thought to occur by
interactions of the chromophore with its protein environment through. for example.
hydrogen bond interactions with nearby amino acid residues or through axial ligation to
the Mg”. In heme proteins, cavity mutants generated by site directed mutagenesis allow
the introduction of exogeneous organic ligands that greatly affect the functional
properties of the protein”. A similar strategy to examine the effect of axial ligands on
magnetic and optical pr0perties of the primary donor of bacterial reaction centers has
been applied by Goldsmith et a1.” However the lack of an adequate spectroscopic probe
has precluded characterization of the ligand. Identification of a histidine ligand to P,Q,+
provides the spectroscopic methodology necessary for characterization of these mutants.
It is possible, however, that the histidine derived nitrogen hyperfine coupling
shown in Figure 3-3 arises from a histidine residue hydrogen bonded to the 9-keto group
of the conjugated rt system of the chlorophyll a. Amino acid residues interacting non-
covalently with the primary electron donors in photosynthetic reaction centers from a
variety of organisms have been suggested as the source of the disparity in their redox
potentials and optical properties.” The role of H-bonded residues as redox potential
"tuning modules" is not unusual in biological systems; it has been suggested as an
explanation for the wide range of redox potentials observed for [nFe-mS] clusters’4 and
has recently been shown to alter the internal charge transfer state in BChl-BPheo

heterodimer mutants in bacterial reaction centers.55 The precedence of histidine residues

103
in H-bonded interactions with P’ has been demonstrated in the bacterial reaction centers.“

Although the H-bonded histidine was not detected directly by spectroscopic methods.
changes in the redox potential and unpaired electron spin density distribution subsequent
to mutations of his L168 led to conclusions that this residue participates in a hydrogen
bond with the 2-acety1 group of the BChl. However, because the spin density distribution
of P+ in bacterial reaction centers differs from that of P70; in PS138 the lack of
spectrdscopic evidence for this H-bonded histidine does not preclude detection in PS1.
Therefore, even without a covalent interaction, ENDOR spectroscopy could detect the
coupling from a H-bonded moiety provided that sufficient dipolar coupling between the
unpaired rt-electron of the radical and the nitrogen nucleus are present.
Hydrogen-bonding interactions of this genre have been studied in other
paramagnetic systems, both in viva and in vitra and can provide information regarding the
form of the hyperfine tensors expected. In powder samples, these interactions were
characterized initially by O’Malley and BabcockS7 who used isotopic labeling and
ENDOR spectroscopy to investigate H-bonds in p-benzoquinone anion radicals. The
proton hyperfine coupling tensor for solvent protons H-bonded to the oxygen of the
semiquinone radical in this system was shown to be axially symmetric. The ratio of -1:-
1:2 for the principal values was determined and corroborated by earlier results from
single crystal work.” In viva, these interactions have been studied extensively in a
variety of systems containing Fe-S centers. The hyperfine tensors of both the proton and
nitrogen involved in the hydrogen bonding interaction have been determined from

electron magnetic resonance experiments. Pulsed l'zH ENDOR and 2H-ZH TRIPLE were

104
used by Doan et al.59 to determine the intrinsic hyperfine coupling of three strong NH-"S

hydrogen bonds in the [Fe,S,(cys),]2' cluster in hydrogenase from D. gigas. Surprisingly.
the hyperfine tensors for the protons hydrogen bonded to the cysteinyl sulfur contained a
substantial isotropic interaction, although no covalent bond exists between the proton and
sulfur atoms. Molecular orbital calculations have shown that the unpaired electron spin
in these systems is highly centralized on the individual Fe atoms, as much as 80% of the
total spin in the case of F e”, and most likely contributes to the isotropic character of the
proton coupling through a spin polarization or hyperconjugation mechanism.60 Spin
density is found on the peptide nitrogen nucleus as well and has been observed
spectroscopically by Cammack et a1."l from an Fe-S cluster of furnarate reductase from
Escherichia coli. In these studies, the hyperfine and quadrupole parameters for a nitrogen
atom assigned to an NH'S type hydrogen bond to the cluster were determined by using
ESEEM spectroscopy and numerical simulations of the ESEEM data. Observation of the
nitrogen coupling was again dependent on the presence of a large amount of unpaired
spin density localized on the Fe of the cluster.

It is unlikely that these spectroscopic methods could be used to detect a H-bonded
histidine in Pm’. The spin density distribution on the 9-keto oxygen, although not
determined definitively. has been estimated to be on the order of 0.4% in the chlorophyll
a cation radical in vitra.62 The majority of the unpaired spin in these systems is
delocalized on the methine carbon atoms and the four nitrogen atoms."3 However, if the
peaks attributed to the nitrogen of the histidine observed in the ENDOR spectra of Figure
3-3 (top and middle) do represent the perpendicular components of a purely dipolar

hyperfine tensor, then the corresponding parallel components would be expected in the 2

105

 

 

Experimental

 

Amplilude (a.u.)

 

Simulation
,//\
/

O 1 2 3 4 5
Frequency (MHz)

 

 

 

 

Figure 3-6: Fourier transformations of three-pulse experimental (top) and simulated
(bottom) ESEEM spectra from P7,; globally labeled with "N. Experimental conditions

were identical to those of Figure 3-5. Simulation parameters: A1 = -0.64 MHz, AII = 1.28
MHz, all other parameters as in experiment.

106
MHz region of the spectra. No discernible peaks can be seen in this frequency range in

either of the spectra. The resolution of the parallel components of the hyperfine tensor
can be difficult as the number of orientations with parallel symmetry contributing to the
overall lineshape is small and the lack of these features is not indicative of a purely
isotropic hyperfine interaction.

To estimate the dipolar contribution to the observed "N-histidine hyperfine
coupling, numerical simulations of ESEEM data fi'om "N enriched Pm,‘ were performed.
Because the modulation depth observed in an ESEEM experiment relies on the quantum
mechanical mixing of the nuclear states that occurs as a result of this contribution, it can
be used to detemrine the dipolar character of a coupling. "N ESEEM simulations using a
purely dipolar tensor, A1 = -0.64 MHz, A, = 1.28 MHz, predict that a broad feature,
centered at the "N Larmor frequency, 1.4 MHz, would be observed under our
experimental conditions (see Figure 3-6). These simulations indicate that the peaks
observed in the ENDOR spectrum are primarily isotropic in character and are consistent
with a nitrogen participating in a covalent bond with the Mg”.

The presence of an isotropic coupling to the nitrogen of the histidine ligand
implies a non-zero spin density on the Mg” atom of Pm]. Molecular orbital calculations
on chlorophyll a monomers have shown that the availability of a low lying excited state
with a decidedly different spin distribution that can influence the electronic structure of
the radical and subsequently its magnetic resonance characteristics."2 The spin density
distribution of this excited state A,u molecular orbital includes spin delocalized on the .
Mg2+ atom.” ENDOR experiments comparing the magnitude of proton hyperfine

couplings in P,Q,+ to those of chlorophyll a model compounds led O’Malley and

107
Babcock"4 to suggest that the ground state orbital for P700’ contains 75% ground state and

25% excited state character. These researchers suggested that interactions of Pm,’ with its
protein environment, either through axial ligation of the Mg” atom or by electrostatic
interactions of the radical with nearby charged amino acid residues, could provide the
perturbation necessary to mix the ground and excited state molecular orbitals. These data
and the presence of a spectroscopically active histidine ligand to P,,,,,’ exhibiting an
isotropic hyperfine interaction are supportive of a model of Pm’ where the excited state
orbital makes a significant contribution to the spin density distribution of the cation
radical and may explain the differences in redox potentials and optical properties

observed for the plant and bacterial primary donor species.

LIST OF REFERENCES

LIST OF REFERENCES

 

I Mitchell, P. (1961) Nature 191, 144.
2 Chance, B. & Nishimura, M. (1960) Proc. Natl. Acad. Sci. USA 46. 19.
3 Devault, D. & Chance, B. (1966) Biophys.. J. 6, 825

‘ Marcus, RA. (1956) J. Chem. Phys. 24, 966; Marcus, R.A. & Sutin, N. (1985) Biachim.
Biophys. Acta 811, 265

5 Levich, V.G. & Dogonadze, RR, (1959) Dokl. Adad. Nauk USSR 124, 123
6 Jomter, J. (1976) J. Chem. Phys. 64, 4860

7 Maser, C.C., Keske, J .M., Warncke, K., Farid, R.S., Dutton, PL. (1992) Nature 355,
796.

8 Michel, H., Deisenhofer, J. & Epp, O. (1986) EMBO J. 5, 2445

9 Deisenhofer, J., Epp, 0.. Miki, K., Huber, R. & Michel, H. (1985) Nature 318, 618-624.

'° Allen, J.P., Feher, G., Yeates, T.O., Komiya, H. & Rees, DC. (1987) Proc. Natl. Acad.
Sci. USA 84, 5730-5734.

” Golbeck, J. (1992) Annu. Rev. Plant Physiol. Plant Mol. Biol, 43, 293-324.
'2 Michel, Beyerle, ME. (1988) Biochim. Biophys. Acta 932, 52.

" Fromme, P., Witt, H.T., Schubert, W.-D., Klukas, O., Saenger, W. & Krauss, N. (1996)
Biochim. Biophys. Acta 1275, 76-83;

" Gouterman, M. (1978) in The Porphyrins, Vol. 3, (Dolphin, D., ed.) Academic Press,
New York, Chapter 1.

" Weiss, C., (1972) J. Mol. Spectrosc. 44, 37.
108

109

 

" Chang, GK, Hanson, L.K., Richardson. P.F.. Young. R.. & F ajer, J. (1981) Proc. Natl.
Acad. Sci. USA 78, 2652. .

‘7 Spangler, D., Maggiora, G.M., Shipman, L.L. & Christoffersen.R.E.. (1977) JAm.
Chem. Soc. 99, 7470;

" Fajer, J. & Davis, MS. (1979) in The Porphyrins, Vol. 4 (Dolphin, D., ed.) Academic
Press, New York, Chapter 4

- ‘9 Davis, M.S., Forman, A. & F ajer, J. (1979) Proc. Natl. Acad. Sci. USA 76. 4170.

2° F ajer, J ., Brune, D.C., Davis, M.S., Forman, A. & Saulding, L.D. (1975) Proc. Natl.
Acad. Sci. USA 72, 4956.

” Fujita, 1., Davis, M. S. & Fajer, J. (1978).]. Am. Chem. Soc. 100, 6280.
22 Callahan, P.M. & Cotton, T.M. (1987) J. Am. Chem. Soc. 109, 7001.

” Hanson, L.K., Thompson. M.A. & Fajer, J. (1987) in Progress in Photosynthesis
Research, Vol. 1, (Biggins, J ., ed.) Martinus Nijhoff, Dordrecht, 311.

2’ Cotton, T.M. & Van Duyne, R. P. (1981) J. Am. Chem. Soc. 103, 6020.

25; Davis, M.S., Forman, A., Hanson, L.K., Thomber, J .P. & Fajer, J. (1979) J. Phys.
Chem. 83, 3325.

2° Edwards, W.D. & Zemer, MC. (1986) Can. J. Chem. 63, 1763.

” Fajer, J., Davis, M.S., Brune. D.C., Spaulding, L.D., Borg, D.C. & Forman, A. (1977)
Braokhaven Symp. Biol. 28, 74.

” Fajer, J., Fujita, M.S., Davis, M.S., forrnan, A., Hanson, L.K. & Smith , KM. (1982) in
Electrochemical and Spectrochemical Studies of Biological Redox Components, ACS
Advances in Chemistry, No. 201 (Kadish, K.M., ed.) American Chemical Society,
Washington DC, 489.

” Homing, T.L., Fujita, E. & Fajer, J. (1986) J. Am. Chem. Soc. 108, 323.
3° Zysmilich, M.G. & McDermott, A.E. (1994) J. Am. Chem. Soc. 116, 8362-8363.
3’ Schenck, C.C., Gaul, D., Steffen, M., Boxer, S.G., McDowell, L., Kinnaier, C. &

Holten, D. (1990) in Reaction Centers of Photosynthetic Bacteria: F eldafing 1] (Michel-
Beyerle, M.E., ed.) pp. 229-23 8, Springer Verlag, Berlin.

110

 

” Bylina, E.J. & Youvan, DC. (1988) Proc. Natl. Acad. Sci. USA 85. 7226-7230.
33 Goldsmith, 1.0., King, B. & Boxer, so. (1996) Biochemistry. 35. 2421-2428.

3" Cui, L., Bingham, S.E., Kuhn, M., Kass, H., Lubitz. W. & Webber. A.N. (1995).
Biochemistry 34, 1549—1558.

3’ Robert, B. & Moenne-Loccoz, P. (1990) in Current Research in Photosynthesis
(Baltscheffsky, M., ed.) Vol. I, pp 65-68, Kluwer, Dordrecht, The Netherlands.

3" Plato, M., Mobius, K. & Lubitz, W. (1990) in Chlorophylls (Scheer, H.. ed) pp 1015-
1042, CRC Press, Boca Raton

37 Commoner, B., Heise, J .J . & Townsend, J. (1956) Proc. Natl. Acad. Sci. USA 42, 710.

3" Davis, I.H., Heathcote, P., MacLachlan, D.J. & Evans, M.C.W. (1993) Biochim.
Biophys. Acta 1143, 183- 189. KaB, H., Bittersmann-Weidlich, E.. Andreasson, L. -E. .,
Bonigk, B. & Lubitz, W. (1995) Chem. Phys, 194419-432. KaB, H. & Lubitz. W.
(1996) Chem. Phys. Lett, 251, 193-203.

3’ Lendzian, F., Lubitz, W., Scheer, H., Hoff, A.J., Plato, M., Trankle, E. & Mobius, K.
(1988) Chem. Phys. Lett. 148, 377-385.

Lendzian, F., Huber, M., Isaacson, R.A., Endeward, B., Plato. M., Bonigk, B., Mobius,
K., Lubitz, W. & Feher, G. (1993) Biochim. Biophys. Acta 1183, 139-160.

‘° Tang, X.-S., Diner, B.A., Larsen, A.S., Gilchrist, M.L., Lorigan, G.A. & Britt,R.D.
(1994) Proc. Natl. Acad. Sci. USA 91, 704-708.

" Rippke, R., DeRuelles, J., Waterbury, J .B., Herdman, M. & Starkier, RY. (1979).].
Gen. Microbial. 111, 1-61.

‘2 Tang, X.-S. & Diner, B.A. (1994b) Biochemistry 33, 4595-4603.

’3 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. (1989) J.
Am. Chem. Soc. 111, 8076-8083.

“ Kevan, L. & Kispert, L.D. (1976) in Electron Spin Double Resonance Spectroscopy, pp
2-57, John Wiley and Sons, New York.

’5 Blinder, SM. (1960), J. Chem. Phys 33, 748-752.

“ McCracken, J ., Shin, D.-H & Dye, J .L. (1992) Appl. Magn. Reson 3, 30.

111

 

‘7 Lin, C.P., Bowman, M.K. & Norris, J .R. (1985) J. Magn. Res. 65, 369-374.
’8 Mehring, M. & Freysoldt, F. (1980) J. Phys. E: Sci. Instrum. 13. 894-895.
‘9 Mims, W.B.(1984) J. Magn. Reson. 59, 291-306.

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

5' El-Kabbani, 0., Chang, C.-H., Tiede, D., Norris, J. & Schiffer, M. (1991) Biochemistry
30, 5361-5369.

’2 DePillis, G.D., Decatur, S.M., Barrick, D. & Boxer, S.G. (1994).]. Am. Chem. Soc.
1 I 6, 6981-6982.

’3 Lin, 8., Murchison, H.A., Nagarajan, V., Parson, W.W., Williams, J .C. & Allen, J .P.
(1994) Proc. Natl. Acad. Sci. USA 91 , 10265-10269.

5‘ Backes, G., Mino, Y., Loehr, T.M., Meyer, T.E., Cusanovich, M.A., Sweeney, W.V.,
Adman, ET. & Sanders-Loehr, J. (1991) J. Am. Chem. Soc. 113, 2055-2064.

’5 Allen, J .P., Artz, K., Lin, X., Williams, J .C., Ivancich, A., Albouy, D., Mattioli, T.A.,
Fetsch, A., Kuhn, M. & Lubitz, W. (1996) Biochemistry, 35, 6612-6619.

’° Rautter, J ., Lendzian, F., Lubitz, W., Wang, S. & Allen, J .P. (1994) Biochemistry, 33,
12077-12084.

’7 O'Malley, P.J. & Babcock, G.T. (1986) J. Am. Chem. Soc, 108, 3995-4001.
’8 Reddy, M.V.V.S., Lingam, K.V. & Gundu Rao, T.K. (1982).]. Chem. Phy. 35, 4398.
’9 Doan, P.E., Fan, C. & Hoffman, BM. (1994).]. Am. Chem. Soc. 116, 1033-1041.

°° Mouesca, J .-M., Noodleman, L., Case, D.A. & Larnotte, B. (1995) Inorg. Chem, 34,
4347-4359. '

6‘ Cammack, R., Chapman, A., McCracken, J., Cornelius, J .B., Peisach, J. & Weiner, J.H.
(1988) Biachim. Biophys: Acta 956, 307-312.

‘2 Petke, J .D., Maggiora, G.M., Shipman, L.L. & Christoffersen, RE. (1980)
Photochemistry and Phatobiolog, 31, 243-25 7.

‘3 Hanson, L.K. (1990) in Chlorophylls (Scheer, H., ed) pp 1015-1042, CRC Press, Boca
Raton

6‘ O'Malley, P. J. & Babcock, G.T. (1984) Proc. Natl. Acad. Sci. US, 81 ,1098-1 101.

Chapter 4

Multifrequency ESEEM Studies of the Primary Electron Donor in PSI:
The Electronic Structure of P,,,’

Background

Oxygenic photosynthesis in higher plants and cyanobacteria requires the interplay
of two pigment-protein reaction .centers, Photosystem l1 and Photosystem 112 (PSI and
PSII). In each of these reaction centers, the primary charge separation event proceeds via
the light-induced generation of a chlorophyll a cation radical species (P700 in PS1 and P,>80
in PSII). This photochemically active chlorophyll is distinguished from the light
harvesting and energy transfer chlorophylls by its spectral and redox properties. After
initial charge separation, subsequent electron transfer events occur on a timescale fast
enough to prevent recombination. This highly effective unidirectional electron transfer
occurs with a quantum yield approaching Lmity. Because these events are controlled by
the spatial and electronic arrangement of the cofactors involved, knowledge regarding the
structure of these radicals is imperative to understand better the electron transfer process
in photosynthetic systems.

Although significant progress has been made3, the lack of a high resolution
crystal structure in the higher plant and cyanobacterial systems has precluded
determination of the nuclearity of the primary donor in PSI. The recent 4.5 A crystal
structure by Fromme et al.3 indicates a dimeric structure for the primary donor and is
supportive of earlier hypotheses that indicate that P700, in its ground state, is a dimer of
chlorophyll a.‘ Additional evidence for a dimeric structure comes from early magnetic

resonance studies by Norris and coworkers5 who found that the EPR spectrum of P,,,,,’ is

112

113
narrower by a factor of 1N2 with respect to the monomeric chlorophyll species in vitro.

Because the EPR spectra arising from chlorophyll aggregate species have been found to
be INN narrower than the corresponding monomeric species. the observed narrowing of
the P70; EPR signal was taken to indicate that the primary donor in vivo is dimeric.S
However, because the internal spin redistributions originating from protein interactions in
chlorophyll species in viva may alter substantially the linewidth by moving spin density
away from the methyl groups onto carbons with no protons, these results are not
conclusive.

Evidence from second moment analysis of these EPR signals, however, is
supportive of a monomeric structure. Wasielewski and coworkers" have found that in
cation radicals of Pm; isotopically enriched with 2H. and 13C, the second moment is
similar to that of monomeric chlorophyll. Although this analysis correctly predicted the
nuclearity of the primary donor in the bacterial systems, the data have questionable
applicability in PS1 since this technique is generally invalid for the non-Gaussian
lineshapes observed for P700’, and, moreover, slight spin density distribution changes in
the system can alter significantly the magnitude of the second moment.

Studies of the triplet state of P7,,” have determined that the zero field splitting
parameter (D) is similar to those obtained for monomer chlorophyll a species in vitra.7
These results do not definitively identify P7,,+ as a monomer since a dimeric species with
plane parallel macrocycle rings would yield D values similar in magnitude to that of the
monomeric species.

The discrepancies in the magnetic and optical properties of P7,,+ when compared

to monomeric chlorophyll a in vitro may stem, ultimately, from interactions with the

114
protein present in vivo. These interactions have been implicated as the source of the

experimentally observed red shifts of the long absorption band of neutral P700 relative to
the monomeric chlorophyll a species.8 These shifts had previously been attributed to a
large exiton interaction between two closely spaced chlorophyll molecules.Q It has
subsequently been shown that interactions with charged amino acid groups can also effect
these types of absorption changes in chlorophyll a and bacteriochlorophyll a model
compounds."

The electronic structure of the cation radical is elusive as well and. although both
monomeric” and dimeric5 unpaired spin density distributions have been suggested for
P700’, neither has been proven conclusively and the ambiguity surrounding its electronic
structure remains. Since the three dimensional structures of the bacterial reaction centers
from Rhodapseudamonas viridis12 and Rhodabacter sphaeroides " have been determined
to high resolution by using X-ray crystallography, these systems have been used as a
model for the oxygen evolving systems of higher plants and cyanobacteria. The bacterial
reaction centers consist of a single photosystem with approximate C, symmetry. Electron
transfer proceeds unilaterally down the L branch of the protein . The primary donor (P*)
has been shown to be a dimer of bacteriochlorophyll, however, the presence of quinones
as terminal electron acceptors in the bacterial reaction centers makes this system more
analogous to PSII than PSI, which uses iron-sulfur centers as terminal electron acceptors.

The presence of an easily trapped light-induced radical species makes each of
these systems (eukaryotic and prokaryotic) especially amenable to the application of
electron magnetic resonance techniques. Electron paramagnetic resonance (EPR) and 'H

Electron Nuclear Double Resonance (ENDOR) studies of single crystals of P’ from

115
bacterial reaction centers have shown that the unpaired electron is distributed

asymmetrically over the two halves of the bacteriochlorophyll dimer. favoring the L
branch by approximately a 2:1 ratio.” This asymmetry has been attributed to structural
differences in the protein environments of the two bacteriochlorophylls. which manifest
themselves as shifts the energy of the highest filled molecular 1r orbitals of the
bacteriochlorophyll monomers. The asymmetry observed may contribute to the high
efficiency of forward electron transfer by moving the hole away from the primary
electron acceptor It has been hypothesized that this asymmetry also occurs in P,,,,,‘.

Recent 1H and "N ENDOR studies on frozen solutions of PSI have indicated
either a very highly asymmetric, or completely monomeric, unpaired electron spin
density distribution over the chlorophyll a dimer for P,,,,,*." However, in systems with
multiple anisotropic hyperfine couplings, unequivocal structural assignments from
powder pattern ENDOR spectra are difficult to make, as often, only the turning points of
the hyperfine coupling tensor are observed and correlation of these components to a
specific "N nucleus is difficult without using specific isotope labeling. Also, the relative
intensities of the ENDOR lines are not proportional to the number of coupled nuclei
contributing.

In contrast, the modulation amplitudes in electron spin echo modulation (ESEEM)
experiments are related directly to the relative populations of spin systems in the sample.
Additionally, ESEEM can be used to probe directly small hyperfine couplings usually not

resolved in conventional, continuous wave EPR, and therefore provides information

116
regarding the electronic structure of these radical species. ESEEM is also ideal for

detection of hyperfine couplings to nuclei like nitrogen with low gyromagnetic ratios.

In ESEEM, the amplitude of the spin echo induced by a series of microwave
pulses is modulated by interactions with nearby magnetic nuclei. The frequency of these
modulations can be related directly to the hyperfine and quadrupole (for nuclei with I 3 1)
interactions of the coupled nucleus. The magnitude of the hyperfine coupling is I
proportional to the fraction of unpaired electron spin residing on the coupled nucleus and
thus yields useful information regarding the radical’s electronic structure. Both one
dimensional and two-dimensional ESEEM (Hyperfine Sublevel Correlation
Spectroscopy, or HYSCORE) experiments have been used to measure the IN hyperfine
and nuclear quadrupole (NQI) interactions in Pm‘." However, spectral simulations of
these ESEEM data have not been successful in determining these parameters
conclusively. This precludes quantization of the number of "N nuclei giving rise to a
particular hyperfine interaction. Here, we report results of multifrequency ESEEM
experiments and corresponding spectral simulations of experimental ESEEM data of P700‘
containing either natural abundance I‘N or isotopically enriched with "N. The
simulations indicate that the three-pulse ESEEM spectra arise from four nitrogens
coupled to the unpaired electron, the hyperfine couplings of these nitrogens are indicative

of a monomeric unpaired electron spin distribution.

117

Materials and Methods
Growth of Cells

Cells of Synechacystis PCC 6803 were grown under photoautotrophic conditions.
with aeration, in BG-l 1. In order to label cells uniformly with "N, KNO3 (18 mM) in the
growth medium was replaced by 18 mM K"NO, (purity >98 %,Cambridge Isotope
Laboratories, Andover, MA). Amounts of IN arising from other nitrogen containing
compounds in the medium accounted for less than 0.002 % of the total nitrogen. When
cell growth had reached the late log phase (OD. 600 >1.2), the cells were harvested by
continuous-flow centrifugation at 4° C, washed twice and resuspended in BG-ll.
Glycerol was added to the cell suspension to 15 % (v/v), the suspensions were mixed

thoroughly and stored at -70° C.
Preparation of Photosystem I Particles

Thylakoid membrane preparation and purification of the photosystems was
performed essentially as described in Noren et al. '° Frozen cells, equivalent to about 20 -
25 mg Chl, were thawed quickly and diluted with break buffer into a 250 mL centrifuge
bottle. Cells were pelleted by centrifugation at 10,000 x g for 10 min, resuspended in a
small volume of break buffer and transferred to the bead-beater chamber. Solubilization

of thylakoid membranes and anion-exchange chromatography were carried out as

1 18
described" except that fractions from the Q-Sepharose Fast Flow column which were

enriched in PS1 (AM > 679 nm) were pooled and precipitated by the addition of an equal
volume of 30 % (w/v) PEG- 8000 in 50 mM MES, pH 6.0. After mixing. the preparation
was incubated ovemight in the refrigerator and the resulting precipitate was recovered by
centrifugation at 20,000 x g for 10 min. The pellet was redissolved in Mono-Q start
buffer and the solution was chromatographed on the Mono-Q column. Fractions enriched
in PS1 were pooled and an equal volume of 30 % PEG-8000 was added followed by
centrifugation at 40,000 x g for 30 min. The supernatant was removed completely and
the pellet was resuspended in 50 mM HEPES pH 7.5, 15 mM NaCl,, 0.05 % dodecyl
maltoside) to a final concentration of > 3 mg Chl/mL. The sample was then transferred to
a 4 mm O.D.quartz EPR tube and stored in liquid nitrogen until further use. Pm,“ was
generated by illumination at 273 K for 30 seconds. The radical was trapped by freezing
immediately to 77 K. The radical’s presence, in both natural abundance l”N and

isotopically enriched "N samples, was confirmed by cw-EPR (see Figure 4-1).

Spectroscopy and Spectral Simulations

Continuous Wave EPR: The cw-EPR data were obtained (10 K) at X-band on a Bruker
ER200D EPR spectrometer outfitted with a Bruker TE102 EPR cavity and an Oxford
ESR-900 continuous flow cryostat. Microwave frequency was determined by using an
EIP Microwave Model 258 frequency counter and the static magnetic field strength was

measured with a Bruker ER03 5N NMR gaussmeter.

119
ENDOR spectroscopy and data analysis: The ENDOR data were obtained at X-band on

a Bruker ESP300e spectrometer with a Bruker ESP360 DS ENDOR accessory and home-
built demountable coils.l7 All ENDOR data were collected at a field position
corresponding to the center of the EPR absorption line. Constant temperature in the
cavity was maintained with an Oxford ESR900 continuous flow cryostat. Microwave
frequency was determined by using an EIP Microwave Model 25B frequency counter and
the static magnetic field strength was measured with a Bruker ER035M NMR gaussmeter
The peaks attributed to "N (I=1/2) electron nuclear hyperfine coupling are
expected to be split symmetrically about the nuclear Larmor frequency for "N (v ,S,, =
1.45 MHz at 3500 G). In solution ENDOR, the position of these peaks is indicative of
the strength of the isotropic hyperfine coupling according to the relationship
v, = v.4: A,./2
where vN 2 A,,o/2. However, in frozen solution, the anisotropy of the hyperfine coupling
gives rise to powder lineshapes described to first order by
v: = vN -_l-_ A/2
where A = A“ c0520 + A, sin2 0
AH and A1 in the above expression are the principal values of the axially symmetric

hyperfine tensor and 0 describes the orientation of the principal axis system of the

hyperfrne tensor with respect to the laboratory magnetic field.

120

 

 

 

 

 

 

 

2500 , 1
P700 all 15N

2000 ..
o
'0
3 -t
E 1500 -
o.
E
<

P700 14N his
1000 - _
500 I I I
3300 3350 3400. 3450 3500

Magnetic Field (Gauss)

Figure 4-1: EPR spectra of Pm’ containing natural abundance l‘N (top) and isotopically
enriched with "N. Spectrometer conditions: microwave power, 20 dB; center field, 3380
G; modulation amplitude, 20 Gpp; modulation frequency, 100 kHz; time constant, 200

ms; sweep time, 200 5; sample temperature, 8 K.

121
ESEEM data collection and analysis: ESEEM data were collected on a home built

spectrometer" by using a reflection cavity where either folded stripline" or slotted-tube
structures” served as the resonant element. A three pulse or stimulated echo (90°-r-90°-
T-90°) pulse sequence was used. Dead time reconstruction was performed prior to
Fourier transformation as described.” Computer simulations of the ESEEM data were
performed on a Sun Sparcstation 2 computer utilizing FORTRAN software based on the
density matrix formalism of Mims.” The analysis software for the treatment of
experimental and simulated ESEEM data was written with Matlab (Mathworks. Nantick.
MA). The experimental dead time was included in the simulations. An isotropic g-tensor
was assumed in all of the calculations. The success of the spectral simulations was based
on the modulation depth and duration observed in the time domain traces as well as the

lineshapes, peak positions and relative peak intensities in the frequency spectra.

Results

The continuous wave EPR spectra from P,Q,+ containing natural abundance l"N (top) and
isotopically enriched with "N (bottom) are shown in Figure 4-1. The spectra are
indistinguishable from one another; both exhibit the characteristic lineshape of a
chlorophyll 1t cation radical species. The spectra consist of a single peak centered at
=2.0025 with a linewidth of 7-7.5 G with no observable fine structure. ESEEM
spectroscopy was used to resolve the small hyperfine couplings not observed in this

inhomogeneously broadened line. The data were collected at magnetic field positions

122

 

    
  
   
   
      
 

 

 

 

" . 1 1 fl
3997091992199 ..................... _
€9,725 GHz,3480G

3
5
Q
'0
.13 .. ........................................ , ........................ ..
% 311.488 GHz,41§1OG
or 2 I
o . .
.C
0
G
E14335 GHz,51§30t3
......;..................j ................... i..................l .................. i ...... —:
o 0.5 1 15 2
time(nsec) x 104

Figure 4-2: Multifrequency stimulated echo ESEEM time domain decay traces from
P,,,,,’ containing natural abundance ”N. Spectrometer conditions: magnetic field strength,
as noted; tau value, A) 250 ns, B) 162 ns, C) 300 ns, D) 400 ns; microwave pulse power,
A) 45 dBm, B) 49 dBm, C) 55 dBm, D) 60 dBm; pulse repetition rate, A) 20 Hz, B) 60
Hz, C) 30 Hz, D) 60 Hz; pulse width (FWHM), 22 ns; sample temperature, 4.2 K.

 

amplitude (a.u.)

14.335 GHZ, 5130 G

 

 

 

 

6
frequency (MHz)

Figure 4-3: Cosine Fourier transformations of three-pulse ESEEM time domain traces at
multiple microwave frequencies. Spectrometer conditions: as in Figure 4-2.

124
corresponding to the zero-field crossing on the EPR spectrum. Typical stimulated echo

ESEEM patterns for P700’ containing natural abundance 14N and their corresponding
Fourier transforms are shown in Figures 4-2 and 4-3, respectively.

The spectra in Figure 4-3 are dominated by the quadrupole coupling of the IN
nucleus (1 = 1). If the hyperfine interaction is approximately equal to twice the nuclear
Zeeman fi'equency, then the energy level splittings in the spin manifold where these two
cancel is determined primarily by the I"N nuclear quadrupole interaction (see Chapter 2
for a detailed description of this interaction). To determine the field position at which
this exact cancellation condition is met, it is essential that the experiment be performed at
multiple microwave frequencies. The exact cancellation ESEEM pattern is characterized
by modulations that are deep and long-lived, while the corresponding Fourier transformed
spectrum exhibits three sharp, low frequency features (two of which add up to the third);
these correspond to the zero field nuclear quadrupole transitions (V: and v0). If the exact
cancellation condition is met, then the quadrupole parameters, the quadrupole coupling
constant equ and the asymmetry parameter, n, which measure the magnitude and the
symmetry of the electric field gradient tensor, may be calculated directly from the
experimental ESEEM spectrum where

v : = 3/4 equ (1 : n/3)
v 0 = 10 equ n.
The presence of a broad AM, = 2 or “double quantum” transition arising from the spin
manifold where the nuclear Zeeman term is doubled by the hyperfine coupling, observed

at approximately four times the nuclear Zeeman frequency is also indicative of this exact

125
cancellation condition, although the lineshape of this double quantum peak depends

heavily on the amount of state mixing that is induced by the hyperfine coupling term of
the Hamiltonian and also by the relative orientation of the NQI tensor with the PAS. The
approximate peak position of the double quantum transition as well as the isotropic

hyperfine coupling can be determined, in the exact cancellation limit, by using

2
vDQ=2\/(vnilAy2) +82 where §=K,/3+n2

However, since both the isotropic and anisotropic part of the hyperfine coupling

 

contribute to the A value in this expression, a large amount of dipolar coupling can make
it difficult to determine A,” directly from the position of the double quantum peak.

Figure 4-3B shows qualitatively that this exact cancellation condition is realized
for at least two sets of nitrogen nuclei at a magnetic field strength of 3480 G. This
spectrum contains five major low frequency components at 1.04, 1.15, 1.47, 1.60 and
2.60 MHz. From the positions of these peaks and by using the equations above, the
nuclear quadrupole parameters can be determined for two sets of nitrogens; set A) equ =
2.70 MHz, n = 0.85 and set B) equ = 2.81 and n = 0.73.

Because the hyperfine contributions to the spectra in Figure 4-3 are obscured by
the dominant quadrupole interaction, the reaction centers were isotopically enriched with
"N (I = 1/2). The "N nucleus has no quadrupole moment and, therefore, the stimulated
echo modulation pattern is determined primarily by the hyperfine interaction. The
multifrequency stimulated echo ESEEM pattern and Fourier transforms of Pm” enriched
with "N are shown in Figures 4-4 and 4-5, respectively. The modulations present in

Figure 4-4 arise as a result of hyperfine coupling between the unpaired 1t electron and

126

 

 

 

 

3167 G
E?
3 4870 G
O
'0
a
.73..
E
<
O
.C
0
Lu
5695 G
0 5000 10000 15000
Time (ns)

Figure 4-4: Multifrequency stimulated echo ESEEM time domain decay traces from
P7,; isotopically enriched with "N. Spectrometer conditions: magnetic field strength, as
noted; tau value, A) 222 ns, B) 318 ns, C) 400 ns; microwave pulse power, A) 45 dBm,
B) 45 dBm, C) 57 dBm; pulse repetition rate, A) 90 Hz, B) 30 Hz, C) 30 Hz; pulse width
(FWHM), 22 ns; sample temperature, 4.2K.

127

 

 

 

 

 

 

 

 

 

3167 G

37 WT 7‘ w A
3
o
'0
:3: 4870 e
E

5695 G

0 2 4 6 8 10
Frequency (MHz)

Figure 4-5: Cosine Fourier transformations of three pulse ESEEM time domain traces at
multiple microwave frequencies. Spectrometer conditions: as in Figure 4-4.

128
nitrogen nuclei in the chlorophyll macrocycle. The increase in modulation depth evident

at 4280 G (Figure 4-4, middle) is indicative of a situation where the hyperfine interaction
is equal to twice the nuclear Zeeman frequency. At this “exact cancellation” like
condition the modulation depth parameter, k (see Chapter 2) is maximized. The Fourier
transform of this time domain decay trace (Figure 4-5, middle), unfortunately only
exhibits two features: a large, sharp feature at 1 MHz and a broader feature near 4 MHz.
The lack of spectral resolution in this, and all of the spectra from the "N enriched
samples, precludes definitive determination of the hyperfine coupling constants and
number of nuclei giving rise to the observed peaks.

Assignments may be achieved, however, by performing spectral simulations of
the experimental data. The success of these simulations depends on the validity of both
the magnitude of the nitrogen hyperfine and quadrupole coupling constants as well as the
number of nuclei giving rise to a particular coupling. To make unambiguous assignments
of nitrogen hyperfine coupling constants for P7007, the nitrogen hyperfine coupling tensors
and nuclear quadrupole parameters determined must simulate successfully all of the
ESEEM spectra in Figures 4-3 and 4-5.

Because the observed modulations may arise from as many as eight interacting
nitrogen nuclei, with each nitrogen requiring eight independent simulation parameters,
determination of the simulation parameters is severely underdeterrnined. To assist in the
assignment of ESEEM features and provide constraints for the simulations, ENDOR
experiments were performed on the "N enriched Pm,+ sample, see Figure 4-6. Features in

the low (1-5 MHz) frequency region of this spectrum are expected to arise from "N

 

 

ENDOR Intensity

 

 

 

1..

l i

1 2 3 4 5 6
Frequency (MHz)

 

 

Figure 4-6: ENDORspectrum collected at 6 K for P,Q,+ globally labeled with "N.
Spectrometer conditions: microwave power, 1.99 mW; magnetic field strength, 3375 G;
RF power, 200 W; RF frequency modulation: 100 kHz.

130
nuclei coupled to the it electron in P7007. Peaks at 1.12 and 1.78 MHz which are split

symmetrically about the nuclear Larmor frequency for "N (1.46 MHz at 3375 G). have
been assigned to the magnetic coupling between Pm’ and an axial histidine ligand (see
Chapter 3).

The features in the 2-3 MHz region are assigned to hyperfine interactions between
P,Q,+ and the pyrrole nitrogens in the chlorophyll macrocycle. This assignment is based
on the fact that these peaks disappear when the sample is specifically labeled with "N
histidine (see Figure 3-3, middle); in this sample, all other nitrogens, except those in
histidine, are 1‘N and, because of their large quadrupole moments, do not contribute to the '
ENDOR spectrum. The peaks in the 2-3 MHz region represent the perpendicular
components of the hyperfine coupling in one spin manifold. The corresponding
transitions fiom the other manifold, expected to be split symmetrically about v“ for "N.
occur at frequencies near zero and, therefore, are not resolved in this experiment. It is
possible, however, to make some predictions on the strength of the hyperfine coupling
based on the peak positions and lineshapes. For example, if the derivative shaped feature
at 2.35 MHz, can be assigned to the perpendicular singularity of a nitrogen hyperfine
tensor, then a corresponding A, value of 1.78 MHz is obtained. Because the number of
orientations that contribute to the perpendicular singularity is larger than the number
contributing to the parallel component, resolution of the corresponding An featureis often
difficult with ENDOR. '

The ESEEM data can be used, in this case, to provide some insight into the

magnitude of the A” coupling expected. At exact cancellation, the hyperfine coupling is

131
equal to twice the nuclear Larmor frequency. Figure 4—2 shows that this condition is

realized at a magnetic field of 3480 G, where the nuclear Larmor frequency is 1.07 MHz.
The hyperfine coupling corresponding to this is 2.14 MHZ. which. when scaled to both
the "N nucleus ("N/"N = 1.403) and to the magnetic field strength at which the
experiment was performed, yields an isotropic coupling of 2.90 MHz. Model chlorophyll
and bacteriochlorophyll compounds have shown that the hyperfine tensor for the pyrrole

nitrogens is of axial symmetry”, therefore. by using the relations

A91 =Aiso +2f
A1 = Aiso "' f
and

A
vobserved = v1 :3

 

 

where f is the dipolar contribution to the hyperfine coupling, and v, is the nuclear Larmor
frequency, the peak positions of the Au singularity can be determined. An isotropic
hyperfine coupling of 2.90 MHz would therefore yield a peak in the 4 MHz region of the
spectrum. There is a small absorptive feature in this region, which is assigned to the All
turning point of one nitrogen hyperfine coupling.

The improved resolution of this ENDOR spectrum can be used to assist in
constraining the parameters needed to perform the numerical simulations of the ESEEM
data. Three additional A, features can be measured from the positions of the peaks in the
ENDOR spectrum. These couplings, with magnitudes of 2.14, 2.30 and 2.42 MHz,
scaled to the IN nucleus and included with the two sets of quadrupole parameters

determined from Figure 4-3B provide a starting point for the simulations.

132
The lack of features in the double quantum region of the ESEEM Spectrum also

provides insight into the magnitude of the dipolar contribution to the total nitrogen
hyperfine coupling. Increased state mixing induced by the presence of significant
anisotropy in the hyperfine term of the spin Hamiltonian affects the transition
probabilities in the spin manifold from which the double quantum feature originates.
Because the energy levels are quantum mechanical mixtures of eigenstates. all transitions
have a non-zero probability of occurring and because they all occur, for ”N, near 4 MHz.
can result in a loss of resolution in this region. This effect is exhibited in Figure 4-7.
Each calculated ESEEM spectrum in this figure has the same isotropic hyperfine
coupling; the dipolar contribution to the total hyperfine coupling in each case. however. is
different. The effect on the features in the double quantum region as the anisotropy
increases is pronounced as is the effect on the linewidth of the low frequency
components. Because the double quantum peaks are missing from the ESEEM spectrum
collected at 3480 G, a considerable amount of anisotropy must be present in the hyperfine
coupling of at least one nitrogen.

Further insight into the form of the hyperfine and quadrupole tensor can be gained
by examination of the ESEEM spectrum collected at 4110 G (see Figures 4-2C and 4-
3C). The depth and duration of the modulations observed in the time domain trace are
indicative of an exact cancellation condition for at least one nitrogen nucleus. The
presence of only two sharp, intense low fi'equency peaks in the corresponding Fourier
transform, however, does not confirm this hypothesis. The presence of two, rather than

three, peaks can be explained by examining the asymmetry parameter, n. This parameter

133

 

ro to
o o
l

amplitude
25

 

O

 

 

 

L
o
o

 

 

N
O

..a
O

 

 

amplitude
o

 

 

 

10 12

I
d
O

O

2 4 6

 

_a
O

01

 

 

amplitude
O

 

 

 

('11
l
L

12

O
N
.5
on
..a
0

frequency (MHz)

Figure 4-7: The effect of hyperfine anisotropy on the ESEEM spectrum. Simulation
conditions for all spectra unless otherwise noted: A ,: 1.84 MHz (top), 1.64 MHz
(middle), 1.44 MHz (bottom); A": 2.17 MHz (top), 2.56 MHz (middle), 2.97 MHz
(bottom); A,,,: 1.95 MHz; equ: 2.69 MHz; 1]: 0.87; magnetic field strength: 3480 G; tau
value: 162 ns; Euler angles (a, B, y): 0,0,0.

134

 

 

 

amplitude

 

 

 

I

—L

O
O

 

N
O

..L
o

 

 

amplitude
O

 

 

 

o 2 4 6 8 10 12

 

 

amplitude

 

 

 

 

o 2 4 6 8 1 0 12
frequency (MHz)

Figure 4-8: The effect of the asymmetry parameter on the ESEEM spectrum. Simulation
conditions for all spectra unless otherwise noted: A ,2 1.64 MHz; A": 2.56 MHz; A,,,: 1.95
MHz; equ: 2.69 MHz; '1]: 0.2 (top), 0.4 (middle), 0.8 (bottom); magnetic field strength:
3480 G; tau value: 162 ns; Euler angles (or, B, y): 0,0,0.

135
describes the deviation from axial symmetry of the electric field gradient originating from

the electrons in the p orbitals. An asymmetry parameter of 1 describes a system with
rhombic symmetry, while a value of zero for n is indicative of axial symmetry. The
magnitude of this parameter affects the position of the low frequency peaks in the
ESEEM spectrum. These effects are shown in Figure 4-8. When n is small (Figure 4-8.
top) the two lowest frequency peaks coalesce into a single feature. This is most likely the
situation observed for the ESEEM spectnnn collected at 4110 G and information
regarding the magnitude of this asymmetry parameter can be utilized in the parameter set

for at least one nitrogen.

Table 4-1: “N ESEEM simulation parameters.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A1 A" Also equ TI (1 B Y

(MHz) (MHz) (MHz) (MHz)
Nl 1.27 3.66 2.06 3.1 0.20 0 0 0
Nn 1.55 3.29 2.13 2.69 0.87 0 90 0
Nm 1.64 2.56 1.95 2.69 0.87 0 90 0
NW 1.73 3.10 2.19 2.81 0.70 0 0 0

 

The sharp features present in the spectrum collected at 4110 G were also found to be
dependent upon the relative orientation of the hyperfine and quadrupole axis systems. By
performing iterative simulations utilizing various angle combinations and subsequent

qualitative analysis of the spectral features as a function of the Euler angles describing

 

136

 

 

 

 

Experimental
3"
a;
0
8
E
a.
E
<
Simulation
0 2 4 6 8 1 0 1 2
Frequency (MHz)

Figure 4-9: Experimental (top) stimulated echo ESEEM spectrum from P70; containing
natural abundance 1‘N and corresponding numerical simulation (bottom). Experimental
conditions: as in Figure 4-3B. Simulation parameters: As in Table 4-1; additional
parameters as in experimental.

 

 

 

 

 

 

3"

9;

§ Experimental

.1:

'5.

E

<

Simulation
0 2 4 6 8 1 0 1 2
Frequency (MHz)

Figure 4-10: Experimental (top) stimulated echo ESEEM spectrum from Pm’
isotopically enriched with IN and corresponding numerical simulation (bottom).
Experimental conditions: as in Figure 4-SB. Simulation parameters: As in Table 4-2

additional parameters as in experimental.

138
this orientation, it was found that this lineshape was present only when the two axes were

collinear.

By utilizing these experimental constraints, successful numerical simulations of
the ESEEM data for both nuclei at multiple microwave frequencies was possible. Figures
4-9 and 4-10 present the simulations and experimental data for P700- containing natural
abundance l‘N (collected at 3480 G) and isotopically enriched with l5N (collected at 3167
G). The parameter sets for these simulations are tabulated (see Tables 4-1 and 4-2).

There is very good agreement between these simulations and the experimental spectra.

Table 4-2: "N ESEEM simulation parameters.

 

 

 

 

 

 

A; An Am r
(MHZ) (MHZ) (MHz) (A)
Nl 1.78 5.13 2.39 1.93
N" 2.17 4.62 2.99 2.13
N", 2.30 2.67 2.73 2.65
va 2.43 4.35 3.07 2.32

 

 

 

 

 

 

The peaks in the experimental spectrum from l5N enriched P700+ that are not reproduced in
the simulation are the result of residual l‘N contamination. Mass spectrometry of the 15N
enriched P70; sample confirms the presence of this "N (see Figure 4-11). Because the

modulations originating from the "N quadrupolar interactions are so strong, even residual

139

1254 r!

50-

    

 

 

865.52‘

 

854358

 

913.78
33
850 855 853 865 87B 875 888 885 890 895 988 905 918 915 920

m/z

Figure 4-11: FAB-MS data on chlorophyll a extracted from PSI isotopically enriched
with ”N. Peaks at 896.65 and 873.68 indicate the presence of residual MN.

amounts of the nucleus can be observed in the ESEEM spectrum and, moreover, the
positions of these features in the "N spectrum corresponds to the observed frequencies of
the peaks in the I‘N spectrum (Figure 4-3B). In the absence of the mass spec data, these
features, which appear in the same low frequency region of the spectrum as those arising
from l5N, may have been erroneously assigned to lsN resonances.

Multifrequency simulations utilizing the parameter sets in Table 4-1 are shown in
Figure 4-12. The agreement between these calculated spectra and the experimental data
shown in Figure 4-3 is consistent at every field value. These multifrequency experiments

and corresponding simulations are necessary for determination of the complete hyperfine

and quadrupole parameter sets that contribute to the ESEEM modulation pattern.

 

140

 

 

 

 

I I I I I
3210 G
34806
ET
El
0
'0
.3
3 41106
<
5130G
I I I I I
o 2 4 s 8 1o 12

Frequency (MHz)

Figure 4-12: Numerical simulations of multifrequency ESEEM spectra from P70;
containing natural abundance “N. Simulation parameters: as in Table 4-1; all other
parameters as in experiment (see Figure 4-3A-D).

1 41
Discussion

The combination of isotopic substitution, ENDOR, multifrequency ESEEM and
numerical simulations of ESEEM data utilized in this study presents a methodology that
can be used to determine definitively the magnetic properties of radicals in protein
systems. Although early ESEEM and ENDOR investigations of chlorophyll a model
compounds and P700 containing natural abundance ”N or isotopically enriched with ISN
were able to identify the limits for these properties", the combination of these
spectroscopies with isotopic substitution was not enough to determine definitively the
hyperfine and quadrupole tensors for all of the nitrogens contributing to the spectra
observed and currently, the ambiguity surrounding the electronic structure of the primary
donor remains.

The utility of ENDOR spectroscopy as a method to study hyperfine couplings
with little or no anisotropy has been previously demonstrated (see Chapter 3). It’s
applicability to systems containing large isotropic proton couplings has been
demonstrated extensively in the biophysics field.25 However, in systems characterized by
small hyperfine couplings with a large dipolar contribution and containing nuclei with
low gyromagnetic ratios (e.g. "N or 15N), useful spectral information is lost because the
low frequency transitions occur near zero and are not resolved with most commercial
spectrometers. Likewise, the turning points of the broad powder pattern lineshape that
results from a large dipolar interaction can be difficult to resolve experimentally, as only
the large perpendicular component is detected. Additionally, because of second order

effects, the detection of couplings to quadrupolar nuclei is difficult and results in line

142
broadening that decreases resolution. Again, isotopic substitution with 15N can be utilized

to remove the quadrupolar contribution, however the low frequency resolution problem
remains as the gyromagnetic ratio of 15N differs only slightly from that of MN.

ESEEM spectroscopy circumvents these problems. The exact cancellation
condition, usually realized for nitrogen nuclei near 3300 G, allows determination of the
pure quadrupole resonances while the hyperfine interactions can be discerned by using
isotopic substitutution with 15N, which eliminates the quadrupole contribution.
Furthermore, the physical requirement for bandwidth sufficient to excite all transitions
simultaneously means that detection of small couplings to nuclei with low gyromagnetic
ratios is possible. This technique is also ideal for measuring hyperfine couplings
containing a large dipolar contribution, since the nuclear modulation effect depends on
the quantum mechanical mixing of states that occurs as a result of the anisotropic
interaction. Moreover, since all coupled nuclei contribute, in a multiplicative manner, to
the modulation pattern observed in an ESEEM experiment, quantitation of the number of
interacting species based on peak intensities is possible. The intensities of the peaks
observed in ENDOR, because they are governed by competing relaxation times, yield no
qualitative information about the number of species contributing.

Analysis of the peaks observed in the ESEEM spectrum, however, can be difficult
as the number of combination peaks arising from multiple nuclei precludes the definitive
assignment of these features. This is especially apparent in the primary donor systems of
both plants and bacteria. In these systems, nitrogen couplings are expected from a

minimum of 4 (monomer) and a maximum of 8 (dimer) pyrrole nitrogens. Additional

couplings from axial nitrogenous ligands are possible as well, although couplings from

 

143
the axial histidine ligand to P100- have been shown not to contribute to the ESEEM

modulation pattern (see Chapter 3). The presence of multiple interacting nuclei with
similar magnetic properties results in the presence of multiple combination lines in the
ESEEM spectrum. Moreover, these peaks appear in the 1-5 MHz region of the spectrum.
Definitive assignment of peaks to specific nitrogen nuclei is therefore impossible without
reliable spectral simulations.

Numerical simulations of experimental ESEEM data can be misleading as well
since the number of independent parameters required for a system like P7001 which
contains multiple quadrupolar nuclei, severely underdeterrnines the calculation. To
constrain the calculation, information regarding the magnitude of these parameters can be
obtained from study of model compounds, examination of the literature or from
theoretical calculations modeling the species in vivo and in vitro. Constraints can also be
provided by experimental-results from other spectroscopic methods or by using isotopic
substitution to eliminate specific contributions.

A combination of these methods has been used to constrain the parameter sets in
this study. Perhaps one of the most useful of these constraints was the ability to obtain
results from ESEEM experiments performed at multiple microwave frequencies. Since
the peak position will change as a function of the Larmor frequency of the nucleus from
which the resonance arises, which is dependent on the external magnetic field strength,
this method can be used to make definitive assignments.

Multiple frequency ESEEM can also be used to find the exact cancellation
condition for systems with quadrupolar nuclei. Because the Larmor frequency is a

function of the applied magnetic field strength, an exact match between the hyperfine

144
coupling, which is constant for the system, can only be determined by changing the

magnitude of the external magnetic field. In the case of P700‘, the ability to perform
multiple frequency experiments proved to be extremely advantageous since an additional
set of quadrupole parameters was discovered at 11 GHz that was not observed at the

lower microwave frequecies.

Analysis and Interpretation of Nuclear Quadrupole Interactions in Pm“

Because the peaks observed in the exact cancellation ESEEM spectrum from P700,
represent the zero field transitions of "N nuclei in the chlorophyll, the nuclear quadrupole
coupling constant, equ and the asymmetry parameter, 1‘], can be determined directly
from the experimental spectrum. The quadrupole coupling constants and asymmetry
parameters for the pyrrole nitrogens in P70; as determined from multifrequency ESEEM
experiments are listed in Table 4-1. The magnitudes of the quadrupole coupling
constants determined for P700‘ are consistent with those determined for chlorophyll and
bacteriochlorophyll species studied in vivo and in vitro (see Table 4-3). The values
determined for the asymmetry parameters for three of the pyrrole nitrogens are somewhat
larger than those found for Chl ai. Because this parameter can be dramatically affected
by small fluctuations in the orbital occupancies, local electron density distribution
differences due to interactions with the protein may significantly alter this parameter.26
The asymmetry parameter for NI (see Table 4-1) is consistent with the value obtained for
Chl a‘, indicating that the environment of this nitrogen differs somewhat fiom the other

three. This hypothesis, and possible explanations for this will be discussed later in this

145

Table 4—3: Quadrupole parameters for chlorophyll and bacteriochlorophyll species
in viva and in vitra.

 

 

 

 

 

 

 

 

 

 

Species equ (MHz) Tl
Pm“ 27 2.73 0.66
p960+27 2.90 0.70
pm+24 1.68 0.68

2.72 0.66

Chl a“ 2.68 0.60

3.20 0.50
BChl ami 2.72 0.70
2.80 0.79

 

 

 

 

chapter. For the most part, though, the values for 11 obtained are mostly consistent with
those previously measured for P700‘ Because equ and 11 describe the interactions of the
electric quadrupole moment of the nucleus with local electric field gradients, knowledge
regarding the magnitudes of these parameters is desirable and can yield information
regarding the local electric field gradients due to interactions with charged amino acid
residues as well as the charge concentration present in a covalent bond.

The model of Townes and Dailey,29 which assumes that the electric field gradient
at the nucleus is due only to electrons in the valence 2p orbitals, is a convenient approach
for the interpretation of these parameters. The axis system in this model assumes that the
Z axis lies along the largest component of the electric field gradient tensor. In a pyrrole
compound, the molecular symmetry ensures that one field-gradient axis lies along the NH
bond, one is perpendicular to the molecular plane and the third, in the molecular plane, is
perpendicular to the other two.30 The principle axis is almost always perpendicular to the

molecular plane since the p, orbital contains the lone-pair of electrons. In contrast, the Z

146

Figure 4-13: Quadrupole tensor axis systems for a) imino nitrogens in imidazole as compared to
pyrrole (b) nitrogens.

axis for the imino nitrogen of imidazole has been found to lie in the plane and is only 4°
off of the geometric z axis, defined as the bisector of the C-N-C angle (see Figure 4-13).“

Because the relative orientations of the nuclear quadrupole and hyperfine tensors
can affect the lineshapes of the observed resonances, an Euler angle rotation relating the
two coordinate systems was included in the simulations. The Z direction of the hyperfine
tensor is usually taken as parallel to the p orbital containing the unpaired electron. For
two of the pyrrole nitrogens in P70; the hyperfine and quadrupole axes were found to be
collinear. The axes of the hyperfine and quadrupole tensors for the other two nitrogens
were found to be related by an Euler angle rotation of B=90° (see Euler angles, Table 4-
1). This rotation is, most likely, the result of the asymmetrical distribution of the
unpaired electron in the it molecular orbital and will be discussed in the following

section.

147

Analysis and Interpretation of Hyperfine Interactions in P700+

A large body of work, both experimental and theoretical, has been dedicated to
the determination of the magnitude of the hyperfine couplings for the pyrrole nitrogens in
chlorophyll compounds both in viva and in vitra.32 Much of the early spectroscopic work
was influenced by the single crystal magnetic resonance experiments performed on the
bacterial reaction center. In this system, the primary electron donor consists of a dimer of
bacteriochlorophyll a with the unpaired electron delocalized asymmetrically over the two
halves of the dimer in a 2:1 ratio.“ The electronic structure has been elucidated based on
extensive lH ENDOR studies utilizing the available single crystals. The asymmetrical
spin density distribution motif observed in these systems has been proposed for the
primary donor in PSI.

A recent spectroscopic investigation of ”N enriched Pm‘ has estimated that the
unpaired electron is 90% delocalized bn a single chlorophyll a macrocycle; the other 10%
of the spin is distributed over the other half of the geometric dimer.” This conclusion is
based on the observation of small features in the ESEEM spectrum of ”N labeled P700+
that previously were not detected. These features correspond to a hyperfine coupling of
0.37 MHz and have been assigned to hyperfine couplings from the nitrogens on the other
(10% spin) dimer half. Subsequent ENDOR and HYSCORE experiments indicated the
presence of couplings of this magnitude. However, the presence of residual I‘N in these

”N enriched samples, which has been shown to contribute to the ESEEM spectrum even

at small concentrations, was not ruled out. The observed peaks, therefore, may result

148
from 14N contamination and not correspond to a hyperfine coupling to the 15N pyrrole

nitrogens on the other half of the dimer.

Single crystal ESEEM studies by the same laboratory”. which detected five sets
of quadrupole couplings in Pm’ seem to corroborate these results, however. single crystal
ESEEM studies are complicated by the presence of multiple combination peaks and
unless couplings are confirmed by simulations and multiple microwave frequency
experiments, do not provide definitive proof of these parameters.

The hyperfine and quadrupole coupling parameters listed in Tables 4-1 and 4-2.
however, are the result of an exhaustive and comprehensive study incorporating samples
containing either natural abundance l4N or isotopically enriched with ”N, followed by
multifrequency ESEEM and ”N ENDOR. The couplings extracted from these
spectroscopic studies were confirmed by performing numerical simulations of the
ESEEM data. Constraints for these calculations were provided by the multiple magnetic
resonance experiments performed and, because simulations utilizing the same parameter
set reproduced the experimental multifrequency ESEEM data for bath isotopes, these data
represent the first definitive assignments of hyperfine and quadrupole coupling constants
to the pyrrole nitrogens in Pm’.

The A 1 elements of the hyperfine tensors of P70; determined from ENDOR and
the corresponding values concluded from simulations of the multifrequency three-pulse
ESEEM spectra presented here are identical. The All components of these tensors were
not resolved in the ENDOR and subsequently, the values utilized in the calculations were

determined primarily through qualitative analysis of the numerical simulations. Limits

149
for these couplings. though. were provided through examination of the multifrequency

stimulated echo ESEEM spectra. Isotropic hyperfine coupling constants. determined
from exact cancellation stimulated echo ESEEM spectra, helped to place these limits.
The use of axial tensors to model the hyperfine interaction of these nuclei is warranted
since previous magnetic resonance studies of chlor0phyll and bacteriochlorophyll model
systems have yielded nitrogen hyperfine couplings of axial symmetry.23 Furthermore.
theoretical molecular orbital calculations with the RHF-INDO/SP (Restricted Hartree-
Fock, Intermediate Neglect of Differential Overlap-Spin Polarization) method on
chlorophyll a cation radicals yielded tensors varying only slightly from axial symmetry
(see Table 44).” Corresponding calculations on the radical in viva have not been
presented in the literature. For comparison, the tensor components for P70; derived from
experiment are included in this table.

These hyperfine couplings can be used to quantify the amount of spin density in
the pp orbital of the individual nitrogen nuclei and therefore provide a probe into the
electronic structure of the radical. A full characterization of this structure, however,
should include spin density calculations for each nucleus in the species. Multiple 'H
ENDOR experiments” on frozen solutions of Pm" have yielded coupling constants for
various positions in the heterocycle. Based on the magnitude of these couplings and their
corresponding spin densities, O’Malley and Babcockll proposed an electronic structure
model for the radical whereby the unpaired electron is delocalized over a single
macrocycle. The decreased proton hyperfine coupling constants and corresponding
carbon spin densities observed for Pm“ in viva relative to those of the chlorophyll a

monomer model were attributed to differences in the composition of the ground state

1 50
Table 44: Experimental and theoretical hyperfine coupling tensor components

 

 

 

 

 

 

 

 

 

 

 

for Pm,+ and Chl a‘
NI Nll NIll NIV
RIIF-INDO/SP*
(Chl a“)
AH 1.90 1.11 0.96 0.88
A22 1.97 1.28 1.06 0.98
A33 4.83 3.01 2.31 2.31
A,” 2.90 1.80 1.45 1.34
Experimental
(P7003

Al 1.73 1.27 1.64 1.55
A1| 3.10 3.66 2.56 3.29
A,” 2.19 2.06 1.95 3.29

 

 

 

 

 

*converted from ”N values.

orbital for the two systems. For P7002 a mixture of 25% of the first excited state and 75%
of the ground state orbital was found to give good agreement between the in viva and in
vitra spin densities.

Analogous spin density calculations were performed for the nitrogen nuclei in
Pm‘. Because the unpaired spin density on the nitrogens in these n-type radicals resides
in the p orbital perpendicular to the plane of the, three sp2 hybrids, the spin density is

determined by the magnitude of the dipolar part of the hyperfine coupling, ff”

 

151
Al =Also "f
A!=A150 +2f

The amount of unpaired spin in this orbital can then be calculated by using a corrected
form of the McConnell equation (see Chapter 2)

p]: (cor) = 1.042 p? (uncor)
where the correction factor accounts for the increased 5 character of the orbital containing
the lone pair of electrons relative to the other sp2 bonding orbitals.33 The spin densities

for the nitrogen nuclei in P70; as well as the dipolar part of the hyperfine coupling used in

the calculation are shown in Table 4-5.

Table 4-5: Dipolar couplings and spin densities for pyrrole nitrogens in Pm+

 

 

 

 

 

 

 

 

NI N11 N111 NW
1' 0.79 0.58 0.31 0.46
(MHZ)
spin density 0.017 0.010 0.007 0.013
(corrected)

 

In the O’Malley and Babcock hybrid orbital model, the experimentally
determined spin densities were compared to the spin densities calculated for the ground
state molecular orbital for chlorophyll a”." These experimentally determined “reduction
factors” were then compared to those of the theoretically derived hybrid orbital. The
reduction factors for C1, C2a, C3, C4, C17, C18 and Ca derived experimentally were in

excellent agreement with the theoretically predicted factors (see Table 4-6).

 

152
A similar approach was taken with the nitrogen spin densities determined from

ESEEM and numerical simulations. As exhibited in Table 4-6, an excellent match
between the experimental and theoretical reduction factors was obtained.

The presence of a hybrid orbital also explains the isotropic hyperfine coupling
observed for the axial histidine ligand to Pm; discussed in Chapter 3. Molecular orbital
calculations predict that there is no localized unpaired spin density residing on the Mgz'
nucleus for an electron in the ground state a“, molecular orbital (see Chapter 3).“ The
first excited state molecular orbital however, is expected to have considerable spin
density at this site. Mixing25% of this excited state orbital with the ground state orbital
provides the spin density necessary to produce an observable hyperfine coupling to the
nitrogen moiety of the axial ligand.

The presence of this axial histidine ligand in P70; also explains the admixing of
the ground and excited state orbitals as it has been shown that the energy splitting
between these orbitals can vary with axial ligation.35 In the bacterial reaction centers, the
presence of an axial histidine ligand, which has been confirmed by the high resolution
crystal structure,”'” is not expected to catalyze this mixing since the energy separation
between the ground and excited state orbitals in bacteriochlorin systems has been found
to be larger than in chlorin systems (see Chapter 3).34

The relative magnitudes of the spin densities determined for the pyrrole nitrogens
in PM“ can be used to explain the quadrupole parameters, and, in particular, the relative
orientations of the hyperfine and quadrupole axis systems. Numerical simulations of
ESEEM data have shown that the quadrupole axes of two of the nitrogens (N1 and N.,)

are collinear with the axis system of the hyperfine tensor. The principle axes of these two

153

Table 4-6: Experimental and calculated reduction factors for P700‘

Atom Ground state, Do. (0.75) D0 x (0.25) Dl RFMC, RFMW
C1 0.023 0.019 1.2 1.1M
C2 0.014 0.0100 1.4
C3 0.037 0.0285 1.3 1.1"
C4 0.020 0.0142 1.4 1.5"
C5 0.006 0.0075 0.8 1.5"
C6 0.078 0.0600 1.3
C9 -0.001 -0.0008 1.2
C11 0.072 0.0485 1.5
C12 0.090 0.0655 1.4
C13 0.115 0.0830 1.4

C14 0.147 0.1082 1.4
C15 0.167 0.1235 1.3
C16 0.105 0.0760 1.4
C17 0.099 0.0734 1.3 1.3"
C18 0.73 0.0505 1.4 1.3"
C01 -0.025 0.0275 0.9 1.1"
CB -0.041 0.0060 6.8

Cy —0.009 0.0322 0.3

C5 -0.004 0.0455 0.1

C2a 0.002 0.0017 1.2 1.1M
C2b 0.002 0.0016 2.0

01 0.030 0.0235 1.3

06 - - -

Nl -0.005 0.0205 0.2 0.3
N2 -0.013 0.0107 1.2 1.3
N3 -0.008 0.0062 1.3 1.1
N4 -0.001 0.0317 0.03 0.07

*data from reference 34
"data from reference 1 1

154
tensors lie parallel to the pz orbital, normal to the molecular plane. In nuclei carrying a

large amount of unpaired spin density (e.g. N, and NW). the electric field gradient in the p
orbital, which contributes to the 7: molecular orbital. would be expected to be larger than
in the s orbital participating in the a bond to Mg. The principle axis of the field gradient
therefore is expected to be along p,; this is the case for N, and NW.

Because the principle axis of the field gradient is sensitive to small changes in the
orbital occupancy,26 particularly in pyrrole systems where the field gradients are similar
in the p orbital and the orbital containing the lone pair, small changes in the unpaired spin
density can effect changes in the direction of the principle axis. This is most likely the
case for N,, and N,,,, the two pyrrole nitrogens carrying smaller amounts of unpaired spin
density. The orientation of the quadrupole axis system for these two nuclei was found to
be 90° shifted from that of the hyperfine tensor. For these nuclei, the s orbital
participating in the coordinate covalent bond with the Mg” atom now carries the largest
electric field gradient therefore, the principle axis of the quadrupole tensor lies along this
orbital, in the molecular plane and perpendicular to All, the principle axis of the hyperfine
tensor.

The asymmetry parameter is also very sensitive to electric field gradients and has
been found to shift dramatically as the result of hydrogen bonding and other electrostatic
interactions. Interactions with charged amino acids or even with the carbonyl or ester
functionalities of the other half of the chlorophyll dimer near N, and N,v may be

responsible for the smaller 11 values determined for these nuclei. The distance between

155
the two chlorin planes has been determined by X-ray crystallographic studies to be about

4.5 A, close enough to affect the electric field gradient near the nitrogens.3

Conclusions

Taken together, these data represent the first definitive assignments of hyperfine
and quadrupole coupling constants to the nitrogen nuclei in P700- . The magnitude of the
hyperfine couplings and their corresponding spin densities are indicative of a monomeric
chlor0phyll a cation radical interacting with its protein environment through axial ligation
or hydrogen bonding. These effects cause an admixing of the ground and first excited
state molecular orbitals to create a hybrid orbital with an altered spin density distribution
when compared to chlorophyll cation radical species in vitra. Additionally, the
methodology utilized to determine these parameters, a combination of isotope
enrichment, electron magnetic resonance spectroscopy and numerical simulations,
provides a means by which complicated couplings in biological systems can be

consistently and systematically evaluated.

LIST OF REFERENCES

LIST OF REFERENCES

 

' For reviews see: a) Golbeck, J. (1992) Annu. Rev. Plant Physiol. Plant Mal. Biol. 43.
293 b) Malkin, R. (1996) in Advances in Photosynthesis (Ort, D. & Yocum. C .F., eds.)
Kluwer, Dordrecht, in press.

2 For reviews see: Diner, B.A. & Babcock, G.T. (1996) in Advances in Photosynthesis
(Ort, D. & Yocum, C.F., eds.) Kluwer, Dordrecht, in press. b) Britt, RD. (1996) in
Advances in Photosynthesis (Ort, D. & Yocum, C.F., eds.) Kluwer, Dordrecht, in press.

3 Fromme, P., Witt, H.T., Schubert, W.-D., Klukas, 0., Saenger, W. & Krauss, N. (1996)
Biachim. Biophys. Acta 1275, 76-83; Krauss, N., Hinrichs, W., Witt, 1., Fromme, P.,
Pritzkow, W., Dauter, Z., Betzel, C., Wilson, K.S., Witt, H.T. & Saenger, W. (1993)
Nature 361, 326-330.

‘ Moenne-Loccoz, P., Robert, B., Ikegarni, I. & Lutz, M. (1990) Biochemistry 29, 4740.
Ratajczak, R. Mitchell, R. & Haehnel, W. (1988) Biochim. Biophys. Acta 933, 306.

5 Norris, J.R., Uphaus, R.A., Crespi, H.L. & Katz, J.J. (1971) Proc. Natl. Acad. Sci. USA
68, 625.

6 Wasielewski, M.R., Norris, J .R., Crespi, H.L. & Harper, .1. (1981).]. Am. ChemSac.
103, 7664.

7 Rutherford, A.W.& Mullet, J.E. (1981) Biochim. Biophys. Acta 635, 225.

‘ Davis, RC. & Pearlstein, RM. (1983) Nature (London) 280,413.

9 Katz, J.J., Shipman, L.L. & Norris, LR. (1981) in Chlorophyll Organization & Energy
Transfer in Photosynthesis, Ciba Foundation Symposium 61, (Wolstenholme, G. &
Fitzsimmons, D.W., eds.) pp 1-40, Excerpta Medica, Amsterdam.

'0 Eccles, J. & Honig, B. (1983) Proc. Natl. Acad. Sci. USA 80, 4959.

“ O'Malley, P. J. & Babcock, G.T. (1984) Proc. Natl. Acad. Sci. US, 81,1098-1101.

'2 Deisenhofer, J., Epp, 0., Miki, K., Huber, R. & Michel, H. (1985) Nature 318, 618.

156

157

 

'3 Allen, J.P., Feher, G., Yeates, T.O., Komiya, H. & Rees, D. (1987) Proc. Natl. Acad.
Sci. USA 84, 5730.

" a) Lendzian, F., Lubitz, W., Scheer, H., Hoff. A.J., Plato, M., Trankle. E. & Mobius, K.
( 1988) Chem. Phys. Lett. 148, 377. b) Lendzian, F., Huber. M.. Isaacson. R.A..
Endeward, B., Plato, M., Bonigk, B., Mobius, K., Lubitz, W. & Feher, G. (1995)
Biochim. Biophys. Acta 1183, 139.

” A) Kat}, H., Bittersmann-Weidlich, 13., Andreasson, L.-E.. Bonigk, B. & Lubitz. W.
(1995) Chem. Phys, 194, 419432. B) Kai}, H. & Lubitz, w. (1996) Chem. Phys. Lett.
251, 193-203.

'6 Noren, G.H., Boemer, RJ. & Barry, BA (1991) Biochemistry 30, 3943.

'7 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. (1989) J.
Am. Chem. Soc. 111, 8076-8083.

‘3 McCracken, J ., Shin, D.-H & Dye, J.L. (1992) Appl. Magn. Reson 3, 30.

'9 Lin, C.P., Bowman, M.K. & Norris, JR. (1985).]. Magn. Res. 65, 369-374.

2° Mehring, M. & Freysoldt, F. (1980) J. Phys. E: Sci. Instrum. 13, 894-895.

2' Mims, W.B.(1984) J. Magn. Reson. 59, 291-306.

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

23 Kass, H., Rautter, J ., Bonigk, B., Hofer, P. & Lubitz, W. (1995) J. Phys. Chem. 99.
436.

I 2‘ Bowman, M.K., Norris, J .R., Thumauer, M.C., Warden, J., Dikanov, S.A.& Tsvetkov,
Yu. D. (1978) Chem. Phys. Lett. 55, 570. Dikanov, S.A., Astaskhin, A.V., Tsvetkov,
Yu.D. & Goldfeld, MG. (1983) Chem. Phys. Lett. 101, 206. Astashkin, A.V., Dikanov,
S.A. & Tsvetkov, Yu.D. (1988) Chem. Phys. Lett. 152, 258.

2’ Barry, B.A., El-Deeb, M.K., Sandusky, P.O. & Babcock, G.T. (1990) J. Biol. Chem.
265, 20139.

26 Ashby, C.I.H., Cheng, C.P. & Brown, TL. (1978) J. Am. Chem. Soc. 100, 6057.

27 Davis, I.H., Heathcote, P., MacLachlan, DJ. & Evans, M.C.W. (1993) Biochim.
Biophys. Acta 1143, 183-189.

158

 

’8 Hoff, A.J., DeGroot, A.. Dikanov. S.A., Astashkin, A.V.& Tsvetkov. Yu.D. (1985)
Chem. Phys. Lett. 118, 40.

’9 Townes, C.H. & Dailey, BR (1949) J. Chem. Phys. 17', 782.

3° Luckens, E.A.C. (1969) in Nuclear Quadrupole Coupling Constants. Chap. 1 1. pp 217-
233, Academic Press, New York.

3' Blackrnan, G.L., Brown, R.D., Burden, F.R. & Elsum, LR. (1976) J. Mal. Spectrosc.
60, 63.

32 Davis, I.H., Heathcote, P., MacLachlan, D.J. & Evans. M.C.W. (1993) Biachim.
Biophys. Acta 1143, 183-189. K'aB, H., Bittersmann—Weidlich. E.. Andreasson, L.-E..
Bonigk, B. & Lubitz, W. (1995) Chem. Phys, 194, 419-432. Kali, H. & Lubitz. W.
(1996) Chem. Phys. Lett, 251 , 193-203.

3’ Gordy, W. (1980) Theory and Applications of Electron Spin Resonance in Techniques
of Chemistry, Vol. XV (West, W., ed.) John Wiley & Sons, New York

3‘ Petke, J.D., Maggiora. G.M., Shipman, L.L. & Christoffersen, RE. (1980)
Photochemistry and Phatobiolog, 31, 243-25 7.

3’ Hanson, L.K. (1990) in Chlorophylls (Scheer, H., ed) pp 1015-1042, CRC Press, Boca
Raton

Chapter 5

The Effect of Sr’+ Substitution on the Oxygen Evolving Complex of Photosystem I]

Introduction

Water oxidation in Photosystem II (PSII) occurs at an oxygen evolving complex
(OEC) that consists, in part, of a cluster of four Mn atoms.“2 In addition to Mn. two Caf’
. ions per reaction center are required to activate oxygen evolution;3 their location in P811
is uncertain (see Chapter 1 for a complete description of PSII). Successive light-induced
charge separations cause the OEC to advance through five oxidation states. SO-S,.‘ A
characteristic multiline EPR signal that arises from an S= '/2 ground state of an exchange
coupled Mn tetramer is associated with the S2 oxidation state of the OEC.5'6 Substitution
of Sr" for Ca” maintains catalytic activity and gives rise to a modified multiline EPR
spectrum." These modifications are manifested as changes in both the number and
average hyperfine splittings of the peaks (see Figure 5-1) and have been attributed to
chagnes in either the structure or ligand environment of the Mn complex.° Electron spin
echo envelope modulation (ESEEM) studies have identified a nitrogen-based ligand'0 to
the OEC, which is assigned to the amino acid histidine." This chapter describes studies
in which the Mn-Nhis ligation is utilized as a probe into the structural changes that may
occur as a result of the Sr". Well-resolved two and three pulse ESEEM data on the
multiline EPR signal show that the electronic and geometric structure of at least one Mn

is unaffected by Sr” substitution.

159

160

{MW
MW

2500 3000 3500 4000

 

Figure 54: Light minus dark multiline EPR spectra from untreated (top) and Sr"
substituted (bottom) RCCs. Spectrometer conditions: microwave power, 20 mW;
modulation amplitude, 20 Gpp; modulation frequency, 100 kHz; sweep time, 100 s; time
constant, 200 ms; sample temperature, 8 K.

l 61
Materials and Methods

The development in 1981 of a PSII enriched thylakoid membrane preparation.
free from other electron transfer proteins, yet still functional in terms of oxygen evolution
facilitated study of the P811 environment.” These particles. BBYs. allow direct
specuoscopic studies of the OEC and serve as the starting point for the isolation of
smaller membrane flee complexes. This preparation relies on the fact that the thylakoid
membranes, with their inherent negative charges, can be “stacked” by using divalent
cations, such as Mg”. This increased lateral concentration can then be pelleted out by
using selective solubilization techniques. This treatment removes approximately 95% of
the PSI as well. Only the light harvesting complex proteins and the PSII/OEC remain
after completion of this process. Moreover, the decreased number of chlor0phyll
molecules per PSII, 400 Chl/PSII in prior preparations to 250 Chl/PSII in BBYs.
facilitates direct spectroscopic examination of the PSIl electron transfer cofactors and, in
particular, the Mn ensemble. The decreased chlorophyll concentration relative to PSII
also results in increased 02 evolution activity by allowing saturating light to reach the
reaction centers.

Even with the increased concentration of active PSII centers and concomitant
increase in 02 activity, spectroscopic investigations of the EPR signal from the S2 state of
the OEC have been inhibited by the small number of spin present in the sample. Even
with saturating light sources and a minimal number of inactive centers, extensive signal
averaging was still necessary to obtain a resolved cwEPR spectrum. The subsequent

development of a preparation lacking the light harvesting complex by Ghanotakis and

162
coworkers” produced a reaction center core preparation (RCC) that contained only the

central PSII components. These preparations. again characterized by increased oxygen
evolution activity, eliminated the necessity for extensive signal averaging and. moreover.
allowed the application of advanced spectroscopic techniques that require larger sample
concentrations.

Sample preparation: PSII membrane samples from market spinach, isolated by using the
method of Berthold et al.12 were diluted to a chlorophyll concentration of 1.0 mg/ml by
using an appropriate volume of SMN buffer (0.4 M sucrose, 50 mM MES. 15 mM NaCl.
pH 6.0) containing 0.8% of the detergent octyl-B-D-thioglucopyranoside (OTG). This
non-ionic detergent was included with the high salt concentration to solubilize the PSI]
membranes and facilitate the removal of the light harvesting complex proteins as well as
the 17 and 23 kDa polypeptides." Following incubation. the solution was centrifuged to
remove the remnants of these polypeptides. The supernatant, containing the reaction
center cores, was incubated in SMN in a 1:0.8 ratio. MgCl, was added to a final
concentration of 10 mM. Following centrifugation, the pellet, consisting of the light
harvesting complex and solubilized membranes, was discarded and the reserved
supernatant was incubated on ice with an equal volume of 30% w/v of polyethylene
glycol (PEG). The PEG precipitation step acts to aggregate the water molecules by
causing “clumping” of the P811 constituents, and represents a modification to the original

published procedure. The P811 cores are easily pelleted out with centrifugation,

resuspended in SMN and reserved for later use.

 

163
Substitution of Srz” for Cai‘ in these RCC samples is accomplished by using Ca3'

free buffers to wash the isolated reaction centers after incubation in a high (1 M NaCl)
salt buffer. Substitution of Sr2+ is easily accomplished by resuspending the final pellet in
a Sr” (10 mM) buffer. Samples were loaded into 4 mm O.D. quartz EPR tubes and
frozen to 77 K until EPR experiments could be performed.

EPR Spectroscopy: Dark EPR spectra of the Sr" and Ca2+ samples at 8 K were recorded
with a either a Bruker ER200D or ESP300e spectrometer fitted with an Oxford
Instruments ESR-900 cryostat. Microwave frequency was determined by using an EIP
Microwave Model 258 fiequency counter and the static magnetic field strength was
measured with a Bruker ER035M NMR gaussmeter. A rectangular TE102 cavity was
used.

Dark adapted samples (5 minutes) were illuminated at 200 K in a liquid
nitrogen/ethanol bath. Illumination time was 2 minutes. Following illumination. the
samples were frozen in the dark and the light spectra were recorded at 8 K. The spectra
from both the dark and the illuminated Ca2+ sample correspond to an addition of 30
accumulations. The spectra from the Sr2+ sample correspond to an addition of 60
accumulations. Signal averaging was accomplished (on the ER200D spectrometer) with
an IBM clone fitted with a data analysis program written in house. The spectra presented
represent a light minus dark difference. Subtraction was carried out with an IBM clone
by using software written in house. The experimental conditions accompany each figure.
ESEEM Spectroscopy: ESEEM data were collected on a home built spectrometer” by

using a reflection cavity where a folded stripline” served as the resonant element. Either

 

 

164
a two-pulse (90°-t-180°) or a three pulse or stimulated echo (90°-t-90°-T—90°) pulse

sequence was used. Dead time reconstruction was performed prior to Fourier
transformation as described.‘7 Computer simulations of the ESEEM data were performed
on a Sun Sparcstation 2 computer utilizing FORTRAN software based on the density
matrix formalism of Mims." The analysis software for the treatment of experimental and
simulated ESEEM data was written with Matlab (Mathworks, Nantick. MA). The
3 experimental dead time was included in the simulations. An isotropic g-tensor was

assumed in all of the calculations.
Results

Difference ESEEM spectra obtained by Fourier transformation of two-pulse time domain
data from untreated and Sr2+ substituted RCCs are shown in Figures 5-2 and 5-3.
respectively. These data were collected at field values corresponding to g = 1.8 and g =
1.9 for the two sample to avoid interference from the intense tyrosine radical signal at g =
2.0. The two pulse ESEEM spectraare identical for the two samples; a feature near 4
MHz is present in both spectra. An additional feature arising from weakly coupled
matrix protons is also present in both spectra. The broad, featureless peak at 4 MHz is
similar in frequency and appearance to the peak assigned, based on specific isotopic
substitution and subsequent ESEEM studies, to the nitrogen moiety of histidine.”
Resolution of low frequency component in a two pulse ESEEM experiment can be poor

due to the short phase memory time Tm. To resolve these components, stimulated echo

I65

 

 

 

 

 

 

Ca2+ sample
’3
31
m
1:
2
‘5.
E
<
0 10 20 30 4O 50
Frequency (MHz)

Figure 5-2: . Fourier transformation of light minus dark twopulse ESEEM data from
untreated RCCs. Spectrometer conditions: magnetic field strength, 3440 G; microwave
pulse power, 44 dBm; microwave pulse width, 22 ns; pulse repetition rate, 90 Hz; tau
value, 200 ns; sample temperature, 4.2 K.

166

 

   

Sr2+ substituted sample

Amplitude (a.u.)

 

 

 

0 10 20 30 4o 50
Frequency (MHz)

Figure 5-3: Fourier transformation of light minus dark two-pulse ESEEM data from Srz‘
substituted RCCs. Spectrometer conditions: magnetic field strength, 33250 G; all other
conditions as in Figure 5-2.

167
experiments were performed. Since the decay of the modulations in these experiments is

governed by the spin-lattice relaxation time. T,.'9 improved resolution of the low
frequency modulations is expected.

Stimulated echo difference ESEEM spectra from the untreated and Sr}
substituted RCC samples are presented in Figures 5-4 and 5-5. respectively. These
spectra were collected at field values identical to those in Figures 5-2 and 5-3.
Unfortunately, no additional frequency components were resolved in these experiments.
The spectra are similar; each exhibiting a single resonance in the 4 MHz region. Because
the two spectra were collected at slightly different magnetic field values. these peaks are
not observed at identical frequencies. However, by taking the difference in fields into
account, these peaks represent identical couplings.

The magnetic field position at which the data are collected can have a profound
effect on the appearance of the ESEEM spectrum. This is especially apparent in the case
of the P811 samples where the overlap of multiple radical signals can make data analysis
complicated and, although difference spectroscopy is employed to eliminate many of
these contributions, care must be taken to avoid erroneous assignments.20 The
photoexcitation utilized to effect Mn oxidation to the S2 state can produce radical species
on the acceptor side of PSII that can not be completely eliminated by difference
spectroscopy.

ESEEM spectra obtained by Fourier transformation'7 of three-pulse time domain data
from untreated RCCs collected at 1.8 K are shown in Figure 5-6. All spectra were

collected at g = 1.95 to avoid interference from the intense tyrosine radical signal at g =

168

 

 

 

 

 

Ca2+ sample
2?
g
m
1::
.3.
'3
E
<
O 5 10 15 20 25
Frequency (MHz)

Figure 5-4: Fourier transformation of light minus dark three-pulse ESEEM data from
untreated RCCs. Spectrometer conditions as in Figure 5-2.

169

 

 

 

 

Sr2+ substituted sample
’5
s;
0
'O
2
E.
E
<
0 5 10 15 20 25

Frequency (MHz)

Figure 5-5: Fourier transformation of light minus dark three-pulse ESEEM data from
Srz‘ substituted RCCs. Spectrometer conditions as in Figure 5-3.

170

 

 

 

 

 

 

 

l
A 1
Q
1:
2
.3
E
< B
Ww/JXJ‘” CW
”I‘M \m/
0 2 - 6 8 10 12 14 16
=reouency (MHz)

Figure 5-6: Three-pulse ESEEM spectra from untreated RCCs after illumination (top),
prior to illumination (middle) and the light minus dark difference spectrum (bottom)
collected at g=1.95. Spectrometer conditions: microwave frequency, 8.955 GHz;
magnetic field strength, 3290 G; microwave pulse power, 20 W; tau value, 213 ns; pulse

repetition rate, 90 Hz; sample temperature, 1.8 K.

171
2.0. The ESEEM spectrum from untreated RCCs in SI (Fig. 5-6. middle) shows

peaks at 2.41, 3.85 and 14.10 MHz that arise from paramagnets present in this state.

When advanced to S, by continuous illumination at 195 K for 10 minutes. the ESEEM

spectrum (Fig. 5-6, top) shows peaks at 0.78, 1.55, 2.41. 3.85. 4.84 and 14.10 MHz.

Following subtraction of the SI time domain data from that of SI. and subsequent Fourier

transformation of the difference (Fig. 5-6, bottom), peaks at 0.78. 1.55. 4.25 (shoulder)
and 4.84 MHz are clearly evident. Because these spectra differ from those obtained at
higher field values, the couplings observed must arise from a contaminating paramagnetic
species that is not removed by using difference spectroscopy. The most likely candidate
for this spurious radical is the quinone anion radical that acts as an electron acceptor in
PSII." Magnetic interactions between this quinone species and a non-heme iron produce
an EPR signal centered at g = 1.9 and overlaps considerably at this g value with the
multiline signal.22 The peaks observed in the difference ESEEM spectrum most likely

arise from a histidine residue interacting with this quinone-iron species.

The difference spectra in Figure 5-6 is indicative of l4N-magnetic couplings close

to the condition of exact Cancellation.” In this condition, the positions of the two narrow,

lower frequency peaks are dominated by “N nuclear quadrupole interactions. The
frequency and lineshape of the 4.8 MHz feature is most sensitive to the superhyperfine

coupling tensor and its orientation relative to the nuclear quadrupole tensor. Numerical
simulations" of this I4N-ESEEM spectrum yields an isotropic hyperfine coupling
constant of 2.2 MHz, an effective dipole-dipole distance of 2.8 A, an l4N nuclear

quadrupole coupling constant, equ, of 1.6 MHz and an asymmetry parameter, 1], of 0.9.

172
Anisotropic hyperfine coupling constant of 2.2 MHz is consistent with values obtained in

a detailed "N ESEEM study of di-u-oxo bridged Mn(III)Mn(IV) bipyridyl dimer
compounds," where scalar couplings of 2.8 MHz for the directly coordinated bipyridyl
nitrogens were found. Therefore, the features of the ESEEM spectrum of Figure 5-6 are
assigned to a histidyl nitrogen terminally coordinated to the non-heme iron and not to the

histidine residue ligated to the Mn complex.
Discussion

EPR spectral shifts in a multinuclear metal cluster without corresponding changes
in the ESEEM spectrum have precedence. For example, when the Rieske iron-sulfur

center23 of cytochrome bbf is treated with 2,5-dibromo-3-methyl-5-isopropyl-
benzoquinone, shifts in the g value occur without affecting the Fe-N interaction observed

by ESEEM. Similarly, studies of copper (II) compounds ligated to substituted imidizoles

show that changes in coordination geometry lead to changes in the Cu(II) hyperfine
coupling without corresponding changes in l‘N-ESEEM.”

The ESEEM results presented here show that the bond length and bond covalency

of the Mn-Nhi‘ bond, as well as the unpaired spin density on the Mn, do not change

appreciably upon Srzl substitution.“27 The modifications in the EPR spectrum suggest,
however, that structural or electronic changes do occur. Manganese hyperfine couplings
in a cluster are a function of the spin projections of the individual ions onto the system

spin as well as the hyperfine coupling values for the individual metals.’ The spin

173
projections are related directly to the ground spin state of the cluster. whereas the

hyperfine coupling constants of the individual metal ions are sensitive to changes in
ligand environment. To affect the hyperfine coupling observed in the multiline EPR
spectrum, one, or both. of these parameters must change. Because the S= ‘2: ground spin

state of the Mn complex in S2 is believed to be well separated from the first excited

state,28 it is unlikely that the Srz‘ induced modifications in the EPR spectrum arise from

changes in the ground spin state.29 However, changes in the ligand environment can
cause the Mn hyperfine coupling to vary from 65 to 84 gauss.3o The substitution of Sr}

for Ca2+ may affect the hyperfine coupling of one, two or three of the Mn in the OEC

without perturbing the Mn-Nhis ligation observed with ESEEM.

Recent extended X-ray absorption fine structure (EXAFS) studies of PSII

show that the Mn-Mn, Mn-O and Mir-terminal ligand distances are the same or only
slightly different upon Sr2+ substitution." The EXAFS data also suggest a close
proximity of Ca2+ (Srzl) to the Mn cluster and a model based on the EXAFS data shows
the Cap bridged to one Mn by a carboxylate ligand." The slightly larger ionic radius of
Sr» over Caz. (1.13 A versus 0.99 A) appears to cause a modification of the structure of

the Cap binding site, which is transmitted to the Mn complex. Taken together, the
ESEEM, EPR and EXAFS results indicate that a shift in hyperfine coupling that arises

from terminal ligand effects remote from the Mn-Nhis detected by ESEEM, rather than
rearrangement of ground spin states, is the source of the EPR spectral modifications

observed upon Sr2+ substitution.

LIST OF REFERENCES

LIST OF REFERENCES

 

‘ For reviews see: Diner, B.A. & Babcock, G.T. (1996) in Advances in Photosynthesis
(Ort, D. & Yocum, C.F., eds.) Kluwer, Dordrecht, in press. b) Britt. RD. (1996) in
Advances in Photosynthesis (Ort, D. & Yocum, C.F., eds.) Kluwer, Dordrecht. in press.
C) Debus, RJ. (1992) Biochim. Biophys. Acta 1102, 269.

2 Yocum, C.F., Yerkes, C.T., Blankenship, R.E., Sharp, R.R. & Babcock. G.T. (1981)
Proc. Natl. Acad. Sci. USA 78,7507.

3 Yocum, CF. (1991) Biochim. Biophys. Acta 1059,l.
‘ Kok, B., Forbush, B. & McGloin, MP. (1970) Photochem. Photobiol. 11, 457.

5 A) Dismukes, G.C. & Siderer, Y. (1981) Proc. Natl. Acad. Sci. USA 78. 274. B)
Hansson, O. & Andreasson, L.-E. (1982) Biochim. Biophys. Acta 679,261.

6 A) Britt, R.D., Lorigan, G.A.. Sauer. K. , Klein, M.P. & Zimmerman, J .-L. (1992)
Biochim. Biophys. Acta 1140,95. B) Koulougliotis, D., Hirsh, D. & Brudvig, G.W.
(1992) J. Am. Chem. Soc. 114, 8322.

7 Boussac, A. & Rutherford, A.W. (1988) Biochemistry 27, 3476.

a Ghanotakis, D.F., Babcock, G.T. & Yocum, CF. (1984) FEBS Lett. 167, 127.

9 Sauer, K., Yachandra, V.K., Britt, R.D. & Klein, MP. (1992) in Manganese Redox
Enzymes, pp. 141-175, VCH Publishers, New York.

'° DeRose, V.J., Yachandra, V.K., McDermott, A.E., Britt, R.D., Sauer, K. & Klein, MP.
(1991) BiOChcmistry 30, 1335.

" Tang, X.-S., Diner, B.A., Larsen, B.S., Gilchrist, M.L., Lorigan, G.A. & Britt, RD.
(1994) Proc. Natl. Acad. Sci. USA 91, 704.

'2 Berthold, D., Babcock, G.T. & Yocum. CF. (1981) FEBS Lett. 34, 231.
174

175

 

” Ghanotakis, D.F., Babcock. G.T. & Yocum. C .F. (1984) FEBS Lett. 143. 175.

M OTG ref.

” McCracken, J., Shin, D.-H. & Dye, J .L. (1992) Appl. Magn. Reson. 3. 30.

'6 Lin, C.F., Bowman, M. K. & Norris, J.R. (1985) J. Magn. Reson. 65. 369.

‘7 Mims, W.B. (1984) J. Magn. Reson. 59, 291.

‘ '8 Mims, W.B. (1972) Phys. Rev. B 5,2409.

'9 Schweiger, A. (1991) Angew. Chem. Int. Ed. Engl. 30, 265.

2° Miller, A.-F. & Brudvig, G.W. (1991) Biochim. Biophys. Acta 1056, 1.

2‘ Nugent, J.H.A., Diner, B.A. & Evans, M.C.W (1981) FEBS Lett. 124. 241.

22 Rutherford, A.W. & Zimmerman, J .-L. (1984) Biochim. Biophys. Acta 767. 168.

23 A) Flanagan, H.L. & Singel, DJ. (1987) J. Chem. Phys. 87, 5606. B) Mims, W.B. &
Peisach, J. (1978) J. Chem. Phys. 69, 4921.

2‘ A) Britt, RD. (1988) Ph.D. Thesis, Department of Physics, University of Califomia-
Berkeley, Lawrence-Berkeley Laboratory, Laboratory Report LBL-25042. B) Britt. R.D.
& Klein, MP. (1992) in Pulsed Magnetic Resonance: NMR, ESR and Optics, pp. 390-
410, Clarendon Press, Oxford.

2‘ A) Britt, RD., Sauer, K., Klein, M.P., Knaff, D.B., Kriauciunas, A., Yu, C.-A., Yu, L.

& Malkin, R. (1991) Biochemistry 30, 1892. B) Malkin, R. (1982) Biochemistry 21,
2945.

26 Jiang, F., Karlin, K.D. & Peisach, J. (1993) Inorg. Chem. 32, 2576.

27 Tsao, J., Sivaraja, M.& Dismukes, GO (1991) Biochemistry 30, 4734.

2’ Hansson, O., Aasa, R. & Vanngard, T. (1987) Biophys. J. 51, 825.

29 A) Chan, M.K.& Armstrong, W.H. (1992) J. Am. Chem. Soc. 113, 5055. B) Kirk,
M.L., Chan, M.K., Armstrong, W.H. & Solomon, El. (1992) J. Am. Chem. Soc. 114,

10432. C) Pal, 8., Chan, M.K. & Armstrong, W.H. ( 1992) J. Am. Chem. Soc. 114,
6398.

I76

 

3° A) Diril, H., Chang, H.-R., Niges, M.J., Zhang. X., Potenza. J .A.. Schugar. H.J.. Isied.
S.S. & Hendrickson, D.N. (1989) J. Am. Chem. Soc. 111. 5102. B) Bonvoison. J..
Blondin, G.,-Girere, J.-J. & Zimmerman, J .-L. (1992) Biophys. J. 61. 1076.

3' Yachandra, V.K., DeRose, V.J., Latimer, M.J., Mukerji. 1.. Sauer. K. & Klein. M.P.
91993) Science 260, 675.

HICHIGAN STATE UNIV. LIBRARIES
ll111111111111111W111111111111111111
31293015644622