PLACE IN RETURN BOX to roman thb checkout from your mood.
TO AVOID FINES Mum on or baton data duo.

DATE DUE DATE DUE DATE DUE

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

VIBRATIONAL STUDIES OF METALLOPROTEINS AND
RELATED MODEL COMPOUNDS

By

Matthew Terrell Gardner

A DISSERTATION

Submitted to
Michigan State University
in partialfulfillment of the requirements
for the degree of

DOCTOR OF PHILOSPHY
Department of Chemistry

1997

VIB

An uliim
understanding 0]
functions. Vibratj
goal. It has me a
WWI inform
Vibrational Spec:
[0 three SFStem:
533131115. in 0Tdi
mechanism by .

Order Of increa,

 

compoufld Syn:

\ibfaiIOna] pr or

cF‘anide moietx

 

ABSTRACT

VIBRATIONAL STUDIES OF METALLOPROTEINS
AND RELATED MODEL COMPOUNDS
By

Matthew T. Gardner

An ultimate goal of biophysical studies of enzymes is to achieve an
understanding of the mechanism by which enzymes carry out their biological
functions. Vibrational spectroscopy is just one method that can be used to achieve this
goal. It has the advantage that it has the potential to elucidate both mechanistic, and
structural information about the system being studied. We have applied two forms of
vibrational spectroscopies, resonance Raman, and infrared absorption spectroscopy,
to three systems, a synthetic model of an enzymatic active site, and two enzymatic
systems, in order to achieve a greater understanding of the structure, function and
mechanism by which they operate. The experiments detailed here, are arranged in
order of increasing complexity and technical challenge. The first system, a model
compound synthesized as an analogue to the cyanide ligated form of the binuclear
center of oxidized cytochrome c oxidase, was studied, and found to exhibit unique
vibrational properties. These properties are attributed to the bridging geometry of the
cyanide moiety between the copper and iron atoms of the model complex.

The second system presented is the enzyme soluble guanylate cyclase. This

enzyme, which is the only known physiological receptor for NO, was found to have a

unique heme
stretching frc
hemeemmz
accounts for
binding kinet
The
photosyntheti
molecular ox;
manganese at
Spectroscopy.
information re

cycle of the en

 

unique heme-binding pocket. The CO-bound form of this enzyme exhibits a CO
stretching frequency higher than has been previously reported for any CO-ligated
heme-enzyme. This, we confirm, is attributable to a highly polar heme pocket, which
accounts for not only the observed vibrational properties of CO, but also the ligand
binding kinetics of the enzyme.

The third system studied is the oxygen evolving complex of the
photosynthetic enzyme, photosystem II. This enzyme, which oxidizes water to
molecular oxygen in plants and cyanobacteria, has an active site consisting of four
manganese atoms. We have developed novel methods whereby, using infrared
spectroscopy, we can look directly at the manganese cluster and thus have access to
information regarding the specific molecular intermediates in the oxygen evolving

cycle of the enzyme.

To my Dad

There are m
made to my u
on way. mac
and Others as it
On the work Sll
~The te.
SPeCific
Ron Ha
and Joh;
and fror

ad\ ice :1

 

 

The Ct

are blex

lhese

 

ACKNOWLEDGEMENTS

There are many folks to whom I owe a great deal. The contributions people have
made to my work and my life over the past six years vary widely. But each, in their
own way, made my life better, my work easier, or both. I would like everyone here,
and others as well, to accept my heartfelt thanks.
On the work side of things, the folks I especially would like to thank include:
-The technical support/professional staff here at Michigan State University.
Specifically I would like to thank Russ Geyer, Sam Jackson, Dick Menke,
Ron Haas, Scott Sanderson, Scott Bankroff, Manfred Langer, Keki Mistry,
and John Rugis. Each one of them has the ability to take some ethereal idea,
and from that, produce exactly what it is that that you need. Their input,
advice and expertise was indispensable
-The Computer gang, Dr. Tom Carter, Dr. Tom Atkinson, and Paul Reed. We
are blessed here at MSU with top-notch computer facilities and networks.
These guys were always available and helpful in many different ways.
-The Business Office. Those folks put up with my endless requests for things
that I needed yesterday. I appreciate their efforts and their understanding of

what was usually my own lack of foresight.

V

On the pew
gratitude:
- The
ever}
intellc‘
beholt
- My
backp;
altkays
~lhe E
shared
definite
how m1
-Profes:

Coming

BO“ lb}

 

On the personal side of things, there’s even more folks to whom I owe a lot of

gratitude:
- The gang at Science Theatre. It was an immense pleasure to work with
everyone at ST. The dedication of this group of people, the intensity of their
intellects, and the fun they pumped into everything we did, was awesome to
behold. Thanks for letting me work with you and for you.
- My friend Chris. Our adventures together learning how to home brew, our
backpacking journeys, and really getting into photography are things I’ll
always cherish. And we’ll do a ton more of them (So, middle of September??)
-The Dagwoods Gang: Chris, Sara Helvoigt, Per Askeland and Michelle. We
shared a lot of good times together. My sanity and my sense of humor were
definitely retained because of the time we spent there. If you thought about
how much money we have spent there. . ..
-Professor Charlie Yocum, for being the person who single handedly
convinced me that being a died-in-the-wool liberal was a good thing. Dr. Neil
Bowlby for having an expansive intellect that I envy, Dr. Matthew Espe for
not taking any shit from this “liberal arts” East-Coaster, Dr. Curt Hoganson
for always asking the right questions, for keeping me honest, and for
providing incredible insight to just about any scientific discussion and last but
not least, Dr. Michelle Mac, for being a fiiend and confidant through it all.
-My current research group: Andy, Curt, John, Kristy, Denis, Costas, Michelle

P., Pierre, Nadia, Heather, Marilena, Catherine, Matt, Kris, Gwynne and Hans.

vi

Es
GO
lo_t l
-M}
my 1
not 0
-Prof.
impor

Somet.

Indispc
d€tai ls

make a

My ad

 

You folks have put up with me above and beyond the call of duty and made
my time here a blast. Thanks.

-Past members of my research group and other colleagues, including Dr. Matt
Espe, Dr. Geurt Deinum, Dr.Yvonne Gindt, Dr. Craig Essenmacher, Dr. Jerry
Godbout, and Dr. Tom Halasinski. The subbasement was a lot more fun and a
lo_t more interesting with those last two guys.

-My fi-iends here in Lansing. Danielle and Rod, Bill and Aimee. You folks are
my friends for life. The fun we had and the games we played, made Lansing
not only tolerable, but a place I look forward to returning to.

-Professor George Leroi who allowed me the use of his lab space, and more
importantly, whose warm and compassionate way of interacting with people is
something ll aspire to.

-Dr. Andy Chu. Andy’s help, friendship, and collaboration was an
indispensable aspect of the photosystem II work. His dogged attention to
details ensured the biological relevance of all that we did. This work is going
make a splash because of him.

-My advisor, Professor Jerry Babcock, for more than a few things. He taught
me how to do science (if he wants to claim credit). He indulged my wide-
ranging interests, yet, when necessary, beat me to a pulp to stay focused. His
enthusiasm for science is infectious, and his attention to us as people, as well
as scientists, made graduate school a complete experience. I will never forget

his poise delivering the plenary lecture at Montpellier. Nor will I forget his

vii

laug
\Vflh
inrn
-hl}
hlor
Pete
each
-Sue
have
learr

mad:

laughing in the face of danger on the rapids of the Sturgeon River, or his
willingness to handover his hard —earned dough on poker nights. I owe a lot,
in many different ways, to Jerry.

-My family. Without their love and support, I couldn’t have done this. My
Mom, Maich, my stepdad, Ed, my stepmom Suze, my greatly missed dad,
Peter, my brothers Chris, Josh and Zack, and my sisters Amanda and Griffin,
each played a huge role in this work, and for that I will always be grateful.
-Sue. She is my partner and friend. We’ve been through a lot together and
have many more adventures coming. We have learned, and will continue to
learn a lot from each other. Her love and support, kindness and affection have

made these years wonderful. We’re a good team.

viii

LIST OF T
LIST OF P

CHAPTER
Abs
Vibi
lnfrz
Rarr.
Four
Sign
Refe

CHAPTER "

Absu
IRUOt
Mate;
Resul

DISCU:

Refere;

TABLE OF CONTENTS

LIST OF TABLES .................................................................................... x

LIST OF FIGURES ................................................................................. xi

CHAPTER 1 Introduction to Vibrational Spectroscopy of Biological Systems
Abstract ........................................ ' ................................................. 1
Vibrational Spectroscopy of Biological Systems ....................................... 2
Infrared Spectroscopy ....................................................................... 3
Raman Spectroscopy ......................................................................... 4
Fourier Transform Infrared (FTIR) Spectroscopy ...................................... 11
Signal-to-Noise Considerations in FTIR ................................................ 16
References .................................................................................... 20

CHAPTER 2 Vibrational Analysis Of A Molecular Heme-Copper Assembly With A
Nearly Linear Fem-Cn-Cu" Bridge: Insight Into Cyanide Binding To Fully
Oxidized Cytochrome C Oxidase

Abstract ....................................................................................... 21
Introduction ................................................................................. 22
Materials and Methods ..................................................................... 26
Results
UV-Visible Spectroscopy ........................................................ 33
Infrared Spectroscopy ............................................................ 33
Resonance Raman Spectroscopy ................................................ 4O
Vibrational Analysis .............................................................. 40
Discussion
Electronic Structure and Vibrational Characteristics .......................... 46
Normal Mode Composition ...................................................... 47
Resonance Enhancement Conditions ........................................... 51
Relevance to Cytochrome c Oxidase ........................................... 52
References ................................................................................... 57

ix

ClltPTER 3 St
Tl
cl
Abstract .
Introduct I
It
R
Material ~

Results o
Referenc

ClllPTER 4 l’
1:
Abstract
lntroduc
llhat \l'
FTIR of
Detector
Windou
Temper;
Conclus
Referen

_-'-_-_-_

l.
ChtPTERS 1
l
Abstrac'
Illlloduk'l
h’laterjg,l
Results.
DlSCUSs
C0DClU>
Referee,

 

CHAPTER 3 Small Molecule Binding To Heme Proteins: Vibrational Spectroscopy of
The Ligand Binding Site in The ONO Binding Enzyme Soluble Guanylate

Cyclase
Abstract ...................................................................................... 60
Introduction ................................................................................. 61
Infrared Spectroscopy of Heme Proteins ....................................... 62
Resonance Raman of Heme Ligand Interactions .............................. 64
Materials and Methods ..................................................................... 74
Results and Discussion .................................................................... 74
References ................................................................................... 78

CHAPTER 4 Probing Low—Frequency Vibrations in Metalloproteins Using Light
Induced Infrared Difference Spectrosc0py: Challenges and Solutions

Abstract ...................................................................................... 80
Introduction ................................................................................. 80
What Will be Found in the Low-Frequency Region? ................................. 83
F TIR of Aqueous Systems: The Low Frequency Challenge ......................... 84
Detectors .................................................................................... 96
Window/Substrate Materials .............................................................. 99
Temperature Control ..................................................................... 102
Conclusions ................................................................................ 108
References ................................................................................. 1 10

CHAPTER 5 Low Frequency Modes in the Light-Induced FTIR Difference Spectrum of

Photosystem II
Abstract .................................................................................... 112
Introduction ................................................................................ 113
Materials and Methods ................................................................... 130
Results ...................................................................................... 133
Discussion ................................................................................. 158
Conclusions ................................................................................ 176
References ................................................................................. 179

lukll
lukLZ
lhhld
loh24
Table 2-5
luh34
loh4l
lnh4q
inhlg

Tok54

lthQ

lhh53

 

Table 2-1

Table 2-2

Table 2-3

Table 2-4

Table 2-5

Table 3-1

Table 4-1

Table 4-2

Table 4-3

Table 5-1

Table 5-2

Table 5-3

List of Tables

Isotopic Dependence of Vibrational Frequencies ............................. 36
Extinction Coefficients of Various CEN Stretches ............................ 36
Vibrational Frequencies for CEN in Selected Structures ..................... 39
Force Constants for NCA of Fe-CN-Cu system ............................... 44
Experimental and Calculated Vibrational Frequencies (cm") ............... 45

Vibrational Frequencies of Various CO-adducts of Heme-enzymes. . . ....73

Vibrational Frequencies of Protein and Metal-Ligand Structures... ........82
Infrared Absorbances of H20 and D20 ......................................... 87
Properties of Infrared Window Materials ..................................... 101
Vibrational Data (1800-1000 cm") From Various LI-FTIR

Measurements of Photosystem II .............................................. 162
Group Frequencies of Protein-based Vibrations ............................. 163
Low-Frequency Vibrations Possibly Present in Photosystem II ........... 168

xi

Figure l-l:
Figure l-Z:
Figure 1-3:

Figure Z-l:

Figure 2-2:

Figure 2-3:

Fllute 24:

Figure 2-5:

figure 2-6

Flgmc 2.7:

(I)

 

 

 

Figure 1-1:
Figure 1-2:
Figure 1-3:

Figure 2-1:

Figure 2-2:

Figure 2-3:

Figure 2-4:

Figure 2-5:

Figure 2-6:

Figure 2-7:

List of Figures

Schematic of Raman scattering spectrometer ................................... 7
Schematic of scanning infrared spectrometer .................................. 12
Schematic of Michelson Interferometer ........................................ 14

Top: Schematic representation of [(1-MeIm)Fe(OEP)-CN-Cu(Me6tren)]2+
Bottom: Model used for normal coordinate analysis. Point masses
represent atoms or fractional representations of ligands ..................... 25

UV-Vis absorption Spectra of “bridged”, [(py)Fe(OEP)-CN-Cu(Me6tren)],
and “unbridged”, [(py)Fe(OEP)-CN], complexes. Sample concentrations
were 2.2 mM (bridged) and 1.9mM (unbridged). Spectra were dissolve in
acetone and spectra were recorded in a €an cell, SOuM pathlength ...... 34

FTIR spectrum of [(1-MeIm)Fe(OEP)-CN-Cu(Me6tren)]2+ and
isotopomers. Spectral conditions described in Materials and
Methods ............................................................................. 37

FTIR spectrum of [(1-MeIm)Fe(OEP)-CN-Cu(Me,,tren)]2+ and
isotopomers. Spectral conditions described in Materials and
Methods ............................................................................. 38

Resonance Raman spectrum of [(1-MeIm)Fe(OEP)-CN-Cu(Me6tren)]2+

and isotopomers. Spectral conditions described in Materials and
Methods ............................................................................. 41

Resonance Raman spectrum of [(1-MeIm)Fe(OEP)-CN-Cu(Me,,tren)]2+
and isotopomers. Spectral conditions described in Materials and
Methods ............................................................................. 42

Plot of calculated vibrational frequencies versus
Fe-C stretching/Fe-C-N bending interaction force constant ................. 49

xii

Figure 2-8:
Figure 29:

Figure 3-1:

Figure 3-2:

Figure 3-3:

Figure 34:
Fitness:
Figure 4.1:
Figure t2:
FiEure 4.3;
Flgurc 44;

figure 4-5

Flgure 46;

 

 

Figure 2-8:

Figure 2-9:

Figure 3-1:

Figure 3-2:

Figure 3-3:

Figure 3-4:

Figure 3-5:

Figure 4-1:

Figure 4-2:

Figure 4-3:

Figure 4-4:

Figure 4-5:

Figure 4-6:

Plot of calculated vibrational frequencies versus
Cu-N stretching/Cu-N-C bending interaction force constant ............... 50

Representation of the binuclear center of oxidized cytochrome c
oxidase .............................................................................. 55

Infrared absorption spectrum of water, Can cell, 50pm pathlength, 25 °C.
Approximate frequencies of X-Y stretches of heme-ligated XY diatomic
ligands are shown (XY = CEN or CEO) ....................................... 65

UV-Vis spectrum of heme-ligand complex ([(py)Fe(OEP)-CN]). Sample
concentration = 2.2 mM, in acetone. Spectrum acquired in 50pm Can
cell, room temperature ............................................................ 67

Correlation between Fe-C stretching frequencies and C-0 stretching
frequencies for a variety of histidine ligated carbonmonoxy heme
complexes. Adapted from [16] ................................................... 68

Resonance Raman spectrum of isotopomers of carbonmonoxy-soluble
guanylate cyclase. Laser excitation 7c=423 nm[20] ........................... 72

FTIR spectrum of isotopomers of carbonmonoxy-soluble guanylate
cyclase. Spectral conditions given in Materials and Methods ............... 76

Infrared absorption spectrum of water. Spectrum acquired in an AgCl cell,
room temperature, 32 scans, ~5 um pathlength ................................ 86

Difference spectrum of water sample showing poor subtraction

characteristics below 1000 cm". 5 pm pathlength, 200K, AgCl sample
cell ................................................................................... 88

Construction of a theoretical light-induced difference spectrum ............ 91

Partially hydrated films of PSII samples, with and without sucrose in the
drying medium. Sample temperature: 25 °C ................................... 95

Difference bands induced by temperature fluctuation of partially hydrated
film of photosystem II. The spectrum was acquired by cooling the sample
to 200K, acquiring one spectrum (500 scans), raising the temperature 0.2
°C, and acquiring a second spectrum (500 scans). The ratio of the two
spectra resulted in the displayed difference spectrum ....................... 104

Temperature control system for low temperature FTIR experiments. . 106

xiii

Figure4
.7-

Figure i

.-1'

Fr

Igurtl
.3

re: a
l 3‘.

F'
igure 54

ft
rum
..5-

Fi
gure i
.-6-

F'
lgurei
.-7'

Fe
12
mi 8
I a‘

F“
We 5

f.
We 5
.9.

Fi
SUI
esio
a
1

Fl
fill!
65.”
a
.l

Figure 4-7: Sample compartment for low-temperature F TIR experiments. . . . . . . . . ....107

Figure 5-1: Schematic of the Photosystem [1 reaction center. Numbers are approximate
molecular weights of the constituent polypeptides. See text for description

of components .................................................................... l 15
Figure 5-2: Schematic depicting the primary photochemical events in photosystem

II .................................................................................... 116
Figure 5-3: Kok model for oxygen evolution .............................................. 119

Figure 5-4: Multiline EPR spectrum of BBY particles, 200K illumination. 8K sample
temperature. Spectrum is the difference between a EPR spectrum acquired
after illumination at 200K, and the EPR spectrum of the sample before
illumination. Microwave power: 2 mW, 20 gauss modulation amplitude,

gain =lx10’ ....................................................................... 120
Figure 5-5: Open “C” model for the manganese cluster of Photosystem II ............ 122
Figure 5-6: Scheme of oxidation states of manganese cluster during the catalytic cycle
of water oxidation ............................................................... 123
Figure 5-7: Proposed sequence of steps in the water oxidation process ................ 126

Figure 5-8a: Dark stable EPR signal from YD in BBY particles. Microwave power
lmW, 2.5 gauss modulation amplitude, 2 minute scan, room
temperature ....................................................................... 136

Figure 5-8b: Tris-washed BBY particles, before, and under illumination. Same spectral
conditions as Figure 5-8a ....................................................... 137

Figure 5-9: Sz-multiline signal from partially dehydrated film of O2 evolving BBY
particles illuminated at 200K. 2mW microwave power, 20 gauss
modulation amplitude, gain=lx105 ............................................ 138

Figure 5-10a) (top) EPR spectrum, 200K, partially hydrated BBY particles, b) tris-
washed BBY particles before (black) and after (red) illumination.
Microwave power = 0.2mW, 4 gauss modulation amplitude, gain =
1x105 .............................................................................. 140

Figure 5-11 a) (top) and b) (bottom) Black: SZQA-/S,QA- light induced FTIR difference
spectra of BBY particles, 200K, 9000 scans. Red: “dark-dark” for noise
reference ........................................................................... 142

xiv

Figure 5-123) (1
\I

Figure 5-l3: .\

a

ll

Figure 5-14a; t

Figure 5-l~lh:

Figures-14c:

Figure 5-1 521:
Figure S-ISb;

Time 5‘16 a’
Flgm‘e 5.173)
FigUIE 5.183)

 

Figure 5-12a) (top) and b) (bottom) Light-induced FTIR difference Spectrum of tris-

Figure 5-13:

Figure 5-14a:

Figure 5-14b:

Figure 5-l4c:

Figure 5-15a:

Figure 5-15b:

washed BBY particles, 200K. 6000 scans .................................... 143

Multiline EPR spectrum of partially-hydrated reaction center cores. 200ug
chlorophyll dried on mylar. Microwave power: 2 mW, 20 gauss
modulation amplitude, gain = 1x10’. Temperature = 8K ................... 146

EPR spectrum of O2 evolving RCCs, before (black) and under (red)
illumination. Integration indicates the formation of chlorophyll radical in
approximately ~50% of the centers. Same spectral conditions as Figure 5-
10 ................................................................................... 147

EPR spectrum of Tris-washed RCCs at 200K, before (black) and under
(red) illumination. Integration of signals shows generation of chlorophyll
radical in approximately 85% of centers. Same spectral conditions as
Figure 5-10 ........................................................................ 148

EPR spectrum of Sr” substituted RCCs at 200K before (black) and after
(red) illumination. integration of signals showed chlorophyll radicals
generated in approximately 75% of the centers. Same spectral conditions
as Figure 5-10 .................................................................... 149

EPR spectrum of Sr” substituted g=4.1 signal in partially hydrated RCC
sample. Spectral conditions same as Figure 5-13 ........................... 150

EPR spectrum of NHl3 substituted g=4.1 signal in partially hydrated RCC
sample. Spectral conditions same as Figure 5-13 ........................... 151

Figure 5-16 3) top and b) bottom. Ll-FTIR difference spectrum of the SZQA-IleA state.

reaction center cores, 200K illumination. 6500 scans total. Red spectrum is
dark/dark for noise reference ................................................... 153

Figure 5-17a) top and b) bottom. (black)LI-FTIR spectrum of tris treated reaction center

cores, 200K.(red-dash) O2 evolving RCCS (Figure 5-16). Tris-spectrum
average of 12 samples, 6000 scans. 200 K ................................... 154

Figure 5-18a) top, and b) bottom.(black) LI-FTIR spectrum of NH20H treated RCCs

(manganese depleted), 200K, 6000 scans. (red-dash) O2 evolving samples
(Figure 5-16) ..................................................................... 155

Figure 5-19a) top, and b) bottom. (black) LI-FTIR spectrum of RCCs soaked in '80-

labeled buffer. 6000 scans, 200K. (red-dash) O2 evolving RCCs (Figure 5-

16) ................................................................................. 156

XV

Figure 5-203)

Figure 5-21 a)

Figure 5-22 3

Figure 5-231a

Figure 5-24:
Figure 5-25

Figure 5-20a) top, and b) bottom. (black) LI-FTIR spectrum of RCCS turned over in
presence of 'SO-labeled buffer. 6000 scans, 200K. (red-dash) O2 evolving
RCCS (Figure 5-16) .............................................................. 157

Figure 5-21a) top and b) bottom. (black) LI-FTIR difference spectrum of NH3 treated
RCCS. (g=4.1 conditions) 3000 scans. (red-dash) O2 evolving RCCS
(Figure 5-16) ...................................................................... 159

Figure 5-22 a) top and b) bottom. LI-FTIR difference spectra of Sr2+ substituted RCCS
(g=4.1) conditions. 4000 scans. (red-dash) O2 evolving RCCS (Figure 5-
16) ................................................................................. 160

Figure 5-23:a) top 800-600 cm’I range of LI-FTIR difference spectrum of tris-washed
and O2 evolving RCC samples. 5000 scans (tris washed) and 6500 scans
(O2 evolving) respectively b) 800-600 cm'l range of NHZOH and O2
evolving RCCS. 4000 scans (NH20H) and 6500 scans (O2 evolving). . ..172

Figure 5-24: 800-600 cm'1 range of LI-FTIR spectrum of RCCS allowed to turnover in
presence of '80 labeled water. (red-dash) O2 evolving RCCS (Figure 5-
16) ................................................................................. 175

Figure 5-25: 800-600 cm’l range of LI-FTIR spectrum of Sr” substituted RCCS. LI-
F TIR spectrum of O2 evolving samples shown for reference .............. 177

xvi

Abstract
Vihral
and the bond
Various t'ibr
Structure of
tihrational
Whoiques;
description
S}‘Stems wi-
Smaller m0]
Sample COn.
fragility of
Pieces of v
plOblemS 0
spectrOSCo

l

CHAPTER 1

INTRODUCTION TO VIBRATIONAL SPECTROSCOPY

OF BIOLOGICAL SYSTEMS

Abstract

Vibrational spectroscopy provides a direct probe of the geometrical environment
and the bonding interactions of a molecular system. One can use the data acquired from
various vibrational spectroscopic studies and draw detailed conclusions about the
structure of the site under investigation. The two primary methods of acquiring
vibrational data from molecules are infrared and Raman spectroscopies. These
techniques, when used in conjunction with one another, can provide a nearly complete
description of the vibrational behavior of the system. However the study of biological
systems with vibrational spectroscopy has, until recently, lagged behind studies of
smaller molecules that used these techniques. This is due to issues arising from the lower
sample concentrations typically achievable with biological samples, as well as the relative
fragility of the samples being studied. In this dissertation are presented several different
pieces of work that demonstrate the utility of vibrational spectrOSCOpy when applied to
problems of biological relevance. These projects are presented in the order of increasing
spectroscopic and technical complexity, and show that biological vibrational

spectroscopy can lend important insight to mechanistic and structural questions.

Vi

3
53.

0TH

bets
can
Qua;

C0114

Vibrational Spectroscopy of Biological Systems:

Vibrational spectroscopy of biomolecules is a field that has seen much progress in
the last twenty years. With the readily accessible technology of lasers, and the wide-
spread acceptance of Fourier-Transforrn spectroscopies, vibrational studies have been
able to shed light on many difficult problems of biological relevance.

A challenge facing all biological spectroscopists relates to signal-to-noise ratios. It
is not uncommon for a biological vibrational spectroscopist to desire information about
several vibrations in a molecule that may have thousands, tens of thousands or hundreds
of thousands normal modes of vibration. Ofien times we wish to seek information about
the active Site of a metallo-protein, or look for changes in the vibrational spectra of an
organic cofactor, or look directly at a single redox-active amino acid in an enzyme. With
enzymes easily reaching molecular weights in the several hundred-thousand Dalton
range, it is necessary to have the ability to detect these small signals with commensurate
sensitivity, while at the same time having methods whereby one can discriminate
between those signals of interest and the remaining (3N-6) signals that constitute what
can be an overwhelming background signal. New developments in the fields of solid-state
quantum detectors for the infrared region of the electromagnetic spectrum and Charge-
Coupled Device (CCD) type detectors for the visible region have been of great use for
infrared and Raman spectroscopists. These technical developments in conjunction with
the judicious use of isotopic labeling and reaction modulated difference spectroscopies
have made the issues of signal detection and signal discrimination much more

manageable.

llhil
utility in ma
properties 01
intestigator

of the particr

lnfrared Sp

lnfrar
molecular 3}
mg differu

If we
haunonic osc
ill for
then at time =
Ihat describes
[hit describes
Places all of
Whiting bet

(lithe meme

 

exFJFESSlOr] ii

dlc.
\
(2) dj

 

3
While both Raman spectroscopy and infrared spectroscopy have proven their

utility in many different biological contexts, they are sometimes limited by the specific
properties of the individual enzyme being studied. The choice as to which techniques an
investigator will utilize for a given biological system depends upon the optical properties

of the particular sample.

Infrared Spectroscopy

Infrared spectroscopy[1] is a technique whereby vibrational transitions in a
molecular system are induced by the application of light of energy that matches the
energy difference between the ground and first excited vibrational states of the system.

If we assume that the total vibrational wave function, ‘P(x,t), of a simple diatomic
harmonic oscillator can be described with a two state equation.
(1) ‘I’(x,t) = co‘I’o(x,t) + c,‘I’,(x,t)
then at time = 0, we will assume that cl = 0 and c0 = l, where ‘I’o(x,t) is the wave function
that describes the system in its ground vibrational state, and ‘I’1(x,t) is the wave function
that describes the system in the first excited vibrational state. This assumption, in effect,
places all of the oscillators in their ground vibrational state, ‘1"). In order to induce a
transition between the two states we need to provide a mechanism that changes the values
of the coefficients c0 and c,. If we apply radiation in order to induce a transition, then the
expression that represents the probability of the system undergoing a transition from
‘I’o to ‘I’, is represented by the transition moment integral

d(°t * cl)

2 2
_°c50 M01

(2) dt

there c1 repres

c,‘ is the comp

represents the
on the particle.
that is indeper

transition inte
M 01

In this case 1h
functions, In

x. ln other it
along a \‘ibra
minim.
Rattan Spec

inflasticallt'

 

SClittered licl

mOIFCUlC is r
I

propolllOnaI

 

4

where cl represents the coefficient of the first excited vibrational state ‘I’, , in equation (1),

c,* is the complex conjugate of c, and M0. is the transition moment integral

(3) Mor = Iexl/lt * l/lodx

we

where

(4) ex = It

represents the dipole moment operator of the molecule. In these equations e is the charge
on the particle, and x is the displacement along the x-axis of the particle. If p. is a constant
that is independent of the vibration of a molecule, then it would factor out of the above

transition integral equation, giving
MOI = I“ IV: 1' Wodx

In this case the transition integral would vanish, as W0 and ([11 are by definition orthogonal
functions. In order that the transition integral be non-zero, u must change as a function of
x. In other words, the dipole moment of the molecule must change as the atoms displace
along a vibrational coordinate in order to observe that vibrational mode in the infrared
spectrum.
Raman Spectroscopy

Raman spectroscopy is a technique whereby incident light is scattered
inelastically from molecules. The energy difference between the incident light and the
scattered light is equal to the energy of a vibrational transition of the molecule. If a
molecule is placed in an electric field then a dipole will be induced in the molecule that is

proportional to the applied electric field.

linen
there a. the
If anelectric
5:
is applied to
it...
Assuming th
of the mole
polarizahilit}
a = t
Where a, is
POlarimbilit}
displacement

for the induc

it...

' IL

With the sub

cost

lhe exPffissio

[11M :

lhis FqUt’itio

i“ l
\l and v.

”induced = as
where a, the proportionality constant, is also known as the polarizability of the molecule.
If an electric field

a = soc0527rvt
is applied to a molecule, then the induced dipole can be represented as

,uind = aaocosvat
Assuming that the polarizability, at, is a property that changes as the internal coordinates
of the molecule change, then we can represent the polarizability in terms of a static
polarizability a0, and a varying part, AOL.

or = a0 + (Aa)c0327rvot
where do is the equilibrium polarizability, AOL is the maximum deviation of the
polarizability from the equilibrium value, and v0 is the vibrational frequency of the
displacement. If we use this factored representation of the polarizability in the expression
for the induced dipole, the result is

pm, = [a0 + (Aar)c0527rvotlg0 cos 27:14]

With the substitution
cos6bos¢ = (-;- [cos(6 + ¢) + cos(6 -— ¢)]
the expression for the induced dipole becomes
Mud = aos‘OCOSZInt + $250530 [cos 27r( v + v0 )t + cos 2;;-( v — v0 )t]

This equation predicts that the induced dipole moment will oscillate at frequencies v,

v+v0 and v-vo. The oscillating electric dipole then radiates electromagnetic radiation of

the same fre
non-zero ca
incident rad
the t'ihratior
a Vibration.
intense. mor

the t'ihratior

samples wit]
source. In or
passed throu
leafs. lasers
already high]
SpFClTOmeter
€0llection 0F
which is the
detected p0
lllfCtroscopy

Oilhe {mm

 

6

the same frequencies, v, v-v0 and v+vo. Thus a vibration (with frequency ya) that causes a
non-zero value for Act scatters incident radiation at l) the same frequency, v, as the
incident radiation (Rayleigh scattering), 2) at the incident frequency minus the energy of
the vibration, v-vo, (Stokes radiation), and 3) at the incident frequency plus the energy of
a vibration, v+vo, (anti-Stokes radiation). The advent of laser technologies to provide
intense, monochromatic radiation, has made Raman spectroscopy an invaluable tool for
the vibrational study of molecular systems.

Early forms of Rarnan spectroscopy were performed by illuminating
samples with single wavelengths of light that originated from a high intensity white light
source. In order to select the single wavelengths of light, the white light continuum was
passed through a grating, and then the dispersed light was passed to the sample. In recent
years, lasers have made the experiment much easier as they readily provide very intense,
already highly monochromatic beams of light. See Figure 1-1 for a schematic of a Raman
spectrometer system. The scattering from the sample is collected with the appropriate
collection optics, then passed to a monochromator which disperses the scattered light
which is then detected with a photomultiplier tube one wavelength at a time, or it is
detected polychromatically with a CCD-type detector. Analogous to infrared
spectroscopy, the probability of Raman transitions is determined by a probability integral

ofthe form
P: IWr.dI/’odx

where a is the polarizability operator for the molecule. If a transition is to be Raman

 

 

 

 

 

 

 

 

 

Laser /R\
CD
collection optics< || 1 ll
Raman
scattering
monochromator
\
l__l
detector

Figure 1-1: Schematic of Raman scattering spectrometer.

active. the '
the molecul
If th
the sample.
scattering a
that is as m
is observed
correspond

scattering is

FOllOMng eq

96

g and if
Electronic Stg
r eSOttance te
PTOVides a 5,
radiation. Vt.
initial and 6X
3” approxin

 

equal liter] &.

 

8
active, the transition from we to w, must induce a change in the overall polarizability of

the molecule.

If the light incident on the sample has the same energy as an optical transition of
the sample, then vibrations that are coupled to that optical transition will exhibit Raman
scattering at much higher intensities. This intensity enhancement can result in scattering
that is as much as 103-10" greater than the scattering from the same vibrational mode that
is observed in an experiment in which the energy of the incident light does not
correspond to an optical transition. The term that governs the intensity of Raman

scattering is (01,”),‘ , the polarizability tensor component for the transition k. In the

following equation

:1 (flMperrlMalg) +(flMaerrlMplgl ‘
mm)" 712i i

v,-vg -v0 +il: Vr‘Vf +v0 +11“,

 

lg) and | f > are the initial and final states of the molecule, and |r> is the excited

electronic state from which scattering occurs. The first term of this equation is called the
resonance term. The second is a nonresonant term that, under conditions of resonance,

provides a small contribution to the overall signal. When the frequency of the incident
radiation, v0 is significantly different from the difference in the frequencies between the
initial and excited states, v, — vg , then the first and second terms of the above equation
are approximately equal in their contribution to the overall scattering phenomenon.
However, when the incident energy and the energy difference v, - vg are approximately

equal, then as the denominator of the first term goes to zero, the first term dominates the

9

scattering process. The result is a greatly increased change in the polarizability, which
results in a large increase in scattering efficiency.

The resonance Raman effect has been used widely in the spectra of heme proteins
and other metalloproteins[2] that exhibit intense absorption in the visible region of the
electromagnetic spectrum. In these systems, Raman scattering that arises from the active
site of the enzyme have been observed by using small ligands that exhibit characteristic
vibrational signals, isotopic labeling, and Site-directed mutagenesis. However, in systems
that have cofactors or pigments that exhibit high extinction coefficients, yet where the
signals of interest are not coupled to the optical transitions of those cofactors, one must
be careful to use an excitation wavelength that avoids regions of any pigment absorbance.
Failure to do this will result in spectra that are dominated by cofactor scattering, and
scattering from the species of interest will be difficult to detect. Photosynthetic systems
are such an example. The light harvesting apparatus of photosynthetic systems is
designed to absorb incident light efficiently. Over millennia, the pigments utilized to
perform the primary light harvesting role have developed absorption spectra that, while
characterized by certain absorption maxima, display significant absorption throughout the
visible region of light. As a result, Rarnan scattering from photosynthetic samples contain
vibrational modes from the pigment bed, and any signals arising from molecules in the
enzyme that are not themselves efficient absorbers of light are expected to be negligible
relative to the more intense resonance enhanced vibrations of the chlorin pigments.

In systems where both infrared absorption and Raman scattering spectroscopies
can be utilized, the spectroscopist has the luxury of being able to obtain a relatively

complete picture of the vibrational properties of the system being studied. Both of these

techniques
situational
transition

dipole moi
for Ramar
elements t
contains tl
transition 5
this symn
SWCtmm

infrared 5]
slttttrum .
SPlectrosco

Wine of

[0

techniques depend upon non-zero values of their respective transition integrals for a
vibrational transition to be observable. The only difference between the respective
transition integrals lies in the operators upon which the phenomena rely, namely, the
dipole moment operator, ,2: , in the case of infrared, and the polarizability operator, 5:,
for Raman spectroscopy. If the triple product of the symmetry of LI", the symmetry
elements of the dipole moment or polarizability operator, and the symmetry of ‘Po ,
contains the totally symmetric irreducible representation of the point group then the
transition integral will have a non-zero value, and thus will be observable. As a result of
this symmetry dependence, it is possible to obtain vibrational data from the Raman
spectrum of a system that is not observable in the infrared spectrum. Likewise, the
infrared spectrum frequently may contain data that are not observable in the Raman
spectrum of the same molecule. Therefore, if it is possible to perform both forms of
spectroscopy on a given system, the data obtained will usually present a more complete
picture of the vibrations of the system than if one of the techniques is used exclusively.

A significant advantage that Raman spectroscopy has over infrared spectroscopy
for the study of biological molecules relates to the water content of the samples. Water
displays strong absorbances throughout the mid-infrared spectrum (Figure 1-1). The
Raman scattering from water however is fairly weak, which makes Raman scattering the
technique of choice for investigating the vibrational properties of biological systems
where the presence of water is necessary to ensure biological activity and relevance. But
in cases in which the vibrational signals of interest are not coupled to the electronic

transitions of a highly absorbing chromophore, and such chromophores are in fact

present infran
are not limiter
able to exanti
from enzymes

challenges nee

Fourier Tran
hwm
SPecific functi
infrared light
developed intt
abSorption 5p,
and accuracy ,
exll'itiments
Slr‘eCUOrneters
“hich Separat
then 59m thrt;
mm“ “It: dit‘

absorbed as n

 

 

speculum? [hi‘

 

c(liltinUous m,

ll was “OI Um:

l 1
present, infrared spectroscopy must be the technique of choice. Infrared spectroscopists

are not limited to samples that contain strongly absorbing chromophores, and thus are
able to examine a wider range of enzymatic samples. However in order to gather data
from enzymes with infrared spectroscopy, several daunting technical and biochemical

challenges need to be addressed.

Fourier Transform Infrared (FT IR) Spectroscopy:

It was first reported in 1882 that a correlation existed between the presence of
specific functional groups in a molecule and the absorption of specific frequencies of
infrared light [3]. However, it was not until 1903 that modern IR spectroscopy was
developed into the form in which it is currently used. W.W. Coblentz investigated the IR
absorption spectrum of hundreds of inorganic and organic substances[4] and the precision
and accuracy of his work allows many of these spectra to still be considered useful. These
experiments and subsequent experiments until the mid 1960’s relied on infrared
spectrometers that consisted of a source of infrared radiation and a dispersive grating
which separated the light into it’s constituent wavelengths (see Figure 1-2). This light is
then sent through the sample and each wavelength is detected one resolution unit at a
time. The difference between the incident and the transmitted light is the light that is
absorbed as result of vibrational activity in the sample. In order to acquire the complete
spectrum, the dispersive grating is scanned, changing the frequency of the light in a
continuous manner, until the complete infrared absorption spectrum is obtained. However

it Was not until the advent of Fourier-Transforrn infrared spectroscopy that it was possible

bro:

Source

 

l2

broadband infrared radiation

\ / grating

    
 
  
 

Source
H incident intensity =Io

sample N
/transmified intensity: I

detector / -

 

transmittance: 1;

 

 

 

Figure 1-2: Schematic of scanning infrared spectrometer

13
to apply infrared spectroscopy to biological systems routinely. The inherent advantages

offered by this technique allow spectra with sufficient signal-to-noise ratios to be
acquired in an acceptably short amount of time.

FTIR spectroscopy is a technique that is based upon the principle of
interferometry which allows a continuum of radiation to be convoluted in to a single
signal that is detected instantaneously. This signal, which contains all of the spectral
information, is then deconvoluted through the use of a Fourier transformation. This
instantaneous detection of all of the frequency components lends certain distinct
advantages in the signal-to-noise ratio of the acquired spectrum. Many FTIR
spectrometers are based upon the Michelson interferometer (Figure 1-3). This is an
interferometer that consists of a moving mirror and a fixed mirror that is held
perpendicular to the displacement vector of the moving mirror. Bisecting the angle
between the two mirrors is a beamsplitter which is assumed to be ideal, i.e. it transmits
50% of the incident light and reflects 50% of the incident light. To illustrate how this
system works, we will first consider a beam of monochromatic radiation[5]. When the
beam of radiation strikes the beamsplitter, 50% of this light is reflected to the stationary
mirror and 50% of the light is transmitted to the moving mirror. If the center of the
beamsplitter is designated point X, and the centers of the moving mirror and the fixed
mirrors are points Y and Z respectively, then the difference between the distances that the
beams of light must follow are 2(XY-XZ). This difference is called the retardation, 6. An

assumption that is made in this treatment is that the incident radiation, while not

infrar
light ft
sourt

beamsi

 

14

fixed mirror

 

 

, ‘1
infrared Z i .
llght from , W . axis of motion
source ,
a X
v .J '
beamsplitter moving mirror
modulated
infrared light
to sample and
detector

Figure 1-3: Schematic of Michelson Interferometer

 

necessarily ir
mirrors are e
recombined v
incident ener;
minor is disp
radiation. Wilt
at the detectt
pathlength is
the detected e
mil equals
at zero dispt,
milmaphot
mOt‘ing mlm
modif- the
Signal being

that is 311 line

til intennedia

 

lid)

 

“”3 Signal c.
mfg)“ of the

ii is [his mo

15

necessarily infinitely narrow, is, at a minimum, collirnated. When the fixed and moving
mirrors are equidistant from the beamsplitter (2(XY-XZ)=0) then the radiation, once
recombined will remain perfectly in phase and will constructively interefere. All of the
incident energy will thus be passed through to the detector. However, if the moving
mirror is displaced by a distance £1. , (hence the difference in pathlength is $4.) then the

radiation, when recombined, will destructively interefere and no energy will be observed

at the detector. If the moving mirror is displaced by %1 then the total difference in

pathlength is a. The result is that the waves will once again constructively interfere, and
the detected energy will be equal to the incident energy. In the case for which the incident

energy equals the detected energy, it is not possible to determine if the moving mirror is
at zero displacement or some multiple of %.t. In practice, however, FTIR instruments

utilize a photoelectric sensor circuit to determine the starting and stopping point of the
moving mirror, thus eliminating this problem. When the mirror is scanned at constant
velocity, the detected signal will be seen to vary sinusoidally, with the maximum of this

signal being detected each time the moving mirror passes through a displacement value
that is an integral multiple of %a (so the total displacement 5 is an integral multiple of a ).

At intermediate values of 6, the intensity of the signal is given by

I'(c5') = O.51(r7){l + cosan} = 0.51(r7){1+ cosZrti76}

This signal consists of two parts, a constant, steady-state component equal to one-half the
energy of the incident radiation, and a modulated component equal to 0.51(r7)c0327r 76.

It is this modulated component that is referred to as the interferogram. In reality, a

‘petiettu btttl
sensitive detet

necessan' to i

 

 

obscned at th
hi i ) in the i

W) =
We can con
beam spectr
It 0“:
llatltematit

calculated

lllf llllerfe'

more den

Signal—[0

Fo
Finns ofsl

16
“perfect” beamsplitter is not possible. Additionally, perfectly reflecting optics, infinitely

sensitive detectors and noise free electronics are also not feasible. With this in mind, it is
necessary to incorporate these realities into the expression for the intensity of radiation
observed at the detector. We must include a wavenumber dependent “reality” coefficient
H( V) in the expression for the signal intensity.
1(5) = 0.51(t7)H( t7)c0527r r75
We can combine the two terms I( V)H(t7)into one B(t7) which we will call the single
beam spectral intensity function. With this, the expression for the interferogram becomes
1(5) = B( V)c0827r r75
Mathematically, 1(5) is the cosine Fourier transform of B( V) and the actual spectrum is
calculated by cosine Fourier transforming 1(5). When a polychromatic source is utilized,

the interferogram can be represented by an integral of the form
1(5) = Camcoszzva . d?

The other half of the Fourier transform pair is
B( v) = 1:1(5)cos2m75 - d5

The spectrum is, as before, computed from this half of the F ourier-transforrn pair. For a

more detailed description of the data treatment methods, see[5].

Signal-to-noise considerations in FTIR
Fourier transform infrared spectroscopy maintains several distinct advantages, in

terms of signal-to-noise ratios, over conventional dispersive infrared experiments. These

ads:
tool
ads;
syst

Fou

tithe

Spt’i

“he
the

oft?

Bec

adt;

of ti

sin

l7
advantages are what has allowed biological infrared spectroscopy to become a promising

tool for the investigation of enzymatic structure/function determination. The first of these
advantages relates to the enhanced throughput achievable with interferometer based
systems, and is known as Jacquinot’s Advantage[5]. The maximum throughput of a

Fourier spectrometer is

O, = 27rAl AV

V

 

cm2 sr

where A1 is the area of the reflective surfaces. The analogous expression for grating

spectrometers is

hAGAT/
fai72

96 = cmzsr

 

where A“ is the area of the grating, f is the focal length of the collimating mirrors, [1 is
the slit height and a is the grating constant. If one compares the throughput advantages

of the Fourier transform instrtunent relative to that of the grating instrument, one obtains

o, _ 27rA'fai72
ea AGhT/m

 

Because of the V2 (where?! is the frequency of light being used to interrogate the
sample) dependence it is clear that Fourier transform spectroscopies have large
advantages in throughput over comparable grating spectrometer-based systems.

Another advantage of Fourier transform spectroscopies relates to the fact that all
of the wavelength components of a given spectral range are examined simultaneously.
Fourier spectrometers detect all of the wavelength components of light simultaneously,

while conventional scanning spectrometers require the grating to be moved each time a

l8
different wavelength of light is to be detected. Because this movement of the grating

takes time, the Fourier instrument is able to acquire the complete spectrum in the same
amount of time that the scanning instrument is able to scan a fraction of the same region
of the spectrum. Because the signal-to-noise ratio of a spectrum is proportional to the
square root of the time spent gathering the signal, this multiplexing allows much higher
signal-to-noise ratios to be achieved in a given amount of time. This advantage, called the
multiplex or Fellgett’s advantage, is another key aspect of Fourier spectroscopies, and it
is another reason that infrared spectroscopy has only recently become useful for studying
samples of biological origin[5].

The first two chapters of this dissertation show how FTIR Spectroscopy in
conjunction with Raman Spectroscopy can provide critical insight into the structure and
firnction of two very different systems. The first system examined is a series of model
compounds that represent the cyanide-ligated form of the oxidized binuclear center of the
respiratory enzyme cytochrome c oxidase. FTIR spectroscopy was utilized in this study in
a technically simple way, but when the infrared and Rarnan data were combined and
analyzed by using sophisticated vibrational analysis techniques, key insights into how
cyanide may be bound in the enzyme emerged. In chapter three, somewhat more
technically challenging FTIR measurements were performed, again in conjunction with
Raman spectroscopy, to analyze the binding mode of carbon monoxide to the enzyme
soluble guanylate cyclase (sGC). In chapters four, five, and six we have extended F TIR
spectroscopy to its technical limits in an effort to elucidate structural and mechanistic
aspects of photosynthetic water oxidation. These studies introduce the concept of reaction

modulated infrared difference spectroscopy and represent one of the most technically

19
challenging FTIR experiments currently being performed. We have developed new

methods whereby we can examine the low frequency region of the infrared spectrum,
where the vibrations of possible intermediates in the water oxidation process and where
the vibrations of the manganese containing active site of the reaction center will be
observed. This technique provides an entirely new genre of information regarding how
plants and cyanobacteria oxidize water. The techniques developed here will allow us to
also extend this type of spectroscopy to other enzymatic systems. Many other
metalloproteins will exhibit vibrations in the low frequency region of their infi'ared
spectrum. These vibrations, as in photosystem 11, may contain information regarding
substrate binding, active site structure, as well as information regarding the catalytic

mechanism. With these techniques, this information is now accessible.

 

U)

20

LIST OF REFERENCES

Harris, DC. and Bertolucci, M.D., Symmetry and Spectroscopy. 1978, New
York: Dover Publications.

Spiro, T.G., Biological Applications of Resonance Raman Spectroscopy, ed. T.G.
Spiro. Vol. 3. 1988, New York: Wiley.

Abney, W.d.W. and Festing, E.R., Phil. Transactions, 1882. 172: p. 887.

Coblentz, W.W., Investigations of Infrared Spectra, 1905, Carnegie Institute:
Washington, DC.

Griffiths, PR. and de Haseth, J .A., Fourier Transform Infrared Spectrometry.
1986, New York: Wiley-Interscience.

Al

ma:
res;
assi

libr

of is
\‘jbra
mod:

T6501]

CHAPTER 2

VIBRATIONAL ANALYSIS OF A MOLECULAR HEME-COPPER
ASSEMBLY WITH A NEARLY LINEAR Fem-CN-Cuu BRIDGE:
INSIGHT INTO CYANIDE BINDING TO FULLY OXIDIZED

CYTOCHROME C OXIDASE

Abstract

The complete vibrational analysis of [(l-MeIm)Fe(OEP)—CN-Cu(Me,,tren)]2+ (l),
which has been constructed as a model for the cyanide-ligated binuclear center in the
respiratory protein cytochrome c oxidase, has been carried out. The resonance Raman
spectra (Am: 647 nm) and the mid-infrared spectra display three cyanide isotope
dependent vibrational modes. Two vibrational frequencies decreased with increasing
mass of the cyanide ligand (2182-2137-2146-2101 cm" and 535-526-526-520 cm“,
respectively, for the l2C”N-”C"’N-”C”N-‘i’C'sN isotopomers), and could thus be
assigned to the CEN and Fe-CN-Cu stretching vibrations, respectively. The third
vibration, detected with resonance Raman, exhibited frequencies more sensitive to 13C
labeling of the cyanide ligand than with 15N labeling (495-487-493-485 cm'1 with the set
of isotopomers above). This pattern of isotopic dependence is characteristic of a bending
vibration. Additionally, with the same laser excitation frequency, the C:-:N stretching
mode was observed, which is the first time that this vibration has been detected in the

resonance Raman spectrum of a synthetic heme-cyanide complex. The normal coordinate

21

3.1”.

lil

In

TEE

prc

is;

22
analysis showed marked differences between bridged and unbridged heme-cyanide

complexes. Internal coordinates that are orthogonal in unbridged systems are significantly
mixed in the bridged model, despite the overall linearity of the Fe-CN-Cu moiety. These
measurements strengthen the proposal that cyanide bridges the two metal atoms in the
cyanide-ligated, oxidized binuclear center of cytochrome c oxidase. A quantitative
consideration of the vibrational characteristics of cyanide bound to the resting enzyme in
light of our model compound results strongly suggests that the binuclear center is flexible
and can undergo structural rearrangement to accommodate exogenous ligands. This is
likely to be of mechanistic importance in both dioxygen reduction and proton

translocation.

Introduction

Cytochrome c oxidase is the terminal electron acceptor in the respiratory electron
transfer chain of prokaryotic and eukaryotic organisms. This membrane-bound enzyme
reduces molecular oxygen to water and utilizes the resultant gain in free energy to pump
protons across the cell membrane [1-6]. The active site of this enzyme, at which oxygen
is reduced to water, consists of a binuclear CuB/heme a3 center. The recent solution of the
crystal structures[7, 8] of cytochrome c oxidase from two sources has provided new
insight into the arrangement and structure of the binuclear center of the enzyme. The
distance between the two metals is 4.5 A in the resting bovine enzyme[7] and 5.2 A in the
bacterial system[8]. The close proximity of the two metals had been deduced from earlier
spectroscopic, biochemical, and molecular biological studies and a number of models

have been advanced for their function in both 02 reduction and in the redox-coupled

23
translocation of protons. A central issue is the extent to which the partially reduced

dioxygen substrate interacts with the two metals. At the peroxide level in the reaction
sequence, for example, a range of structures, from a u-peroxo firlly covalent linkage to a
terminal hydroperoxo (Fe-O-O-H) with minimal interaction between the CuB and the
heme a3 moiety, has been proposed[1-4].

Cyanide, a potent inhibitor of oxidase activity, has often been used as an alternate
ligand to the heme of the binuclear center[9-14] and can provide insight into the modes
by which small molecules bind. The Soret band in the optical spectrum of heme a3 shifts
upon cyanide ligation as the heme iron undergoes a high-spin to low-spin transition. This
shift in the optical and magnetic properties of the site has formed the basis for a variety of

optical, electron paramagnetic resonance, magnetic-circular dichroism and magnetic

susceptibility studies of the CN- ligated enzyme[l-4]. A firrther useful spectroscopic

probe is the CEN stretching vibration[15], which is characterized by a strong, sharp

absorption around 2030-2200cm". This has allowed the use of infrared spectroscopy to

study the interactions of bound CN- in the binuclear center[9-14]. Although the

prevailing view of the mode of cyanide binding in the fully oxidized enzyme invokes a
bridging conformation[9], models have been advanced that postulate monodentate
binding[10]. If cyanide does, in fact, bind in a bridging configuration, the vibrational
behavior of the ligand in terms of its geometry remains to be elucidated. These are
questions that can be addressed through careful vibrational spectroscopic analysis,

isotopic labeling, normal coordinate analysis, and model compound studies.

A series t
been prepared
cttochrome c 0\
llelm). L is a
coordinate C uill
molecular bridge
characterized b}
porphyrin (S 0.‘
anariant bond le
dependent on 1
investigation of
llelm)Fe(0EPl,-
lTigure 2-1. The
from the closelt
mil ligands is
Which With the
internal“ Ot'er 5.
because of the p
and [he simpitt
enz.l'trte[19.2 l l
Culhleénenflz. [l

budging O"“3123.

 

24
A series of molecules of the general type [(N-base)Fem(OEP)-CN-Cu"LL'] have

been prepared as analogues of the cyanide-bound binuclear center of oxidized
cytochrome c oxidases[16-l8]. Here the N-base is pyridine or l-methylirnidazole (1-
Melm), L is a tridentate or tetradentate polyamine ligand bound to four- or five-
coordinate Cu(II) and L' is a unidentate anion, when present. A number of these
molecular bridged assemblies have been structurally defined by crystallography. They are
characterized by a six-coordinate low-spin Fe(III) atom in the mean plane of the
porphyrin (S 0.1 A displacement), an essentially linear Fe-C.=_N group (176-180°) of
invariant bond lengths, variable Cu-N distances (1.88-2.17 A), Cu-N-C angles (147-174°)
dependent on ligand L, and Fe---Cu separations of 4.93-5.11 A. For this initial
investigation of the vibrational properties of Fe-CN-Cu bridges we have examined [(1-
MeIm)Fe(OEP)-CN-Cu(Me6tren)]2+ (l) [18],whose structure is presented schematically in
Figure 2-1. The structure of this complex has not been determined; structural data are
from the closely related molecule with a pyridine axial ligand[18]. This difference in
axial ligands is expected to have a negligible effect on the Fe-CN-Cu bridge parameters
which, with the exception of the Cu-N bond length and the Cu-N-C angle, are essentially
invariant over some six structures. The imidazole-ligated bridge assembly was chosen
because of the presence of a proximal imidazole at heme a3 in the binuclear center[7, 8]
and the simplifying feature of the nearly linear bridge. As in the cyanide-inhibited
enzyme[19-21], the metal spins are ferromagnetically coupled in [(py)Fe(OEP)-CN-
Cu(Me,,tren)]2+ to afford an integer spin ground state[22]. Analogous assemblies with

bridging oxo[23, 24] and hydroxo[25] groups have been prepared and structurally

 

Figure
C“(Me
maSSeS

25

Me C
N!
Fell]
L
,N, No Ni.
179 174’
- p A b , ,
Im~ Fe- .0.-- N ‘m - -.N
" . 1.92 +1 ”1.14 “1.88 -: ~L
Np Npl

N, N.

Figure 2-1: Top: Schematic representation of [(l-MeIm)Fe(OEP)-CN-
Cu(Me6tren)]2* Bottom: Model used for normal coordinate analysis. Point
masses represent atoms or fractional representations of ligands

characterized. Tl
enzyme. These
cyanide coordine
of the binuclear t

in the \\
Spectroscopies tc
used the \‘ibratic

metalligand cor

 

refined set of fo
well as possible

details of cyanic‘

Materials and

The co
from RH. H01,
eqUtmola; lCt

MEPKOCI

Set-era] VOlum
violet “EEdleg

an Spear;
Stream of Colt

liith dry KBr

26

characterized. Their magnetic properties are similar to those of the oxidized resting
enzyme. These bridged assemblies, being structurally proven examples of bidentate
cyanide coordination of a type possible in the enzyme, can lend insight into the structure

of the binuclear center of the cyanide-inhibited protein.

In the work reported here, we have applied resonance Raman and infrared
spectroscopies to l and varied the isotopic composition of the cyanide ligand. We then
used the vibrational data to carry out a normal coordinate analysis[26] of the binuclear
metal/ligand core of the complex. The calculated potential energy distributions and the
refined set of force constants give insight into the nature of the bonding interactions as
well as possible non-bonding, or steric constraints. The results are used to probe the

details of cyanide ligation to resting cytochrome c oxidase.

Materials and Methods

The compound [(1-MeIm)Fe(OEP)-CN-Cu(Me6tren)](C10,)2 (l) was obtained
from R.H. Holm and co-workers at Harvard University. It was prepared by the reaction of
equimolar [Cu(Me6tren)(CN)](ClO,)[18] (containing the desired cyanide isotopomer),
[Fe(OEP)(OClO3)], and l-methylirnidazole in acetone solution, followed by diffusion of
several voltune equivalents of ether over several days. The product separated as dark
violet needles. [(1-MeIm)Fe(OEP)CN] (2) was synthesized as described[l8]. Resonance
Raman spectra were obtained with solid state samples, cooled to -80 °C in a dewar by a
stream of cold dinitrogen gas. The samples were prepared by grinding the compounds

with dry KBr powder and pressing to form a pellet. To obtain Raman Spectra, a Coherent

Kr' laser
on to the
dispersed
cooled C I
were acqt
nitrogen t
DTGS de
were coa
functions

.\'
funnalisn
modificat

\ihration;

27

Kr’ laser (Coherent K-90) was used for excitation at 647 nm. The laser light was focused
on to the sample with an incident power of 30 mW. Resonance Raman scattering was
dispersed with a Spex 1877 Triplemate spectrometer and detected by a liquid nitrogen
cooled CCD detector (EG&G OMA 4, model 1530-CUV-1024S). Mid-infrared spectra
were acquired on a Nicolet 520P FTIR spectrometer, equipped with a DTGS or liquid-
nitrogen cooled MCT—B detector, or a Nicolet IR—40 FTIR spectrometer equipped with a
DTGS detector. The samples were prepared as CsI pellets. Double sided interferograrns
were coadded at a resolution of 2 cm'I and apodized with Happ-Ganzel apodization
functions.

Normal coordinate analysis was performed by using the Wilson GF matrix
formalism[26]. All calculations were performed on VAX/V MS systems, using a
modification of the Shimanouchi program[27]. The GF matrix method of interpreting
vibrational spectra consists of constructing and solving the secular equation
(1) |GF - Ell = 0
where G is the matrix that describes the geometry of the molecule being studied in terms
of its internal coordinates, F is the matrix of force constants used to represent the bonding
and nonbonding interactions between the atoms of the system, E is the unit matrix and ,1
is the frequency of the vibration in question. The construction of such a secular equation
is a non-trivial task that has been greatly facilitated by the use of computers and
appropriate programs designed for just such a purpose.

In order to perform a normal coordinate analysis, one must calculate both the

kinetic and potential energy functions for the molecule being studied[28]. In terms of the

llOlTI

mole

ti'hert

mam".i
canesi
system
energy .
t4; 2

and

28

normal coordinates of a molecule we will define the kinetic and potential energies of a

molecule as
(2) 2T = 2Q} and 2V = 2,1,9}
(where i = l to 3N-6). In matrix form, these expressions are

(3) 2T = Q'Q and 2V = gig

. d .
where Qis the column vector of the normal coordinates, Q = -dQT' , and i is the diagonal

matrix of ,1 values. It is more convenient, however, to construct a molecule within a
cartesian coordinate system initially, as the atomic positions within this coordinate
system are easy to interpret. In this case the expressions for the kinetic and potential
energy functions of a molecule are quite Simple and easily recognizable

(4) 2T = _)_(_='_I_\4__X

and

(5) 2v = X'FXX

 

where T is the kinetic energy of the system, V is the potential energy, g is a diagonal

matrix of the masses of the constituent atoms and F__: is a non-diagonal matrix of the

force constants between the atoms. The molecule treated here is assumed to be a
diatomic; expanding this treatment into three dimensions is straightforward. However,
because the modes of vibration of a molecule are combinations of motions along the x, y,
and z axes of the Cartesian coordinate system, this matrix of force constants is not as

intuitively interpretable as if the molecule and its force constants were described in terms

[De

29

of bond stretches and angular bending motions. To make the force constant matrix as
intuitive as possible, it is necessary to convert from an external frame of reference to an
internal description of the molecule. This is accomplished by developing a set of internal

coordinates for the molecule such that

(6) 2T = R'KR

and

70

0) 2V=WF

where gis the column matrix of the internal coordinates and E and __F_‘ are the kinetic and

potential energy matrices respectively. Internal coordinates can be thought of as
coordinates that are unaffected by rotational or translational motions of the entire

molecule. These internal coordinates obey the relation

3N
(8) R, = Zena, r =1,2,3,....,3N —6

i-l

where the coefficients Bti are determined by the geometry of the molecule. Both the __K
and the E matrices, in the context of the internal coordinates, are non-diagonal, but the g

matrix can now be thought of in terms of primary and interaction force constants. With
this strategy in place, we merely need to evaluate the respective contributions of the
kinetic and potential energies in terms of the internal coordinates in order to arrive at the
solution to the vibrational problem.

The momentum of a diatomic molecule can be expressed in terms of a column
matrix

(9) P. are;

LIKE“

andm

Note

mome
transfc

(ll)

where
coordin.
H2) _,

and

30
Likewise, the kinetic energy of the molecule can be thought of as functions of the mass

and momentum of the system

(10) 2T = P;M" P,

Note the similarity to the standard expression for kinetic energy T=%mv2. The

momentum matrix can be also expressed in terms of the internal coordinates, via the

transformation matrix g

(11) P.

ll
"‘7?
ll "e

where the 2 matrix defines the internal coordinates, g , in terms of the cartesian

coordinates of the system, or,

(H) am
and
(n) R=BX

From equations (9) and (11), we get

(M) a; = a:

Solving for g , and inserting the result into (13) gives us

(5) §=BM4aP

Substituting equation (11) into (10) we can rewrite the kinetic energy expression as
(16) 2T = P'BM'l B'P

We can now define the g matrix as

(W) g=8M4E

Using this we ca

 

OI

H I»)
A H
O it
12; ”w

l

H
=9
CD

 

H7:
H
O

This reSUlI ShO
Coordinates. In

the Vibrational

31

Using this we can rewrite (16) as
(18) 27' = P’GP
Using equations (17) and (15), we can now write

R=_Q__13

01'

"'1:
ll
9.
2:

Therefore we can express the kinetic energy in these terms
2T = 2:95
= (flr'gtfl)
= Re" R

 

When we compare this last result to equation (6) we can see that

(19) i = 9.7".
This result shows that we can express the kinetic energy in terms of the internal
coordinates. In order to derive the secular equation that we must solve in order to solve

the vibrational problem, and thus determine the normal coordinates of the molecule, we

must define a transformation matrix, ; , such that
(18) a = __L_Q

This allows us to relate the internal coordinates to the normal coordinate Q. As a result of
this transformation, we can now define the potential energy in terms of these normal
coordinates. Taking the previous expression for the potential energy, and inserting our

definition for the normal coordinates we get

(19) 2V = R'FR =(gg)'__1f(LQ)= ggflg

32

By comparison with the potential energy expression in equation (3) it can be seen that
<20) '11 = a

With equations (6) and (19), we can now rewrite the kinetic energy term.

 

 

 

 

Comparing this equation with the kinetic term in equation (3), one sees that
(22) L'G’I L =£
where 1 is the identity matrix. Using equation (20) we see

(23) ;g = gg = L'G" La

Multiplying through on the left by (gf' gives us
(24) as = %
Lefi multiplying by G, leaves
<25) are = 22
This is the secular equation must be solved for the vibrational problem. Diagonalization
of the GF product will yields the diagonal matrix A , whose diagonal elements are the
fiequencies of the normal modes of vibration.

The initial set of force constants was based upon the modified quadratic fields
found in found in [12] and [29]. The dependence of the calculated vibrational frequencies
were evaluated as a function of the various force constants employed. Those force

constants that were found to have a negligible effect upon the calculated frequencies were

not varied as part of the iterative process. Those force constants found to affect the

 

calcula
betyyee

each 01

Result
UV-V’i
The L“
absorpt
absorpt
Cyanide
include
absorpt
absorpt
Show a
are 5110

Ci’anide

[Mitre

ln

hate fre
of these

{Willem

33

calculated frequencies were systematically adjusted to achieve the optimal agreement
between the calculated frequencies and the experimentally determined frequencies of

each of the isotopically labeled species.

Results
UV-Visible Spectroscopy:
The UV-visible spectrum of the title complex (Figure 2-2) is characterized by a Soret
absorption maximum at 403 nm, visible bands at 526 and 553 nm and a near ultraviolet
absorption at 345 nm. When compared to the absorption spectrum of the unbridged heme
cyanide complex 2, despite general similarity, several differences are notable. These
include a small (3 nm) blueshifi in the Soret maximum, a slight (3 nm) redshift of the UV
absorption, and subtle bandshape differences of the visible peaks[l8]. The visible
absorption bands of both the bridged binuclear complex and the unbridged heme complex
show a trailing edge and small features at approximately 650 nm. These absorption bands
are shown below to be important in the context of resonance Raman detection of the

cyanide stretching and bending modes of this molecule.

Infra red Spectroscopy:
Infrared and resonance Raman spectra of 1 display three vibrational modes that
have fiequencies sensitive to isotopic labeling of the cyanide moiety (Table 2-1). The first
Of these appears at 2182 cm'I in the mid-IR spectra of the molecule (Figure 2-3). From its

li'eq uehcy and relative intensity, this mode can be assigned to the CEN stretching

34

8.0x10’

------------- [(py)Fe(OEPrCNr
. —— [(py)Fe(OEP)-CN-Cu(Me6tren)]

6.0x10’

4.0x10‘

8 (l-mol'l-cm'l)

2.0xio“

 

 

0.0

 

400 500 600 700

wavelength (nm)

Figure 2-2: UV-Vis absorption spectra of “bridged”, [(py)Fe(OEP)-CN-
Cu(Me6tren)], and “unbridged”, [(py)Fe(OEP)-CN], complexes. Sample
concentrations were 2.2 mM (bridged) and 1.9mM (unbridged). Spectra were
dissolve in acetone and spectra were recorded in a €an cell, SOuM pathlength.

 

 

yibr
min
the

ligar

St’CO

and
Sir 61E

isotot

stretc
comp.
the in‘
are cc
COmpl
detern
€11an

SOll‘en-

3S

vibration of bound, bidentate cyanide. When the cyanide is substituted with 'ZC‘SN the
vibration shifts to 2146 cm"; with l3C"’N it shifts to 2137 cm" and with l3C'5N cyanide
the mode appears at 2101 cm". This frequency decrease with increasing mass of the
ligand is expected for a CaN stretching vibration in simple heme-cyanide complexes. The
second vibrational mode that displays an isotope dependence appears in the infrared
spectrum at 535 cm" (Figure 2-4). With l3C”N or 12ON the mode appears at 526 cm",
and with the doubly labeled cyanide the mode shifts to 520 cm". As with the CEN
stretching mode, this absorption displays a mass-dependent decrease in frequency upon
isotopic labeling and is assigned to a stretching vibration of the Fe-C-EN-Cu bridge unit.
The extinction coefficients for the infrared absorbances attributable to the CEN
stretches of l and 2 were determined by dissolving known amounts of the crystalline
compounds in acetone. Infrared absorption spectra of these solutions were acquired, and
the integrated intensity of the cyanide stretching mode absorptions were calculated. These
are collected in Table 2-2, along with relevant extinction coefficients from related
complexes[30]. The effect of hydrogen bonding on the vc.N stretching frequency was
determined by dissolving 2 in neat methanol. A 3 cm'1 shift to higher frequency was
eVident in methanol, indicating a distal interaction between the hydrogen bond-donating

solvent, and the nitrogen atom of the cyanide ligand (see Table 2-3).

 

‘

a
”mi

Egon
mom)

Table

pgnc
anc
RdhC

36

Table 2-1 Isotopic Dependence of Vibrational Frequencies'l

 

 

VIBRATIONAL IR/RR
MODE DETECTED? ”ON ”ON ”ON ‘3C'5N
VCEN IR/RR 2182 2137 2146 2101
v IR 535 526 526 520
5 RR 495 487 493 485

 

8 samples in solid matrix. (Csl for IR and KBr for R. Csl facilitated concurrent measurement of Far-IR

region.) No isotope dependent modes were found between 200cm" and 400 cm". See individual spectra for
information on specific experimental conditions.

Table 2-2 Extinction Coefficients of Various CEN Stretches

 

 

VcsN(Cm'l) 8 (M—l cm—l reference
F em-CEN 2129 17 this work
F em-CEN-Cun 2182 150 this work
Rum-CaN 2053 ~2500 l 1

Run-CEN-RuIII 21 18 ~50 l 1

 

37

 

 

 

 

 

 

 

 

 

 

2182
12Cl4-N
"C"N 2146
0
o _
I:
M
E
o 13 I
3 CN ‘/’2137
a
"C"N 2101
l 1 I n I r I 1 I 1 l r I
2000 2050 2100 2150 2200 2250 2300

frequency (cm'l)

Figure 2-32FTIR spectrum of [(l-MeIm)Fe(OEP)-CN-Cu(Me6tren)]2’ and
isotopomers. Spectral conditions described in Materials and Methods.

38

absorbancc

 

 

I I I I I I I I I I I I I I I I
400 450 500 550
frequency (cm")

Figure 2-4: FTIR spectrum of [(1-MeIm)Fe(OEP)-CN-Cu(Met,tren)]2+ and
isotopomers. Spectral conditions described in Materials and Methods.

Table 2-

 

Cb‘ geo

linear

bent‘

39

Table 2-3 Vibrational Frequencies for CEN in Selected Structures

 

 

CN geometry/N-terminus example ranges of Vcn reference
interaction (curl)
linear/none [(l-Melm)Fe(OEP)(CN)] 2129 this work
bent/none SP-14 (“strapped-heme”) 2118-2130 16
linear/hydrogen-bonded oxidized HRP, 2 in methanol 2131, 2132 12, this work
bent/hydrogen-bonded no published examples - -
linear/Cu-bonded l, Fe-CN-Cu 2140-2182 5, this work
bent/Cu-bonded Fe-CN-Cu (0 < l40-l70°)' 2120-2181 5, this work
- oxidized cytochrome c oxidase 2146-2152 3

 

 

' 0 = angle defined by Cu-NEC

Resona:

spectrur

cm". ll
cyanide
\‘ibratic
ligated
29. 32]

53516111:

an isot
substitt

Mil“

40

Resonance Raman Spectroscopy:

The third cyanide isotope sensitive vibration of 1 appears in its resonance Raman
spectrum. When 647 nm excitation is used, a vibration is present at 495 cm'I in the
Raman spectrum (Figure 2-5). In the l3C“N labeled complex this vibration shifts to 487
cm". With ”ON the vibrational mode occurs at 493 cm'1 and with doubly labeled
cyanide this mode shifts to 485 cm". This isotopic dependence is diagnostic of a bending
vibration of the Fe-CEN unit[31]. The observation of heme-ligated cyanide and heme-
ligated CO bending modes in enzymatic systems is well precedented in the literature[l -4,
29, ’32], however this vibration has not previously been observed in synthetic porphyrin
systems[33, 34].

When higher frequency Raman scattering was examined with 647 nm excitation,
an isotope sensitive mode was observed at 2182 cm'l (Figure 2-6). Upon isotopic
substitution, the shifts correlated with those observed in the infrared spectrum for the
C-=_N stretching mode, allowing its assignment as the normally infi'ared active/Raman
inactive CEN stretching mode. Resonance Raman detection of the CEN stretching mode

has no precedent in cyanide-ligated heme systems.

Vibrational Analysis:

Normal coordinate analysis (N CA) was performed to assign the vibrational modes
0f 1 and to determine the contributions of individual internal coordinates to the
Vibrational modes of the complex. The model used to represent 1 is shown in Figure 2-1.

The heme was represented by four point masses, arranged equatorially around a six-

41

 

Intensity

  
      

. l . l
400 450 500

550 600

Raman shifi (cm")

Figure 2-5: Resonance Raman spectrum of [(l-Melm)Fe(OEP)-CN-

Cu(Me6tren)]2+ and isotopomers. Spectral conditions described in
Materials and Methods.

42

 

2057 2239
2145 #87 new
2154 2240 12 .
2056 C5”
.9
g 2057 2144 2240 ”UN
0 r
E
0
2 57 2107 224° "C”N
2145
L M l n l n l 1 I n J 4 J
2000 2050 2100 2150 2200 2250 2300

Raman shift (cm")

Figure 2-6: Resonance Raman spectrum of [(1-Me1m)Fe(OEP)-CN-

Cu(Me6tren)]2+ and isotopomers. Spectral conditions described in Materials
and Methods.

 

 

obse
fit it
peel
tXpe
mod

labi

that

obse

font
0f 11

bitl

43

coordinate iron atom. The four non-cyano ligands to the five-coordinate copper atom
were treated as point masses, with each weighing one-fourth of the total ligand mass. The
l-Me-irnidazole was also approximated by a point mass representing the total weight of
the ligand.

The NCA was performed to find the best set of force constants that duplicated the
observed cyanide isotopic shifts of the vibrational frequencies in 1 (Table 2-1). The peaks
fit were the infrared observed modes at 2182 cm" and 535 cm", and the observed Raman
peak at 495 cm". The optimized force field for this molecule is listed in Table 2-4. The
experimental and calculated vibrational frequencies for the cyanide isotope sensitive
modes, along with the potential energy distributions for these modes, are collected in
Table 2-5.

The potential energy distribution (PED) for the 2182 cm" band (Table 2-5) shows
that this mode is comprised entirely of the CEN stretching coordinate. The Shifts
observed are duplicated by a simple, two body harmonic oscillator, further supporting the
potential energy distribution calculated in the complete analysis. An interesting result is
found for the potential energy distribution of the 535 cm’l peak. This mode is comprised
of roughly equal contributions of the F e-C (35%) and Cu-N (30%) stretching coordinates,
with smaller contributions from the F e—CCNENCN bending mode (16%). Earlier results[29,
32-34] indicate that when ligands are bound to hemes in a linear geometry, the stretching

and bending modes are orthogonal to each other, with each mode contributing little to the

Fe-t

 

Cu-
"It
0111-
111-;

Out.

 

Table 2-4 Force Constants for NCA of Fe-CN-Cu system

 

 

 

 

 

 

 

Primary: Off-Diagonal

Stretches: Value (mdyn/A) Stretch-stretch: Value (mdyn/A)
Fe-CCN l .69 Fe-CCN/CCN-NCN 0.35
CCN-NCN 1 7.91 Ir‘e-CCN/Cu-NCN -0. I9
lmd-Fe 1 .55 Fe-CCN/imd-Fe 0.1
NCN-Cu l .26 Cu-NCN/CCN-NCN 0.1
Fe-NW l .4 CCN-NCN/imd-Fe 0.09
Cu-NLlGAND 1.0 Cu-NCN/ Cu-Lu 0.1
Bends: Value (mdyn-A/radz) Stretch-Bend: Value (mdyn/rad)
Fe-CCN-NCN 0.705 I'e‘e-Cm/Fe-CCN-NCN 0.3
Cu-NCN-CCN 0. l 9 Ir‘e-CC,.,/Cu-1~lc,.,-CCN -0. l7
in-plane irnd-Fe-CCN 0.2 Cu-NCN/Fe-CCN-NCN 0.07
out-plane imd-Fe-CCN 0.2 Cu-Nc,.,/Cu-NC,.,-CcN -0.55
in-plane Cu-NCN-CCN 0.2 Bend/Bend Value (mdyn-A/rad’)
out-plane Cu-NCN-CCN 0.2 Fe-CCN-Nc1../Cu-NC,,,-CCN 0.17

 

 

 

PED 0
the str
comrflf
constal
mdynfi.
(Table
the Fe-

cm'

495

Mi, 1]
bending
the Cu”.
on eith.

V1. bran 0:

45
PED of the other. For 1, however, despite the linear geometry, substantial contribution of

the stretching coordinates to the bending modes was observed; conversely the bends
contribute significantly to the predominantly stretching modes. The interaction force
constant for coupling between the Fe-C stretch and the Cu-N stretch has a value of -0.19
mdyn/A (Table 2-4). The potential energy distribution (PED) for the 495 cm’1 peak
(Table 2-5) shows that this mode is comprised of the F e-CEN bending coordinate, (68%),
the Fe-C stretch (18%) and the Cu-N=:=C (13%) bending mode. As in the case of the 535

cm"

Table 2-5 Experimental and Calculated Vibrational Frequencies (cm")

 

 

 

IICIIN IZCISN IJCIJN IICIKN PED

exp. Cale. exp. Calc. exp. Calc. exp. Calc.

2182 2182 2146 2148 2137 2136 2101 2102 99%CCN-NCN stretch

535 535 526 527 526 527 520 519 35% Fe-CCN stretch,
30% NCN-Cu

stretch, 19% Fe-

CCN-NCN bend

495 496 493 492 487 487 485 483 68% Fe-CcWNcN

bend, 18% lie-CCN
stretch, 13% Cu-
NCN-CCN bend

peak, there is an appreciable contribution of the stretching coordinate to what is overall a
bending vibration. A series of control calculations demonstrated that substantially varying
the Cu/non-cyanide ligand or the F e/porphyrin force constants did not have a large effect
on either the frequency or the potential energy distribution of the isotope sensitive

vibrational modes

DISCUSSIOIN
Electronic Str

The vi"
labeled cyanid
picture of the
porphyrin 5351.

o-donor. lowe

Which is antibe
a higher CsN
isoelectronic e
is a poor o-dc
3W orbital, u

“hen <
Vt“; V1111 funm
This can be rat
interactiOn at t.
Inolecuiar Orbi
all OV'Eral] 0“ E
CSN Stretchin.
argument is tha
Case of the brig

stretching mode

46
DISCUSSION

Electronic Structure and Vibrational Characteristics:

The vibrational properties of l were investigated with the use of isotopically
labeled cyanide. To understand the bonding nature of cyanide, one must have a clear
picture of the electronic structure of this ligand. When cyanide is bound to a metal
porphyrin system, the cyanide ligand binds preferentially via the carbon end and acts as a

o-donor, lowering the electron density in its 50' highest-occupied molecular orbital,

which is antibonding[35]. Consequently a terminally bound cyanide moiety generally has
a higher CEN stretching frequency than does free cyanide. This is in contrast to the
isoelectronic example of carbon monoxide, which acts preferentially as a n-acceptor and
is a poor o-donor. When terminally bound, it readily accepts electron density into its
2p1t* orbital, which lowers the CEO vibrational frequency relative to free CO.

When cyanide acts as a bidentate ligand, forming units of the type M-CaN-M’,
vc.N will further increase, and the metal-ligand stretching frequency will decrease[35-3 7].
This can be rationalized with simple molecular orbital arguments. Introducing a bonding
interaction at the nitrogen atom of the bound cyanide will deplete electron density in the
molecular orbital carrying the electron density at that end of the molecule. This orbital is
an overall 0"“ antibonding orbital and depleting electron density leads to a higher overall
CEN stretching force constant. Another prediction made by the molecular orbital
argument is that the M-CN bond will be weaker in a bridging geometry. However, in the
case of the bridged system presented here, the vibration attributed to the metal-cyanide

stretching mode appears at a significantly higher frequency, 535 cm", than that of the

corresp
obserx'z
of the l
for the
cones;
b60166
the unl
in stre

followi

Norm;

compo
have i
l‘ibratit

mode |

47
corresponding unbridged system, 2, whose F e-CN vibration appears at 450 cm’”33' 34]. This

observation can be rationalized by examining the force constants and the resulting PED
of the normal coordinate analysis (Tables 2-4 and 2-5). It is seen that the force constant
for the Fe-C stretch is indeed lower (1.69 mdyn/A) than that found in studies of the
corresponding unbridged system (~2.0l mdyn/A)[33, 34], indicating a weaker interaction
between the iron and the carbon end of the cyanide ligand in the bridged as compared to
the unbridged systems. The basis for this paradoxical behavior, 1'. e. an observed increase
in stretching frequency despite a lower stretching force constant, is discussed in the

following section.

Normal Mode Composition:

Previous studies have undertaken careful analyses of the normal mode
compositions of the vibrational frequencies of heme/cyanide complexes. These studies
have identified characteristic interdependencies for certain bending and stretching
vibrations. For example, there is little or no interaction between the Fe-CEN bending
mode and the orthogonal Fe-C stretching internal coordinate in systems in which the
ligand binding is parallel to the normal of the heme plane. An example of this is
cyanoferric horseradish peroxidase[29] in which the cyanide binds linearly. For this
complex, the stretch/bend (S/B) interaction force constant is small, 0.079 mdyn/rad. In 1,
however, in which the ligand binds in an analogous, almost linear, fashion (179° Fe-CEN
angle), an appreciable interaction force constant (IF C) of 0.3 mdyn/rad between these two

coordinates is necessary to maintain the relative ordering of the bending and stretching

 

 

Sll

be

V;

the

nei

lll

Valt

dim

Chm

48

modes. The effect of changing the value of this interaction force constant on the relative
ordering of the bending and stretching vibrations is shown in Figure 2-7. At vanishing
values of this interaction force constant, the stretching mode is higher in frequency than
the bending mode. As the IFC value is increased, the bending and stretching modes
approach each other, then cross at a value of ~0.42 mdyn/rad. At values for the IFC above
0.42 mdyn/rad the bending mode is predicted to be higher in frequency than the
stretching mode. The value that best reproduces the observed bending and stretching
bands at 495 cm‘I and 535 cm'I respectively, is 0.30 mdyn/rad. The value of the
interaction force constant also affects the mode composition of the two normal modes.
Values for the IFC of approximately 0.3-0.6 mdyn/rad increase the amount of mixing of
the Fe-CEN bending coordinate with the Fe-C stretching coordinate. At values near the
point where the ordering of the bending and stretching modes switch (0.42 mdyn/rad),
neither mode exhibits the isotopic dependence characteristic of a bending vibration.

When the bending and stretching mode compositions and frequencies were examined as a
function of the Cu-N/Cu-NEC stretch/bend IFC, similar behavior at the limits of the
range examined was observed (Figure 2-8). However, instead of the modes switching
their ordering around a specific value for the IFC, they diverge, maintaining the
stretching mode at a higher frequency than the bending mode. There are two possible
values for this IF C, -0.55 mdyn/rad and -0.30 mdyn/rad, that give the observed frequency
difference between the bending and stretching modes. Only with the more negative value
of -0.55 mdyn/rad however, did the absolute frequencies for the two modes match the

experimentally observed frequencies.

.|)

frequency (cm

 

Figur
Fef

 

frequency (cm 4)

580 ..
560 -
540 ..
520 -
500 .I
480 -

460-

440 _

 

420

49

Anding mode

stretching mode . .

 

bending mode

 

V

I r l i
0.0 0.2 0.4 0.6 0.8

interaction force constant

Figure 2-7: Plot of calculated vibrational frequencies versus
Fe-C stretching/Fe-C-N bending interaction force constant.

frequency (cm

U!
8
L l A

A
00
O
I

460 -

440 .

 

420 h

Figu‘
C u)

 

frequency (cm ")

580

560

540

520

500

480

460

440

420

l

d
u
q
q
-
d
—

q

 

50

\ o
\ O/ stretching mode
: /

K mixed
§ stretch/bend

  
  

 

 

‘ I
bending mode
_ bending mode
I ' l ' T ‘ l ‘ l j
-0.8 -0.6 -0.4 -0.2 0.0

interaction force constant

Figure 2-8: Plot of calculated vibrational frequencies versus
Cu-N stretching/Cu-N-C bending interaction force constant

ln cy
bending coord
constant. Addi
the relative li
According to
produce a stre
observed here
force constant.i

constants are it

ReSOnance E:

In this
the resOnance
“11h laser e,“
isompe'depen

hate been Sh(

51

In cyanoferric myeloperoxidase[29], the ordering of the v, stretching and the
bending coordinates is strongly dependent upon the choice of the Fe-C stretching force
constant. Additionally, the relative splitting of these modes was found to correlate with
the relative linearity of the ligand. In 1, the Fe-CEN angle is approximately 179°[18].
According to the trends discussed in previous work[29, 32], this linear geometry would
produce a stretching frequency higher than that of the bend, which is, in fact what is
observed here for l. The higher stretching frequency is not due to an increased stretching
force constant, but rather to a coupling of two stretches, where the two individual force

constants are lower than those in unbridged complexes.

Resonance Enhancement Conditions
In this work, the normally infrared active CEN stretching vibration is observed in
the resonance Raman spectrum of this complex when he, = 647 nm. Measurements made
with laser excitation in resonance with the 1t,1r* transitions did not exhibit any cyanide
isotope-dependent modes. In carbon monoxide bound systems, distal steric constraints
have been shown to influence the Raman/IR activity of the CEO stretch[31]. The fact that
the CEN vibration is observed with red excitation indicates the presence of an absorption
band with charge transfer character in the red region of the electronic absorption
Spectrum. The failure to observe vc.N in the red excitation Raman spectrum of 2 indicates
that an analogous CT band is apparently not present in this wavelength range in the

absorption spectrum of the unbridged complex. This suggests that copper in the model

complex and.

significant rol

Relevance to

The q
the binuclear
possibility h
hydrogembg
neighboring
the vm 0f2
and other lin

Cyar
fem heme:
eXhibited b:
of Steric 0r
coneSDOndj
Chang et a
geommyd
The new i
Cthibited l

allowed to

horseradisl

52

complex and, by implication, CuB in the heme/copper terminal oxidases, plays a

significant role in defining the electronic structure of the binuclear center.

Relevance to Cytochrome c Oxidase

The question remains as to whether cyanide covalently bridges the two metals of
the binuclear center in cyanide treated, fully-oxidized terminal oxidases. An alternative
possibility has been raised by Caughey and coworkers[10] that invokes a strong
hydrogen-bonding interaction between the N-terminus of the cyanide ligand and a
neighboring amino acid residue in the distal pocket. These models are based partly upon
the vc.N of 2146-2151 cm" observed in the cyanide inhibited enzyme[9, 10, 13, 14]. This

and other lines of evidence need to be considered when evaluating these possibilities.
Cyanide, when not subject to steric or electronic perturbations, prefers to bind to
ferric hemes in a linear fashion. An example of this situation is 2. The VON frequency
exhibited by this linear system is 2129 cm". When the cyanide ligand is forced, by virtue
of steric or electronic constraints, to adopt a non-linear binding geometry, the v0,N is
correspondingly lower, as seen in the hindered, strapped heme system developed by
Chang et al[38]. The VON for these systems ranges from 2130 cm'I for a linear binding
geometry, to 2118 cm'1 for a CN ligand with a Fe-CEN binding angle of 130° (Table 2-3).
The trend is that linear cyanide groups have higher CEN stretching frequencies then those
BXhi bited by non-linear systems. When the terminal nitrogen of a linear cyanide group is
allowed to interact with an electron acceptor, VON increases. For example, in oxidized

horseradish peroxidase, in which groups capable of donating hydrogen bonds are present

.. . . .szT. “k._ ,7 It; 4....
&. ... ..

 

.

1‘. til.

2 yr .. ,I.A;..., « .V . . u 01.14....0. Acl‘ .I‘OJGI Ia Incan‘l ‘11 .aI. z. I ..l...
3.3 . x q... a? main}; 3}»... . .2 P43. 9.... .2. fliéwvfiflfil . peter. M... .l l; .. .. $1.... .

 

53
in the ligand binding pocket, vC.N occurs at 2131 cm’l (Table 2-3). When 2 is dissolved in

methanol, a hydrogen bond donor, the VON shows a small (3 cm"), but discernible,
upshifi in value (Table 2-3). This shift is attributed to a hydrogen bonding interaction
between the solvent and the N-terminal end of the CN ligand. In complexes that are both
non-linear and have the N-terminus interacting with an electron acceptor (such as a
hydrogen bond donor or a copper ion), a range of VON values is displayed. These
frequencies are convolutions of both the specific binding angle and the N-terminus
interaction.

As shown in Table 2-3, the only systems whose VON values bracket that observed
in cyanide-treated cytochrome c oxidase are those with a copper(II) ion coordinated to the
N-terminus of the cyanide ligand. These models are characterized by a relatively linear
Fe-CEN geometry, and varying angles at which the Cu(II) ion is coordinated to the N-
terrninus of the cyanide ligand. The model complex studied here presents one example of
essentially linear (176°) Fe-CEN and CaN-Cu (174°) binding angles. The VON observed
in this case is 2182 cm". Clearly, the observed value of VON in cyanide-treated
cytochrome oxidase necessitates invoking more than a simple distal hydrogen bond and a
relatively linear binding geometry to rationalize the relatively high observed frequency

for this vibration in the enzyme. We conclude that the vibrational analysis presented here
provides strong evidence for a bridging structure in the cyanide complex of oxidized
c.Ytocl'lrorne oxidase.

It has been suggested that the intensity of the CEN stretch in bridging complexes

is Significantly lower than the CEN stretch in unbridged systems[14]. Because of this

apparent trendl
oxidase were 1
weakened by t
in scope and t
the C 2N stret
solutions. II c
much lower

However, in t

stretch '13 gig

 

lherefore, arg
“P011 the inter
An in
h“ emerged
heme 0:. plat
diSIanceg for
Structure to
138°. From
Mame of
clTOchIOme
of all A a

displaCed f“

which requr;

54
apparent trend, it was argued that the CEN stretching bands observed in cytochrome c

oxidase were far too intense to arise from a bridging geometry. However, this argument is
weakened by the fact that the authors based their conclusions on studies that were narrow
in scope and of questionable relevance. Table 2-2 presents the extinction coefficients of
the CEN stretch of some previously reported complexes[30] and l and 2 in acetone
solutions. It can be seen for the ruthenium systems that the bridged cyanide exhibits a
much lower extinction coefficient than does the corresponding unbridged system.
However, in the case of the models studied here, the extinction coefficient of the CEN
stretch is significantly higher in the bridged case, relative to the unbridged case.
Therefore, arguments for or against the bridging of cyanide in cytochrome oxidase based
upon the intensity of the cyanide stretching vibrations are not well founded.

An interesting feature of the binuclear center in the bovine resting enzyme that
has emerged in the crystal structure[7] is that CuB is positioned 1 A off the normal of the
heme a3 plane (Figure 2-9). From simple geometric arguments (assuming reasonable
distances for the Fe-C and C.=_N distances, and the 4.5 A distance provided by the crystal
structure to accommodate a linear Fe-CEN angle), the CEN-Cu bond angle would be
138°. From the correlation published by Scott and Holm[l7] this would lead to a Cu-N

distance of z2.4 A and v0." z 2120 cm". However, the vC.N observed for cyanoferric
Cytochrome c oxidase is 2151 cm’I [14]. This frequency correlates with a Cu-N distance
Of ~2.1 A and a Cu-NEC angle of ~155-160°. This implies that the CuB ion must be
displaced from the position seen in the crystal structure of the oxidized bovine enzyme[7]

WhiCh requires a certain degree of structural flexibility in the protein structure near the

Fig
oxi

55

\lCu/$ ' .
1 Cu
heme at3 B 7”

Figure 2-9: Representation of the binuclear center of oxidized cytochrome c
oxidase.

binuclear cer
aFe-C u dist:
detected betu
ligated solver
resting. oxidi

center has the

56

binuclear center of cytochrome c oxidase. Indeed, this is supported by the observation of
a Fe-Cu distance of 5.2 A in the azide-treated bacterial oxidase[8]. Electron density is
detected between the two metals of the binuclear center, due either to the presence of a
ligated solvent or azide molecule, while no intermetal electron density is detected in the
resting, oxidized bovine system. This variable Fe-Cu distance indicates that the binuclear

center has the ability to alter its geometry to accommodate exogenous ligand binding.

F‘)

Maln‘.
Babcr
Einar
Rich.
Fergu
Kitigt
Tsuki
Shinz
269: 1
[Wata

Palm:

Caug
BiOer

Oenn
WOOC

Kim?
and()

YOShi
3518.

YOShl
I995.

10.

ll.

12.

13-

I4.

57

LIST OF REFERENCES

Malmstrom, B.G., Acc. Chem. Res., 1993. 26: p. 332.

Babcock, G.T. and Wikstrom, M., Nature, 1992. 356: p. 301.

Einarsdottir, 0., Biochim. Biophys. Acta, 1995. 1229: p. 129.

Rich, P.R., J.Plant Physiol., 1995. 22: p. 479.

Ferguson-Miller, S. and Babcock, G.T., Chem. Reviews, 1996. 96: p. 2889.
Kitigawa, T. and Ogura, T., Prog. Inorg. Chem, 1996. 45: p. 431.
Tsukihara, T., Aoyama, H., Yamashita, E., Tomizaki, T., yamaguchi, H.,
Shinzawa-Itoh, K., Nakashima, R., Yaono, R., and Yoshikawa, S., Science, 1995.
269:p.1069.

Iwata, S., Ostermeier, C., Ludwig, B., and Michel, H., Nature, 1995. 376: p. 660.

Palmer, G., J. Bioenerg. Biomembr., 1993.25: p. 145.

Caughey, W.S., Dong, A., Sampath, V., Yoshikawa, S., and Zhao, X.-J., J.
Bioenerg. Biomembr., 1993. 25: p. 81.

Oertling, W.A., Surerus, K.K., Einarsdottir, 0., Fee, J .A., Dyer, R.B., and
Woodruff, W.H., Biochemistry, 1994. 33: p. 3128.

Kim, Y., Babcock, G.T., Surerus, K.K., Fee, J.A., Dyer, R.B., Woodruff, W.H.,
and Oertling, W.A., Biospectroscopy, Submitted for publication.

Yoshikawa, S., O'Keefe, DH, and Caughey, W.S., J. Biol. Chem, 1985. 260: p.
3518.

Yoshikawa, S., Mochizuki, M., Zhao, X.-J., and Caughey, W.S., J. Biol. Chem,
1995. 270: p. 4270.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

58

Nakamoto, K., Infiared and Raman Spectra of Inorganic and Coordination
Compounds. 1987, New York: Wiley-Interscience.

Lee, S.C., Scott, M.J., Kaufi‘mann, K., Munck, E., and Holm, R.H., J. Am. Chem.
Soc., 1994. 116: p. 401.

Scott, M.J. and Holm, R.H., J. Am. Chem. Soc., 1994. 116: p. 11357.
Scott, M.J., Lee, SC, and Holm, R.H., Inorg. Chem, 1994. 33: p. 4651.

Kent, T.A., Munck, E., Dunham, W.R., Filter, W.F., Findling, K.L., Yoshida, T.,
and Fee, J.A., J. Biol. Chem, 1982. 257: p. 12489.

Kent, T.A., Young, L.J., Palmer, 6., Fee, J .A., and Munck, E., J. Biol. Chem,
1983. 258: p. 8543.

Barnes, 2., Babcock, G.T., and Dye, J.L., Biochemistry, 1991. 30: p. 7597.

Kauffrnan, K.E., Doy, 1E.P., Munck, E., Lee, SC, and Holm, R.H., Inorg. Chem.
Submitted for publication.

Lee, SC. and Holm, R.H., J. Am. Chem. Soc., 1993. 115: p. 11789.

Karlin, K.D., Nanthakumar, A., Fox, S., Murthy, N.N., Ravi, N., Huynh, B.H.,
Orosz, R.D., and Day, E.P., J. Am. Chem. Soc., 1994. 116: p. 4753.

Scott, M.J., Zhang, H.H., Lee, S.C., Hedman, B., Hodgson, K.D., and Holm,
R.H., J. Am. Chem. Soc., 1995. 117: p. 568.

Wilson, E.B., Decius, J .C., and Cross, P.C., Molecular Vibrations. 1955, New
York: McGraw-Hill.

Shimanouchi, T., Computer Programs for Normal Coordinate Treatment of
Polyatomic Molecules. 1969, Tokyo: Tokyo University.

McIntosh, DP. and Michaelian, K.H., Canadian Journal of Spectroscopy, 1979.
24: p. 1.

Lopez-Garriga, J .L., Oertling, W.A., Kean, R.T., Hoogland, H., Wever, R., and
Babcock, G.T., Biochemistry, 1990. 29: p. 9387.

Doom, S.K., Dyer, R.B., Stoutland, P.O., and Woodruff, W.H., J. Am. Chem.
Soc., 1993. 115: p. 6398.

31.

Spht
Spin

‘Yu.
81:;

lino
3215

Kerr
Sane

Dow
33.

Alta

123.

Tana

31.

32.

33.

34.

35.

36.

37.

38.

59

Spiro, T.G., Biological Applications of Resonance Raman Spectroscopy, ed. T.G.
Spiro. Vol. 3. 1988, New York: Wiley.

Yu, N.-T., Benko, 8., Kerr, E.A., and Gersonde, K., Proc. Natl. Acad. Sci., 1984.
81:p.5106.

Uno, T., Hatano, K., Nishimura, Y., and Arata, Y., Inorg. Chem, 1988. 27: p.
3215.

Kerr, E.A., Mackin, HQ, and Yu, N.-T., Biochemistry, 1983. 22: p. 4373.
Sano, M. and Yamatera, H., Chem. Phys. Lett., 1979. 60: p. 257.

Dows, D.A., Haim, A., and Wilmarth, W.K., J. Inorg. Nucl. Chem, 1961. 22: p.
33.

Alvarez, S., Lopez, C., and Bermejo, M.J., Transition Met. Chem, 1984. 9: p.
123.

Tanaka, T., Yu, N.-T., and Chang, C.K., Biophys. J., 1987. 52: p. 801.

CHAPTER 3

SMALL MOLECULE BINDING TO HEME PROTEINS:
VIBRATIONAL SPECTROSCOPY OF THE LIGAND BINDING SITE

IN THE oNO BINDING ENZYME SOLUBLE GUANYLATE CYCLASE

Abstract

Infrared spectroscopy is a useful technique to investigate the modes of binding of
small molecules to heme proteins. Carbon monoxide and cyanide are two potential
ligands to heme proteins that, due to their high molar absorptivity, are ideally suited to
interrogation by infrared spectroscopy. These ligands also present very characteristic
vibrational frequencies that depending upon the specific chemical environment in which
they are located. In this chapter, a study is described whereby we investigated the infrared
spectrum of carbon-monoxide bound soluble guanylate cyclase (sGC). Soluble guanylate
cyclase is the only known physiological receptor for nitric oxide. This heme-containing
enzyme also binds carbon monoxide, and to a much lesser extent, dioxygen. The CEO
vibrational frequency for the CO bound form of sGC was found to be 1987 cm", which is
the highest reported frequency for a CEO stretching vibration in an enzymatic system.
This vibration was conclusively identified through the use of isotopic labeling. These
vibrational studies of CO bound sGC have elucidated specific details of the binding
pocket of this enzyme[ 1] and subsequently provided information about the mechanism by

which NO activates this enzyme.

60

6 1
Introduction

Hemes are porphyrin macrocycles that contain iron as their central atoms. Heme
containing proteins play a large number of diverse roles in biological enzymatic
processes. These proteins, of which there are hundreds, if not thousands of examples,
perform roles in various organisms that range from very complex, such as biological
metabolism of molecular oxygen or peroxidic species, to more simple roles such as the
storage and transport of reducing equivalents that provide the electrons necessary for
other enzymatic systems to carry out their tasks. Many heme proteins also have the ability
to bind small molecules such as dioxygen, carbon monoxide, cyanide, nitric oxide or
azide, reversibly. While only some of these ligands are physiologically relevant, they all
offer unique spectroscopic signals that can provide insight into the modes and
mechanisms of ligand binding to these enzymes.

The spectroscopy of small ligand/heme interactions has been studied extensively
and reviewed over the last decades [2]. One of the principle tools for studying the
interactions of small molecules with heme-proteins has been vibrational spectroscopy.
Both infrared and resonance Raman spectroscopies provide opportunities to acquire
information about how these small molecules bind to the heme macrocycle. When Raman
and infrared spectroscopies are used in conjunction with one another, the complementary
data they can determine, can provide a complete description of the vibrational properties
of the heme-ligand complex.

There are three normal modes of vibration expected for an Fe-X-Y unit, as is
found when a diatomic ligand binds to the iron of a heme macrocycle. These are the F e-

XY Stretching mode, the Fe-X-Y bending mode and the X-Y stretching mode. The

62
vibrational frequency of each of these modes contains particular information about the

geometry and environment of the F e-X-Y unit. The technique that is best suited to
observe the various vibrations depends upon the specific experimental conditions.
Typically, infrared spectroscopy is utilized to observe the X-Y stretching mode. This is
because the X-Y stretching vibration is not electronically coupled into the heme
molecular orbital system, which results in a lack of adequate resonance enhancement in a
Raman experiment. The Fe-XY stretching mode, and the Fe-X-Y bending mode,
however, are usually resonance enhanced with excitation into the heme absorption bands,
which, when combined with the chromophore-specificity of Raman spectroscopy, makes

that a very potent tool for studying heme-ligand vibrational behaviors.

Infrared Spectroscopy of Heme Proteins

It is fortuitous that many of the small molecules that bind to hemes exhibit large
mOIar absorbtivities for infrared radiation. This allows the determination of the
Vi brational frequencies of these ligands without needing to utilize unreasonably large

quantities of samples. Many studies have been published that use the well characterized

Vibrational frequencies of CEO, or CEN as markers to provide information regarding
Vari ous types of interactions that can be present in the heme binding pocket of the
81123, tne[3-13]. The vibrational frequencies of these ligands have been shown to be highly
SetlSitive to various types of interactions that in vivo could play a significant role in the

chemistry of these ligand-binding sites. Previously, we have shown that the vibrational

Frequency of a CEN ligand to a molecular heme-copper assembly can be used to support

63
the proposal that CEN exists in a bridging conformation in the oxidized, cyanide ligated

form of the respiratory enzyme cytochrome c oxidase[S]. Likewise, the isoelectronic CEO

molecule also exhibits CEO stretching frequencies that are highly dependent upon the
interactions with the surrounding protein or solvent environment. Thus, accurately
determining these vibrational frequencies can aid in parameterizing the molecular
interactions in a ligand-heme complex.

It has been shown that, when there is a distal interaction present in a heme binding
pocket, the vibrational frequency of a bound cyanide molecule is shifted to significantly

higher frequency (see previous chapter and references therein) when compared to the

C"=—N stretching frequency of an unperturbed heme-cyanide complex, or to the frequency
of free cyanide ion. An example of the type of distal interaction that may be encountered
in the heme pocket of an enzyme is hydrogen bonding to the distal end of the ligand
mOIecule. In the case of cyanide, which is primarily a o-donating ligand, this distal

interaction will remove electron density from the slightly antibonding HOMO of the

cyanide ion, effectively increasing the CEN force constant, and thus the vibrational

fl‘ecluency. Carbon monoxide, when subjected to the same hydrogen bond, however, will

e"(ITIibit a relative redshift in the vibrational frequency of the CEO stretching mode of
vi l31‘ation. The reason for the opposite response of CO to the presence of the distal
interaction originates in the different electronic configuration of CO versus CN. This
di fl‘erent configuration results in CO acting as a n-acid, rather than a o-donor. When CO
is ligated to a Fe atom, the ligand accepts electron density in to its antibonding LUMO,

w -
hlch reduces the carbon-oxygen force constant. The result is that CO exhibits a lower

64

C-0 stretching frequency when it is ligated to a heme, than when it is unbound. If an
additional interaction, such as a hydrogen bond, is introduced at the distal end of a heme-
ligated CO molecule, the effect will be a ftu'ther increase in the electron density in the n*-
orbital, which will result in a further decrease in the CEO stretching fi'equency[14, 15].
The presence (or lack) of a hydrogen bonding interaction, therefore, will be reflected in
the observed vibrational frequency of the bound ligand. As a result, by comparing the
fi'equencies of bound ligands to those of unbound ligands, one can draw some detailed
conclusions about interactions between the heme molecule, the ligand, and the
surrounding protein matrix in which the heme resides.

Infrared spectroscopy is typically hampered by the presence of water in the
sample. However, when cyanide or carbon monoxide is the ligand, their primary
Stretching vibration is observed in a region of the infrared spectrum where water has
minimal absorbances (Figure 3-1). This fact, coupled with the large extinction coefficient
for the primary X-Y stretching mode, allows the use of a cell with a relatively long
pathlength (~50 um), which, in conjunction with moderately concentrated samples

(>50pM) and highly sensitive MCT-type infrared detectors, allows for relatively facile

collection of data.

l{esunrance Raman of Heme Ligand Interactions

Resonance Raman spectroscopy is a technique that is ideally suited for the study

0 . . . . .
f l"heme contarmng systems. Because hemes exhibit very strong absorbances 1n the

65

3:5.
31)-
215-

2-0 4 CN stretching region

LS-u

 

 

absorbance

11)-
0:54

01)q
. CO stretching region

 

'05 I I f T r I ' l ' I ' l ' l
4000 3500 3000 2500 2000 1500 1000 500

 

frequency (cm")

Figure 3-1: Infrared absorption spectrum of water, Can cell, 50pm pathlength, 25
°C. Approximate frequencies of X-Y stretches of heme-ligated XY diatomic ligands
are shown (XY = CEN or CEO)

66
visible region of the electromagnetic spectrum, there are many Opportunities to utilize the

resonance enhancement effect of Raman spectroscopy. Additionally, Rarnan spectroscopy

is not subjected to the same limitations with respect to water content of the samples, as is

infrared spectroscopy.

The vibrational modes of the axial ligands of heme macrocycles are typically

observed with Rarnan spectroscopy using excitation frequencies that correspond to the

so-called B and Q electronic transitions of the porphyrin moiety (Figure 3-2). These

 

absorptions in the optical spectrum are characteristic for heme containing systems. The
observed absorptions are primarily the result of 1t-1t“ transitions, and it has been seen that

the excitation profiles of heme-ligand modes (when the ligand is a rt-acid such as CO, NO

or 0;) correspond directly to these electronic transitions.

Much effort has been expended in detecting the various normal modes of
Vibration for heme-ligand systems. Vibrational measurements on CO-bound heme
Sy Stems are particularly useful to infer the status of the proximal ligand and the
environment of the distal pocket because the correlation between the fundamental
Stfetching frequencies, v(Fe-C) and v(CEO), is well established in both model porphyrin
and heme-protein systems[16]. This correlation is understood by considering that as the
metal d, backbonding to the CEO ligand increases, the Fe-C bond strength increases.
IFlo‘t’vever, as a result of this, the electron density in the CEO 1t* orbital also increases,
lo“’Vering the bond order, and thus the CEO vibrational frequency. This correlation can be
Seen clearly in Figure 3-3, where the Fe-C vibrational frequencies from a number of

heme-protein and model compounds are plotted against the CEO vibrational frequencies

s (l-mol"-cm")

67

8.0x10‘ —

———[(PY)F€(OEP)-CN]

6.0x10‘

4.0xto‘

2.0x10‘

 

 

 

0’0 n I n I n I 1 I l 1 ‘ '
350 400 450 500 550 600 650 700
nanometers

Figure 3-2: UV-Vis spectrum of heme-ligand complex ([(py)Fe(OEP)-CN]).
Sample concentration = 2.2 mM, in acetone. Spectrum acquired in 50pm Can
cell, room temperature.

68

Fe-C stretching frequency
53
o
i

 

 

 

r I 1

I ' I I I ' I ' I ' I
1930 1940 1950 1960 1970 1980 1990

C-O stretching frequency

Figure 3-3: Correlation between Fe-C stretching frequencies and C-0 stretching
frequencies for a variety of histidine ligated carbonmonoxy heme complexes.
Adapted from [16].

69

from the same systems. When the trans-ligand is constant, in this case histidine, there is a
distinct inverse linear relationship between the vibrational frequencies.

The Fe-CEO bending vibration is typically observed in the Raman spectra of
hemoproteins at around 575 cm", while the Fe-CEN bending mode typically occurs in the

range of 400 cm" to 450 cm". An interesting observation is that the Fe-CEO bending

mode has not been observed in a synthetic heme-ligand system where the CEO ligand

binds parallel to the normal of the heme plane. With CN' as the ligand, the bending mode

has also been detected in a relatively limited number of cases. It has been successfully
resonance enhanced, however, in hindered systems, such as the strapped heme systems
developed by Chang and co-workers [17]. This mode of vibration has also been detected
in numerous heme-protein systems where there is the possibility of a distal interaction
arising from the protein matrix or where the protein may have forced the ligand to adopt a
nonlinear binding geometry.

The Fe-CO stretching vibration, on the other hand, is typically detectable with all
heme-ligand systems as is the Fe-CN stretching mode. This difference between the
relative enhancements of the bending versus the stretching modes has been attributed to
the geometry of binding in a number of different studies [2]. The argument states that
When the ligand is forced to adopt a bent configuration, it’s empty 1t* orbitals can more
6&8in interact with the electron density at the pyrrole nitrogens in the porphyrin 1t*
01' bital. It is this interaction that gives rise to the porphyrin-to-ligand charge transfer that
is responsible for the observed resonance enhancement. However, there is a case in which

the ligand is bound in a strictly linear fashion, yet resonance enhancement of the bending

70
mode is observed[5]. In this case, the resonance enhancement may be attributed to the

presence of a ligand-to-metal charge transfer transition in the 650 nm region of the
electronic absorption spectrum. The precise mechanism by which bending modes of
heme-ligand modes are resonance enhanced, however, requires further study in order to
be understood fully.

Soluble guanylate cyclase is the only known physiological receptor for the
intercellular signaling agent nitric oxide (NO). When -NO is bound to the enzyme, sGC
catalyzes the conversion of GTP to cGMP, which has been implicated as a factor in many
important physiological events such as vascular smooth muscle relaxation, platelet
aggregation and neuronal communication. The rate of GTP to cGMP conversion by sGC
increases by 400-fold over the basal level of activity upon ONO binding. sGC is a
heterodimer, and each unit contains one protoporphyrin IX type heme which is the target
for -NO binding. The heme Fe" is 5-coordinate high-spin with histidine as the axial
l i gand. However, unlike other 5—coordinate high-spin heme proteins, sGC shows unusual
behavior toward small ligands such as 02, CO, nitrosomethane, and -NO. The affmity for
02 is so low that in air-saturated buffer the heme Fe remains 5-coordinate. This is
physiologically important because most hemes will rapidly bind oxygen, and then react
With ONO. CO binds to the heme of sGC, but the affinity is much lower than for typical 5-
Coordinate ferrous imidazole-ligated heme proteins such as hemoglobin and myoglobin.
This reduced affinity for sGC is primarily due to an increased off rate (3.5 s"). -NO binds
to the heme Fe of sGC to form a transient 6-coordinate heme. In this complex, however,
the Fe—His bond is unstable and the breakage of this bond is believed to be linked to the

actiVation of the enzyme. While the binding of -NO to heme is virtually irreversible in

 

71
both model systems and heme proteins, dissociation of 0N0 from heme of sGC appears to

be fast, as hypothesized from whole organ studies[l8]. Recent kinetic studies with
purified sGC provide molecular evidence to support this speculation. A similar ligand,
nitrosomethane (CH3NO) which typically binds irreversibly to high-spin ferrous heme
proteins to form a 6-coordinate complex, also binds to the heme of sGC but in a
reversible manner due to a fast off rate[l9]. The observations for all of these ligands
deviate from what is commonly observed for heme-proteins and indicate that the heme
environment in sGC is distinct from that of other heme-containing proteins. Vibrational
Spectroscopy can provide insight into how the heme-binding site of the enzyme affects
the ligand binding properties of this enzyme, and can thus shed light upon the mechanism

0f oNO activation.

Recently, Deinurn et. al. reported resonance Raman data for the CO and -NO
1 igated form of SOC [20]. These data are presented in Figure 3-4 and are tabulated in
Table 3-1. The Raman results show the Fe-CO stretching frequency at 472 cm", which is
the lowest frequency yet observed for a CO-bound heme complex. It is assigned as the
F e-CO stretching mode due to the fact that, as the mass of the ligand is increased through
the use of isotopically labeled ligand, the observed frequency decreases. There is also a
Vibrational mode observed at 562 cm’l that exhibits isotope dependent frequencies.
Because the stretching mode is conclusively assigned through the use of isotopes, this
mode, by default, is assigned to the Fe-CEO bending vibration. A mode was observed at
204 cm" that was assigned to a proximal histidine residue, present in the 6-coordinate

C:()-species. This is particularly low for a Fe-His bond. On the basis of the trans-effect

72

7x10‘

6x10‘

5x10‘

4x10‘

Raman intensity (counts)

 

3x104

 

 

2x10‘ . . .
200 400 600 800 1000

 

frequency (cm")

Figure 3-4: Resonance Raman spectrum of isotopomers of carbonmonoxy-soluble
guanylate cyclase. Laser excitation A=423 nm[20].

73
Table 3-1 Vibrational Frequencies of Various CO-adducts of Heme-enzymes

 

 

v(Fe-CO) 6(Fe-CEO) v(CEO) ref
(cm") (cm") (cm")
CO-sGC 472 562 1987 this work
CO-Mb 512 577 1944 [21, 22]
CO—Hb A 507 578 1951 [22]

 

described above, a weak Fe-proximal ligand bond will lead to a strong Fe-C bond.
However, exactly the opposite is observed which , in addition to the distinct ligand
binding properties of sGC, implicates distal effects as playing a major role in defining the
1 i gand binding properties of sGC.

By comparison to the data correlating Fe-C and CEO stretching frequencies

(Figure 3-3), predictions about the vibrational properties of sGC become clear. From the

data obtained'from the Raman spectra regarding the Fe-C stretching frequency, the CEO
Stretching frequency is predicted to occur close to 2000 cm", which would be the highest
ever observed for a heme-CO system with a proximal histidine ligand. This prediction,
due to the unanticipated nature of the results regarding the low frequencies of both the
Fe-C and the F e-His bonds, warrants testing. The results of these experiments are

described here.

74

Materials and Methods

All enzyme samples were isolated in the laboratory of Professor Michael Marletta,
University of Michigan, Ann Arbor. Soluble guanylate cyclase was isolated from bovine
lung as described by Stone and Marletta[23] except that the last desalting step was done i
with 25mM TEA, 5mM DTT, 50mM NaCl, pH 7.4 and that no glycerol was added to the

sample. The heme content of this preparation was determined to be 1.5 hemes per

heterodimer.

FTIR spectra were acquired on a Nicolet 740 F TIR spectrometer equipped with a
MCT detector. CO was bound to the sGC sample by exposing it to a positive pressure of
CO. The sample was then placed in a Can cell (50 uM pathlength). CO binding was
confirmed by monitoring the shifi in the Soret band from 431nm to 423nm. The cell was
then thermostated at 3°C in a homebuilt cryostat system equipped with AgCl and Ge
Windows. Double-sided interferograms were collected at 4 cm’l resolution and apodized
by using a Happ-Ganzel apodization function. 20,000, 11,000 and 20,000 scans were
averaged for the l2C"’O, I3C"’O, and '3C'8O isotopically labeled samples, respectively.

S mooth polynomial functions were subtracted from the data to baseline correct the data.

Results and Discussion

Figure 3-5 shows FTIR spectra of CO-bound sGC in the 1800-2050 cm’1 range.

With natural isotopic abundance C20, a single peak is observed at 1987 cm". This mode

Shifis to 1943 cm’1 with I3C"’O and to 1895 cm'l with '3C‘80. The observed isotope shifts

are as expected for a CEO harmonic oscillator, and therefore, the peak can be

75
unambiguously assigned to the CEO stretching mode of a heme-CEO complex. When this

result is examined in the context of the Fe-C/CEO correlation published by Ray et.
al. [16], it can be seen that the CEO stretching frequency is indeed higher than has ever
been seen for a heme-CEO complex with a proximal histidine. In addition to this, the fact
that this data point clearly falls on the same line as for other proximal histidine/heme-
C 50 systems lends support to the proposal that the heme in sGC is, in fact, ligated
proximally by a histidine residue.

That a histidine proximally ligates the heme in sGC has previously been inferred
from absorption data[24] and the mode observed in the Raman spectrum at 204 cm’l
Which had been assigned to the Fe-His stretch[20]. This is consistent with the previous
i nterpretations that the histidine is neither anionic nor hydrogen bonded, because, in these
cases, the v(FeC)/v(CaO) points fall below the imidazole correlation curve due to the
Weakened F e-Co bond, as shown in a model compound and in horseradish peroxidase at
high pH[25]. The position along the correlation curve is mainly determined by the
polarity of the distal pocket[16]. sGC falls on this curve at the lowest point thus far
Observed. The protein nearest to sGC on the curve is the porcine myoglobin mutant
[H 64V-V68T], in which back-bonding is inhibited due to the negative polarity of the
distal pocket in the absence of the hydrogen bonding donor (His 64)[26, 27]. Since sGC
f a1 ls on this same backbonding curve we infer that there is negative polarity in the distal
POCket of sGC. Introducing negative polarity increases not only the stretching frequency

of CEO bound to the heme but also the off-rate of both 02 and CO binding, as has been

 

 

 

 

1987
12C160
1943
13C160
1895
13C180
r l ' I T l 1 l
2050 2000 1950 1900 1850

frequency (cm")

Figure 3-5: FTIR spectrum of isotopomers of carbonmonoxy-soluble guanylate
cyclase. Spectral conditions given in Materials and Methods.

77
shown in the mutant porcine myoglobin samples[26, 27]. The general trend in sGC is the

same, showing an increased v(CaO) and a very high off rate for CO (3.5 s")[19], which
further supports the idea that there is negative polarity in the distal pocket of the heme in
sGC and that it is a major determinant for the ligand binding properties of sGC.

In general, ligand binding affinity to hemes increases when going from O2 to CO
and to NO. For sGC, a strong decrease for O2 binding, an increased dissociation rate for
CO, and the apparent relative ease for NO dissociation are observed, indicating an
apparent overall decreased affinity for these small ligands;. It is very likely that sGC uses
the negative polarity to achieve the reduced 02 binding and the relative ease of NO

dissociation. Both of these are necessary for proper functioning of the enzyme.

10.

ll.

12.

13.

78

LIST OF REFERENCES

Kim, S., Deinum, G., Gardner, M.T., Marletta, M.A., and Babcock, G.T., J. Am.
Chem. Soc., 1996. 118: p. 8769.

Spiro, T.G., Biological Applications of Resonance Raman Spectroscopy, ed. T.G.
Spiro. Vol. 3. 1988, New York: Wiley.

Bogumil, R., Hunter, C.L., Maurus, R., Tang, H.-L., Lee, H., Lloyd, E., Brayer,
G.D., Smith, M., and Mauk, A.G., Biochemistry, 1994. 33: p. 7600.

Caughey, W.S., Dong, A., Sampath, V., Yoshikawa, S., and Zhao, X.-J., J.
Bioenerg. Biomembr., 1993. 25: p. 81.

Gardner, M.T., Deinum, G., Kim, Y., Babcock, G.T., Scott, M.J., and Holm, R.H.,
Inorg. Chem, 1996. 35: p. 6878.

Hill, J., Goswitz, V.C., Calhoun, M., Garcia-Horsman, J.A., Lemieux, L., Alben,
J.O., and Gennis, R.B., Biochemistry, 1992. 31: p. 11435.

Palmer, G., J. Bioenerg. Biomembr., 1993. 25: p. 145.
Tsubaki, M. and Yoshikawa, S., Biochemistry, 1993.32: p. 174.
Tsubaki, M. and Yoshikawa, S., Biochemistry, l993. 32: p. 164.

Tsubaki, M., Mogi, T., Hori, H., Sato-Watanabe, M., and Anraku, Y., J. Biol.
Chem., 1996. 271: p. 4017.

Yoshikawa, S., O'Keefe, DH, and Caughey, W.S., J. Biol. Chem., 1985. 260: p.
3518.

Yoshikawa, S. and Caughey, W.S., J. Biol. Chem., 1990. 265: p. 7945.

Yoshikawa, S., Mochizuki, M., Zhao, X.-J., and Caughey, W.S., J. Biol. Chem.,
1995.270:p.4270.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

79

Nakamoto, K., Infiared and Ramon Spectra of Inorganic and Coordination
Compounds. 1987, New York: Wiley-Interscience.

Lopez-Garriga, J.J., Hanton, S., Babcock, G.T., and Harrison, J.F., J. Am. Chem.
Soc., 1986. 108: p. 7251.

Ray, G.B., Li, X., -Y., Ibers, J.A., Sessler, J.L., and Spiro, T.G., J. Am. Chem.
Soc., 1994. 116: p. 162.

Tanaka, T., Yu, N.-T., and Chang, C.K., Biophys. J., 1987. 52: p. 801.

Palmer, R.M.J., Fenige, A.G., and Moncada, S., Nature, 1987. 327: p. 524.

Stone, JR. and Marletta, M.A., Biochemistry, 1995. 34: p. 16397.

Deinum, G., Stone, J.R., Babcock, G.T., and Marletta, M.A., Biochemistry, 1996.
Rimai, L., Salmeen, 1., and Petering, D.H., Biochemistry, 1975. 14: p. 378.
Tsubaki, M., Srivastava, R.B., and Yu, N.-T., Biochemistry, 1982. 21: p. 1132.
Stone, J .R. and Marletta, M.A., Biochemistry, 1995.34: p. 14668.

Stone, J .R. and Marletta, M.A., Biochemistry, 1994.33: p. 5636.

Evangelista-Kirkup, R., Smulevich, G., and Spiro, T.G., Biochemistry, 1986. 25:
p. 4420.

Li, T., Quillin, M.L., Phillips, J ., G.N., and Olson, J .S., Biochemistry, 1994. 33: p.
1433.

Cameron, A.D., Smerdon, S.J., and Wilkonson, A.J., Biochemistry, 1993. 32: p.
13061.

CHAPTER 4

PROBING LOW-FREQUENCY VIBRATIONS IN
METALLOPROTEIN S USING LIGHT INDUCED
INFRARED DIFFERENCE SPECTROSCOPY:

CHALLENGES AND SOLUTIONS

Abstract

Several critical issues surrounding the successful acquisition of vibrational
information in the low frequency region of the infrared spectrum of biological molecules
are discussed. These issues include the development of methods to overcome the
presence of highly-absorbing water in the sample, the design of systems to control the
temperature of the sample within close tolerances and the choice of detectors and window
materials that will provide the best opportunity to detect the signals of interest. In the
following chapter, data is presented from experiments where these methods were utilized

to acquire data possibly originating from the manganese cluster of the photosynthetic

enzyme, photosystem 11.

Introduction

Spectroscopists have long recognized the utility of vibrational information when
seeking to answer structural or mechanistic questions about metalloproteins. In particular,
when studying an enzymatic system, it is desirable to acquire information about low

frequency (<1000 cm") vibrations, as they may contain a wealth of information directly
80

81
related to bonding interactions between the metal site and ligands to the metal site

whether the ligands are substrate derived or protein derived. Information about these
vibrations could provide direct insight into the structure and, possibly, the specific
molecular steps that occur in the catalytic cycle of a metalloenzyme. Table 4-1 shows
some examples of structures that are present (or proposed to be present) in certain
metallo-enzymes along with their estimated vibrational frequencies. Close examination of
the normal mode compositions of these vibrations shows that they will also display very
characteristic isotopic dependencies upon isotopic substitution. It has been shown in
numerous studies that isotopic labeling allows for the unequivocal assignment of
vibrational modes. The usefulness of vibrational information is clear. If low frequency
vibrational information can be acquired from the enzymatic systems, then specific
questions regarding the mechanistic details of these systems can be addressed.

Until recently, the experimental techniques available to gain access to the low-
frequency vibrations were primarily limited to forms of resonance Raman spectroscopy.
Although a powerful technique for studying some systems, namely those that contain
highly absorbing chromophores, resonance Raman spectroscopy is certainly not capable
of addressing all of the needs of the biological vibrational spectroscopist. The other
primary technique for gathering vibrational spectra is infrared (IR) spectroscopy. This
technique, which relies upon detecting absorbances of IR radiation, is not a priori suited
for studying biological systems, let alone the low frequency vibrational modes exhibited
by these systems. This is due to the fact that proteins, in their native environment, are
solvated by a large molar excess of water. Because water exhibits very strong

absorbances at several characteristic regions through out the infrared spectrum, its

82
Table 4-1 Vibrational Frequencies of Protein and Metal-Ligand Structures

 

Structure Approximate Vibrational
Frequency (cm")

 

   

Protein Backbone Modes

 

 

 

 

 

 

Amide V (NHob, CN,)‘
Amide IV (Cob, CCS,CNCd) 625
Amide VI (COob, CN,) 600
Amide v11 (er0,, CN, CO...) 200
Metal-Ligand Modes F:::.:i.,'._':.'.‘.If”1.1;;'m’ ,“z lllllllll ;
Fe—CO 45°
ctr—co 400
l N—Cu 250-350
N
Ni—CO 450-500
Mn=o 75°
MT—G'l 400-500
O—O
/ 0 \ 600-800
Mn=::::' ::::Mn
‘0’
Mn<g>Mn 600400
M'I—Cl 300-400

0 s, stretch; d, deformation, t, torsion; ib, in-plane-bend; ob, out-of-plane-bend

83

presence could obscure the detection of signals originating from solute molecules. In this
chapter, we will describe methods we have developed by using reaction-modulated
infrared difference spectroscopy (particularly light-induced FTIR difference
spectroscopy) whereby solvent subtraction is handled automatically and that also allow
the acquisition of vibrational information, in the low frequency (<1000 cm") region of the
infrared spectrum.

Aside from the issues surrounding solvent water absorbances, there are a number
of other technical challenges that need to be addressed in order to s interrogate sample
successfully, and obtain reliable information regarding low frequency modes of vibration.
These challenges include detector sensitivity, the choice of suitable sample substrate
materials, and the necessity for precise sample temperature control. These challenges and

some solutions to them, are addressed in the following chapter.

What Will be Found in the Low Frequency Region?

The low frequency infrared spectra of proteins are an under-investigated
spectroscopic region. While much work has been done to assign vibrational modes in the
region above 1000 cm", very little foundation has been laid for the assignment of modes
below 1000 cm". From studies of N-methyl acetamide, some predictions can be made
about the vibrational modes that polypeptides may exhibit in their infrared absorbance
spectra[Arrondo, 1993 #78]. These are summarized in Table 4-1. This table predicts that
proteins should display several vibrations, in the low frequency region, that arise from

Primarily out-of-plane vibrations of the peptide backbone.

 

84

Aside from modes having their origin in the protein backbone, there are other
vibrations that should appear in the low frequency region of the spectrum. Metal ligand
complexes, in proteins and synthetic model compounds, also will display vibrations in the
low frequency region of the spectrum. Some of these are also listed in table 4-1. It has
been shown previously [1] that heme-ligand complexes exhibit low fi'equency IR-
detectable modes. Metal-oxygen species, such as those proposed to exist in the catalytic
cycles of cytochrome c oxidase, or photosystem II will exhibit infrared absorptions, that,
when substituted with the appropriate isotope, will show very characteristic frequency
shifis. This technique of using isotopes to shift the vibrational frequency of a mode to aid
in its assignment, plays a critical role in biological vibrational spectroscopy. This is
especially true when there are a large number of peaks in a spectrum, and it is desired to
assign one or several of them to a specific molecular species. From Table 4-1, it is clear
that developing methods whereby one can detect low frequency vibrational modes in

enzymatic systems will be of great use in helping to elucidate mechanistic details of their

catalytic cycles.

FTIR of Aqueous Systems: The Low-Frequency Challenge

The primary issue facing the biological infrared spectroscopist pertains to water
content of the samples. Water, the ubiquitous solvent for biological systems, hampers
typical infrared spectroscopy due to the fact that it exhibits strong absorbances in several
different regions throughout the mid infrared spectrum and completely obscures the
region below 1000 cm’I (Figure 4-1). The strong absorbance centered at approximately

3400 cm" is due to the O-H stretching modes of water, the peak at ~1650 cm" is

85
attributed to the H-O-H bending mode, and the intense, broad feature below 1000 cm‘1 is

attributed to the strongly coupled lattice vibrations of aggregates of water molecules. In
order that one may observe vibrational modes of the enzyme that overlap any of the water
absorbances, ways must be found to either eliminate these absorbances entirely, reduce
them to levels where they can be accurately subtracted out of the spectrum, finesse the
issue by using reaction modulated difference spectroscopies, or use some combination of
the above strategies so that the signals of interest become detectable. In particular, we are
interested in finding methodologies by which we can observe modes of vibration below
1000 cm" in the infrared spectrum of enzymatic systems.

The extinction coefficients of the various water absorbances have been well
characterized[2, 3]. Table 4-2 lists these vibrational modes, and the associated extinction
coefficients. With this information, it is possible to perform a simple Beers Law type
calculation to determine the pathlength necessary to return a water absorbance that is
reproducible within the technical limitations of the instrumentation. Various detectors and
instruments will have different limitations with respect to the linearity of their response to
a given absorbance. The generally accepted rule-of-thumb states that in order for a
solvent spectrum to be accurately subtracted from a protein sample, the overall
absorbance of a given peak must be less than one absorbance unit (a.u.)[5]. With this
limitation for the absorbance of the sample, the pathlength of an aqueous solution must
not exceed 10pm, if we are considering the extinction due to the water bending vibration
centered at 1645 cm". This method of using simple Beers Law calculations to address the

water absorbance problem has been successfully applied in numerous different infrared

86

L2

 

1.1 _
1.0 -
(L9 _

0.8 -

absorbance

(L7 -
(16 b

(15 -

 

 

1 l 1 l 1 l r l 1 l 1 l L J

4000 3500 3000 2500 2000 1500 1000 500

 

frequency (cm")

Figure 4-1: Infrared absorption spectrum of water. Spectrum acquired in an AgCl
cell, room temperature, 32 scans, ~5um pathlength.

87

Table 4-2 Infrared Absorbances of H20 and D,O[4]

 

 

Vibration H20 (cm") s(M"cm") DzO(cm") 8(M'rcm")
O-X stretch 3920 0.83 2900 0.60
3490 62.7 2540 59.8
3280 54.5 2450 55.2
Association 2125 3.23 1555 1.74
X-O-X bending 1645 20.8 1215 16.1
Libration Broad absorbance below 1000 cm‘1

 

studies. By using a combination of highly concentrated samples that reduced the effective
water concentration from 55M, and a short optical pathlength, infrared spectra have been
acquired that showed very satisfactory subtraction of the interfering water signals[6].
However, these studies have been limited to the region above 1000 cm". This method is
not applicable, if one wishes to acquire information below 1000 cm". This is due to the
fact that the large water lattice vibrations below 1000cm", even when measured with
pathlengths as short as 5pm, do not show adequate subtraction characteristics. The bands
are so exquisitely sensitive to minute changes in pathlength, or sampling conditions
(temperature, ionic strength, etc.)[3], that large subtraction artifacts remain even after

careful subtraction (Figure 4-2)

1 Ir rte EDS-9359*.

40‘. 17. .
_V.a v

11.
3.4

«w .1322...

.111 .

 

 

 

   

.l. .. WqJ-Hdt'ulp‘mt ._ . _ a— .

 

88

 

0.0006 _
0.0004 _
0.0002 _

0.0000 ,

AA.U.

-0.0002 ..
-0.0004 ,—
-0.0006 ..

-0.0008

 

 

 

. 1 . 1 . 1 . 1 . 1 t 1 .
2000 1800 1600 1400 1200 1000 800 600

frequency (cm")

Figure 4-2: Difference spectrum of water sample showing poor subtraction
characteristics below 1000 cm". 511m pathlength, 200K, AgCl sample cell.

89
A technique that elegantly addresses the issues of solvent subtraction is reaction

modulated F TIR spectroscopy[7]. This technique consists of interrogating a sample and
digitally storing the spectrum from that sample. Then, exactly the same sample, through
some method, is induced to undergo some reaction (light is the reaction-inducing agent in
this set of studies, but other reaction modulating techniques have been used, such as
exposure to an electric field) and another spectrum is acquired from the same sample in
this new state. The pre-reaction and post-reaction spectra are then subtracted, and the
resulting difference spectrum reflects the vibrational changes that have taken place as a
result of the chemistry that has occurred in the sample.

This technique allows the issue of solvent subtraction to be fmessed, in that
exactly the same sample is interrogated for each of the spectra. While there will be
(hopefully!) slight vibrational differences that arise from the reaction that was induced,
there will be no gross perturbation of the solvent concentration, ionic strength or
pathlength. Thus the large scale baseline distortions that are common in solvent
subtracted data are eliminated. In addition, this technique allows for the examination of
very small signals, superimposed upon very large background absorbances. It is common
for samples to exhibit an infrared absorbance spectrum that approaches one absorbance
unit in intensity. However, the slight perturbations that result from the reaction of
interest, may affect only a few amino acids, or just the substrate molecule. Therefore the
absorbance changes arising from these perturbations may constitute a fraction of a
percent of the overall signal. With sufficient signal to noise ratios, it is not uncommon,
using this technique, to be able to observe difference signals with intensities of

approximately 10" a.u. These small signals, which may reflect changes in the enzyme

90
that are directly related to the catalytic cycle, are the type of information we are seeking

to acquire. Using modern instrumentation (including detection schemes), it is possible to
detect changes in single amino acid residues in enzymes with molecular weights
approaching one million daltons.

This reaction modulated difference technique provides a very detailed look at the
changes that occur as an enzyme system undergoes it catalytic cycle. In Figure 4-3, the
theoretical construction of such a difference spectrum is presented. The Figure shows a
possible step in the catalytic cycle of Photosytem II, the enzyme in plants responsible for
oxidizing water, to molecular oxygen. In the initial state (4-3a), the substrate water, in the
form of a hydroxide molecule, is bound to the enzyme active site, but no chemistry (bond
breaking/bond making) has occurred with regards to the substrate. An infiared spectrum
of the enzyme in this “prereaction” state is acquired and stored digitally. This spectrum
contains the vibrational information from the enzyme and the substrate molecule
(including, for example, the absorbance due to the p-oxo bridges of the manganese
cluster, and the Mn-OH bond). Then, the enzyme is induced to perform a step (or
multiple steps) of the catalytic cycle. When the enzyme has reached a new point in the
catalytic cycle (the “stopping point” depends upon the thermodynamic, kinetic, and
biochemical control that the investigator can exert over the sample), another spectrum is
acquired (4-3b). This spectrum contains the vibrational information indicating the

presence of the new species (the Mn-oxo species). This second spectrum is then

91

Structures IR absorbance spectrum

/0\
a) Mn\0 /Mn

hv

 

 

c) (spectrum b) -(spectrum a)

 

 

 

Frequency —->

Figure 4-3: Construction of a theoretical light-induced difference spectrum.

92
subtracted from the first spectrum. The difference spectrum (4-3c) displays both positive

and negative peaks that correspond, in the case of the positive going-modes, to oscillators
that appear afier the enzyme acts on the substrate, and the negative-going modes
correspond to vibrational modes that disappear as the substrate is reacted. Any vibrational
modes that are remote from the active sight, including the vast majority of solvent
molecules, will be unaffected by the distant reactions, and thus will display exactly the
same vibrational frequency before and after the enzyme-substrate reaction. They will be
subtracted out, leaving a flat baseline. Any vibrational modes that are slightly perturbed
(an amino acid residue gaining or losing a hydrogen bonding interaction, a solvent
molecule that is in the active site of the enzyme, or something less dramatic such as a
vibration whose normal mode composition changes subtly) will appear as difference
peaks.

With the advent of reaction modulated difference spectroscopy, the issue of
solvent subtraction was addressed in a very elegant manner. However, in order for solvent
subtraction to be successful, it is necessary to have samples that do not absorb the
infrared radiation too highly, or else the detectors will measure absorbances at levels at
which their responses cannot be considered linear and therefore reliable. Any difference
spectra acquired under these highly absorbing conditions will contain systemic errors.
Short pathlengths is one method that can be used to ensure that fully hydrated samples do
not exhibit absorbances that exceed the limits of the instrumentation. This method has
been used in several studies of photosystem II, the enzyme responsible for the oxidation
of water to molecular oxygen in plants and cyanobacteria,[8-l6] and the bacterial reaction

center[7] where highly concentrated pellets of the reaction centers were pressed between

93
the sampling windows to produce both a short pathlength, and an extremely high

concentration of sample. The resulting difference spectra were of high quality in the high
frequency regions. However, this technique is not sufficient, due to the non-linearities of
the water absorbances below 1000 cm‘l , to observe signals in the low frequency region of
the spectrum. As a result of this, other methods of handling water absorbances need to be
considered.

An obvious solution, and one with which we have had good success, consists of
physically removing the majority of the bulk solvent water from the sample. This method
has been used previously to obtain high quality light-induced infrared difference spectra
of bacteriorhodopsin[17-21], the photosynthetic bacterial reaction center[22] and
photosystem II[23-27]. However, drying biological samples raises some very legitimate
biochemical questions regarding physiological relevance. Careful control experiments
must be performed to ensure that the samples retain their biochemical integrity after
water removal. These control experiments may consist of UV-visible measurements,
electron paramagnetic resonance, or any other technique that has been used to
characterize the enzyme system in its fully hydrated state. If the samples maintain their
activity after partial drying, then this technique allows spectroscopic access to the entire
mid-infrared region of the spectrum, including, in particular, the region of metal-ligand
vibrations below 1000 cm". There are several different ways that controlled partial
dehydration of protein samples can be accomplished.

The easiest method to achieve partial dehydration of samples consists of
depositing the sample on to a suitable substrate material (the choice of suitable substrate

materials will be covered in a subsequent section), and then drying it under a stream of

94

dry nitrogen gas. In buffers that contain no hygroscopic materials such as sucrose,
glycerol, or ethylene glycol (common components of biochemical buffering systems),
almost complete dryness can be obtained this way (see Figure 4-4), based upon the
relative intensities of the O-H stretching mode at 3400 cm’l and the so-called amide I
protein backbone mode at 1650 cm". However, many enzymes, when desiccated, undergo
extensive structural perturbations[6], which may effect their catalytic chemistry. If
control experiments show that the samples cannot handle arbitrary levels of dehydration,
then it may be necessary to remove the water in the sample in a highly controlled manner,
so that a balance can be achieved between enough dehydration to ensure spectroscopic
utility, while maintaining a high enough level of hydration to ensure biochemical
integrity.

We have found that if a small amount of hygroscopic materials (such as those
named above) are present in the sample, then the samples can usually be dried to a point
where the infrared spectra are spectroscopically acceptable, yet the proteins retain enough
water whereby they exhibit native biological activity (Figure 4-4). In our experiments it
has been found that 150mM sucrose in the drying buffer, allows enough water to be
removed to reduce the water absorptions to an acceptable level, yet, as is shown in the
next chapter, control EPR experiments exhibit all of the characteristic signals associated
with an intact reaction center. It is important to determine a set of quantitative
benchmarks that can be used to evaluate sets of samples with regard to their levels of
hydration in order to ensure that the spectra acquired show high levels of reproducibility.

In the case of bacteriorhodopsin, it was detennined[28] that the integrated area of the

95

——150mM sucrose in drying buffer

1 5 _ DO SUCl'OSC

 

sucrose absorbances

L0 -

absorbance

(15 _

. M:
01) -_Ah——-—._ze

4000 3500 3000 2500 2000 1500 1000

 

 

 

 

frequency (cm")

Figure 4-4: Partially hydrated films of PSII samples, with and without sucrose in the
drying medium. Sample temperature: 25 °C.

96
water peak at 3300 cm'l should be at least 6 times the integrated area of the amide I

protein mode at 1650 cm'I in order for the protein to be fully competent to undergo a
photochemical reaction. In our case, we routinely achieve a 10:1 ratio of the integrated
area of the same peaks. Under these conditions, the water absorbances are well within the
linear range of the detector, and thus we are able to acquire high-quality, light induced

difference spectra of Photosystem 11 across the entire mid-infrared range.

Detectors

A critical issue that must be addressed when performing FTIR spectroscopy on
biological systems concerns the choice of detectors. This is particularly true when one is
interested in detecting low frequency vibrations. There are a wide variety of infi'ared-
sensitive detectors available, and with the development of semiconductor detectors,
namely the mercury-cadmium-telluride (MCT) detectors, it is possible for investigators to
acquire data across the entire infrared spectrum with great sensitivity. Additionally, with
sufficient signal averaging, random noise can be reduced to levels where the small signals
from the enzyme system can be observed.

Infrared detectors are commonly evaluated in terms of their specific detectivity,

or, D“ values[5]. The sensitivity of a given infrared detector is usually expressed in terms
of the noise-equivalent power (N EP)
1

NEP = 1:" W112“?

pl

 

97

In this expression V" is the root mean square (ms) of the detector noise voltage in volts-
Hz"”, and RV is the voltage responsivity of the detector to signals in volts/watt. The

specific detectivity, or D‘, of a detector is given by

l
5 l
D*=-(—;—g)?cm-sz -W’l

where AD is the area of the detector element. This expression factors into the overall
expression for the signal to noise ratio (SNR) of the FTIR instrument which is

l
_U,m-e-Av-zz .5
' NEP

SNR

 

1
_U,m-o.AT/-:2 -§-D*
- 1

A05

 

where U ;(T) is the spectral energy density, (9 is the throughput of the system, A? is the

resolution, tis the time, and 5 is an instrumental efficiency factor. This expression
clearly shows that the signal-to-noise ratio of a given spectrum depends linearly upon the
D“ value of the detector, the resolution, AV, and the throughput, (9. It is, therefore, in
the best interest of the spectroscopist to optimize these values as much as is possible in
order to acquire the highest quality spectra in the shortest amount of time.

There are several types of detectors available for FTIR measurements. The most
common type of detector with the most linear response function are the “thermal” type
detectors. These detectors operate by sensing the change in temperature of an absorbing
material as more (or less) infrared radiation impinges upon it. The most common class of

this type of detector is called a pyroelectric bolometer. This type of detector utilizes a

98

material for its sensing element that exhibits a change in its electrical polarization when
its temperature changes. Deuterated triglycine sulfate (DTGS) is one of the more
common materials used in this type of detector. These detectors have the advantage that
they can be operated at room temperature, and that they display highly linear responses to
varying intensities of incident radiation. However, for biological spectroscopy, these
detectors are of limited utility because they are not highly sensitive (D* = ~107 - 108),
thus making the detection of small experimental signals over large background signals
difficult.

Biological infiared spectroscopy underwent a renaissance with the advent of
semiconductor-type quantum detectors. This class of highly sensitive detectors allowed
for the detection of minute signals, in otherwise highly absorbing samples. These
detectors, of which the mercuryocadmium-telluride (MCT) detectors are the most popular,
have D“ values ranging from 109 to 10'l and higher. These detectors operate by using a
ternary semiconductor detector element whose bandgap corresponds to the energies
present in infrared radiation. Because of the large background of infrared radiation
present at room temperature, these detectors must be cooled, usually to liquid nitrogen
temperatures (77K), to eliminate thermal background contributions to the detected
signals. The wavelengths to which these detectors respond depends upon the specific
stoichiometric ratio of the constituent elements. However, an MCT detector optimized for
low frequencies (peak response at ~500cm") exhibits a D“ value approximately an order

of magnitude lower than an MCT detector that is optimized for higher frequency

responses.

99

Another choice for low frequency IR detection is the liquid helium cooled
quantum detector. This technology relies upon a germanium or a doped germanium
detector element, and operates by the same principles as the MCT type detector. A key
difference is the need for a cooling system that is able to maintain the detector element at
a temperature below that readily attainable with liquid nitrogen as the coolant. This
necessitates the use of liquid helium and the associated hardware necessary for a helium
refrigeration system. Because of this, this type of detector is considerably more expensive
than the simpler nitrogen cooled MCT type detectors. Because investigations of low
frequency modes in the infrared spectra of biomolecules is still in its infancy, these
detectors have not been tested for their usefulness for these studies. However, they do
have comparable D" values to MCT detectors, and the fact that they respond to radiation
further into the far-IR than any MCT detector, shows their potential utility for these types

of studies.

Window/Substrate Materials

A question that needs to be addressed when performing any type of spectroscopy
regards the types of window materials through which the interrogating radiation must
pass. The material must not only be transparent to the radiation, but must also pass
enough of the radiation to ensure that a high enough optical throughput is attained to
achieve suitable signal-to-noise ratios in the resulting spectra. There are numerous
window materials available to the infrared spectroscopist. However, when the
spectroscopist is interested in utilizing samples that contain water and is interested in

observing the low frequency region of the spectrum, then the list of candidate

41114

4 . . . . . . . ..
0“ Cd: . (r . _ .

 

.fi . .I._..q.... .

100

materialsthat meet this criteria becomes limited. Additionally, because there is a lot of
usefirl information in the high-frequency region of infrared difference spectra, it would be
beneficial to use a material that has a clear window of transmission from 2000 cm‘l down

to a low frequency cutoff that will ensure that all of the vibrational modes of interest are
detectable.

A list of the common infrared window materials, the frequencies of radiation that they
pass, and their water solubility characteristics are listed in table 4-3. From this table there
are several materials that fit the criteria for transmitting radiation below 1000 cm", as
well as showing limited or no solubility in water. Silver halide salts are a class of
materials that have been used successfully in several different lR experiments. The two
most common materials are silver chloride and silver bromide. These materials transmit
radiation from the visible region of the electromagnetic spectrum to 435 cm'l and 285
cm'1 respectively. They both have a tendency to cold flow (physically distort), and they
are also prone to photoreduce upon exposure to UV light, so their transmission properties
tend to degrade over time. This photoreduction process needs to be considered when
samples are deposited upon the surface of one of these materials. The reduction potential
of Ag+ is such that some reactive intermediates in certain enzymatic processes may react
with the silver ions in the substrate rather than reacting in their normal physiologically
relevant manner. Proper control experiments must be performed to ensure that this is not
an issue. If these side reactions are, in fact, found to not play a role in the experiment,
then the silver halide salts are good materials for these types of experiments. Their IR
transmission properties coupled with the visible light they can transmit (for light induced

experiments) make them very useful for light induced FTIR experiments. Zinc selenide

101

Table 4-3 Properties of Infrared Window Materials

 

 

Material Water Solubility Transmission Range
(g / 100g H20, 20°C) (cm")
NaCl 36.0 40,000-625
KBr 65.2 40,000-400
AgCl 0.00015 25,000-435
AgBr 0.000012 20,000-285
ZnSe insoluble 10,000-500
CdTe insoluble 5,000-320
Ge insoluble 5,000-600
Polyethylene insoluble 625-30

Diamond insoluble 1 ,7 00-33

 

I .. ..a .ne..Lro‘..u.cv_

. . . ..r. . . .1U“J.Vnnl.r 744.3 , . .. . . . . o 1“)“ 1.. o . ... I. . .11 0 .
._ .. .. ....... . I . o . . . . .2 .. . . .. , ,. T . r ..rb. LE Pal-(PP). Ibrwt. . .. 1.P (Elev . . . . .I 1.3.01..§u.,.0

 

102

and cadmium telluride are two materials that have not enjoyed widespread use for
biological FTIR experiments, due in part to their higher cost and the availability of
cheaper materials that show similar transmission properties. Germanium is a material that
is suitable for some studies, yet its high index of refraction leads to greatly attenuated
throughput of the IR radiation. It has served the light-modulated difference spectroscopy
field well in that it passes IR radiation (albeit at levels too low for some purposes), yet it
completely blocks visible light. Most FTIR spectrometers have as part of their design, a
helium-neon laser that is coaxial with the IR beam. This laser light, which is used to
provide an internal frequency standard to the instrument, must be prevented from
impinging upon the sample if the reaction that is to be studied is going to be induced by a
controlled source of visible light. Germanium is often the material of choice to perform as
a visible light filter since it still passes the IR radiation. One serious shortcoming of
germanium is that it does not transmit 1R radiation below 600cm". Many metal-ligand
vibrations fall below this cutoff, thus making germanium a liability for these studies.
Polyethylene and silicon are two materials that appear attractive for low-
frequency work, yet they are restricted in their utility due to their limited transmission

range and, in the case of silicon, the high index of refraction.

Temperature Control

Once the sample conditions are under control it is crucial to evaluate the method
by which the temperature of the sample will be controlled. It has been shown that infiared
spectra of aqueous systems display large baseline distortions when the temperature is not

maintained at a constant value[3]. Additionally when two spectra, of the same partially

103

dehydrated protein film, taken immediately consecutively to one another, experience
temperature drifts of as little as 0.3°C , large differential signals appear in the difference
spectrum (see Figure 4-5). Care must be taken when evaluating data, as these “thermal
bands” can easily be mistaken for real peaks.

There are several technologies available to achieve a high level of temperature
control with a sampling system. Closed-cycle helium refi'igeration systems allow samples
to be cooled to l-2K. These systems display very stable temperatures, yet they do not
allow for easy sample switching. These cryostat systems place the sample into a high-
vacuum environment to reduce or eliminate thermal contact with the ambient
temperatures. Each time a sample is removed this vacuum must be broken, and
reestablished before a new set of data, with a new sample, can be acquired. This greatly
increases the amount of time necessary to achieve a given signal-to-noise ratio in a final
spectrum. It is not uncommon for spectra from multiple samples to be averaged together,
thus a system whereby one can change samples easily and quickly is desirable. In
addition to this, the windows on the refrigeration unit must be able to not only pass
infrared radiation, but also have the mechanical strength necessary to withstand the great
pressure differential between the high vacuum that surrounds the sample and the
atmospheric pressure outside of the apparatus. Typically more brittle materials, such as
calcium fluoride, are used, but, because of their IR transmission properties, they are not
of any use for studies of low-frequency modes of vibration. The materials that are capable

of transmitting the lower frequencies are typically softer, and thus very thick windows are

1x10J-

5x104—

Al\ll

 

-5x10‘

 

I l I l l I l I l I
2000 1800 1600 1400 1200 1000 800 600

-1x104

frequency (cm“)

Figure 4-5: Difference bands induced by temperature fluctuation of partially
hydrated film of photosystem II. The spectrum was acquired by cooling the sample
to 200K, acquiring one spectrum (500 scans), raising the temperature 0.2 °C, and
acquiring a second spectrum (500 scans). The ratio of the two spectra resulted in the
displayed difference spectrum.

105

necessary to meet the mechanical specifications. Thicker windows reduces the overall
throughput on a system, thereby also degrading the signal-to-noise ratio of the acquired
data.

A system that we have designed to accommodate our experimental needs is
presented in Figure 4-6. It maintains a high level of temperature stability, yet allows for
facile sample changing, while also allowing for high optical throughput. The system is
based upon the simple concept of flowing a stream of gas at a precisely controlled
temperature around the sample. In our system, nitrogen gas, at room temperature, is
passed through two heat exchangers to cool the gas to a temperature below the desired
target temperature. The first heat exchanger uses an acetone/dry ice bath, the second uses
liquid nitrogen. The gas is then passed through a glass dewar transfer arm, that contains a
nichrome heating element that is used to heat the gas to the desired temperature in a
controlled manner. The gas, now at the desired temperature, then flows into the sampling
chamber that is located in the sample compartment of the FTIR spectrometer. The
sampling chamber is a simple aluminum box, with windows to allow the IR radiation to
pass (Figure 4-7). The first window is a bandpass filter (either germanium or a custom
dielectric-layer bandpass filter discussed later) that acts to block the coaxial helium-neon
laser beam, while still transmitting the IR radiation in the fiequency range of interest. The
second window is made of silver chloride. Care is taken to place the actual enzyme
sample at the focal point of the IR radiation. The sample, a partially-dehydrated film of
protein, is on a silver chloride window, that is sealed with another silver chloride

window. Incorporated into the sample window holder is a thermocouple used to monitor

 

 

.537. M... ‘3‘ 2111. ..prrm .. 12.1.4.4».
— t

I . «9.1.. “If”! .9 v1.01 . . .»1WVI»J ..iw . ,. I...r. ._
..v........ . ...a .a . I ~ . 1

.. . . . “we
1 11’.FL I-Irlll. . .. .

  

. .

106

 

 

 

 

 

 

 

 

 

 

 

 

Dewar Transfer Arm
Gas Flow .9 1 Gas Flo? exhaust
l ’ 3 ]
a L I

 

 

 

 

 

 

 

 

 

 

Sample compartment

Heat Exchanger #1 Heat Exchanger #2

Figure 4-6: Temperature control system for low temperature FTIR experiments.

107
TOP VIEW

 

Germanium
window

1R hv K

 

/

 

 

 

 

 

 

wi‘rigdcoiv fiber optic probe
I 1
Clear window
for optional external
illumination
gas flow in gas flow out
” SIDE VIEW
0 o Germanium
AgCl window

 

 

 

window /

‘(fi)

0 ¥ 0

Thermocouple jacks

 

 

 

 

 

 

 

 

 

Figure 4-7: Sample compartment for low-temperature FTIR experiments.

108
the temperature of the sample. The temperature is controlled controlled to +/-0.1°C by a

LakeShore Cryogenics Model 321 temperature control unit. With this level of control, the
thermal contributions to the difference spectra are eliminated, and the design of the
system allows large numbers of samples to be investigated in a relatively short period of
time. Most importantly, there is no limit upon the choice of window materials used on the
sampling chamber. Therefore, the high and low frequency cutoffs of the acquired data are
not limited by the temperature control system. In order to induce the reaction with actinic
light, a fiber optic probe is positioned next to the sample holder. A 150W halogen bulb,

filtered to remove IR components, can provide 20mW of visible light to the sample.

Conclusions

With the systems described above, we have developed ways whereby we can
acquire data using infrared spectroscopy, that gives us information about the low
frequency vibrations present in photosystem II. This is the first time that information such
as this has been reported, and the techniques developed here show great potential for
application not only to photosystem II, but to metalloenzymes in general. We are
currently adapting these techniques to study the binuclear iron-copper center of the
respiratory protein cytochrome c oxidase. Other systems for which this technique may be
useful include methane monooxygenase, ribonucleotide reductase, and the manganese
catalase systems. Studies of these systems have been limited, until now, to relatively

indirect methods of spectroscopy. The information acquired here is direct, in that it can

109

confirm the presence or lack of certain specific chemical bonds during the various steps

of the catalytic cycles of the enzymes.

10.

ll.

12.

110

LIST OF REFERENCES

Gardner, M.T., Deinum, G., Kim, Y., Babcock, G.T., Scott, M.J., and Holm, R.H.,
Inorg. Chem., 1996. 35: p. 6878.

Arrondo, J.L.R., Muga, A., Castresana, J ., and Goni, F.M., Prog. Biophys. Molec.
Biol., 1993. 59: p. 23.

Rahmelow, K. and Hubner, W., Appl. Spec., 1997. 51: p. 160.

Fringeli, UP. and Gunthard, H.H., in Membrane Spectroscopy, E. Grell, Editor.
1981, Springer: Berlin. p. 270.

Griffiths, PR. and de Haseth, J .A., Fourier Transform Infrared Spectrometry.
1986, New York: Wiley-Interscience.

Koenig, J.L. and Tabb, D.L., Infiared Spectra of Globular Proteins in Aqueous
Solution, in Analytical Applications of FT -IR to Molecular and Biological
Systems, J .R. Durig, Editor. 1980, D. Reidel Publishing Co. p. 241.

Mantele, W., Infi'ared Vibrational Spectroscopy of the Photosynthetic Reaction
Center, in The Photosynthetic Reaction Center, J. Deisenhofer and J .R. Norris,
Editors. 1993, Academic Press: New York. p. 239.

Nabedryk, E., Andrianarnbinintsoa, S., Berger, G., Leonhard, M., Mantele, W.,
and Breton, J., Biochim. Biophys. Acta, 1990. 1016: p. 49.

Nabedryk, E., Leonhard, M., Mantele, W., and Breton, J ., Biochemistry, 1990. 29:
p.3242.

Noguchi, T., Ono, T., and Inoue, Y., Biochim. Biophys. Acta, 1995. 1228: p. 189.
Noguchi, T. and Inoue, Y., FEBS Letters, 1995. 370: p. 241.

Noguchi, T. and Inoue, Y., J. Biochem., 1995. 118: p. 9.

13.

14.

15.

l6.

l7.

l8.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

111
Steenhuis, J.J. and Barry, B.A., J. Am. Chem. Soc., 1996. 118: p. 11927.

Hienerwadel, R., Boussac, A., Breton, J., and Berthomieu, C., Biochemistry,
1996.35:p.15447.

Hienerwadel, R. and Berthomieu, C., Biochemistry, 1995. 34: p. 16288.
Berthomieu, C. and Boussac, A., Biochemistry, 1995.34: p. 1541.

Maeda, A., Sasaki, J., Shichida, Y., Yoshizawa, T., Chang, M., Ni, B.,
Needleman, R., and Lanyi, J .K., Biochemistry, 1992. 31: p. 4684.

Rothschild, K.J., He, Y.-W., Sonar, S., Marti, T., and Khorana, H.G., J. Biol.
Chem., 1992. 267: p. 1615.

Bousche, O., Braiman, M., He, Y.-W., Marti, T., Khorana, H.G., and Rothschild,
K.J., J. Biol. Chem., 1991. 266: p. 11063.

Braiman, M., Mogi, T., Stern, L., Hackett, N.R., Chao, B.H., Khorana, H.G., and
Rothschild, K.J., Proteins: Str'uct. Func. Gen., 1988. 3: p. 219.

Engelhard, M., Gerwert, K., Hess, B., Kreutz, W., and Siebert, F., Biochemistry,
1985.24:p.400.

Mantele, W., Nabedryk, E., Tavitian, B.A., Kreutz, W., and Breton, J., FEBS
Letters, 1985. 187: p. 227.

MacDonald, GM. and Barry, B.A., Biochemistry, 1992. 31: p. 9848.

MacDonald, G.M., Bixby, K.A., and Barry, B.A., Proc. Natl. Acad. Sci., 1993.
90:p.11024.

MacDonald, G.M., Steenhuis, J.J., and Barry, B.A., J. Biol. Chem., 1995. 270: p.
8420.

Bernard, M.T., MacDonald, G.M., Nguyen, A.P., Debus, R.J., and Barry, B.A., J.
Biol. Chem., 1995.270: p. 1589.

Tavitian, B.A., Nabedryk, E., Mantele, W., and Breton, J., FEBS Letters, 1986.
20l:p.151.

Roepe, P., Ahl, P.L., Gupta, S.K.D., Herzfeld, J., and Rothschild, K.J.,
Biochemistry, 1987. 26: p. 6696.

Chapter 5

LOW FREQUENCY MODES IN
THE LIGHT -INDUCED FI‘IR DIFFERENCE SPECTRUM

OF PHOTOSYSTEM ll

Abstract

We report the first light-induced FTIR difference spectra of photosystem II which
include data from the low frequency (600-1000 cm“) region of the spectrum. This is the
region where metal-ligand vibrations and metal-substrate vibrations are expected to be
observed. These spectra, which show the vibrational differences as the enzyme is
advanced from the SOA state to the SZQA- state, display a multitude of vibrational modes
that are attributable to 1) amino acid vibrations that are perturbed by the advancement of
the redox state of the enzyme, 2) protein backbone modes, and 3) metal-ligand species
whose bonding interactions are altered upon advancing the manganese cluster from the Sl
state to the S2 state. Both sub-chloroplast BBY-type particles and highly-refined reaction
center core preparations were utilized for these studies. Upon manganese depletion of the
samples, the difference spectra showed reproducible changes in the ~630 cm'l region of
the spectrum where manganese-substrate bonds are expected to appear. Additionally,
when the samples were induced to oxidize water, in the presence of lsO-labelled water,
reproducible changes of the ~632 cm" peak were observed. The frequency to which the
vibration is expected to shifi, based upon model compound studies, is currently below the

range of our instrumental capabilities. Modifications are planned to extend the low-
112

 

I

 

Ii!“ 1! 3...... 03...... at. .0 7‘ r. f
J1... .u. .. 4." u. ..r: ... _. __ ..r;

. . .
.. .1 . u 6,-1.1: . I 1.
~1.?....._.. .:o ..o... .34 .4 “up. . I3... . urrplli.

113
frequency capability of the instrumentation, which will allow the observation of modes

down to 350 cm". Additionally, photosystem 11 samples which display the g=4.1 EPR
signal of the 82 state of the enzyme were studied. Vibrational differences in the low
frequency region of the spectrum were observed, which is consistent with the occurrence
of a structural rearrangement of the manganese cluster under the conditions in which the

g=4.1 EPR signal is exhibited.

Introduction

Photosynthesis is the process whereby higher plants, cyanobacteria, and some
bacteria harvest light, and use that energy to separate charges across cellular membranes.
Ultimately the energy is used to drive metabolic processes. The infrastructure that all
photosynthetic organisms utilize to generate this initial charge separation is universal.
Energy, in the form of light, is harvested and used to create a charge separated state
between a dimer of chlorophyll molecules (the “special pair”), and a quinone molecule,
QA. The mechanism by which the positively charged “hole” on the special pair of
chlorophyll molecules is filled differentiates non-oxygen evolving photosynthetic
bacteria, from those organisms that evolve oxygen, such as cyanobacteria and higher
plants. In the non-oxygen evolving systems, where the photosynthetic enzyme is called
the bacterial reaction center, the hole on the special pair of chlorophyll molecules is filled
by reducing equivalents from cytochrome c molecules. In cyanobacteria and higher
plants, the enzyme that performs the initial steps of photosynthesis is known as

Photosystem 11 (P811). In PSII, the hole on the chlorophyll dimer (called P680), is filled

114

by an electron which originates from the oxidation of a solvent water molecule. This
oxidation of water results in the production of molecular oxygen (see Figure 5-1).

In 1937, it was determined that the element manganese was intimately involved in
the water oxidation process[l]. In these studies it was shown that, if manganese were
removed from the growth media of plants and algae, the organisms were no longer
capable of oxygen evolution. If the manganese was replaced, oxygen evolution was
rapidly restored. Over time, the debate regarding the stoichiometry of manganese in PSII
has converged, and it is widely agreed upon that there are four manganese atoms per

photosystem II reaction center. In addition to the four manganese ions, it is known that

Ca2+ and C1' are essential cofactors; their absence abolishes oxygen evolution.(For a

comprehensive review of PSII, see [2-4]). A cartoon that depicts the primary
photochemical events of PSII is shown in Figure 5-2. Upon photoexcitation, an electron
is ejected from the special pair of chlorophyll molecules called P680. This electron
travels through a pheophytin molecule, to the primary plastoquinone molecule Q A, and
then to a second plastoquinone molecule QB, which can accommodate two electrons. QB,
once doubly reduced, dissociates to the stromal side of the chloroplast membrane. The
reducing equivalents on this quinone molecule are ultimately used in the reduction of
NADP” to NADPH. P680+ is one of the strongest oxidants known in biological systems
(~1.1eV). Once oxidized, it is immediately reduced by an electron from a redox active
tyrosine residue of the protein backbone, YZ, to give a neutral tyrosine radical, Yz- . This
radical is then reduced by electrons derived from a substrate water molecule that is

interacting with the manganese cluster.

115

Photosystem II

, quinone pool

 

Stroma

Lumen

(Mn),, Ca”, cr

Figure 5-1: Schematic of the Photosystem II reaction center. Numbers are approximate
molecular weights of the constituent polypeptides. See text for description of
components.

116

Photosystem II

, .4, quinone pool

  
 
  
 

hv“
Stroma \1.

3‘, l l rl ‘1] l l ‘5}.

l. ‘,
illitelrylldda:
1.. =1 ‘l' 1‘

mo _
O,+ 4H*+ 4e-

Figure 5-2: Schematic depicting the primary photochemical events in photosystem II.

After four
released a
active tyru
radical fo
Tl

sink as 5

action. T

remains t

of water

require 11

(1)

k)

this proc
SUbstrate

c“H’entlj

Played b
OXYgEn
Subjects
OXYgen

117

After four electrons are removed from the manganese cluster, molecular oxygen is
released and the manganese cluster is reduced back to the So state. There is another redox
active tyrosine residue, YD, whose physiological role is unknown, that is stable in the
radical form, and exhibits a dark stable EPR signal.

That an organism has developed an enzymatic method whereby a thermodynamic
sink as stable as water can be oxidized is a remarkable achievement of evolutionary
action. The step-by-step molecular mechanism of oxygenic photosynthesis, however,
remains one of the puzzling mysteries of biophysical chemistry. From the stoichiometry
of water oxidation (1) it is clear that this process involves multiple electrons, which must

require the enzyme that carries out this

(1) 2H20= 02 + 4H“ + 4e“ E° = 0.93 V @ pH 5.0

this process to exert considerable biochemical and physiological control. Whether the
substrate water is oxidized in a sequential, step-wise fashion, or in one concerted step, is
currently the subject of much debate.

Several key experiments have provided critical insight into the specific role
played by manganese ions in the water oxidation process. In 1969 an experiment where
oxygen evolution from chloroplasts of plants was measured as the samples were
subjected to short, saturating flashes of light[5]. It was noted that there was a burst of
oxygen released upon the third flash given to the dark-adapted samples, and subsequently

on every fourth flash thereafter. This observation led Bessel Kok, to propose the Kok

cycle for (
there are

These statn
the most 1
samples. I:
state is re:
process. g
thinking rt
Specific, 1
manganeSt

OVer the 01

 

 

118
cycle for oxygen evolution [6](Figure 5-3). This model for the “S-states” proposes that

there are five distinct oxidation levels involved in photosynthetic oxygen evolution.
These states, labeled So-S4, represent different redox levels of the enzyme, with S0 being
the most reduced form of the enzyme, and S4 being the most oxidized. Dark adapted
samples, primarily in the SI state, are consecutively oxidized by one electron until the S1
state is reached. At this point water is oxidized to dioxygen in a concerted four electron
process, giving back the So state with the release of Oz. The Kok model inspired much
thinking regarding the exact mechanism of oxygen evolution. However, with a lack of
specific, molecular-level information regarding the structure and geometry of the
manganese cluster, little evidence was available to prove the feasibility of any one model
over the others.

Another significant advance in the study of oxygen evolution in photosystem 11,
occurred in 1981 when Dismukes and co-workers reported observing an EPR signal from
dark adapted PSII samples given a single flash of saturating light [7]. As with oxygen
evolution, the amplitude of this signal also exhibited a periodic four flash dependence.
This behavior is indicative of the EPR signal arising from the S2 state of the enzyme. This
“multiline” signal (g=1.98 @ 9 GHz) consists of 18-22 lines, each separated by 75-90
gauss, and was assigned to a mixed-valence cluster of at least two exchange coupled
manganese atoms (Figure 5-4). When l7O-labeled water was used as the substrate, slight
line broadening in the multiline signal was observed[8] which was attributed to the
interaction between the S=5/2 l7O nucleus and the S=l/2 spin state of the manganese

cluster. In PSII samples treated with the substrate analog NH3, the multiline signal was

ll9

 

Kok Cycle

Figure 5-3: Kok model for oxygen evolution.

dx "/1111

 

120

 

 

dx'VdH

 

 

l r l 1 l 4 l 1

0 1000 2000 3000 4000 5000

 

gauss

Figure 5-4: Multiline EPR spectrum of BBY particles, 200K illumination. 8K sample
temperature. Spectrum is the difference between a EPR spectrum acquired after
illumination at 200K, and the EPR spectrum of the sample before illumination.
Microwave power: 2 mW, 20 gauss modulation amplitude, gain =1x105.

alterec
manga
being

oxidati

manga
g=4.1 .
signal.
which
switch
widely

(IIIJV.

the org

OEC) .

 

121

altered, which suggests a direct interaction between the ammonia molecule, and the
manganese cluster[9, 10]. Both of these studies implicate the manganese ensemble as
being both the site of substrate binding as well as the site of subsequent substrate
oxidation.

Under conditions of Cl' depletion, or substitution with F', NH, or Sr”, the
manganese cluster exhibits an EPR signal attributed to the S2 state that is observed at
g=4.1. The intensity of this signal appears at the expense of intensity of the 82 multiline
signal. The g-value at which it appears is consistent with an S=3/2 or S=5/2 spin state,
which implies that a magnetic reorganization must occur for the manganese cluster to
switch between displaying the S=1/2 multiline signal and the g=4.1 signal. Because it is
widely agreed that the oxidation states of the individual manganese ions in the S2 state are
(III,IV,IV,IV), the g=4.1 signal most likely arises from a S=5/2 state.

Other techniques that have contributed to the advancement in the understanding of
the organization of the manganese cluster (also called the Oxygen-Evolving Complex, or
OEC) of photosystem II are the various X-ray absorption spectroscopies, particularly X-
ray Absorption Near Edge Spectroscopy (XANES) and Extended X-ray Absorption Fine
Structure spectroscopy (EXAFS). EXAFS, when applied properly, has the potential to
provide information regarding the structure and coordination environment of a metal
ion[11-15]. While the validity of the various experiments that have been performed are
vigorously debated within the EXAFS community, one point of agreement that has been
reached is the existence in the manganese cluster of a metal-metal distance of 2.7A and a
second distance vector 3.3 A in length. Whether this 3.3 A distance is a Mn-Mn distance

or a Mn-Ca2+ distance is not known. The 2.7 A distance, when compared to data acquired

fron
ator
the I
that
man

(to 2

respt

Slates
eitct -
561 fr

abou

122
from model compound studies, is diagnostic of a an2 structure, where the two oxygen

atoms form a bridge between the two manganese atoms in a di-u-oxo arrangement. On
the basis of these measurements and the EPR results, a model for the OEC was proposed
that consisted of two di-u-oxo dimers ( the so—called “dimer of dimers” model) of
manganese atoms (to accommodate the 2.7 A distance), linked by a mono-u-oxo bridge
(to accommodate the 3.3 A distance)[16]. Further experiments on layered, oriented PSII
samples, showed evidence for an open “C” geometry of the two di-u-oxo cores with

respect to one another (Figure 5-5) [17].

\ ....... own...
}Mn O/ \
O
/ /
/C O 3.3A
0...,\ . I
'Mn ............ 0- ......... . ........

Figure 5-5: Open “C” model for the manganese cluster of Photosystem II

XANES is a technique that can be used in the determination of the oxidation
states of metal atoms[18, 19]. By examining the energy of the X-ray photon necessary to
eject an inner-shell electron from a metal atom, and then comparing that energy to a data
set from model compounds of known oxidation level, it is possible to draw conclusions

about the oxidation states of metal centers in proteins. While these experiments also are

1101

spec
ques
som
som

oxid

Flgm
WHIEr

0165:

disa 4

123

not without controversy, there are several points regarding the oxidation levels of the
manganese cluster in the various S-states that are agreed upon by a majority of the X-ray
spectroscopy practitioners[20]. These oxidation schemes are depicted in Figure 5-6. A
question remains as to whether manganese is oxidized on the S2 to S3 transition, or if
some organic cofactor is oxidized to generate an organic radical. Manganese XANES, in
some studies, does not show the characteristic small edge shifi associated with metal

oxidation[20]. However, the data acquired from these experiments, and more particularly,

s, 82
(Mn’v, Mn”, Mn”, Mn”) 4— (Mum, Mn”, an, Mn”)

(Mn'i Mn' (Mnm, Mn‘“, Mn”, Mn”)

V, Mnlv, Man)
s, \ s.

(Mnll, Mnlll, Man, Man)
So

Figure 5-6: Scheme of oxidation states of manganese cluster during the catalytic cycle of
water oxidation.

the subsequent treatment of these data, is subjective, and small subtle shifts can appear or

disappear depending upon the specific data treatment method utilized[21]. Under certain

non-oxyg
observed
fluoridefi
the S2 to
effective
that this

induced

OEC ha
incorpo:
COmprei
Figure 1
a hydro
four m;
acts 11

result 1'
and th

Olher

 

anoth
leVel,
dime

of S:

 

of g;

 

124

non-oxygen evolving conditions, a 90-180 gauss EPR signal centered at g=2 can be
observed in PSII samples that have been calcium depleted, or substituted with
fluoride[22-24]. Some have attributed this signal to a histidine residue that is oxidized on
the S2 to S3 transition in lieu of manganese [25, 26]. However, this assignment has been
effectively challenged by Britt and co-workers who, using pulsed EPR methods, showed
that this 90-180 gauss radical can be assigned to tyrosine Z in an altered environment,
induced by the removal of calcium[27].

All of the information regarding the possible structure and oxidation states of the
OEC have been assimilated in a model for oxygen evolution[28]. This model attempts to
incorporate the available structural and chemical information regarding the OEC, into one
comprehensive mechanism for water oxidation. A schematic of the model is presented in
Figure 5-7. A key aspect of the model is that the redox active tyrosine residue, Yz, acts as
a hydrogen atom abstractor. In the S0 state, substrate water is bound to the OEC, whose
four manganese atoms are at the (II,III,IV,IV) or (III,III,IIl,IV) oxidation level. YZ then
acts upon the substrate water, abstracting a hydrogen atom from the bound water. The
result is the SI state, in which the manganese cluster has been oxidized by one electron,
and the substrate now consists of a bound hydroxide ligand oriented cofacially to the
other substrate water molecule. Upon the next oxidation, which results in the S2 state,
another hydrogen atom is abstracted, leaving the manganese in a (III,IV,IV,IV) oxidation
level. This is the state that gives rise to the multiline EPR signal. A manganese (III, IV)
dimer in which the two manganese are antiferromagnetically coupled, has an overall spin
of S=1/2. An antiferromagnetically coupled (IV,IV) dimer of manganese atoms has a spin

of S=0. Coupling these two spin states gives an overall spin state of S=l/2, which is the

spin sta
observe
reorgan
dimer c
an oxo

0f the

necessi
0n the

the tyr<
pair, w
the din
the 0pc
oxoxhy

S130nta

 

125

spin state attributed to the manganese ensemble when the multiline EPR signal is
observed[20]. An interesting proposal in this model is that, in order to avoid large
reorganization energies, Yz must only abstract hydrogen atoms from the top half of the
dimer of dimers cluster. Upon initial formation of the 82 state, a species that is formally
an oxo species is generated, while there is still a bound water molecule on the bottom half
of the dimer. However, due to the basicity of this terminal oxo group, or due to the
necessity for the cluster to remain charge balanced, a proton from the substrate molecule
on the bottom dimer moves to the oxo moiety on the top half of the dimer. This allows
the tyrosine residue to continue abstracting hydrogen atoms from the top half of the dimer
pair, which obviates radical motion to reach the substrate molecule on the bottom half of
the dimer. The result, in the S2 state, is two hydroxide ligands cofacially oriented across
the open “C” end of the manganese cluster. Further oxidation to the S3 state results in an
oxo/hydroxo substrate configuration, and the S4 state transiently forms cofacial oxos, that
spontaneously condense to give dioxygen, while the four electrons from the oxos (still
formally at the oxidation level of water) reduce the manganese cluster back to the So state.
The role of calcium in this model is not clear, however, it could act as a substrate handler,
ferrying new substrate molecules into the active site of the enzyme. It is known that if
Sr“ is substituted for Ca”, oxygen evolution is supported, albeit at reduced rates (~40%),
while other divalent or univalent cations are ineffective at supporting O2 evolution[29].
This observation demands further experimentation before divalent cation activation of

PSII becomes well understood.

126

 

| Water Oxidation — Proposed Reaction Cycle I

 

Figure 5-7: Proposed sequence of steps in the water oxidation process[28].

Tl
pieces of
mangane
substrate
hyperfin
phenolic
perform
this. the
freedom
cofacto

amino

reducta

fOund
kcal/
hydro.

 

 

 

&

11de

no“

 

 

of it

 

127
This hydrogen abstraction model for oxygen evolution is supported by many key

pieces of evidence. Electron spin echo modulation studies place Yz within 3.5-5 A of the
manganese ensemble, well within range for Yz to act directly upon manganese bound
substrate molecules[27, 30]. It was determined by examining the line widths of the
hyperfine couplings of the unpaired spin of the tyrosine Z radical to the B-protons of the
phenolic side chain that Yz has significant rotational mobility, which implies that it
performs a role more complex than that of a simple electron transfer cofactor. Contrary to
this, the other redox active tyrosine residue in PSII, YD does not enjoy the rotational
freedom of Y2, indicating it is more likely to play the role of a pure electron transfer
cofactor[31]. There are several 02 dependent metalloradical enzymes that utilize an
amino acid radical as a hydrogen-atom abstraction agent. These include ribonucleotide
reductase, galactose oxidase, and prostaglandin synthase[32, 33].

The bond dissociation energy (BDE) of the phenolic O-H group on Yz has been
found to be approximately 86-87kcal/mol. This value, when compared to the ~120
kcal/mol BDE for water would indicate the incompetence of tyrosine to perform
hydrogen abstraction upon substrate water. However, the BDE of water ligated to
manganese falls to ~82-94 kcal/mol, and to ~82-83 kcal/mol for hydroxide bound to
manganese. These values indicate that, indeed, Yz could be competent in performing
hydrogen atom abstraction on substrate water or hydroxide molecules[34-37].

While the Hoganson model incorporates much of the experimental information
now known about PSII, it is not widely accepted as the “correct” answer to the question

of the mechanism of photosynthetic water oxidation. In fact, several other mechanisms

have beer
been prot
the mang
fonn an
oxidatim
a clear.

outright
contain

fifquen

obtain 1
in phot
lack oi
II. An

utiliZe

PTO'tei]
exper

regan

Cthr

 

lend

ma

128
have been proposed that disagree with the mechanism described above. One model has

been proposed that utilizes two oxyl radicals in a di-u-oxyl bridging geometry in one of
the manganese dimers[20] as the species present in S4. The oxyl radicals then condense to
form dioxygen. Despite the differing theories for the precise mechanism of water
oxidation, it is clear that the various models that have been proposed need to be tested in
a clear, conclusive manner so that they may be fine tuned, overhauled, or rejected
outright. One aspect of the new series of models is that the S-states are predicted to
contain molecular structures that will present unique and characteristic signals in the low
frequency region of their vibrational spectra.

It has been a goal of ours to develop experimental methodologies whereby we can
obtain information about the low frequency vibrations exhibited by the manganese cluster
in photosystem II, and thus contribute specific information regarding the presence of (or
lack of) specific chemical species at certain points in the catalytic cycle of photosystem
II. An example of the utility of this type of information comes from the experiments that
utilized time resolved resonance Raman spectroscopy to investigate the respiratory
protein cytochrome c oxidase, which catalyzes the reduction of O2 to water[38, 39]. These
experiments provided vibrational data that facilitated the formulation of proposals
regarding dioxygen reduction in the cytochrome c oxidase. However, the presence of
chlorophyll molecules in PSII obscures any signals arising from the manganese, and
renders a resonance Raman approach ineffective.

Because Raman spectroscopy is not a viable option. for the study of the

manganese cluster in PSII, an obvious alternative is infrared spectroscopy. However, as

was discu
is not inf
reaction-1
chahengi
high-qua
differenc
spectra.

Specific

Hogans
Water b
two lig
11 Sam
adV’anc
electr
0CCur
Onet
bOun
diffe
vibr.
the K
ener

Orig

129
was discussed in chapter 4, low-frequency information in the infrared spectra of proteins

is not information that is easily obtained. But, as we have found, the judicious use of
reaction-modulated infrared difference spectroscopy, can give us a foothold into this .
challenging area of spectroscopy. We have developed methods whereby we can obtain
high-quality Light-Induced FTIR difference (LI-FTIR) spectra that reflect the vibrational
differences between the S, and the S2 state of the OEC. Within the vibrational difference
spectra, we intend to find information that will provide direct evidence of the existence of
specific molecular structures in the reaction center.

As an example of the type of information we plan to obtain, consider the model of
Hoganson et a1 presented in Figure 5-7. In the S, state, it is postulated that there will be a
water bound to one manganese, and a hydroxide bound to the second manganese. These
two ligands to manganese will have characteristic vibrational frequencies. If photosystem
11 samples are illuminated at 200K the S, state will advance to the S2 state. Further
advancement to the higher S-states is not possible due to the fact that at 200K or lower,
electron transfer from Q A to Q,, is inhibited and thus only a single charge separation can
occur[40]. In the S2 state it is theorized that there are two hydroxide molecules bound,
one to a Mn(IV) ion, and one to a Mn(III) ion. In 8,, there was postulated a hydroxide ion
bound to a Mn(III) ion, and a water bound to another Mn(III) ion. In an infrared
difference spectrum of the S2 state minus the S, state, negative modes will correspond to
vibrations that disappeared upon formation of the 82 state, i. e. those that corresponded to
the S, state. Assuming that the Mn(IIl)-OH bond in S,, and the Mn(III)-OH bond in S2 are
energetically degenerate, then the difference spectrum will show a positive mode

originating from the Mn(IV)-OH and a negative mode from the loss of the Mn(III)-HZO;

the Mi
the id
hr1n=C

vibrat

Where
neces
in 0rd
(EPR
their

result

Matt

‘0 thc

cOnsi

 

130
the Mn(Ill)-OH modes cancel each other out. With the use of isotopically labeled water,

the identification of these modes should be straightforward, as a Mn-HZO, Mn-OH, or
Mn=O structure will each present a unique and identifiable isotope dependence in its
vibrational frequency.

In order to gain access to the low frequency region of the infrared spectrum,
where these intermediates in the water oxidation process will vibrate, it has been
necessary to devise techniques to remove most of the bulk solvent water from the samples
in order that the low frequency vibrational modes are not obscured. Control experiments,
(EPR and oxygen evolution) have been performed to ensure that the samples retained
their biological integrity, and are thus fully competent to perform water oxidation. The

results of these experiments are described here.

Materials and Methods

Photosystem ll enriched subchloroplast particles (BBYs) were prepared according
to the methods of Berthold et al [41]. The samples were suspended in SMN buffer which
consisted of 0.4M sucrose, 50mM 2-(N-Morpholino)ethanesulfonic acid (MES) (pH 6.0),
and 10mM NaCl. The particles exhibited oxygen evolution rates of 600-700
umoles/Oz/mg chl/hour. For FTIR measurements, the samples were resuspended in 5mM
MES, 5mM CaClz, 5mM NaCl. The samples exhibited no decline in oxygen evolution in
this medium. Reaction center cores (RCCS) were prepared according to the methods of
Ghanotakis et al [42, 43]. Briefly, BBY particles were solubilized with the non-ionic

detergent octylthioglucoside. The solubilization served to strip away membrane

fragments.
The RC C:
for FTIR
(FTIR bu
incubatic
l
the pres
EDTA‘
the pr:
manga
remai,

evolu‘

With
04:,
kDa
Src

buf

131

fragments, and the light harvesting proteins, leaving an oxygen evolving core preparation.
The RCCs exhibited O, evolution rates of 1400-1600 umoles O,/mg chl/hour. Samples
for FTIR were resuspended in 150mM sucrose, 5 mM MES, 5mM NaCl, 5mM CaCl2
(FTIR buffer). Oxygen rates were not affected by resuspension is such a medium, nor by
incubation in the dark, at 0°C, for ~24.

Manganese depleted samples were prepared by 1) incubating BBYs or RCCs in
the presence of 0.8M Tris(hydroxy-methyl)aminomethane (pH 8.0) (“tris”), 1.00 mM
EDTA for 30 minutes under room light at 0° C[44], or 2) by incubating BBYs or RCCS in
the presence of 10mM NH,OH, for 30 minutes, at 0°C, in the dark[45, 46]. After
manganese depletion, samples were washed twice in FTIR buffer to remove any Mn2+
remaining after manganese depletion. Samples so treated exhibited no detectable oxygen
evolution.

For the g=4.1 studies, samples were prepared by either substitution with Sr” or
with ammonia. The Sr2+ samples were prepared by washing the reaction center cores in
0.4M sucrose, 50mM MES (pH 6.0), 2 M NaCl. This treatment removes the 17 and 23
kDa polypeptides. The samples were then resuspended in FTIR buffer containing 40mM
SrCl,. Ammonia treated samples were prepared by resuspending the samples into FTIR
buffer and adding 100mM NH4C1.

FTIR samples were prepared by depositing an aliquot of PSlI’s onto 25mm
diameter AgCl IR windows (Wilmad). The amount of sample deposited was calculated to
give a total quantity of chlorophyll of 30ug, or ~10"° moles of reaction center cores,

based upon 250 chlorophyll molecules per BBY particle, and 75 chlorophyll molecules

per reactior

of 1.3 cm

for 15-20
integrated
routinely
mode at

(

 

partially
FTIR 52
Oxford

control

beams
l .00)

(see
abso
Sam
teml
for
C01
a9

Si

132

per reaction center core particle. The samples were deposited to give a sample “spot size”
of 1.3 cm in diameter. The samples were then dried under a stream of dry nitrogen gas,
for 15-20 minutes, until no liquid water was visible under visual examination. The
integrated intensity of the infiared water absorption band centered at ~3500 cm" was
routinely 8-10 times that of the integrated intensity of the Amide 1 protein absorption
mode at 1650 cm".

Control EPR experiments were performed on either fully aqueous samples, or on
partially hydrated samples deposited on to mylar, and dried in the same manner as the
FTIR samples. EPR spectra were run on a Bruker ER-300 Spectrometer equipped with an
Oxford Systems Liquid Helium cryostat for cryogenic studies, or at 200K for FTIR
control studies.

F TIR experiments were performed with a Nicolet 740 spectrometer with a KBr
beamsplitter and a Graseby Infrared MCT detector (Graseby Infrared Model MCT-16-
1.00). The sample was cooled to 200K by using a home-built nitrogen gas flow cryostat
(see chapter 4 for a description of the sample temperature system). The absolute
absorbance of the amide I band of the sample did not exceed 1.0 absorbance unit. The
sample temperature was controlled to i0.l°C with a LakeShore Cryogenics Model 321
temperature controller for the experiments using reaction center cores. The temperature
for the experiments with BBYs was controlled with an Omega model 76000 temperature
controller to 102°C. Double sided interferograms were collected at 4 cm‘l resolution, and
apodized using a Happ-Ganzel apodization function. 500 scans were averaged for each

spectrum. Sample illumination was carried out with a 150W halogen bulb. Infrared

component
copper sulL
fiber optit
sampling
spectra is,
directly 1
the “clan.
Polynon
Offset tl
Specific
Flat b3,

the Say

Dfiaks

electxl

for B

Resl

de V-
ref]

Chl

 

133

components of the actinic light were removed by using two thick glass filters, 1 cm of 5%
copper sulfate solution, and a low frequency filter that passes visible light D650nm. A
fiber optic probe positioned next to the sample window transmitted the light into the
sampling chamber. 20mW of light was incident upon the sample. S,Q,.,-/S,QA difference
spectra were obtained by collecting a dark spectrum at 200K, illuminating the sample
directly in the cryostat, then collecting a spectrum of the illuminated sample. The ratio of
the “dark” to the “light” provided the S,Q,,—/S,QA difference spectrum. A smooth
polynomial function was subtracted from each light-minus—dark pair to correct for an
offset to the baseline. Multiple spectra (6-15 individual spectra, depending upon the
specific spectrum) were then averaged together to produce the final difference spectra.
Flat baselines were verified by ratioing two “dark” spectra together prior to illuminating
the sample. These “dark/dark” spectra were collected and examined and no reproducible
peaks were found in the region below 1000 cm'I ( for example, see Figure 5-16).

Oxygen evolution measurements were performed with a Clark-type Oxygen
electrode. 100 uM or lmM 2,6 dichlorobenzoquinone was used as an electron acceptor

for BBYs, and RCCs respectively.

Results

Spectroscopic investigations of photosystem II rapidly progressed with the
development of refined preparations of photosystem 11. Until 1981, the most highly
refined preparation consisted of plant chloroplasts where the membranes of the

chloroplast had been mechanically or biochemically broken. In 1981, however, Berthold

and cov
0f the c
100. T
biochei
Photos
memb:
molec
chlorc
prepa
Woult
EPR
signa
meth
YEtai

thes:

stud

 

134
and coworkers [4 1] reported a preparation of photosystem II in which a large percentage

of the chloroplast membrane had been digested away by using the detergent Triton X-
100. The remaining membranes were. highly enriched in photosystem II. This
biochemical breakthrough gave spectroscopists access to samples in which the
Photosystem 11 concentration was double what was previously attainable in thylakoid
membrane preparations. Thylakoid membranes typically contain 400 chlorophyll
molecules per reaction center, whereas BBY particles contain 220-250
chlorophylls/reaction center[41]. More importantly, however, is the fact that these BBY
preparations contain minimal photosystem I contamination, which, under illumination,
would contribute to any light induced difference spectra.

EPR Control Experiments-881’s: Because we are ultimately interested in observing FTIR
signals from the low frequency region of the infrared spectrum, it is necessary to devise
methods whereby we can remove the bulk of the solvent water from the sample, yet still
retain the biochemical activity of the samples. To establish the biological relevance of
these samples, it is critical to verify their biochemical integrity. Photosystem II has been
studied extensively by EPR[2]. Thus, we can use the well-characterized EPR signals as
diagnostic benchmarks to establish that our samples can withstand the partial dehydration
necessary for the infrared experiment. Figure 5-8a shows an EPR spectrum of BBY
particles in solution. The signal observed shows the dark stable radical signal attributed to
Y,,, a redox active tyrosine residue in photosystem II[47]. Figure 5-8b shows Tris-treated
(manganese depleted) BBY particles, in solution, at room temperature. In Tris treated
samples it can be seen that under illumination, there is a a light induced increase of the

tyrosine signal because there is no longer a reductant immediately available to reduce the

radicz
from
There
samp
035%
the 1
Street
F igu
parti

illun

belo
man
gen.
sigr
the

TESL

 

135
radical generated at Y,. Thus the radical signal from Y, is added to the tyrosine signal

from the YD. YD exhibits a highly stable EPR signal in O, evolving samples of BBYs.
Therefore it can be used as an intensity standard against which other signals from the
samples can be referenced. As discussed above , it is well known that the S, state of the
oxygen evolving cluster of photosystem II exhibits an EPR multiline signal. Therefore,
the multiline signal is an excellent control signal to verify that the S, state can be
successfully generated in the partially dehydrated samples needed for FTIR spectroscopy.
Figure 5-4 shows the multiline EPR signal generated in an aqueous sample of BBY
particles by illumination of the sample at 200K. It has been shown that continuous
illumination of PSII samples at 200K advances the S-state of the manganese cluster to the
S, state, but no further due to the fact that electron transfer from QA to Q, is inhibited
below this temperature[40]. Because Q, can only accommodate one electron, the
manganese cluster is limited to a single turnover. Figure 5-9 shows the S, multiline signal
generated in BBYs, by 200K illumination, in partially dehydrated samples. The multiline
signal generated in the partially dehydrated samples is equal in amplitude (normalized to
the amplitude of Y,,) and appearance as that generated in fully hydrated samples. The
results of this EPR experiment show that we can successfully generate the S, state in
dehydrated samples of BBY particles.

Because we will generate the S, state by illumination at 200 K, in the F TIR
experiment, it is important that we also characterize the formation of any other light
induced species formed in the dried samples under these conditions. This will ensure that

we are fully aware of the species present when making assignments in the F TIR

136

 

dx"/dH

 

 

 

I . l 4 l A l . l

 

3360 3380 3400 3420 3440
field (gauss)

Figure 5-8a: Dark stable EPR signal from YD in BBY particles. Microwave
power lmW, 2.5 gauss modulation amplitude, 2 minute scan, room temperature.

21?):

Fig

C01

 

137

 

under illumination

dx"/dH

 

 

 

 

1 . 1 1 . 1 .
3360 3380 3400 3420 3440

field (gauss)

Figure 5-8b: Tris-washed BBY particles, before, and under illumination. Same spectral
conditions as Figure 5-8a.

F i g,
Part1
gain

 

 

138

 

dx"/dH

 

 

 

 

. 1 . .
0 1000 2000 3000 4000 5000

field (gauss)

Figure 5-9: S,-multiline signal from partially dehydrated film of O, evolving BBY
particles illuminated at 200K. 2mW microwave power, 20 gauss modulation amplitude,
gain=lx105.

differ:
10a) ir
a smal
amour
signal.
light it
washe
induce
appro:
FT 1R

Partial
Preser.

ShOWs

139
difference spectra. When the dark stable tyrosine signal is examined at 200K (Figure 5-

10a) in partially dehydrated, O, evolving BBYs, it is clear that under illumination there is
a small change in the signal. This change can be attributed to the generation of a small
amount of chlorophyll radical, based upon the linewidth and g value of the light induced
signal. Integration of the signal before and under illumination allows quantification of the
light induced radical, and shows that it is present in 10-15% of the reaction centers. Tris
washed, manganese depleted, BBY samples illuminated at 200K show a very large light
induced increase of the signal, indicative of the generation of chlorophyll radicals in
approximately 70% of the centers (Figure 5-10b).

FT IR Spectra of BBY Particles: We have developed conditions under which we can
partially dehydrate BBY samples and still retain biological integrity as evidenced by the
presence of EPR signals seen in native, fully hydrated reaction centers. Figure 5-1 la
shows a LI-FTIR difference spectrum of oxygen evolving BBY particles between 1800
and 1000 cm". The spectrum was collected by recording the dark spectrum of the samples
at 200 K, illuminating the sample for 10 seconds, allowing the sample to dark adapt for
two minutes to allow any unstable, transiently formed, chlorophyll radical to decay, then
recording a light spectrum. The ratio of the dark and light spectra was taken directly
against each other to obtain the final difference spectrum. Difference spectra from 18
samples were then averaged together to generate the final difference spectrum. This
spectrum, the S,Q,,-/S,QA difference spectrum show vibrations attributable to the S,Q,,-

state as positive modes, and vibrations attributable to the S,QA state as negative modes. In

FigUl
Wash
mod,

140

 

A After illumination

 

 

 

1.! ]
» 1
I ‘Q
l— I / ‘ '
E _ Dark .
r»; l .
l l
.. l :
: l
+ if
1 l 1 l a I 1 I a I L
3340 3360 3380 3400 3420 3440 3460

field (gauss)

 

m A er illumination
,1 ' l
l l

1L1 1

l'il} I

dx"/dH
I

 

 

 

I Y

I I I I I
3340 3360 3380 3400 3420 3440 3460

‘ l ' I

field (gauss)

Figure 5-10a) (top) EPR spectrum, 200K, partially hydrated BBY particles, b) tris-
washed BBY particles before and after illumination. Microwave power = 0.2mW, 4 gauss
modulation amplitude, gain = 1x105

 

1. Exit... .2. T...........,..fin Iii. t3 ...xa.....a.._.......: p.73; ., , . .......r1e..7.. . -. 4... . ., 1 . r .e...1..:........ . «marital... a 11.11 . j n , _ . .. t... .1 . 1.... ... ..

I .I all pi .r? . I . . , . . .. . l

141

red is shown the average of the “dark/dark” spectra taken before illumination, which
gives an indication of the noise levels and any systematic fluctuations in the experiment.

In Figure 5-11b is shown the lower frequency range (1000cm"-600 cm") of the
same sample as Figure 5-11a. The technical advances described in the previous chapter
have allowed us to acquire and present data from the low frequency region for the first
time. A readily apparent feature of this spectrum is the presence of numerous difference
bands. This fact is encouraging in that it generates confidence that we are capable of
detecting absorptions in the low frequency region of the infrared spectrum and that the
low frequency region of the spectrum may contain data relevant to the mechanism of
photosynthetic water oxidation.

In Figure 5-l2a is shown the LI-FTIR difference spectrum of Tris-treated BBY
particles. In Tris-treated BBY particles, the manganese cluster has been removed, and
thus any vibrations in the difference spectrum of O, evolving samples that are attributable
to the manganese cluster, should not be present in the difference spectrum of Tris-treated
samples. Figure 5-12b shows the spectrum of the same sample in the 1000cm'l to 600
cm’l range. In Figures 5-11b and 5-12b it can be seen that the signal to noise ratio for the
peaks in the 680 cm" region is approximately two. In order to make clear, definitive
observation of the peaks, the signal to noise ratio must be increased. Because of this, it
would be advantageous to utilize a more highly resolved preparation of photosystem 11,

such as reaction center cores. This would increase the effective concentration of

7

142

 

0.0004 , , . , . , .

1479

0.0003 _
0.0002

0.0001

AA.U.

-0.0001

 

.00002 - =2

-0.0003 _

 

 

 

. . 1 . 1 .
1800 1600 1400 1200 1000

frequency (cm")

0 M07

 

 

 

 

 

. . . . .
1000 900 800 700 600

frequency (cm“)

Figure 5-11 a) (top) and b) (bottom) Black: S,Q,,-/S,QA- light induced FTIR difference
spectra of BBY particles, 200K, 9000 scans. Dash: “dark-dark” for noise reference.

I43

 

 

 

 

 

 

 

0.0004 . 1 . , . ,
' Tris-washed BBY, 200K
0.0003 -
........ O2 evolving BBY, 200K
0.0002 - "‘
0.0001—

; 0.0000 MAI», h it ,4. \ \... JV vim/4t. t”, “Aw ,- m
-0.0001 :'
-0.0002 -

-0.0003 -
l l A l . l 1
1800 1600 1400 1200 1000
wavenumbers
n mmn
l ‘7 I ' I

000015 TI’IS-Wflhcd BBY, 200K

. ........ O2 evolving BBY, 200K
0.00010-
0.00005 .. ,l .g

   

. é
%
)
a)
{if
S

 

 

 

. , . , .
1000 900 800 700 600

wavenumbers

Figure 5-12a) (top) and b) (bottom) Light-induced FTIR difference spectrum of tris-
washed BBY particles, 200K. 6000 scans.

144

oscillators in our samples, thus increasing the signal to noise ratio. However, the more
highly resolved preparations are more delicate, and thus require special conditions where
they can handle the partial dehydration process required for the FTIR work.

EPR Control Experiments-RCCS; Because of the increased reaction center concentration
in samples of reaction center cores as compared to subchloroplast preparations, it was
desirable to develop conditions where the more fragile reaction centers could withstand
the partial dehydration process and still exhibit the EPR signals attributed to fully native
reaction centers. Resuspension of the samples in a lightly buffered medium that contains
reduced amounts of sucrose, allows us to detect the various EPR signals with the dried
samples, that have been previously attributed to intact reaction center core preparations. If
the samples were suspended in a sucrose free buffer, such as that used for the BBY
experiments, then the dried samples exhibited the multiline EPR signal, but were not
competent in generating an alternate signal from the S, state, the g=4.1 signal. This g=4.1
signal is attributed to a magnetically different conformer of the S, state of the manganese
cluster. Additionally, it was found that with RCCS, the small amount of sucrose greatly
enhanced the signal amplitude in the Ll-FTIR difference spectra In the case of BBY
particles, the LI-FTIR signal amplitudes were not dependent upon the sucrose content of
the medium.

Figure 5-13 shows the multiline EPR spectrum generated in the partially hydrated
films of reaction center cores deposited on mylar strips. As in the case of the BBY
samples, the amplitude of the multiline signal generated in the partially hydrated films of
RCCs is equal in amplitude on a per chlorophyll basis as the multiline signal generated in

the firlly aqueous reaction center core samples. Figures 5-14a-c show the dark stable

145
radical YD spectra of the dried O,-evolving, tris-washed and Sr” substituted RCC films,

respectively, at 200K. Under illumination there is a significant increase in amplitude in
each spectrum, indicating that at 200K, a chlorophyll radical is generated in the samples.
Integration of the signals before and after illumination shows that chlorophyll radicals are
present in 50%, 85% and 75% for O,-evolving, tris-washed, and Sr2+ substituted samples
respectively. In Figure 5-15 is shown the EPR spectrum of the g=4.1 signal generated in
reaction center core samples by substituting Sr2+ for Call, or by introducing NH3 to the
samples.

FT IR Spectra of Reaction Center Cores: Figures 5-16a and 5-16b show the light induced
FTIR difference spectrum at 200K of partially hydrated O,-evolving samples of reaction
center cores, from 1800 cm‘I to 1000 cm". This spectrum is the average of 13 different
samples. For reference, the average of the dark/dark spectra are also included to allow
visualization of the random noise levels present in the spectra. Immediately apparent is
the increased signal-to-noise ratio of the spectrum acquired from RCCS versus the same
spectrum generated with BBY particles. Part of this can be attributed to the increase in
reaction center concentration in the RCCs over the BBYs. RCCs typically contain
approximately 80 chlorophyll molecules per reaction center[42] while BBYs typically
contain 3 times that number. Hence, the increase in reaction center concentration should
substantially improve the signal to noise ratio of the light induced difference spectrum.
As in Figures 5-11a and 5-11b positive modes correspond to the vibrations induced upon
formation of the S,QA- state and negative modes correspond to modes from the S,QA

state. In Figure 5-l7a and 5-17b, is presented the difference spectrum from Tris-treated

146

 

 

 

1 . 1 . 1 . 1
1 000 2000 3000 4000

field (gauss)

 

Figure 5-13: Multiline EPR spectrum of partially-hydrated reaction center cores. 20011g
chlorophyll dried on mylar. Microwave power: 2 mW, 20 gauss modulation amplitude,
gain = 1x105. Temperature = 8K.

147

15000 —

. A er illumination
10000 _

Before illumination

     

5000 -

  
 

dx"/dH

-5000 -

~10000 _

-15000 -

 

I l l l 4
3340 3360 33 80 3400

1
3420 3440 3460

1

field (gauss)

Figure 5-14a: EPR spectrum of O, evolving RCCS, before and after illumination.
Integration indicates the formation of chlorophyll radical in approximately ~50% of the
centers. Same spectral conditions as Figure 5-10.

148

 

/\ After illumination

 

 

 

 

33 _
13
’x
'O
1 L L l 1 l 1 l 1 l 1
3340 3360 3380 3400 3420 3440 3460
field (gauss)

Figure 5-14b: EPR spectrum of Tris-washed RCCs at 200K, before and after
illumination. Integration of signals shows generation of chlorophyll radical in
approximately 85% of centers. Same spectral conditions as Figure 5-10.

149

 

dx"/dH

-

 

 

 

I I I I f I I T I l I
3340 3360 3380 3400 3420 3440 3460

field (gauss)

Figure 5-14c: EPR spectrum of Sr2+ substituted RCCs at 200K before and after
illumination. Integration of signals showed chlorophyll radicals generated in
approximately 75% of the centers. Same spectral conditions as Figure 5-10.

150

 

dx"/dH
I

 

 

 

 

 

l 1 l 1 l 1 I 1 l

1000 2000 3000 4000 5000
field (gauss)

 

Figure 5-15a: EPR spectrum of Sr2+ substituted g=4.1 signal in partially hydrated RCC
sample. Spectral conditions same as Figure 5-13.

151

 

5x10‘

1

 

dx' /dH
c>
1

I

-5x10‘

 

 

 

I 1 I 1 I 1 I

0 1000 2000 3000 4000 5000

 

field (gauss)

Figure 5-15b: EPR spectrum of NH, substituted g=4.1 signal in partially hydrated RCC
sample. Spectral conditions same as Figure 5-13.

152
RCC particles. Because the tris-treated samples are manganese depleted, we expect that

any absorbances in the spectrum of O, evolving samples that are due to vibrations of the
manganese cluster should not be present in the spectrum of the manganese depleted
samples. As a control experiment, RCCs were manganese depleted using an alternative
method. Low concentrations (~1-10mM) of hydroxylamine (NH,OH) is known to
effectively reduce the manganese cluster to Mn”. EPR measurements of dried NH,OH-
treated RCCS exhibited similar quantities of light-induced chlorophyll radical as the tris-
treated samples. LI-FTIR spectra were acquired (Figure 5-18) to check for agreement
with the spectra acquired with Tris treated data.

Figure 5-19a and 5-19b shows the light induced FTIR spectrum of RCCs that
were resuspended in FTIR buffer using 97% '80 labeled HZO. The samples were allowed
to incubate on ice, in the dark, for 12-14 hours before they were deposited onto AgCl
infrared windows and dried. Oxygen evolution of these samples was unaffected by the
prolonged soak period. Not shown is the EPR control spectra of partially dehydrated
samples treated in this way. The signals were identical to fully active RCC samples.
Figure 5-20a and 5-20b shows the light induced F TIR difference spectrum of RCCS that
were resuspended in 97% '80 labeled FTIR buffer with lOOuM 2-6
dichlorobenzoquinone, a quinone molecule that can accept electron from the primary
quinone acceptor, QA, of PSII. The samples, before being dried on to the IR windows,
were incubated on ice, under room lights for 30 minutes to allow the manganese cluster to

cycle through the S-states several times. Control samples, in l6O buffer, were tested to

153

0.0010

 

“IO

02 evolving RCCs
........... dark/dark

 

0.0008 _
0.0006

0.0004

AA.U.

 

 

 

 

800 1600 1400 1200 1000

frequency (cm")

 

. —— O, evolving RCCs

 

 

 

 

l I ‘l I
1000 900 800 700 600

frequency (cm")

Figure 5-16 a) top and b) bottom. LI-FTIR difference spectrum of the S,Q,,-/S,QA state.
reaction center cores, 200K illumination. 6500 scans total. Dashed spectrum is dark/dark
for noise reference.

 

. . , 1 I3 2 . .l . I . 1. 1 1.5.4. ‘1
.n . 1:1 t. .. .1. .. . .r. u... a... a if. t. .L... I)... .. u’. , . a .._,.VI . _
m a ‘.m‘...__ ' -.I_- I .u... ‘.4 ’ . «.2. #‘IW‘fll. j. W .’.W‘..‘. h .‘L b .. .- awn .. . . . . .. . -. ,nu 1- .— i .. . y,Ve 4. .‘ . o L . J_‘r.r b I g a: J I n 1 . ‘ l 4 v

 

154

 

0.0010

 

 

Tris-washed RCCS
. ....... O2 evolving RCCS

   

 

 

 

 

 

 

   

 

 

 

 

 

0.0005
.. \Jfix .
:; 0.0000 iii-“:9. "WWW
<' "'1 " ”j "
< 2
-0.0005
-0.0010 1 1 l
1800 1600 1400 1200 1000
frequency (cm")
0.0004
Tris-washed RCCS
. ------- Oz evolving RCCS
0.0002- g T
. T \| '3 l,
:5 00000-37”; '2: ' -’§-"\,'-,.l\ ”if“ ,~":' "l1 . ' 5‘.
<1 ' ', ‘
-0.0002..
.00004 I r I , I ,
1000 900 800 700 600
frequency (cm")

Figure 5-17a) top and b) bottom. (black)LI-FTIR spectrum of tris treated reaction center
cores, 200K.(dash) O, evolving RCCs (Figure 5-16). Tris-spectrum average of 12

samples, 6000 scans. 200 K.

155

0.0010

 

- ——NH,OH treated RCCS
....... 0, evolving RCCS

00008 -
0.0006 .
00004 -
00002
:3 (10000
< .
< -00002

-0.0004

 

-0.0006

-0.0008 .

 

 

00010 - . . 1 - . -
1800 1600 1400 1200 1000

 

frequency (cm")

0.0004

 

__ NH,OH treated RCCS
. ....... O, evolving RCCS

750

0.0002 ..

U
.°
0
§
1

AA

 

-0.0002 _

 

 

-o.0004 . , . . . 1 .
1000 900 800 700 600

 

frequency (cm")

Figure 5-l8a) top, and b) bottom.(black) LI-FTIR spectrum of NH,OH treated RCCS
(manganese depleted), 200K, 6000 scans. (dash) O, evolving samples (Figure 5-16).

34......» .__ t... .

.3 II...-
‘9. to.

   

«

Adz: .: .v...ao :1 rr 0
p . 0 . .

b 1! I'douybr‘a It. LL

,’1.

W- Iii

its.

#4.; H... i

t. 41,» r

.114 .9.

. :1... ..7 o ..b...,...

I \Ilao

4ml-1An—

 

156

0.0010

 

h l.
0.0008 .- —— Osoaked RCCS

- ....... l"O (native) RCCS
0.0006 ,.
0.0004 :-
0.0002 ’

:4 0.0000

< -0.0002 ’

-0.0004

 

-0.0006 ..

-0.0008 r-

 

 

.0ch0 1 l I l l L l
1800 I600 1400 1200 1000

 

frequency (cm")

0.0004

 

____"0 soaked RCCS
7 ....... "'0 (native) RCCS

0.0002 e

 

0.0000 .>

AA.U.

 

-0.0002 —J

 

0.0004 , , , , t
1000 900 800 700 600

 

 

frequency (cm")

Figure 5-19a) top, and b) bottom. (black) LI-FTIR spectrum of RCCS soaked in 180-
labeled buffer. 6000 scans, 200K. (dash) O2 evolving RCCS (Figure 5-16).

157

0.0010

 

)- - ll
0.0008 _ ——RCCs cycled With 0

- ....... RCCs ("0)
0.0006

0.0002

 

 

 

.00010 . l . 1 . 1 .
1800 I600 I400 1200 1000

 

frequency (cm")

0.0004

 

——RCCs cycled with I'0
‘ ....... RCCs ("0)

0.0002 ..

0.0000 - ‘

AA.U.

 

 

-0.0002 -

 

 

00004 . , I , ,
1000 900 800 700 600

 

frequency (cm")

Figure 5-20a) top, and b) bottom. (black) LI-FTIR spectrum of RCCS turned over in
presence of 18O-labeled buffer. 6000 scans, 200K. (dash) O2 evolving RCCS (Figure 5-16)

158

ensure that samples treated this way, showed oxygen evolution rates and EPR spectra that
were consistent with native, untreated samples.

Figures 5-21 and 5-22 show the FTIR spectra acquired by 200K illumination of
samples that were treated with either a) ammonia (Figure 5-21) or b) Sr” (Figure 5-22).
EPR studies have shown that these conditions will generate the g=4.1 signal in lieu of the
multiline signal. When samples exhibit the g=4.1 signal, it has been proposed that the
manganese cluster is in a significantly different geometric arrangement compared to the
cluster under conditions that the 82 multiline signal is displayed. These spectra may shed

light on the origin of this difference.

Discussion

We have developed methods whereby we can acquire data on the vibrations
exhibited in the low frequency region for photosystem II preparations. The vibrations
associated with the specific molecular structures thought to be present during the catalytic
cycle of the oxygen evolving complex will be contained in this region of the infrared
spectrum Data in this frequency range have not been previously reported in the literature.
The specific advancements that we have made that allowed the acquisition of these data
include sample cooling techniques that allow the use of window materials that do not
obscure the low frequency region of the spectrum and biochemical techniques that allow
us to utilize the more highly resolved reaction center core preparation of PSII rather than

the cruder subchloroplast preparations utilized by other research groups.

l59

 

....... O2 evolv. RCC's
_— NH3 treated RCC's

A.U.

A

 

 

 

 

 

 

 

I ' I ' I ' l '
I800 I600 1400 1200 1000
frequency (cm")
0.0004
------- O, evolving RCCS
. _NH3treated(g=4.l)RCCs
0.0002-
-0.0002_
0.0004 . I . I . I .
l000 900 800 700 600

 

 

frequency (cm")

Figure 5-21a) top and b) bottom. (black) LI-FTIR difference spectrum of NH3 treated
RCCs. (g=4.1 conditions) 3000 scans. (dash) O2 evolving RCCS (Figure 5-16).

5:12.. 7.. . . ...: 2..
.....

i .. . ; 03-2....

 

 

. 0.5%

.I.

. .._ 3. niftrusi

:9

o.

..7.

L. 4...

 

v F. x... .

 

I60

 

0.0008 ------- O2 evolving RCCS
J Sr” subst. RCCS (g=4.1

 

 

 

 

 

 

 

1800 I l6I00 r 1400 I l2'00 I 1000
frequency (cm")
0.0004
------- O2 evolv. RCCs
Srzisubstituted RCCs (g=4. I)
0.0002 -

 

~0.0002 ..

00004

 

 

 

. . . . . . I . . . I . . .
IO 00 900 800 700 6| l0

frequency (cm")

Figure 5-22 a) top and b) bottom. LI-FTIR difference spectra of Sr“ substituted RCCs
(g=4.1) conditions. 4000 scans. (dash) O2 evolving RCCs (Figure 5-16).

I61

The first step in developing these methods, however, was establishing
experimental conditions whereby we could reproduce the infrared difference spectra
reported by other groups, for the SzQA-/S,QA transition of the enzyme. The data acquired,
shown in Figure 5-11, show the infrared difference spectra of partially dehydrated BBY
particles illuminated at 200K. When compared to previously reported spectra[48-50], it
can be seen that we have essentially identical spectra (Table 5-1). It is very difficult to
assign modes in this region of the spectrum due to the presence of numerous difference
bands and the overlapping predictions made by group frequency arguments. Table 5-2
shows a compilation of vibrational information and expected vibrational frequencies for a
number of protein backbone modes of vibration and amino acid side chain vibrations.
Despite the complexity of this region of the spectrum it is possible, however, to make
some very general assignments. The amide I mode of vibration from the protein
backbone(primarily composed of the C=O stretch of the backbone carbonyl moieties) is
centered around 1635cm°' - 1670cm“. These modes of vibration are sensitive to the
specific secondary structure of the protein, and thus large scale changes in the secondary
structure of the protein would be manifested as changes in the amide I modes of
vibration. While there are definitely difference peaks in the spectrum in the range of the
amide I vibrations, the fact that the peaks observed are no larger than the others in the
difference spectrum indicates that the protein does not undergo large scale conformational
rearrangement upon advancement of the S-state. If any differential peaks in that region of

the spectrum do arise from changes in the Amide I modes of vibration, the

162

Table 5-1 Vibrational Data (1800-1000 cm“) From Various LI-FT IR
Measurements of Photosystem II

 

 

SzQA'/SIQA SZQA'/SIQA SzQA'/SIQA SZ/sl QA" QA' Ch] +
This work This work [49] [5]] [49] [52] [53]
(RCCS) (BBY) (BBY) (RCC)
+1751 +1748 +1749 +1747
-l744 -1743 -l742 -l744 -l743 -l744
+1740 +1740 +1739 +1739 +1740 +1738
-l734 -1734 -l736 -l735 -l735 -l736
+1730 +1730 +1731 +1729 +1731
-l724 -l724 -l725 -1725 -1725 -1724
+1719 +1720 +1719 +1719 +1720 +1719
+1714 +1714 +1714
-l707 -l710 -l710
+1703 +1703 +1704 +1706 +1707
-l696 -l696 -l695 -1694 -l695
+1690 +1690 +1688 +1689 +1688 +1687
-l684 -1684 -1683 -l683 -l682 -l684
-1674 -1674 -l675 -l679 -1673 -l672
+1665 +1669 +1666 +1667 +1663 +1660 +1665
-l657 -1657 -l659 -1657 -1658 -l657 -l660
+1649 +1649 +1650 +1650 +1649 +1650
-l645 -l643 -1643 -l645 -1644 -l644
+1638 +1638 +1637 +1637 +1637 +1637
-1632 -l632 -l632 -l630 -1632 -l630
+1620
-1614
+1626 +1626 +1625 +1622 +1624 +1625
+1570
~1561 -1560 -1560 -1560 -1560
-1551
+1553 +1553 +1551 +1551 +1550
-1545 -1545 ~1544 -1544 -1542
-1533
+1532 +1537 +1532 +1532 +1530
-1522 -1524 -1522 -152l -1519
+1505 +1505 +1505 +1506 +1506
-1493
+1480 +1479 +1478 +1478 +1478
+1456 +1456 +1458 +1456
+1441 +1439 +1437
-l422 -1422 -1421
+1416 +1414 +1417 +1417 +1417
-l404 -l402 -l402
+1366 +1366 +1364 +1365 +1365

163

Table 5-2 Group Frequencies of Protein-based Vibrations

 

 

Frequency Mode Origin
(cm-1)
1750-1725 carboxyl C=O stretch Aspartate/glutamate side chain
~168O side chain C=O Asparagine
1670-1630 C=C quinone
1670-1635 Amide l (C=C, backbone) Protein secondary structures
1670-1680 CN3H,+ . Arginine side chain
1670-1640 H-O-H bend water
~1670 side chain C=O Glutamine
~166O N112+ deformation Histidine
1640-1610 NH,". asymm. deform. Lysine
1640-1630 CN,H,+ symm. stretch Arginine
~l620 NH2 deformation Asparagine
1620-1570 carboxylate C-O, asymm. stretch Aspartate/glutamate
~1610 C=C stretch quinone
~1550 N-H bend, Amide 11 Protein secondary structures
~1480-1460 C-O stretch semiquinone
~1480-1460 C-C stretch semiquinone
~1450-1350 carboxylate C-O symm. stretch

Aspartate/glutamate

 

164
amplitudes indicate that conformational changes in the protein must be subtle or

localized to the immediate vicinity of the manganese cluster.

Upon generation of the 82 state, an electron travels from the manganese cluster,
through the electron transport chain, and, at 200K, is inhibited from proceeding further
than the plastoquinone molecule, Q A, thus generating the anion radical Q ,3. Therefore, in
FTIR samples studied before and afier illumination at 200K, the QA-- species should
contribute to the infrared difference spectrum. This Q A“ spectrum has been observed, and
some of the modes were assigned previously by F TIR difference spectroscopy of
photosystem II samples[54]. The difference peaks at 1724 cm"/ 1719 cm'1 have been
assigned to the reduction of the primary quinone acceptor, QA. The peak at 1478 cm’1 has
been assigned to the C-0 stretch of the anion radical form of Q A. Additional peaks in the
1650 cm" to 1750 cm" range may arise from side chain carbonyls, or carbonyls on
chlorophyll molecules that undergo changes as the S-state advances. Between lSOOcm"
and 1650 cm", the antisymmetric stretches of carboxylates will be observed, as will the
N-H stretches of amino groups, such as are found in glutarnine or lysine residues. The
Amide 11 protein backbone mode (composed of N-H bending modes) usually appears
near 1550cm". Therefore it is reasonable to deduce that, analogous to the case of the
Amide I modes, large scale conformational changes do not occurr in the protein as the
manganese cluster is oxidized. The pair of peaks at 1364cm"/1403cm’l have been
attributed to the symmetric stretching vibration of a carboxylate-containing residue that is
involved with the coordination of the Ca2+ associated with the manganese cluster. The

antisymmetric stretch of the same residue is located at 1560 cm"/1587cm". This

I65
assignment is based upon studies of Cab-depleted samples, and the fact that upon lsN

labeling of the samples, these peaks were unaffected. The frequency difference between
the symmetric and antisymmetric modes in the SI and 82 states was used to propose that
the carboxylate in the SI state is bound in a bridging bidentate conformation, and in the S2
state, adopts a unidentate conformation. This is based upon the empirical observation
that, in model systems, the difference in frequency between the symmetric and
antisymmetric modes of a carboxylate ligand is a reflection of the specific mode of
binding of the ligand to a metal ion[50].

It has been shown previously that it is possible to detect the presence of as little as
5% chlorophyll radical in photosynthetic systems with FTIR [55]. Therefore, in the F TIR
spectra of BBY or RCC samples, we expect some contribution from chlorophyll radical
to the light-induced FTIR spectrum. Some modes from chlorophyll radicals in PSII have
been assigned previously[53]. Several vibrations in the SZQAJSIQA spectrum of BBY
particles presented in Figure 5-11 are attributable to chlor0phyll radical species present in
the sample. The negative mode at 1684 cm", and a positive shoulder at 1714 cm”1 are
both vibrations that have been assigned in chchhl difference spectra[53]. However, only
the most intense difference peaks from the chlfl/chl difference spectra are apparent in our
S,Q,I-/S,QA spectrum of BBY particles.

The low frequency region of the SZQA-ISIQA spectrum also exhibits a number of
modes of vibration. However in order to identify conclusively absorptions that arise from
the manganese cluster, it is necessary to acquire spectra that have better signal to noise
ratios. By using reaction center core preparations the effective concentration of reaction

centers is increased by a factor of three to four, which will improve the quality of the

166

spectra substantially. One issue regarding the use of reaction center cores is the relative
fragility of these preparations relative to the more robust BBY type particles. It was
necessary to develop conditions whereby these samples can be partially dehydrated, yet
still exhibit all of the signals associated with native, fully aqueous systems.

In Figure 5-13, the multiline EPR spectrum obtained from samples partially
dehydrated in FTIR buffer, which consists of 5mM MES (pH 6.0), 5mM NaCl, 5mM
CaCl,, and 150mM sucrose, is shown. It was found that normal multiline spectra could be
acquired under conditions when samples were dried without any sucrose present, but the
amount of chlorophyll radical generated by illumination was increased. Additionally, it
has been shown that, under a variety of conditions, native, O2 evolving PSII samples,
generate a different signal associated with the S2 state of the manganese cluster centered
at g=4.1[2]. Without sucrose present, it, however, was not possible to generate the g=4.1
signal in the dried PSII films. Apparently, the appearance of the g=4.1 signal depends
upon the presence of sucrose in the sample. A possible reason for this is the samples that
include small amounts of sucrose in the drying medium retain approximately double the
water after drying, than the samples that are dried in sucrose-free medium. It is possible
that this increased level of hydration is what allows the formation of the g=4.1 signal.

In Figure 5-16, the FTIR spectrum of partially hydrated samples of reaction center
cores is presented. 13 individual sample spectra were averaged to produce this final
spectrum. The signal to noise ratio of this spectrum is approximately 80, for the peak at
1478 cm". The signal-to-noise ratio of the same peak in the spectrum fi'om BBY

particles, is approximately 12. This is despite the fact that the spectrum of RCC’s was

167
acquired on the same total quantity of chlorophyll (300g), and was the result of fewer

total scans (6500 versus 9000 scans). The number of difference peaks appearing in the
low frequency region is striking. Unfortunately, few studies have been performed with
the goal of assigning the low frequency modes of proteins and amino acids. Table 5-3
lists the molecular structures, along with their projected frequencies, that may appear in
the SzQA-ISIQA difference spectrum of PSII. As these structures form, or are depleted in
the S,—)S2 transition, we expect difference peaks to appear in our spectra.

The protein backbone modes Amide 1V, V and VI are all expected to vibrate within the
range of 600 cm" to 725 cm". Each of these could contribute to the difference spectrum.
However, because we do not see large scale rearrangements in the protein matrix (as
would be evidenced by large Amide I or Amide II difference peaks),the protein structural
modes contributing to the difference spectrum in the low frequencies are, most likely, the
result of small, localized changes of the protein in the immediate vicinity of the
manganese cluster.

Other modes that could be contributing to the spectrum in the low-frequency
region include the bending motion of carboxylate residues. It has been shown that
carboxylates that bind metals, including manganese, display shifts in their infrared
absorption spectra as the metal to which they are bound undergoes a change in redox
level[56]. It is universally accepted that the manganese cluster is oxidized directly on the
SI to S2 transition, thus we can expect that any carboxylate residues that are ligated

directly to the manganese will display redox dependent shifts of their bending vibrations.

168
Table 5-3 Low-Frequency Vibrations Possibly Present in Photosystem II

 

 

   

 

 

 

 

 

 

 

Structure Approximate Vibrational
Frequency (cm")
Protein Backbone Modes HI-Tfl::.s.. ..1 ”A“ em- :0 5...” -.~' “ VIN-”m:
Amide v (NHob, CN,)‘ 725
Amide IV (Cob, CCS,CNCd) 625
Amide VI (C001,, CN,) 600
Amide VII CNHob, CN,, C00,) 200
Ligand Side Chain Modes ‘ F f ._,
T
//C\ - 500-720
0 ._._. O
bending
l/AN -
NV 600 650
Out-of-plane bending
I 600-720
0\
H
C-O-H bending (in plane)
Metal-Ligand Modes :_ g
Nh=0 75°
M‘t—O-I 400-500
O—O
/ 0 \ 600-800
Mrrzijf ::::Mn
‘0’
600-700
Mn<O\Mn
o/
Nh—Cl 300-400

‘ s, stretch; d, deformation, t, torsion; ib, in-plane-bend; ob, out-of-plane-bend

169
The EPR control studies performed on the dried samples demonstrated the

presence of chlorophyll radicals generated by illumination with visible light. Because
these are a light induced species it is conceivable that their presence contributes to the
overall light induced difference spectrum. However no infrared studies have examined
the changes exhibited in the low frequency region as neutral chlorophyll molecules are
oxidized to the cation radical form. One possible indication of the relative contribution of
chlorophyll radical to the difference spectrum comes from the control EPR studies. The
light induced increase of the dark stable tyrosine EPR signal observed in Oz-evolving
versus tris-washed material shows that the tris-washed material exhibited a much larger
light induced increase in amplitude after illumination. Based upon the line-width and g
value of this new signal in the EPR spectrum, it is clear that a chlorophyll radical species
is generated under illumination. Therefore, any signals in the FTIR spectrum that arise
from the chlorophyll radical species should exhibit increased signal amplitude in the tris-
treated material versus the O2 evolving material. In Figure 5-17, the light-induced FTIR
difference spectrum of tris-treated RCC samples, at 200 K, shows one mode, at 687 cm",
that has a slight increase in amplitude relative to the same peak in the O2 evolving
samples. This indicates that this peak may originate from a chlorophyll radical species, or
from a new species that coincidentally vibrates at approximately 687 cm". However, by
the same argument, the fact that there are no substantial increases in signal intensity in the
higher (1800-1000cm") region of the spectrum indicates that chlorophyll radicals
contribute only minimally to the overall light-induced difference spectrum.

The peaks observed below 1000 cm“I could also have their origin in manganese-

oxygen vibrations. In model compounds, vibrations that arise from manganese-oxygen

170

structures, such as di-u-oxo structures, are seen in the range of 600 cm'1 to 700cm". in
order to have confidence in our ability to detect these signals it is important to consider
the extinction coefficients of the vibrations that we intend to observe. By measuring the
infrared absorption spectrum of a model compound that contains the Mn202 structure, we
calculated an extinction coefficient of approximately 10‘5 mole". We can assume 80
chlorophyll molecules per reaction center core particle[42], and that 3 ug of chlorophyll
are interrogated in each experiment. Each oxygen evolving complex, of which there is
one per PSII, contains two an2 units. With these approximations, we calculate an
expected absorbance of ~10‘ absorbance units due to the an2 structures present in the
OEC. A can be seen from Figure 5-1 1, and the noise levels present in the spectra, that
these levels are well within our instrumental detection capabilities.

A legitimate question that can be asked regards the peaks that might be observed
from these manganese oxygen structures. As the di-u-oxo core is oxidized by one
electron, it might seem logical that the force constants of the individual bonds that make
up this structure will change, leading to peaks in the difference spectrum of the system.
However, one brief study on manganese-containing model compounds showed minimal
change in the resonance Raman spectrum of these compounds as a function of oxidation
state of the compound[57]. That does not preclude any di-u-oxo structures in the oxygen
evolving complex from displaying redox induced vibrational changes, however. It is
known that the spectroscopic behavior of metal ligand structures can change dramatically
when studied in highly asymmetric protein matrices as compared to the more symmetric

conditions of small inorganic coordination compounds. More comprehensive studies of

 

d _..A¢4 .

ib,! ..r

 

“- It. . la. . 2.2-.131; *3... c it. h... .2 (”Au/:2“). rifaoosxmufllt: s . . . n... .fVF .43....13. al.—35...: ml

..vu .

It“.

 

4
H

 

171

the redox dependent vibrational properties of structurally relevant manganese model
compounds need to be performed in order to draw conclusions about the contributions of
these vibrations to the low frequency difference spectra of photosystem 11.

As a set of control experiments, and ones that can start shed light on possible
assignments of the low frequency modes, manganese depleted RCC samples were
investigated under the same conditions as were the O2 evolving samples. When the light-
minus dark spectra of the Tris-treated samples were examined, the low frequency range
of the spectrum was notable for its similarity to the low frequency region of the oxygen
evolving samples. However, there were some potentially important differences. If any of
the low frequency peaks observed in the O2 evolving samples arise from manganese-
oxygen vibrations, then samples that are depleted of manganese should not show those
modes in a light induced difference spectrum. In Figure 5-23a, the 800-600 cm’1 range of
tris treated samples is overlaid with a spectrum of O2 evolving samples. It can be seen
that a peak at 632 cm'I in the O2 evolving samples is not present in the spectrum of the
tris treated samples. This observation could be attributed to one of two possible scenarios.
The first is that the peak at 632 cm‘I is, in fact, due to a manganese-oxygen or manganese
ligand vibration. Under manganese depleted conditions, if this vibration is the result of a
manganese-oxygen interaction then under manganese depleted conditions, it would not be
observed. If the vibration were from a manganese-ligand interaction, then it would either
not be observed if the vibration were primarily composed of normal coordinates from the

manganese-ligand interaction, or it would be perturbed if it was composed of primarily

0.0004

172

 

 

Tris-washed RCCS
. ....... O2 evolving RCCS

0.0002 -

720

A

 

'— 677

i

’63!

 

 

 

 

~0.0002 -
-0.0004 1 '
800 700 6 0
frequency (cm")
0.0004
— NH,OH treated RCCs
4 ....... 02 evolving RCCS

o
m
h

AA.U.

-0.0002 _

-0.0004 f

 

 

 

800

 

I .
700 6 l0

frequency (cm")

Figure 5-23za) top 800-600 cm’l range of LI-FTIR difference spectrum of tris-washed
and O2 evolving RCC samples. 5000 scans (tris washed) and 6500 scans (O2 evolving)
respectively b) 800—600 cm‘l range of NHZOH and O2 evolving RCCS. 4000 scans

(NH,OH) and 6500 scans (02 evolving)

173

ligand-based normal coordinates. As a control for these possible assignments, samples
that were manganese depleted in an alternative way, by exposure to hydroxylamine, were
investigated. It can be seen from Figure 5-23b that the same result was obtained. In
manganese depleted samples, the vibration at 632 cm" that is present in oxygen evolving
samples, is no longer present in samples that have had the manganese removed from the
system.

If the vibration in question is the result of a manganese-oxygen interaction then its
vibrational frequency should exhibit characteristic dependence upon the oxygen isotope
present in the system. Previous work has shown that manganese atoms linked by a di-u-
oxo core, show a pronounced isotope shift of some of its low frequency modes of
vibration upon labeling of the u-oxo bridges with '80[58]. Our first attempt to label the
PSII samples isotopically consisted of soaking the samples in FTIR buffer that was 97%
enriched with '80. The samples were simply resuspended in the labeled buffer, and
allowed to incubate for 12-14 hours, on ice in the dark. The sample did not perform any
water oxidation chemistry during this dark incubation. The spectra from these samples
appeared identical to the spectra obtained with "’0 labeled buffer (Figure 5-19). This is
not surprising in light of the reported slow exchange kinetics of the bridging oxo groups
proposed to be present in the OEC[S 8, 59]. In order to induce the bridges to exchange, we
incubated PSII samples in the presence of "’0 labeled buffer, and an exogenous electron
acceptor, 2,6 dichlorobenzoquinone. By incubating the samples on ice, under room light
for 30 minutes, the samples undergo several cycles of catalytic chemistry. If the di-u-oxo

oxygen atoms are involved with the catalytic chemistry, then we expect them to exchange

I74
easily under these conditions. Figure 5-24 shows the 800-600 cm‘1 range of the FTIR

spectrum of the samples treated in this manner. The peak at 632 cm‘l appears altered in
these spectra. Unfortunately, our current experimental conditions do not allow us to
examine modes that appear below 600 cm", owing to the cutoff of the germanium filter
we use to block the helium-neon laser beam that is coaxial with the IR probe beam. The
results of the work we have performed on di-u-oxo manganese complexes suggest that,
under conditions of isotopic exchange, an isotopic shift of approximately 30 cm" to lower
frequencies is expected, when the heavier isotope is exchanged into the bridging
positions. So, if the 632 cm" peak is in fact due to a manganese-oxygen vibration, then,
with the heavier isotope, the vibration would appear at the threshold of our current
detection limits. As a result, definitive assignment is not yet possible.

Figure 5-22 shows the LI-FTIR spectra from Sr2+ substituted samples that exhibit
the g=4.1 EPR signal. Recently, it has been proposed that the manganese cluster, under
conditions in which it exhibits the g=4.1 EPR signal, is structurally different from the
cluster under conditions that generate the multiline EPR signal. EXAFS studies have
shown a significant lengthening of a Mn-Mn distance upon generation of the =4.1
signal[l4]. It has been shown in model compound studies that protonation of one of the
bridging u-oxo ligands of a an2 dimer not only causes a significant lengthening of the
Mn-Mn distance, but also significantly alters the J-coupling between the two manganese
ions. In this case, the study showed a decrease of antiferromagnetism (i.e more

ferromagnetic coupling) upon protonation[60, 61], while another study showed an

175

 

 

 

 

0.0004
——RCCs cycled with '80
‘ ....... RCCS ('60)
0.0002 .. N
:3: 0.0000 _
<
< '.
-0.0002 -
-0.0004 . T .
800 700 600

frequency (cm")

Figure 5-24: 800-600 cm’l range of LI-FTIR spectrum of RCCS allowed to turnover in
presence of ”0 labeled water. (dash) O2 evolving RCCS (Figure 5-16).

 

“u— .‘J-i — . . .. r... .u . . . . I ._ '- . .. .... .. 6 w h... . 1.01. . 1‘. .4. I: I .... v1 al.“..n. ‘ .Imr. . Q. ... . ., .
. .3Q uni . . .n'n «Jitrh d“: r..~..i . . . . a a... ..bTIAvrx. '4‘ e .. '0 ~ ...Or *1.“ Ibo...» ICU-I..- .r. Lutle ’ A” s .A... SH .1 )u’.- I. .u ._ . .-

. . n o If . . ..Ill..l..

 

176

increase in the antiferromagnetism upon protonation of a bridging oxo[62, 63].
Protonation, or any other structural rearrangement of the manganese cluster to form the
g=4.1 signal, should be accompanied by an alteration of the vibrational properties of the
manganese cluster. The FTIR difference spectrum of the Sr2+ substituted, g=4.1 samples
shows some interesting differences when compared to the normal multiline generating
samples. The differences include the addition of a shoulder at approximately 1560 cm"
and the appearance of a new derivative shaped feature overlaying the peaks attributed to
the calcium-binding carboxylate at 1364/ 1403 cm". In the low frequency region, there are
some subtle differences, and the addition of some new peaks. In Figure 5-25 we can see
new peaks appear at 650 cm", 625 cm’1 and 619 cm’I in the Sr2+ substituted samples.
Also, there appear to be some pronounced differences between the Sr2+ substituted
spectrum, and that of the O2 evolving samples at approximately 675 cm'l in the difference
spectra. All of these peaks are in the region where an2 cores will exhibit vibrations. If
the manganese cluster undergoes structural rearrangement as it forms the g=4.1/S2 state,

then this is the region where those changes would manifest themselves vibrationally.

Conclusions

We have reported here, for the first time, light-induced FTIR difference spectra of
photosystem II of vibrations that occur below 1000cm". This region of the infrared
spectrum is expected to contain information regarding vibrations due to 1) protein
structural rearrangements upon changes of oxidation state of the manganese cluster 2)

Protein derived metal ligands whose vibrations are perturbed by oxidation of the

I77

 

0.0004

....... O2 evolv. RCCs
Srzisubstituted RCCS (g=4. 1)

 

0.0002 _

AA.U.

 

-0.0002 _

 

 

-0.0004

 

I
800 700 600

frequency (cm")

Figure 5-25: 800-600 cm" range of LI-FTIR spectrum of Srz+ substituted RCCS. LI-FTIR
spectrum of O2 evolving samples shown for reference.

I78

manganese cluster, and 3) metal-oxygen vibrations from species in which the oxygen
atoms are either substrate derived, or are non-catalytically active atoms that are structural
in nature. While we have not been able to make definitive assignments as of yet, the fact
that we can observe these modes with adequate signal to noise ratios shows the promise
of this technique for helping determine the mechanism of photosynthetic water oxidation.
We are planning experiments where we will dismantle, then reassemble the manganese
cluster in the presence of '8O-labeled water. This will guarantee the incorporation of "’0
into the manganese cluster, which should provide definitive answers regarding the origin
of the spectral features present in the LI-FTIR difference spectrum of PSII samples.We
are currently in the process of developing the technology to extend the spectroscopic
range of sensitivity of our instrumentation to 300 cm", where almost all of the potential

vibrations of structures proposed in the mechanism of water oxidation will be observable.

10.

ll.

12.

13.

179

LIST OF REFERENCES

Pirson, A., Z. Bot., 1937.31: p. 193.
Debus, R.J., Biochim Biophys. Acta, 1992. 1102: p. 269.

Britt, R.D., in Oxygenic Photosynthesis: The Light Reactions, D.R. Ort and CF.
Yocum, Editors. 1996, Kluwer: Dordrecht. p. 137.

Diner, BA. and Babcock, G.T., in Oxygenic Photosynthesis: The Light Reactions,
D.R. Ort and CR Yocum, Editors. 1996, Kluwer: Dordrecht. p. 213.

Joliot, B., Barbieri, G., and Chabaud, R., Photochem. Photobiol., 1969. 10: p. 309.
Kok, B., Forbush, B., and McGloin, M., Photochem. Photobiol., 1970. 11: p. 457.
Dismukes, G.C. and Siderer, Y., Proc. Natl. Acad. Sci., 1981. 78: p. 274.
Hansson, 0., Andreasson, L.-E., and Vanngard, T., FEBS Lett., 1987. 185: p. 151.
Andreasson, L-E. and Hansson, 0., NH3 treated ml samples, in Advances in
Photosynthesis Research, J. Biggins, Editor. 1987, Martinus Nijhoff/Dr. Junk
Publishers: Dordrecht, The Netherlands. p. 503.

Beck, W.F., de Paula, J.C., and Brudvig, G.W., J. Am. Chem. Soc., 1986. 108: p.
4018.

Penner-Hahn, J.E., Fronko, R.M., Pecoraro, V.L., Yocum, C.F., Betts, S.D., and
Bowlby, N.R., J. Am. Chem. Soc., 1990. 112: p. 2549.

MacLachlan, D.J., Hallahan, B.J., Ruffle, S.V., Nugent, J.H.A., Evans, M.C.W.,
Strange, R.W., and Hasnain, S.S., Biochem. J., 1992. 285: p. 569.

DeRose, V.J., Ph. D. Thesis, University at California, Berkeley. Lawrence
Berkeley Laboratory Report LBL-16901. 1990.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

180

Liang, W., Latimer, M.J., Dau, H., Roelefs, T.A., Yachandra, V.K., Sauer, K., and
Klein, M.P., Biochemistry, 1994. 33: p. 4923.

Rompel, A., et al., J. Am. Chem. Soc., 1997. 119: p. 4465.

Yachandra, V.K., DeRose, V.J., Latimer, M.J., Mukerji, I., Sauer, K., and Klein,
M.P., Science, 1993. 260: p. 675.

Dan, H., Andrews, J .C., Roelefs, T.A., Latmier, M.J., Liang, W., Yachndra, V.K.,
Sauer, K., and Klein, M.P., Biochemistry, 1995. 34(5274).

Cramer, S.P., Eccles, T.K., Kutzler, F., Hodgson, K.D., and Mortenson, L.E., J.
Am. Chem. Soc., 1976. 98: p. 1287.

Kirby, J.A., Goodin, D.B., Wydrzynski, T., Robertson, AC, and Klein, M.P., J.
Am. Chem. Soc., 1981. 103: p. 5537.

Yachandra, V.K., Sauer, K., and Klein, M.P., Chem. Reviews, 1996. 96: p. 2927.
Hoganson, C.W., Personal communication

Baumgarten, M., Philo, J.S., and Dismukes, G.C., Biochemistry, 1990. 29: p.
10814.

Boussac, A., Zimmerman, J .-L., and Rutherford, A.W., Biochemistry, 1989. 28: p.
8984.

Tso, J., Sivaraja, M., Philo, J.S., and Dismukes, G.C., Biochemistry, 1991. 30: p.
4740.

Boussac, A., Zimmerman, J .-L., Rutherford, A.W., and Lvergne, J., Nature, 1990.
347:p.303.

Ono, T. and Inoue, Y., FEBS LEtters, 1991. 278: p. 183.

Gilchrist, M.I., Ball, J.A., Randall, D.W., and Britt, R.D., Proc. Natl. Acad. Sci.,
1995.92:p.9545.

Hoganson, C.W., Lydakis-Simantiris, N., Tang, X.-S., Tommos, C., Warncke, K.,
Babcock, G.T., Diner, B.A., McCracken, J., and Styring, S., Photosynth. Res.,
1995.46:p.177.

Ghanotakis, D.F., Babcock, G.T., and Yocum, C.F., FEBS Letters, 1984. 167: p.
127.

 

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

181

Tang, X.-S., Randall, D.W., Force, D.A., Diner, B.A., and Britt, R.D., J. Am.
Chem. Soc., 1996. 118(7638).

Tommos, C., Tang, X.-S., Warncke, K., Hoganson, C.W., Styring, S.,
McCracken, J., Diner, B.A., and Babcock, G.T., J. Am. Chem. Soc., 1995. 117: p.
10325.

Sigel, H. and Sigel, A., Metal Ions in Biological Systems. Vol. 30. 1994: Marcel
Dekker, Inc., USA.

Frey, P.A., Chem. Rev., 1990. 90: p. 1343.

Blomberg, M.R.A., Siegbahn, P.E.M., Styring, S., Babcock, G.T., Akerrnark, B.,
and Korall, P., J. Am. Chem. Soc., submitted, .

Baldwin, M.J. and Pecoraro, V.L., J. Am. Chem. Soc., 1996. 118: p. 11325.
Candle, MT and Pecoraro, V.L., J. Am. Chem. Soc., submitted, .

Gardner, K.A. and Mayer, J.M., Science, 1995. 269: p. 1849.

Varotsis, C. and Babcock, G.T., Biochemistry, 1990. 29: p. 7357.

Varotsis, C., Woodruff, W.H., and Babcock, G.T., J. Biol. Chem., 1990. 265: p.
11131.

Brudvig, G.W., Casey, J .L., and Sauer, K., Biochim. Biophys. Acta, 1983. 723: p.
366.

Berthold, D.A., Babcock, G.T., and Yocum, C.F., FEBS Letters, 1981. 134: p.
231.

Ghanotakis, DR and Yocum, C.F., FEBS Letters, 1986. 197: p. 244.
Mishra, R.K. and Ghanotakis, D.F., Photosyn. Res., 1994. 42: p. 37.
Cheniae, G.M. and Martin, I.F., Plant Physiol., 1971. 47: p. 568.

Miller, A.-F. and Brudvig, G.W., Biochemistry, 1990.29: p. 1385.

Tamura, N. and Cheniae, G.M., Biochim. Biophys. Acta, 1987. 890: p. 179.

Barry, BA. and Babcock, G.T., Proc. Natl. Acad. Sci., 1987. 84: p. 7099.

48.
49.
50.

51.
52.

53.

54.
55.
56.

57.
58.
59.
60.

61.
62.

63.

182
Noguchi, T., Ono, T., and Inoue, Y., Biochim. Biophys. Acta, 1993. 1143: p. 333.
Noguchi, T., Ono, T., and Inoue, Y., Biochemistry, 1992. 31: p. 5953.
Noguchi, T., Ono, T., and Inoue, Y., Biochim. Biophys. Acta, 1995. 1228: p. 189.
Steenhuis, J.J. and Barry, B.A., J. Am. Chem. Soc., 1996. 118: p. 11927.
Hienerwadel, R., Boussac, A., Breton, J ., and Berthomieu, C., Biochemistry,
l996.35:p.15447.
Noguchi, T. and Inoue, Y., F EBS Letters, 1995.370: p. 241.

Berthomieu, C., Nabedryk, E., Mantele, W., and Breton, J., FEBS Letters, 1990.
269:p.363.

MacDonald, G.M., Steenhuis, J.J., and Barry, B.A., J. Biol. Chem, 1995. 270: p.
8420.

Smith, J.C., Gonzales_Vergara, E., and Vincent, J.B., Inorg. Chim. Acta, 1997.
255:p.99.

Dave, BC. and Czernuszewicz, R.S., Inorg. Chim. Acta, 1994. 227: p. 33.
Cooper, SR. and Calvin, M., J. Am. Chem. Soc., 1977. 99: p. 6623.

Carroll, J.M. and Norton, J.R., J. Am. Chem. Soc., 1992. 114: p. 8744.

Baldwin, M.J., Gelasco, A., and Pecoraro, V.L., Photosyn. Res., 1993. 38: p. 303.

Baldwin, M.J., Stemmler, T.L., Riggs-Gelasco, P.J., Kirk, M.L., Penner-Hahn,
J.B., and Pecoraro, V.L., J. Am. Chem. Soc., 1994. 116: p. 11349.

Dube, C.E., Wright, D.W., and Armstrong, W.H., J. Am. Chem. Soc., 1996. 118:
p.10910.

Hagen, K.S., Westmoreland, T.D., Scott, M.J., and Armstrong, W.H., J. Am.
Chem. Soc., 1989. 111: p. 1907.

HICHIGQN STATE UNIV. LIBRARIES
llHIllllllll11111111111111"llllllllllllll
3129301577791

 

.an I

4—

 

'fi'.‘

 
 

I
J

I