 

grov-

V.

.1:

.A

v

V. .3:

4.26.5.0:

‘ .-;:'-‘Vo.

 

» ,

.47.. , .
‘ “ﬁg

Ar

an .
Evy,"

v .3

$me \.

a...» . 23.2 ..
9.3. .

AC2},

., u .E ' .Pﬁr. DE?
.3. u u... 1. .

Il"

 

 

 

 

 

my LIBRARY
_ Michigan State
200 ‘6 University
This is to certify that the

dissertation entitled

EPR Studies of Electron and Proton Transfer in'
Cytochrome c Oxidase

presented by

Shujuan Xu

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Chemistry, Biochemistry and
Molecular Biology

 

 

QZZW£W

Major Professor’s Signature

$2M AS, J’ 2008

Date

MSU is an afﬁnnative-acﬁon, equal-opportunity employer

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/ClRC/DateOue.indd-p.1

EPR STUDIES OF ELECTRON AND PROTON TRANSFER IN
CYTOCHROME C OXIDASE
By

Shujuan Xu

A DISSERTATION

Submitted to
Michigan State University
In partial fulﬁllment of the requirements
For the degree of

DOCTOR OF PHILOSOPHY
Chemistry, Biochemistry and Molecular Biology

2008

ABSTRACT

EPR STUDIES OF ELECTRON AND PROTON TRANSFER IN
CYTOCHROME C OXIDASE

By
Shujuan Xu

An understanding of the mechanism of electron and proton transfer in cytochrome c
oxidase (CcO) is essential for a full understanding of the process of energy transduction
in this key respiratory enzyme. Combining site-directed mutagenesis and EPR methods,
some mutants at residue S44 associated with heme a were created and studied to
determine its inﬂuence on the spectral characteristics of CcO. EPR data on 8446
supports the prediction that this hydrogen bonding partner of the histidine ligand of heme
a is one determinant of the EPR spectral difference between bovine and bacterial CcO.
Of additional interest, the pH-dependence of the EPR spectrum of S44D and fast electron
transfer studies showing dramatically inhibited electron transfer from Cu to heme a at
high pH, indicate a proton-coupled electron transfer mechanism. This mutant could
provide a new tool for studying proton and electron transfer reactions in cytochrome c

oxidase.

Using the non-redox active Mg site of CcO, substituted with Mn, as a probe,
proton/water movement was studied in its vicinity. Electron nuclear double resonance
(ENDOR) studies provided us with detailed information on hyperﬁne coupling
interactions between the Mn site and its liganded water molecules. Further studies of
oxidized and reduced CcO, plus and minus cyanide or azide, with Electron spin echo

envelope modulation (ESEEM) technique showed cyanide binding to the Mn in the

reduced enzyme. This ﬁnding supports the hypothesis that reduction of Cu causes one of
the water ligands of the Mn to become a hydroxide ion, playing an important role in

proton pumping in CcO. ESEEM simulations also provide support for this hypothesis.

As a more sensitive technique, high-ﬁeld EPR was used to study CcO with the hope
of gaining more information with higher resolution. The difference in Mn spectra of the
reduced and the oxidized CcO was observed in Rs. CcO, which agrees well with ﬁndings
in Pd. CcO. Further studies of super hyperﬁne coupling interactions between Mn and its
water ligands in the oxidized and reduced enzymes showed some interesting data. The
narrowing effect due to deuterium exchange was observed in the oxidized CcO, which
can be potentially used to quantify the number of coupled protons. However, experiments
with the reduced enzyme in H2170 and 2&0 failed to Show supposedly broadening and

narrowing effects respectively. The possible reasons were explored and explained.

To my dear parents

iv

Acknowledgements

I would like to express my sincerest gratitude to all those who made my Ph.D.
period enjoyable.

I want to thank my advisor in Biochemistry, Dr. Shelagh Ferguson-Miller, who
introduced me to the fascinating world of cytochrome c oxidase, deserves my deepest
gratitude for all the help, guidance and encouragement she has given me during these
years. With her enormous amounts of integrity, determination and dedication she has
gained my greatest admiration and uppermost respect. I have furthermore to thank my
advisor in Chemistry, Dr. John McCracken for agreeing to be my advisor when my
previous advisor moved to another university. John has generously shared with me his
vast knowledge of EPR theory and instrumentation and has been an endless source of
advice and encouragement. I also wish to express my gratitude to my committee
members: Dr. Robert, Cukier, Dr. Mike, Garavito, and Dr. David Weliky. They gave me
great suggestions during my oral exam, annual report presentations.

There are also many numbers of coworkers I would like to thank. Many discussions
I had with Dr. Denise Mills were invaluable and collaboration with her was a great
experience I would never forget. Dr. Carrie Hiser helped me with molecular biology part
to ensure my project go well. Her many suggestions helped me do well with my
presentations. Dr. Martyn Sharpe taught me many things and “pushed” me to do better in
my research. Dr. Denis Proshlyakov provided me assistance using his experiment setup.
He is also a good source of knowledge and advice. And would also like to thank Dr.
Rajendra Muthukumaran, who spent many hours teaching me how to use EPR machines

and shared with me his knowledge of EPR theory. I also had many useﬁ11 discussions

with other lab members in both Chemistry and Biochemistry: Dr. Bryan Schmidt, Dr.
Ling Qin, Namjoon Kim, Jian Liu, Xi, Zhang, Leann Buhrow, Monika, Shannon, Morey,
and Tom Casey as well as all the others who I interacted with through my Ph.D. Special
thank comes to my lab member, Mattew, Krzyaniak who helped me a lot with ENDOR
and ESEEM simulations and EPR instrumentation.

I would also like to acknowledge all my friends here at Michigan State University
for their help and support. Without their ﬁiendship, I would not make it. My deepest
thank to my husband, Zhenwei, for his love and support.

My deepest thank to my family!

TABLE OF CONTENTS

List of Tables .................................................................................................................. x
List of Figures ................................................................................................................ xi
List of Abbreviations ..................................................................................................... xv
Chapter 1
An introduction of cytochrome c oxidase and EPR spectroscopy .............................. l
1. The respiratory chain ................................................................................................... 2
2. Cytochrome c oxidase ................................................................................................. 6
2.1.The overall structure .......................................................................................... 8
2.1.1 The redox-active metal centers ................................................................... 12
2.1.2 The non-redox active centers ...................................................................... 14
2.2. The oxygen chemistry ..................................................................................... 14
2.3 Electron and proton transfer theory .................................................................. 17
2.3.1 Electron transfer in biological systems ....................................................... 17
2.3.2 Proton transfer theory ................................................................................. 23
2.3.3 Proton transfer in CcO ................................................................................ 28
2.4 Water channels in proteins ............................................................................... 30
2.5 Questions to be addressed ................................................................................ 35
3. Introduction of the electron paramagnetic resonance (EPR) ....................................... 37
3.1 Electron spin echo envelope modulation (ESEEM) .......................................... 41
3.2 Electron nuclear double resonance (ENDOR) .................................................. 46
4. Conclusion ................................................................................................................ 50
List of references ........................................................................................................... 52
Chapter 2
cw EPR investigation of the heme a center in several S44 mutants ........................... 59
1. Introduction ............................................................................................................... 6O
2. Experimental methods ............................................................................................... 64
2.1 Protein expression and puriﬁcation .................................................................. 64
2.2 Reconstitution of CcO into phospholipid vesicles (COVS) ............................... 64
2.3 Activities of COVs under steady-state and stopped-ﬂow proton pumping
measurements ................................................................................................................ 65
2.4 UV-visible spectroscopy .................................................................................. 66
2.5 Electron paramagnetic resonance (EPR) ........................................................... 66
3. Results ....................................................................................................................... 67

vii

3.1 Optical spectra and activity .............................................................................. 67

3.2 X-band cw EPR ............................................................................................... 69
4. Discussion ................................................................................................................. 82
4.1 Spectroscopic measurements ............................................................................ 82
4.2 Effect of mutation on electron transfer rate and redox potential ........................ 91
List of references ........................................................................................................... 95
Chapter 3
Pulse electron nuclear double resonance (ENDOR) study of aqua protons
coordinated to Mn(II) in cytochrome c oxidase ......................................................... 98
1. Introduction ............................................................................................................... 99
2. Experimental methods ............................................................................................. 103
2.1 Protein expression and puriﬁcation ................................................................ 103
2.2 Sample preparation ........................................................................................ 105
2.3 X-band pulsed electron nuclear double resonance spectroscopy (ENDOR) ..... 105
2.4 Simulation of EPR and ENDOR spectra ......................................................... 107
3. Results and discussion ............................................................................................. 109
3.1 Electron spin-echo-detected absorption EPR spectroscopy ............................. 109
3.21H ENDOR spectroscopy ............................................................................... 109
3.3 ENDOR Simulation ....................................................................................... 114
4. Conclusion .............................................................................................................. 114
List of references ......................................................................................................... 117
Chapter 4
Electron spin echo envelope modulation (ESEEM) study of water ligands bound to
the Mn(II) site in cytochrome c oxidase .................................................................... 119
1. Introduction ............................................................................................................. 120
2. Experimental methods ............................................................................................. 122
2.1 Protein production and puriﬁcation ................................................................ 122
2.2 Sample preparation ........................................................................................ 124
2.3 Electron spin echo envelope modulation (ESEEM) spectroscopy ................... 124
2.4 ESEEM Simulation ........................................................................................ 125
3. Results ..................................................................................................................... 126
3.1 The use of cryoprotectant ............................................................................... 126
3.2 The deuterium standard curve for Mn2+(HzO)6 and Mn-DTPA ....................... 128
3.3 ESEEM spectra of oxidized CcO in deuterated solvents with cryoprotectant.. 130
3.4 ESEEM studies of the oxidized and reduced CcO plus or minus cyanide ....... 133
3.5 ESEEM simulation ........................................................................................ 136
4. Discussion ............................................................................................................... 138
* Note ......................................................................................................................... 142
List of references ......................................................................................................... 145

viii

Chapter 5

High ﬁeld/frequency cw EPR study of Mn(II) in cytochrome c oxidase ................. 147
1. Introduction ............................................................................................................. 148
2. Experimental methods ............................................................................................. 151
2.1 Protein production and puriﬁcation ................................................................ 152
2.2 UV-visible spectroscopy ................................................................................ 152
2.3 Sample preparation ........................................................................................ 152
2.4 W-band EPR spectroscopy ............................................................................. 153
2.5 Strategy for determination of the number of inner sphere waters .................... 153
3. Results and discussion ............................................................................................. 154
3.1 EPR spectra of CcO in H20 ........................................................................... 154
3.2 EPR spectra of the oxidized CcO in H20 and 75% 2H20 ................................ 157
3.3 EPR spectra ofthe reduced CcO in 75% 2H20 and 30% H2‘7O ...................... 157
3.4 Discussion ..................................................................................................... 157
List of references ......................................................................................................... 164

LIST OF TABLES

Table 2.1:The steady-state activity of S44G, S44D and wild-type at different pH .......... 70

Table 2.2: EPR parameters for heme a in cytochrome c oxidase .................................... 88

LIST OF FIGURES
Images in this dissertation are presented in color

Figure 1.1: The mitochondrial respiratory chain ............................................................ 4

Figure 1.2: Overall structure of CcO showing arrangements of metal centers and proton

uptake pathways. ............................................................................................................. 7
Figure 1.3: Comparison of mammalian and bacterial CcO crystal structures ............... 10
Figure 1.4: Overall structure othodobacter sphaeroides CcO .................................. 11
Figure 1.5: Proposed water channel in bovine CcO ..................................................... 13
Figure1.6: Structure showing the arrangement of metal centers and associated amino

acids in bacterial CcO .................................................................................................... 15
Figure 1.7: Proposed catalytic reactions at the active site of CcO ................................ 16

Figure 1.8: Potential energy diagram showing parameters ﬁ'om Marcus electron transfer
equation ........................................................................................................................ 18

Figure 1.9: S-cycle ofPhotosystem II ......................................................................... 22

Figure1.10: Schematic depiction of the hop-and-tum mechanism for proton transfer
through a hydro gen-bonded water chain ........................................................................ 25

Figure1.11: Photocycle scheme showing the main steps in light-driven proton transport
in bacteriorhodopsin ...................................................................................................... 27

Figurel . 12: Key amino acid residues in well-deﬁned D and K proton uptake pathways
in Rs.CcO structure ........................................................................................................ 29

Figure1.13: (A) The most possible proton exit pathways in bovine CcO and the energy

diagram. (B) Lysine17lB proton pathway ..................................................................... 31
Figure1.14: X-ray crystal structure of AQPl ............................................................... 33
F igure1.15: Snapshots of four representative P450 structures (1phc, lcpt, 2hpd, and

loxa) in which cavities forming the aqueduct can be observed ...................................... 36
Figure1.16: EPR energy diagram for S=1/2 system ..................................................... 39

Figure1.17: Energy level scheme and eigenvectors for the S=l/2, and I=1/2 system 43

xi

Figure1.18: The diagram of pulse sequence for two- and three- pulse ESEEM
experiments. ................................................................................................................. 47

Figure1.19: A) Davies ENDOR pulse sequence, B) Evolution of spin state population

during the Davies ENDOR pulse sequence .................................................................... 49
Figure 2.1: The RS CcO crystal structure (PDB ID: 1M56) ......................................... 61
Figure 2.2. The ligands to heme a are shown for bovine (1V54 structure) and Rs
(1M56) CcO .................................................................................................................. 63
Figure 2.3: The visible spectra of reduced wild type, S44D, and S44G CcOs. ............. 68
Figure 2.4: The steady-state activity assay ofwild type, S44G, and S44D CcOs ......... 71
Figure 2.5: Overview of the EPR spectral changes with wild-type R.s CcO (black),
S44G (red), and S44D (blue) ......................................................................................... 73
Figure 2.6: Effect of pH on the amplitude of the heme a gz EPR spectra of S44D ....... 74

Figure 2.7: Amplitudes of peaks at 2.72, 2.78 for heme a versus buffer pH for S44D .75

Figure 2.8: Effect of ionic strength and cytochrome c binding on heme a gz EPR

spectra of S44D ............................................................................................................. 77
Figure 2.9: EPR spectra overlay of WT, S446, S44D, S44E, and S44N in the

gz region ........................................................................................................................ 78
Figure 2.10: The pH proﬁle of Mn EPR spectra of S44D ............................................ .80
Figure 2.11: The cw EPR spectrum of mutant S44N .................................................... 81
Figure 2.12: The temperature-dependence EPR spectra of S44D ................................. 83

Figure 2.13: Crystal ﬁeld parameters of bovine wild type CcO, Rs. wild type CcO, and
S44 mutants in the frame of the Blumberg and Peisach truth diagram ............................ 86

Figure 2.14: The ligands to heme a in crystal structures and simulated structures of CcO

...................................................................................................................................... 90
Figure 3.1: Proposed water channel in bovine CcO ................................................... 100
Figure 3.2: Diagram oan(II) energy level and EPR transitions ............................... 102
Figure 3.3: Pulse sequence for Davies ENDOR experiment ...................................... 104

xii

Figure 3.4: The visible spectrum of Mn—substituted CcO at resting (blue) and
dithionite-reduced (red) states ...................................................................................... 106

Figure 3.5: Experimental (black) and simulated (red) electron spin-echo-detected
absorption EPR spectra of Mn-substituted cytochrome c oxidase ................................. 110

Figure 3.6: Davies ENDOR spectra of oxidized (red) and reduced (black) cytochrome c

oxidase at central ﬁeld 3.7kG ...................................................................................... 112
Figure 3.7: Davies ENDOR spectra of reduced cytochrome c oxidase in pH 7.4 water
buffer (red) and 75% HO (black) buffer .................................................................... 113
Figure3.8: Experimental (black) and simulated (red) ENDOR powder pattern spectra at
central ﬁeld H=3.7kG .................................................................................................. 116
Figure 4.1: The formation of a hydroxide/hydronium pair upon CuA reduction; modeled
on bovine oxidase oxidized and reduced structures ...................................................... 123
Figure 4.2: Effect Of the cryoprotectant on CcO three-pulse ESEEM data. ................ 127

Figure 4.3: Three-pulse ESEEM time-domain data of Mn(HzO)6(A), Mn-DTPA (B),
and Mn(H20)(,/ Mn-DTPA ratio (C) ............................................................................ 129

Figure 4.4: Three-pulse ESEEM time-domain spectra (A), frequency-domain spectra (B)
of Rs. CcO in varying levels of 2H20 buffers ranging ﬁ'om 25% to 75% in the absence of
cryoprotectant ............................................................................................................. 13 1

Figure 4.5: Fit of Fourier-transformed three-pulse ESEEM spectra of oxidized CcO in
different concentration of 2H20 buffers with cryoprotectant with a linear ﬁmction ....... 132

Figure 4.6: cw EPR spectra of oxidized and reduced CcO in the presence or absence of
cyanide ........................................................................................................................ 134

Figure 4.7 Effect of different radionucleotide of cyanide on the three-pulse ESEEM
spectra of the Mn center ofR.s. oxidase ....................................................................... 135

Figure 4.8: Plot of the amplitude at 2.42MHz versus 2H20 concentration for the reduced

CcO ratio and the reduced CcO plus cyanide ratio ....................................................... 137
Figure 4.9: The ESEEM simulated spectra of CcO assuming different numbers of

deuterons bound to the Mn Site .................................................................................... 139
Figure 4.10: Comparison of the simulated and the experimental data ......................... 143

Figure 4.11: Plots of the amplitude at 2.42MHz versus 2&0 concentration for the
oxidized CcO ratio, reduced CcO ratio and the reduced CcO plus cyanide ratio ........... 144

xiii

Figure 5.1:
Figure 5.2:
Figure 5.3:
Figure5.4:

Figure5.5:

The HF —EPR spectra of the oxidized and the reduced CcO at 15K ......... 155
The temperature-dependence HF—EPR spectra of the oxidized CcO ......... 156
The HF -EPR spectra of the oxidized CcO in H20 and D20 buffers at 15K158
The HF-EPR spectra of the reduced CcO in H20 and H2170 buffers at 15K159

The HF-EPR spectra of the reduced CcO in H20 and 2H20 buffers at 15K16O

xiv

Chapter 1:
AN INTRODUCTION TO CYTOCHROME C OXIDASE
AND EPR SPECTROSCOPY

1 The respiratory chain

Oxidative phosphorylation is the process by which energy is produced in aerobic
organisms. In eukaryotes, oxidative phosphorylation occurs in the mitochondria, where
02 is reduced to H2O by electrons donated by NADH or FADH2. A well-accepted
chemiosmotic theory was proposed by Peter Mitchell in 1961 and was the basis for
understanding ATP synthesis. The theory states that oxidative phosphorylation involves
the ﬂow of electrons through a chain of membrane-bound electron carriers. The electrons
are used to reduce oxygen to water resulting the release of free energy as an exergonic
process, which is coupled to the transport of protons against the positive outside
membrane potential, establishing a transmembrane electrochemical potential. The
transmembrane ﬂow of protons down their concentration gradient through the membrane

protein ATP synthase, provides the energy for ATP synthesis.

The discovery that mitochondria are the site for oxidative phosphorylation in
eukaryotes marked the beginning of biological studies into energy transduction. The
mitochondria have two lipid membranes. The outer membrane is permeable to small
molecules (MW <5000 Da) and ions, which move freely through transmembrane
channels formed by porin proteins. The inner membrane is impermeable to most small
charged molecules, including protons, except through speciﬁc transport systems. The
inner mitochondrial membrane bears the respiratory chain and ATP synthase. The
mitochondrial matrix, enclosed by the inner membrane, contains many enzymes such as

the pyruvate dehydrogenase complex, enzymes Of the TCA cycle, and the fatty acid 13-

oxidation pathway. During the proton translocation process, protons are pumped from the
mitochondrial matrix. This becomes negatively charged due to the uptake of protons, and
is referred to as the negative/N-side. The interrnembrane space becomes positively

charged due to the pumped protons, and is referred to as the positive/P-side.

The respiratory chain is made up of four separate complexes (Fig. 1.1). They are:
NADH dehydrogenase (Complex 1), succinate dehydrogenase (Complex II), cytochrome
bcl (Complex III), and cytochrome c oxidase (Complex IV). Complexes I and II catalyze
electron transfer to ubiquinone from different electron donors: NADH (complex I) and
succinate (complex 11). Complex IH transfers electrons from reduced ubiquinone to a
soluble cytochrome c, and complex IV completes the reaction by transferring electrons

from cytochrome c to O2.

Complex I (NADH dehydrogenase) is a large enzyme composed of forty-two
different peptide chains, including an FMN-containing protein and at least six iron-sulfur
centers. Complex I catalyzes two simultaneous and obligatorily coupled reactions: it
removes two electrons from NADH and transfers them to a lipid-soluble carrier,
ubiquinone (Q). The reduced product, ubiquinol (QH2), is free to diffuse within the
membrane. At the same time, Complex I pumps four protons (H+) across the membrane,
producing a proton gradient. High-resolution electron microscopy shows that complex I
is L-shaped with one arm of the L in the membrane and the other extending into the

matrix.

  

+H‘ Succinate Fanﬁate
Mmlguase deffiggifue k‘cm’l” ”$3.1m” ATP ”“1“"

Figure1.1: The mitochondria respiratory chain. The structure represents the
crystallographically deﬁned structure of each enzyme M. This ﬁgure is a ﬁgure courtesy

of Dr. Denise Mills.

Complex II (succinate dehydrogenase) is the only membrane-bound enzyme in the
citric acid cycle (TCA cycle). It is smaller and simpler than complex I but still contains
ﬁve prosthetic groups of two types and four different protein subunits. Subunits A and B
contain three 2Fe-ZS clusters, bound FAD and a binding site for the substrate, succinate.
Subunits C and D are integral membrane proteins. They contain a heme b and a binding
site for ubiquinone, which is the ﬁnal electron acceptor in the reaction catalyzed by
complex II. It serves to funnel additional electrons into the quinone pool (Q) by removing
electrons from succinate and transferring them (via FAD) to Q. Complex II is not a

proton pump, therefore it does not produce a proton gradient.

Complex 111 (bc, complex) transfers two electrons ﬁom ubiquinol (QH2) to two
molecules of the water-soluble electron carrier cytochrome c, releasing two protons from
ubiquinol on the outside, and translocates an additional two protons from the N side to
the P side of the membrane per pair of electron transferred via the Q-cycle, producing a
proton gradient (a net four protons are translocated). Complex bc; contains four electron

transfer centers, two hemes, cytochrome c], and an iron-sulfur protein as well.

Cytochrome c is a soluble protein of the intermembrane space. It plays a role as an
electron carrier which accepts electrons ﬁom complex III and then transfers them to

complex IV (cytochrome c oxidase).

Complex IV (cytochrome c oxidase (CcO)) is located at the end of the respiratory

chain. It catalyzes the reduction of molecular oxygen to water at the binuclear heme

a 3/CuB center. The reaction is coupled to proton translocation across the mitochondrial or
bacterial membrane, contributing a transmembrane pH gradient, which provides energy
for adenosine triphosphate (ATP) synthesis. The ATP produced is then delivered to
different locations throughout the cell, providing the necessary energy supply to a broad
variety of biochemical reactions. Because of the extreme importance of CcO for the
bioenergetic reactions in living cells, it has been an important research topic for the past
decades from its ﬁrst discovery by MacMunn 5, and further conﬁrmation by Keilin 6. It
took scientists more than a century to have a better, yet not ﬁlll, understanding of

cytochrome c oxidase (CcO).

2. Cytochrome c oxidase (CcO)

CcO catalyzes a seemingly simple reaction: the four-electron reduction of
molecular oxygen to two waters. In the reaction, four equivalents of ferrocytochrome c
provide four electrons, one at a time, and the enzyme takes up four substrate protons. The
reaction is highly exergonic and the ﬁee energy released by this reaction is stored in the
form of an electrochemical gradient across the membrane7. The active site of the enzyme
is buried within the membrane, and the electrons and protons that are used to form water
come from the opposite sides of the membrane, which results in charge separation across
the membrane (Fig.1.2). The reduction of one dioxygen molecule to two waters results in
the net movement of four charges across the membrane (four electrons consumed from
the outside, four protons consumed from the inside), and therefore the generation of a

transmembrane potential gradient. This deceptively simple vectorial mechanism makes

Positive side ”iv-a. ‘Cuh 11 mp
'0... .Mg ...‘

m I @.BOCuB i

\l’ O,
D path K path
Negative side I \ H+
H

Figure 1.2: Overall structure of CcO showing arrangements of metal centers and
proton uptake pathways. The picture is from Schemidt B. et a1

cytochrome c oxidase a powerful bioenergetic machine. What is more amazing is its
ability to translocate protons across the membrane. For each turnover, four additional

protons are pumped across the membrane (Eq. 1 .1).

4 cyt. c2+ + 8 Hi, + 02 —> 4 cyt. c3+ + 2H20 + 4H“out 1.1

Before the advent of the ﬁrst crystal structure of CcO, many biochemical and
biophysical methods were extensively used to understand its structure—function
relationship such as UV-visible, Fourier Transform Infrared (FTIR), Resonance Raman
(RR), and Electron Paramagnetic Resonance (EPR) spectroscopic analyses, site-directed
mutagenesis, and fast kinetics. The breakthrough in studying CcO was the atomic level
X-ray crystal structures from bovine CcO9 and the bacterium Paraccocus denitriﬁcans 3.
These crystal structures provided a full picture of the enzyme, and also veriﬁed many
ﬁndings on the positions of the metal sites by other biophysical and biochemical
methods. Upon careful inspection of the ﬁrst two crystal structures, aqueous channels,
suggested previously by mutagenesis, were clearly conﬁrmed. These two channels were
later established to be two proton uptake pathways. In 2002 and 2006, two more crystal
structures from the bacterium Rhodobacter sphaeroides 4’ '0 were determined at 2.3 A
/2.0 A resolutions. With the aid of high-resolution crystal structures, the Ferguson-Miller
lab is able to address many fundamental questions, such as the proton/water transfer

I'OUICS.

2.1 The overall structure

Mammalian cytochrome c oxidase contains four redox-active metal centers and
thirteen subunits. The complexity of its physical structure, and the difﬁculty of
manipulating the genome using site-directed mutagenesis, has made structure-function
analysis a difﬁcult task. Rhodobactor sphaeroides (Rs.) CcO, on the other hand, has only
four subunits but contains the same redox-active metal centers. It shows a strong
sequence homology to the mammalian CcO and is functionally equivalent to its
mammalian counterpart; therefore it is a useful model for structure-function analysis11
(Fig.1.3). Of the four subunits, subunit I is the largest subunit, with twelve
transmembrane helices. It contains the redox-active heme a, the heme a3/Cu3 active site,
and a non-redox active calcium ion whose role is still not clear. Subunit II has only two
transmembrane helices with an extra-membrane beta-pleated region that contains the
dinuclear CuA center, the initial electron acceptor from ferrocytochrome c. At the
interface of subunits I and II is another metal center, a Mg”, liganded by residues ﬁom
both subunits. Subunit 111 contains seven transmembrane helices which form a V-shape,
and has no metal centers. The ﬁmction of subunit III is not well understood, but it has
been shown that the absence of subunit III promotes ‘suicide’ inactivation of the enzyme,
which may be due to the loss of the structural integrity of the CuB center during turnover.
Subunit IV of Rs. CcO has only one transmembrane helix that interacts with subunit I via

lipids, but its ﬁmction, unfortunately, is still unknown (Fig.1.4).

Six phospholipid molecules were identiﬁed in the bovine structure (PDB ID: locc).
Based on the electron density map, they were assigned as phosphatidylethanolamines

(PE). Two of the PE molecules were found in a cleft formed by the helices of subunit III.

 

 

Figure 1.3: Comparison of mammalian and bacterial CcO crystal structures. The
bacterial CcO structure ﬁ'om Rs. (right) contains subunits that are highly similar to three
core subunits of bovine CcO (left), showing in yellow, blue, and red. They both contain
the same spatial alignment of the redox active metal cofactors (hemes are shown in red,
Cu», and Cu]; are shown in green). Rs. CcO contains subunit IV consisting of only one
transmembrane helix (not shown in this picture). Bovine CcO contains additional ten
subunits (shown in gray in the picture). The ﬁgure is made using VMD software
program.

 

\\ Kputh
I
1‘

Figure 1.4: Overall structure of Rhodobacter sphaeroides CcO. Diagram of four
subunit structure of Rhodobacter sphaeroides cytochrome c oxidase showing two proton
uptake pathways, The D (red), and K (blue) paths. The three redox-active centers are
labeled: dinuclear CuA center, heme a, heme a3/Cu3 binuclear center. The two non-redox
centers, Mg”, and Ca2+ are also included. This ﬁgure is a modiﬁcation of ﬁgure from
Mills, D. et at12

11

The other four PE molecules were found at the interface between subunits I, III, and IV.
Subunit IV has little direct interaction with the other three subunits, as revealed by the X-
ray structure, but maintains its contact via lipid molecules. In the bovine crystal
structures, over two hundred water molecules per monomer were resolved. Most of them
were found on the surfaces exposed to the solvent. Still, quite a few water molecules
were within the enzyme molecule forming hydrogen-bonded networks that are believed

to facilitate proton pumping ‘3 (Fig.1.5).

2.1.1 The redox-active metal centers

Cytochrome c oxidase has four redox-active metal centers: heme a, heme a3, the
dinuclear copper center (Curl), and a type 11 copper center (CuB) which is magnetically
coupled with heme a 3. Time-resolved spectroscopies have shown that electrons are
transferred from a reduced cytochrome c to heme a, then to the heme a3/Cu3 binuclear
center, the catalytic site of the enzyme, where molecular oxygen binds and is reduced to
water. For the Cut dinuclear center, the crystal structure 13 shows that two copper atoms
are bridged by two cysteines, with two histidines for the terminal ligands of each copper;
a methionine residue ligates one copper atom and another glutamate ligates the other
copper atom loosely (Fig.1.6). The planes of both the heme a and the a 3 are perpendicular
to the membrane plane and facing each other. Heme a has two axial histidine ligands and
is a low-spin heme. Heme a 3 is also ligated by a single histidine residue, and is a high
spin heme. The CuB site is coordinated by three imidazoles of histidines, in agreement

with mutagenesis studies. In the bovine crystal structures, Cup, is 19A from heme a and

12

 

V“‘

CUA‘ H2 01H+ channel?

A l \ W /
lSubunit I/II interface m CcO WY Outside
\

 

Figure1.5: Proposed water channel in bovine CcO. Blue ribbons on the top are from
subunit II which houses dinuclear CuA site (blue spheres). The green ribbons on the far
right represent subunit IH, which does not have any metal centers and its ﬁmction is
unclear. The purple ribbons represent subunit I which contains two hemes (red) and C113
site (light green sphere). The proposed water channel is shown as the gap between
subunits I and 11, beginning near heme a3 propionates and the Mg/Mn site (brown
sphere), with residues lining the channel in ball-and stick form The three small blue
spheres represent crystallographically vis1ble water molecules. The ﬁgure is from
Florens, L. et a1.14

13

22 A from heme a 3. Both hemes a and a3, and CuB are located at about the same depth in

the membrane, which is about 1/3 of the membrane spanning region.

2.1.2 The non-redox active metal centers

Besides the redox-active metal centers, crystal structures of bovine and bacterial
CcO revealed two non-redox active metal centers. The Mg site ileA from the surface of
the protein, at the bottom of a proposed water channel It bridges subunits I and II, shares
one ligand with the Cup, center, and is bonded through a histidine ligand and a water
molecule to the propionate of heme a3. The role of this site is not clear. However, it is
located in the vicinity of an apparent water channel, which hints that it may function as a
regulator of proton/water exit 7. The other is the Ca/Na binding site, which is located near
the P side of the membrane in subunit I and appears to be accessible to solvent. It is
formed by residues on the top of helix I of subunit I and in the loop following it, with

possible additional contacts from the loop (Fig.1.6).

2.2 The oxygen chemistry

During the catalytic cycle, electrons are transferred from reduced cytochrome c to
Cu, ﬁ'om Cu to heme a and from there to the binuclear catalytic center heme a3/Cu3.
The measured steady-state activity of the Rs.CcO is ~1600 to 2,000 s'1 at pH 6.5 ”.
Spectroscopic and chemical methods have provided a quite clear picture of the
mechanism of oxygen chemistry carried out by CcO (Fig.1.7). CuA and heme a are the
electron carriers which transfer two electrons from reduced cytochrome c, one at a time,

to heme a and then to the binuclear center. Once the heme a3 and CuB center are reduced,

14

Met263 HisZ17

 

ASP229
Cy5256
Hi3260
Glu254
mHiMﬁ

Figure1.6: Structure showing the arrangement of metal centers and associated
amino acids in bacterial CcO. The copper centers are shown as black spheres, Mg is
shown as green sphere, heme a and heme 03 are both in green.

 

 

 

 

 

 

 

 

 

 

 

 

 

2133+ CuBZ+-0H' l a32+ CuB+ l a32+-O-O CuB+
YOH l YOH YOH
H20
2e', 2H+—>
ll

 

 

 

+
a34+=02’ (3032+ T a34+=02‘ C032 -OH‘
Y'O 1 YO
F “20 P

Figure 1.7 : Proposed catalytic reactions at the active site of CcO. Abbreviation used
as follows: Y, Y288 (R.sphaeriodes numbering); R, reduced; A, oxy; P, oxoferryl; F,
protonated form of P; O, oxidized. Pumped protons are not shown in the view. The two
phases of the catalytic cycle merge to a certain extent, but the metal reduction is
essentially from O to R. The ﬁgure is reproduction of a ﬁgure from Mills, D. et al.12.

 

 

 

 

 

 

 

16

O2 binds to the binuclear center and transiently forms the intermediate A (FeZ+-O2). The
0-0 bond is rapidly broken in a concerted fashion requiring a total of four electrons and
at least one proton. Three of the four electrons are donated by the F e332+ and Cu3+ center,
forming the ferryl-oxo state, Fe4+=O, and cupric ion Cu32+. If heme a is oxidized, the
electron required to break the 0—0 bond appears to come from a nearby tyrosine forming
a radical. However, if heme a is in its reduced state, then the electron comes from heme

a.
2.3 Electron and proton transfer theory

2.3.1 Electron transfer in biological system

Numerous processes in biology involve electron transfer reactions, including
photosynthesis, and respiration. Many metabolic pathways as well as detoxiﬁcation
routes also involve electron transfer. Classical theory (Marcus theory) was used to
explain inner-sphere and outer-sphere electron transfer and the electron transfer rate is

given by the expression15 :

272' 1 — x1+AG° 2
kET =_HjB ( )

6X
72 ./47r2k,T p{ 421ka } 1-2

Where kET is the electron transfer rate, HA3 is the electronic coupling between the initial

 

 

and ﬁnal states, it is the reorganization energy, and AG0 is the total Gibbs free energy
change for the electron transfer reaction, and kb is the Boltzmann constant. The key

parameters are depicted in F ig.1.8.

l7

 

 

 

 

 

 

 

 

 

Nuclear Coordinate

Figure 1.8: Potential energy diagram showing parameters from Marcus electron
transfer equation. The left hand parabola represents the potential energy surface for the
nuclear motion of the reactants in the initial state, and the right hand parabola represents
the potential energy surface for the nuclear motion of the products in the ﬁnal state. The it
represents the reorganizational energy of the eletron transfer reaction and AG° is the
driving force for the electron transfer reaction, and HA3 represents the electronic coupling
between the initial and ﬁnal states. The ﬁgure is the modiﬁcation of the diagram from the
following link: http://www.life.uiuc.edu/croftS/hioph354/lect19.html

18

According to the Marcus theory, the activation barrier for electron transfer depends
on three factors: the driving force (AG°), the reorganization energy (it), and the distance
between the redox centers. The reorganization energy reﬂects the extent of inner sphere
(Am) and outer sphere (2.0m) nuclear rearrangement upon electron transfer. The balance

between AG° and 2 result in different electron transfer rates.

Protein structure is an important determinant of the reorganization energy in
biological electron transfer reactions. Molecular dynamic simulations using dielectric
continuum models indicate that embedding a redox center inside a cavity with low
dielectric constant can greatly lower the outer-sphere reorganization energy”.
Furthermore, it is found that constraining the coordination environment around metal
centers could reduce the inner-sphere reorganization energy 17. This is demonstrated by
an experiment in which a small water soluble protein, cytochrome c, was treated with
guanidine-HCI to unfold the protein and alter the environment such that the redox
potential of the heme was changed, as well as the reorganization energy, altering the

electron transfer rate1 8.

Another factor in the Marcus equation is the H23, the electronic coupling. It
reﬂects the strength of the reaction between reactants and products at the nuclear
conﬁguration of the transition state. When donors and acceptors are separated by a long
distance (210 A), the interaction between donors and acceptors are weak. In electron
transport enzymes, because of the folding of polypeptides, electron transfer centers are

often buried deeply in polypeptides and cannot come into contact. In 1974, Hopﬁeld

l9

proposed a tunneling model in which the coupling depends exponentially on the distance
(r) between redox centers 19. Combining Hopﬁeld’s tunneling model with McConnell’s
model 20, it is seen that the coupling depends on both the length and the nature of the
spacing elements between the donor and acceptor. Electron transfer between the donor
and acceptor, spaced by the spacing elements is referred to as tunneling, and is generally
found to have an exponential dependence on the distance between the donor and the
acceptor described in equation 1.4:

km (rDA) = koexp(-BTDA) 1.3
where k0 is a temperature prefactor, roll is the distance between the donor and the

acceptor, and B is the exponential decay constant and reﬂects the nature of the
intervening matter. Different donor—acceptor systems have different 0 values due to the
variant nature of the bridge elements, therefore they have different electron transfer rates.
For example, conjugated polyenes exhibit smaller [3 values, as low as 0.04 A" and show
remarkably efﬁcient electron transfer over long distances 2‘ while in proteins possessing

or-helix or B-sheet, the [3 values are about 1.0-1.3 A‘l.

Long-range electron transfer is commonly found in many enzymes and can still
show high efﬁciency. For example, electron transfer in C60, ﬁ'om Cu to heme a, is over
19 A with an electron transfer rate of 2-9 x104s'1. Some possible electron transfer
pathways were proposed and identiﬁed by mutagenesis”. The very efﬁcient long-range
electron transfer from CuA to heme a (2x104s‘1) proceeds at a very low driving force
(AGO: -0.05eV) 23. Three electronic coupling routes have been proposed to explain the

fast electron transfer 24'27. Some electron coupling pathways between heme a and heme a3

20

are also considered in which the axial histidine ligands of heme a and heme a 3 are
involved 25’ 28. It is still controversial as to whether there are particular routes for electron
transfer facilitated by the protein, or whether the [3 factor for electron movement through

protein is approximately the same as the reorganization energy.

Redox reactions can involve both electron and proton transfers. The coupling
between electrons and protons is an important feature in the complexes of the respiratory
chain. Proton-coupled electron transfer (PCET) mechanisms have been studied in pH-
dependent rate studies of some self-exchange reactions in model systems, in which a rate
is increased with the increasing pH. Reviews concerning PCET in metal complexes came
out in the 19903 29. In the same period, a series of theoretical calculations of the PCET
mechanism were published by several researchers 30'32. As more insights are gained ﬁ'om
the study of enzymes and their biological processes, it became more evident that
biological systems take great advantage of PCET. Over the time span of evolution,
complex structures have evolved so that the PCET mechanisms are utilized to drive
vectorial proton translocation. The coupling of electron and proton transfer is at the heart
of photosynthesis, respiration, nitrogen ﬁxation, and many other key bioenergetic
processes. Photosystem II is a good example of how the biological system uses PCET. In
the photosynthetic water oxidation reaction depicted in Fig.1.9, a substantial body of data

provide evidence that both the oxidation of tyrosine and the reduction of the tyrosine

21

u

Figure 1.9: S-cycle of Photosystem II. The jagged line represents photon radiation.
Oxidation of two molecules of water requires removal of four electrons. Absorption of
each photon results in the transfer of only one electron. So the oxygen evolving complex
must oxidize four times in a row to form two dioxygen molecules. The Mn in Mn4 cluster
undergoes ﬁve different oxidation states from Soto S4.

22

radical by a tetranuclear manganese-0x0 cluster, the active site for water oxidation,

involve coupled electron and proton transfers 3’3 .

2.3.2 Proton transfer theory

To achieve biological energy conversion, membrane-spanning protein assemblies
use the energy provided by photochemical or redox reactions to pump protons against the
electrochemical gradient, which in turn creates the energy to drive ATP synthesis. To
understand the mechanism of proton pumping, a high level of detail is needed. In addition
to knowing the driving forces behind proton movement in enzymes, the structure and
ﬁmction of proton translocation pathways must also be elucidated. Cytochrome c oxidase
(C00) is a bioenergetic enzyme that functions as a proton pump, driven by reducing
oxygen to water. While advances in understanding proton transfer in CcO have been
greatly aided by a number of crystal structures, along with site-directed mutagenesis, the
functional and structural complexity of CcO presents a formidable obstacle to
understanding how, when and where protons move through the protein. Therefore,
studies on simpler, well-understood systems could provide us with some insights into

how protons are taken up and released from the protein.

More so than electron transfer, proton transfer is extremely dependent on the
distance. Since tunneling is inversely proportional to the mass, m%, the distance over
which a proton can tunnel is less than a covalent bond length. For example, long-range
electron transfers (over 10-20 A) are commonly found in biological systems with

competent rates, whereas the range for proton transfer is restricted to less than 1. A. To

23

achieve a long-distance transfer, protons appear to move via hydrogen-bonded chains

formed by water and amino acid residues within the protein.

Proton transfer is achieved, presmnably, via a hopping mechanism that is similar to
the relay theory proposed by Grotthuss34 200 years ago. The remarkable thing about the
Grotthuss mechanism is that it was proposed prior to the understanding of atomic theory,
and at that time, the composition of water was not determined. In modern molecular
terms, the Grotthuss mechanism can be summarized as a “hop-and-tum” mechanism
(Fig.1.10) involving the chemical exchange of hydrogen nuclei along successive
hydrogen-bond donor and acceptor groups, sometimes involving an extensive network,
followed by the reorientation of these water bonded networks. The true signiﬁcance of
the mechanism was not appreciated until quantum mechanics was developed and ions
could be described in more detail. Grotthuss’s theory and later Onsager’s theory 35 of fast
proton transport along hydrogen—bonded chains were further developed into a proton wire
theory by Nagle and Morowitz, when they revolutionarily applied it to proton transfer in
proteins36. Proton wires are thought to be important mechanisms to account for the fast
conductance of protons along chains of hydrogen-bonded protonable groups in biological

systems.

Redox-driven proton transfer in proteins can be considered to occur by two general
mechanisms: a direct mechanism or an indirect mechanism or a combination of both. A
direct mechanism is one where proton movement depends on the

protonation/deprotonation of the redox center involved, in response to a change in

24

 

F ('f-H (l)-H (')-H ('3-H (|)-H

—>H* (ﬁr-H (l)-H (IJ-H (IF-H ('3-H
HOP
H * +(I)-H->(?-H—>(|)-H->(I)-H->(l)—H —>

H-? H-? H-? H—(|) H-(I) H+->

 

 

4.) 8-? H-? H-9 H-9
M U M \J U

Figure1.10: Schematic depiction of the hop-and-turn mechanism for proton transfer
through a hydrogen-bonded water chain. Protons (blue) approaching an oriented
hydro gen-bonded water chain can hydrogen bond to the terminal water. The extra charge
now on the water can be released as the terminal proton (red) on the other side,
effectively transferring a charge across the chain. In order to accept another proton from
the same 3slide, a turn step is required. The ﬁgure is a modiﬁcation of a ﬁgure from Pomes,
R. et a1.

25

oxidation state. An indirect mechanism can be characterized as one where proton uptake
and release occur through conformational changes, leading to pKa changes at a distance

from the redox center.

A good example of a direct mechanism is demonstrated in the Mitchell hypothesis,
as seen in the Q-cycle in cytochrome bcl of the respiratory chain”. When quinone is
reduced to quinol, it is required to take up two protons; if the reduction (and proton
uptake) on one side of the membrane is followed by oxidation and deprotonation on the

opposite side, the vectorial proton transfer is accomplished.

The classical example of an indirect mechanism is the Bohr effect in hemoglobin.
When oxygen binds to the heme, the protein undergoes a conformational change in its
structure, leading to a change in pKa of the participating residue at a distance from the
heme and release of a proton. Conversely, the acid pH in tissues causes protonation of a

key histidine and a conformational change, favoring unloading of the 02 from the heme.

A combination of both direct and indirect mechanisms are used by some proton
pumping proteins such as bacteriorhodopsin (BR)39. In BR, light absorption drives the
conversion of retinal from all-trans to 13-cis (Fig.1.11). This conversion leads to the
deprotonation of its Schiff base and protonation of an aspartate (D85). D85 is connected
to a glutamate (E204) near the protein surface through water molecules and an arginine.
This connectivity leads to the cooperative pKa change between D85 and E204; when D85

is protonated, E204 releases one proton to the extracellular side. To reprotonate the Schiff

26

Ht (H20)u

H: I33 haw

 

 

 

 

 

 

 

 

 

 

X
0 . K
D96
H7
Cytoplasm
N L

 

 

 

 

 

 

 

Extracellular medium

Figure1.11: Photocycle scheme showing the main steps in light-driven proton
transport in bacteriorhodopsin. BR, ground state bacteriorhodopsin; PSB, protonated
Schiff base. The absorption of light hv induces an all-trans to 13-cis isomerization of
retinal and leads to a metastable protein state termed the K intermediate. L, M, N, and O

are all groansition intermediates during the proton transfer. The ﬁgure is from Garczarek,
F. et a1. .

27

base, another conformational change takes place, creating a hydrogen-bonded pathway
ﬁom the inside. Protonation of the Schiff base returns the retinal to its initial all-trans
resting state, ﬁnishing the cycle. By using both direct and indirect mechanisms during its

cycle, bacteriorhodopsin achieves proton translocation.

2.3.3 Proton transfer in CcO

The catalytic center of CcO is buried deeply within the protein and is situated about
30A from the internal bulk aqueous phase where substrate protons are taken up to form
water at the active site. To effectively deliver the substrate protons to the active site, the

enzyme must have a proton-delivering path Mutation studies 4‘45

and X-ray
crystallography results 3’ 46 have established two pathways by which protons are
transferred into the protein: the D and K pathways (Fig.1.12). The D and K pathways are
named after two key conserved amino acids, aspartate (D132) and lysine (K362). The D
channel starts at D132, and continues through a hydrogen-bonded network consisting of
approximately ten ordered water molecules and polar residues, to another highly
conserved amino acid E286 which is about 24A ﬁom D132. The K channel contains two
important residues E101 and K362 as well as T359 and Y288. Depending on the steps in
the catalytic cycle, the protons can be delivered via either the D- or the K-channel.
Mutation studies showed that the K channel is essential for providing substrate protons to
make water“. The D channel supplies protons for both water formation and for the
pumped protons (1H+/e'). The uni-directional proton transfer could be achieved by either

conformational changes of the E286 side chain during turnover or by changes in water

chain structure.

28

 

0 Path

D132 5*

Figure1.12: Key amino acid residues in well-deﬁned D and K proton uptake
pathways in Rs.CcO structure. The metal centers are represented by different colorﬁil
spheres: organge, dinuclear CuA; purple, Mg; green, Cug. Heme a and heme a3 are green
and brown color-labeled. D pathway is indicated as red dashed line and K pathway is as
blue dashed line. The ﬁgure is a modiﬁcation of ﬁgure from Mills, D. et a1“.

29

High-resolution crystal structures greatly facilitate our understanding of CcO at a
molecular level, however, the stationary snapshots of the protein structure cannot provide
any dynamic information. The mechanism of proton pumping and how this is coupled to
electron transfer remains controversial, particularly regarding the proton exit routes.
Close examination of the crystal structures, however, reveals a hydrophilic region located
at the interface between subunits I and 11. Previous work done by Florens et al suggested
that the Mn/Mg Site was part of the water exit channel and/or involved in a proton exit
pathway '4. A systematic mutagenesis study of amino acids R481/482 (Rs.) in this region,
which interact with the D ring of the propionates of both hemes, suggested that these
arginines might be part of a proton exit or backﬂow channel”. Yoshikawa proposed E51
as a proton release group that is part of the proton exit pathway on the basis of the bovine
crystal structure“. Electrostatic calculations were employed to examine the possﬂale
proton exits by evaluating the electrostatic potential, energies, and protonation state of
bovine CcO for different redox states of the enzyme (Fig.1.13). The results indicate that
K171B/Dl73B in bovine (R408 and K227B in Rs.CcO) form the most likely proton exit
pathway”. This has not yet been resolved as there are considerable differences in the Rs.

and bovine oxidases in this region above the hemes.

2.4 Water channels in proteins

The ubiquitous nature of water in living systems could lead one to think that water
charmels are not necessary in proteins. Indeed, both experimental and computational data
show that even the most hydrophobic regions of proteins are accessible to water

molecules 51. However, the control of water movement in some proteins is necessary and

30

A B

 

 

 

 

 

 

 

 

 

 

 

 

 

II
_1 maul ails—«,3: . .1 (.1
<3p51dl/uﬂvAsp2waﬁ——+zeo 9 N4"
\. "“291 +140 Pump Stroke
"9‘38‘ 9 —*130 I AG*=uomv
PRO 59" PM Asp364 "*1:
6 .e—o-e-o-e_'.tr 1L
0 :
BNC Hemea; are-0+0 0 IV (J! (c)
eta—0+0_.._ .30 ‘
N ‘ ackﬂew (“leak”)
o-o-oc—e ___Jr_ -260 “am“

 

 

Figure1.13: (A) The most possible proton exit pathways in bovine CcO and the
energy diagram. (B) Lysine17lB proton pathway. (a) Energy diagram of different
protonation states of key residues of the binuclear center (BNC). (b) Sequence of proton
transfer steps in the pump stroke. (c) Sequence of proton transfer steps leading to the
backﬂow. The numbers represent. 1, BNC; 2, His291; 3, PRA a3; 4, Asp364; Lys171b. In
both cases, energy barriers are evaluated. The ﬁill or empty circles represent protonated
or deprotonated sites. The ﬁgure IS a modiﬁcation of ﬁgures from Popovic, D. et also

31

important. It is likely that speciﬁc mechanisms for water transport occur in many
organelles. Since the ﬁrst water channel, aquaporin, was discovered in 1992, research on
aquaporins has been a rapidly developing ﬁeld. Aquaporins are crucial for life and they
are found in all organisms, from bacteria to plants to human. Aquaporins facilitate rapid,
highly selective water transport, thus allowing the cell to regulate its volume and internal
osmotic pressure according to hydrostatic and/or osmotic pressure differences across the
cell membrane. The particularly remarkable thing about aquaporins is that the selectivity
of aquaporin water channels is so high that even protons cannot pass. The discovery of
aquaporin-l (AQP-l) was serendipitous and occurred during studies of red cell Rh blood
group antigens, but turned out to be an important chapter in water channel studies 52.
Since AQP-l was discovered, ten more aquaporins in this family were identiﬁed. So far,
the best understood members of this family are AQP-l and the bacterial channel protein
aquaglyceroporin, Glp-F, which preferentially transfers glycerolS3.

Before crystal structures of AQP-l 54 and Glp-F 53 were reported, little was known
about the origin of water selectivity. Both of these proteins are made up of six
transmembrane tit-helical domains arranged in a right-handed bundle, with the amino and
the carboxyl termini located on the cytoplasmic surface of the membrane. The amino and
carboxyl halves of the sequence show similarity to each other, in what appears to be a
tandem repeat (Fig.1.14). There are also ﬁve interhelical loop regions (A — E) that form

the extracellular and cytoplasmic vest1bules. Loops B and E are hydrophobic loops which

32

Loop A

(a) (b) Extracellular /\

Loop C

   

H2/

 

w -COOH

C-terminal NH?
hollx

Cytoplasmic

   

Figure1.14: X—ray crystal structure of AQPl. (a) AQPl tetramer viewed from the
extracellular surface. Blue part in one monomer representing the N-terminal and yellow
part representing the C-terminal tandem repeat. The red dashed line indicates the pseudo-
2-fold axis relating the two halves of the monomer, and the red square indicates the 4-
fold axis of the tetramer. The asterisks indicate the water pore in each of the four
subunits. (b) AQPl monomer viewed in the direction parallel to the membrane.
Membrane-spanning helices are denoted as Hl—H6, loops are denoted as A—E, and the
two pore helices formed by_ loops B and E as HB and HE, respectively. The ﬁgure is a
ﬁgure from Walz, T. et a1. 5“

33

contain the signature motif Asn-Pro-Ala (NPA), and these two loops overlap in the
middle of the lipid bilayer of the membrane, forming a 3-D 'hourglass' structure where
the water ﬂows through. This overlap forms one of the two well-known channel
constriction sites in the peptide, the NPA motif and a second and usually narrower
constriction known as the “selectivity ﬁlter”. The crystal structures of aquaporins explain
how aquaporin allows water to permeate while excluding other cations. Through a
combination of channel sterics and solute binding Sites, AQPl facilitates the rapid and
highly selective throughput of water.The steric limit of ~2.8 A is established at the
constriction region, and the chemical properties of the residues forming this structure
provide additional criteria for solute selection. Different aquaporins have different
peptide sequences, which allows for the size of the pore in the protein to differ between
aquaporins. The resultant size of the pore directly affects what molecules are able to pass
through the pore, with small pore sizes only allowing small molecules like water to pass

through the pore while excluding ions such as H+or KL

In CcO, both protons and waters are transferred through the protein, but these two
processes may take place in separate channels. However, water is produced at the active
site and it has been suggested that it may exit close to the Mg/Mn site, exiting at the
interface of the subunits I/II/III. The pumped protons may exit through the same path, or
via a different path. The selective mechanism for water transport in CcO via the Mg/Mn
site may be similar to aquaporin and may prevent the ﬂow of protons in this region. A
proposed water channel was studied in CcO by rapid-ﬁeeze quench-EPR (ESEEM) in the

laboratory of Ferguson-Miller. The studies showed that bulk water can exchange quickly

34

within the region near the Mg/Mn site and suggest that water exits the protein via a

discrete pathway, not by random diffusion 8.

Water channels are not necessarily static. Members of the cytochrome P450 family
catalyze mono-oxidation of a wide variety of endogeneous or xenobiotic compounds. The
typical reaction involves the insertion of one atom of oxygen into an organic substrate
(RH) while the other oxygen atom is reduced to water. The exclusion of water from the
active site has been shown to be essential for efﬁcient enzymatic ﬁmction“. Oprea et al.
proposed a two-state model in which a conserved arginine undergoes a conformational
change from the initial stable state to another metastable state“. In the new state, a new
ﬁrnctional water channel is formed. This water cluster communicates with the surface in
the closed state and is partially replaced by the ﬂipping arginine side chain in the open
state, allowing water molecules to exit to the surface or to reaccess the active site

(Fig.1.15).

2.5 Questions to be addressed

Although considerable progress has been made on understanding the mechanism of
CcO, with the help of high-resolution crystal structures, mutations, and kinetic analysis,
the mechanism of coupling between proton translocation and electron transfer has not yet
been resolved and many ﬁmdamental questions regarding regulation of activity and
efﬁciency remain to be addressed. In my thesis, some efforts are made to try to

understand the following aspects: the proton coupled electron transfer events in some

35

 

 

 

 

 

 

____l

 

Figure1.15: Snapshots of four representative P450 structures (lphc, lcpt, 2hpd, and
loxa) in which cavities forming the aqueduct can be observed. The grid points
comprising the cavities are shown in blue. Black arrowheads indicate where the cavities
reach the protein surface. The substrate-access channel is not shown. (a) In P4500am
(lphc), rotation of the R299 side chain opens a wide channel directly connecting the
substrate cavity to the protein surface. (b) In the R398-modiﬁed P4SOBM3 (2hpd), the
substrate cavity extends into the region previously occupied by water. However, the
guanidinium group of the rotated R398 blocks a direct path from the substrate cavity to
the protein exterior. (c) In P450terp, substrate and solvent cavities are only connected if
the probe radius is reduced to 0.82 A in the cavity calculations (purple). Dynamic
ﬂuctuations are expected to widen the narrow channel formed between W317 and T103,
which separate the two cavities. The second cavity, extending to the left, does not
connect with the surface. (d) In P450eryF (loxa), the solvent channel extends ﬁ'om the
substrate cavity to the surface, with the narrowest part between the heme propionate A
and T291. The ﬁgure is the ﬁgure from Oprea, T. et 3.1.57

36

site-directed mutants in the vicinity of heme a, the proton/water environment of the
Mg/Mn site in the proton exit region, and the changes in that environment depending on

the redox state of the protein.

3. Introduction of Electron Paramagnetic Resonance (EPR)

Since its discovery in 1944 by Russian scientist Yevgeny Zavoisky from Kazan
State University, EPR spectroscopy has gradually developed into a powerﬁrl tool for
studying species with unpaired electrons. Examples of centers that have unpaired
electrons are organic radicals and transition metals in appropriate oxidation states such as
Cu2+, low Spin ferric ion in hemes, and Mn2+. For the past three decades, EPR has been a
major tool in efforts to characterize metal centers in model compounds or metalloenzyme
active sites in bioinorganic chemistry and biology. By analyzing hyperﬁne-coupling
interactions between the electron spin and the nuclear spin associated with either the
ligands or the nearby nuclei, valuable information regarding the composition, structure,
and bonding of paramagnetic centers in proteins can be gained. The application of this
technique is the major focus of this thesis, and therefore a more detailed account of the

underlying theory will be explored.

Every electron has a magnetic moment and this makes every electron behave like a
tiny bar magnet. The spin does not contribute to the overall energy of a system when the
external magnetic ﬁeld is absent, resulting in a degeneracy of energy levels. However,

this degeneracy can be broken when an external magnetic ﬁeld with strength B0 is

37

applied. It will have the lowest energy state when the magnetic moment of the electron,
u, is aligned with the magnetic ﬁeld and a highest energy state when u is aligned against
the magnetic ﬁeld. The magnetic energy is given by the expression: E= - p Hcosf) where
E is the magnetic energy, 11 is the magnitude of magnetic dipole moment, H is the
magnitude of the magnetic ﬁeld, and 0 is the angle between the magnetic moment and the
external ﬁeld. The two states are labeled by the projection of the electron spin M5, on the
direction of the magnetic ﬁeld. Because the electron is a spin ‘/2 particle, the parallel state
is designated as E = gepBBo where ge is so-called “g factor” and us is the electron Bohr
magneton. This equation implies that the energy difference is directly proportional to the
strength of the magnetic ﬁeld, as shown in the diagram below (Fig1.16). The absorption
or emission of electromagnetic radiation of energy = hv causes a transition of unpaired
electron between two energy levels. When combining a = hv and AB = gepBBo, the
fundamental equation of EPR spectroscopy hv = gepBBo is obtained. From the equation,
we can see two possible alternative instrumental conﬁgurations. We can apply a constant
magnetic ﬁeld and scan the frequency of the electromagnetic radiation as in conventional
spectroscopy. Alternatively, we can keep the electromagnetic radiation ﬁequency
constant and scan the magnetic ﬁeld. A peak in the absorption will occur when the
magnetic ﬁeld tunes the two spin states, so that their energy difference matches the
energy of the radiation. This ﬁeld is called the ﬁeld for resonance. To solve sensitivity
issues in an EPR experiment, commercial instruments often have the ﬁ‘equency ﬁxed

while scanning the magnetic ﬁeld.

EPR has two categories: continuous-wave (cw) EPR and pulse EPR. In cw EPR,

38

MS: +112

AE

\

B H M, = 412

B 2 0 Resonant
Field

 

 

Figurel.l6: EPR energy diagram for S=1/2 system. As the external magnetic ﬁeld is
applied to a paramagnetic system with spin S=1/2, the degenerate energy levels are lifted
and split into two spin manifold Ms: -1/2 and Ms: +1/2. When the resonance condition
AE =hv =gBB is met, the absorption of the microwave radiation occurs. The magnetic
ﬁeld strength at the resonant point is called the resonant ﬁeld.

39

molecules are irradiated continuously by low-power microwave radiation; while in pulse

EPR, microwaves are applied in very short and high-powered pulses. Most EPR
Spectrometers are operated at 9GHz (also known as X band), a ﬁ'equency that provides a
compromise between sensitivity and ease of sample handling. Some other ﬁequencies are

also used to achieve higher resolution.

Measurement of g-factors can give us some useﬁrl information; however, they do
not tell us much about the molecular structure of our sample. Fortunately, the unpaired
electron, which gives us the EPR spectrum, is very sensitive to its local surroundings.
The nuclei of the atoms in a molecule or complex often have a magnetic moment, which
produces a local magnetic ﬁeld at the electron. The magnetic coupling between the
electron and nuclear magnetic moments is called the hyperﬁne interaction. It gives us a
wealth of information about our sample, such as the identity and number of atoms
involved in a coupling, as well as their distances from the unpaired electron. Two
mechanisms by which electrons and nearby nuclei interact are by through-bond
interactions and through—space interactions (dipolar interaction). The former applies to
the case of isotropic interactions (spectra independent of sample orientation in a magnetic
ﬁeld) and the latter to the case of anisotropic interactions (Spectra dependent on sample
orientation in a magnetic ﬁeld). In principle, the hyperﬁne couplings are manifested in
EPR spectra, however, in practice, these couplings cannot be observed directly because
of the inhomogeneous broadening of EPR lines in protein samples. Two techniques can
reduce these static inhomogeneous broadening contributions and dramatically increase

the spectral resolution so that coupling information can be extracted. These are: l)

40

electron nuclear double resonance (ENDOR) and 2) electron spin echo envelope
modulation (ESEEM). In both methods, the NMR transition frequencies of nuclear spins

are monitored through EPR electron spin transitions in an indirect way.

3.1 Electron spin echo envelope modulation (ESEEM) spectroscopy

In the ESEEM experiment, the NMR transitions are observed due to the mixing of
the nuclear states within each electron spin manifold. This state mixing is derived ﬁom
the anisotropic hyperﬁne coupling and yields non-zero EPR transition probabilities for
semi-forbidden (11mg: i1, Am; = i1) and allowed (Ams= 21:1, Am; =0) EPR transitions.
Because of the constraints placed on ESEEM by the requirements of nuclear state mixing
(each energy level must be involved in at least two different microwave transitions) and
the microwave pulse bandwidth (the microwave pulse bandwidth needs to be broad
enough to cover the spread in ﬁequencies needed to fully excite the "branching"),
ESEEM is particularly USCﬁll for studying systems with weak hyperﬁne coupling

interactions.

ESEEM is a pulse version of the electron paramagnetic resonance (EPR) teclmique.
It is a powerful tool for measuring spin transitions of nearby nuclei with relatively weak
magnetic couplings to electron spins, and is used widely to study the electronic and
geometric structure of paramagnetic centers in biological systems. ESEEM is mostly used
in determining hyperﬁne couplings between a paramagnetic center and its ligands, which
will help identify the types of interacting ligands, and determine their orientation with

regard to the paramagnetic center’s magnetic axes. In the ESEEM experiment, electron

41

spin echoes are generated by applying suitable short and high-power microwave pulses to
induce semi-forbidden transitions, in addition to allowed transitions of nuclear spins
magnetically coupled to the electron spins. The electron spin echo decay envelope is then
measured by plotting the integrated intensity of these echoes as a ﬁmction of the time, 1:.
The resulting spectrum shows an overall decay of the electron spin magnetization that is
modulated by electron-nuclear hyperﬁne interactions 58. Here two ESEEM experiments,
two-pulse Hahn echo and the three-pulse stimulated echo, will be discussed in detail. For
simplicity, an electron spin of 1/2 and nuclear spin of 1/2 system will be used as the

example in this chapter.

For a S=1/2 and I= 1/2 system, the spin Hamiltonian can be expressed with the
assumption that the g is isotropic and hyperﬁne coupling is axial as follows:
H/h=wSSZ+AzzSZIZ+AxZSZIX—w,IZ 1.4
Where the ﬁrst term is the electronic Zeeman, the second and third terms are electron-
nuclear hyperﬁne interaction and the last term is the nuclear Zeeman. The Hamiltonian
matrix is easily constructed in a general basis set consisting of electron and nuclear spin
product states, lms, m,>. Diagonalization to yield the eigenvalues and eigenvectors of

eqn.1.4 can be carried out independently for the two electron spin manifolds since the

only term in the operator that gives rise to off-diagonal matrix elements is that involving
Ix (McCracken, John, Encyclopedia of Inorganic Chemistry). The results are summarized

in the energy level diagram (Fig.1.17).

42

|3>

 

|4>

 

 

Eigenvectors:
|l> = cos(cpa/2) |+ +> + sin(cpa/2) l+ ->

|2> = -Sin(q)a/2) |+ +> + cos((pa/2) |+ ->

13> = cos(ch/2) l- +> + sin(ch/2) l- ->

|4> = -Sin(ch/2) l- +> + cos(<pB/2) l- ->

 

wa-J((OI — A/2)2+ B2/4

 

wﬁsﬂwl + A/2)2+ 32/4

Figurel.l7: Energy level scheme and eigenvectors for the S=1/2, and I=1/2 system.
This ﬁgure is a ﬁgure courtesy of Dr. John McCracken

43

The lul and IV] in Fig. 1.17 are normalized probability amplitudes for the EPR

transitions. The angles, (pa and (pp respectively, deﬁne the axes of quantization for the or
and B electron spin manifolds, and are expressed as sinth = B/2toa and sinth = B/2wg,

where (pa and (DB are the nuclear transition frequencies. Four EPR transitions are possrble

between the upper and lower doublets. When microwave pulses are applied to the spin
system, the allowed and forbidden transitions are induced with different probabilities

determined by the corresponding matrix elements.

Mims 59 derived the originally analytical expression for two-pulse ESEEM ﬁmction
as equation 1.6:
Emod (r) = 1-k/2{'1-cos (war) —cos(a)pzj + ’/2 cos[(a)a+wg)z] + ’/2 cos[(a)a-w,a) 7]} 1.5
Where I is the time interval between the microwave pulses and k is the modulation
amplitude factor with the expression: k= Sl112((t)a-0)p) = [wlB/(tuacom 2 1.6
The observed ESEEM spectrum is a composition of modulation function and a decay
ﬁrnction that describes the loss of magnetization due to spin relaxation. From equation
1.6, we can see that modulations of the two-pulse echo amplitude occur at both
ﬁmdamental hyperﬁne frequencies and their sum and difference combination frequencies.
The amplitude of the modulations is directly related to the transition probabilities for the
two different transitions. In a spin system with multiple nuclei contribute to the
modulation of a single paramagnetic center. The modulation ﬁrnction in this instance is
given by the expression:

N
5(7): Vdecayn Erlnod(z') 1-7

i=1

44

where N represents the number of coupled nuclei. In equation 1.7, the observed ESEEM
spectrum is the product of contribution from each nucleus, leading to frequencies that are
combinations of the frequencies arising from different nuclei. This increased complexity,
combined with the lowered resolution due to short phase memory time, makes two-pulse

ESEEM analysis a formidable challenge.

Three—pulse ESEEM has the pulse sequence of 90°-t-90°-T-90° in which two
microwave pulses separated by ‘l.’ are applied, followed by a third pulse after time T, and
the stimulated echo is observed at time I after the third pulse (Fig.1.18). Unlike two-pulse
ESEEM, no sums and differences of fundamental nuclear frequencies are observed in
three-pulse ESEEM spectrum, greatly simplifying the whole spectrum. Moreover, the
short phase memory time limitations associated with two-pulse ESEEM are largely
alleviated in three-pulse experiments, therefore, the modulation echo envelope stays
longer, resulting in narrower lines in the spectrum. For a typical three-pulse
measurement, 1 is ﬁxed to a value shorter than the spin-spin relaxation time of the
electron spin center, and the interval (T) between the second and third microwave pulses
ranging from 40 ns to ~ 10 us is scanned. For the speciﬁc case of the Mn(II) site in CcO,

t was set to 204 ns and T was stepped in 16 ns increments from 40 ns to > 8 us.

For a three-pulse ESEEM, the analytical expression for a S=l/2, I=1/2 system is

given in equation 1.8:

Emod(1:, T) = l-k/4{[1-cos(wat)][l-cosmBT)]+[ 1-cos(t1)Bt)][l-coswaT)]} 1.8

45

The equation shows that three-pulse ESEEM has a simpler form, in that there are only
lines from fundamental nuclear frequencies. The equation also Shows that the modulation
is affected by the I value. If carefully chosen, different T values can either enhance or
suppress certain ESEEM signals from contributions of different nuclei. This :-

suppression effect is very useful in making spectral assignments.

3.2 Electron nuclear double resonance (ENDOR) spectroscopy

ENDOR is another pulse EPR technique in which the ENDOR spectrum is obtained
by recording the echo intensity as a ﬁmction of the frequency of the RF pulse. A change
in the echo amplitude occurs when the RF is on resonance with a NMR transition,
therefore, generating an ENDOR signal. There are many ways to run pulse ENDOR.
Mims ENDOR 60 and Davies ENDOR 61 are the most frequently employed methods so
far. Davies ENDOR is a selective method and is suitable for systems with medium to
large hyperﬁne couplings (A>2MHz). Mims ENDOR, on the other hand, is nonselective
and most efﬁcient for systems with small hyperﬁne couplings (A<2MI-Iz) but suffers

blind spots. If carefully chosen, proper 1: values can place blind spots well outside the
1
ENDOR spectrum. Generally speaking, Davies ENDOR is a better way to record H

2
while Mims is best for H. Davies ENDOR is employed in the projects in this thesis and

will be given more space for discussion.

Fig.1.19 shows a schematic description of the Davies ENDOR sequence and the

evolution of spin state populations during a Davies ENDOR pulse sequence. “In the

46

 
     

2—pulse: E=f (1')

r.
‘.

 

342111“: E=f(1')

Figurel.lS: The diagram of pulse sequence for two- and three- pulse ESEEM
experiments. In the two-pulse experiment at time t after the second pulse, a spin echo is
formed. The modulation envelope is obtained as t is incremented. In the three-pulse
sequence, two pulses separated by ‘t are applied, followed by a third pulse after time T,
and the stimulated echo is observed at time t after the third pulse. The modulation

envelope is obtained as T is incremented. The ﬁgure is a modulation of the ﬁgure from
Deligiannakis, Y. et a1. 62

47

preparation step of the pulse sequence, a selective microwave 7t pulse is applied to one of
the electron spin transitions to transfer the initial thermal polarization of the electron Spin

to the nuclear spin polarization. In the mixing step, a single RF pulse is applied, when the
RF frequency matches the frequency Of one of the polarized NMR transitions for a
nucleus that is hyperﬁne coupled to an electron, nuclear spin polarization is transferred
between nuclear levels. This transfer of nuclear spin polarization also changes the
electron spin polarization of those electron Spin states connected by the EPR transition in
the preparation period, which can be detected using a two-pulse (Hahn) echo pulse
sequence. The RF -induced nuclear spin polarization transfer is manifested as a change in
the electron spin echo mtensity”63. For nuclei with different hyperﬁne couplings, one can
vary the width of the microwave pulse to achieve enhancement in amplitude. In general,
longer microwave pulse width enhances ENDOR amplitude from nuclei with small

hyperﬁne couplings, while amplitudes ﬁ‘om nuclei with large hyperﬁne couplings can be

enhanced by using shorter microwave pulse widths with higher magnetic ﬁeld strengths.

“ENDOR is a nonlinear phenomenon. Intensities of ENDOR can be markedly
altered by, not only the calculable effects, such as spin concentration, but also by spin
relaxation, by the degree of EPR and RF power saturation, and the size of holes burned
by the microwave pulse“. So ENDOR is not suitable for quantiﬁcation, but is useﬁrl in
providing hyperﬁne-coupling interactions between the paramagnetic center and its
surrounding nuclei. Obtaining hyperﬁne-coupling constants using the ENDOR method is

very common.

48

S=1/2, I=1/2 Davies ENDOR

 

 

   

 

 

 

1t
T Fl, 1; _ - tr>>tn , t,>>_t_e
5. repetition dela
— RF L y
B __ 1;: i=1
‘wﬁm3%_u%—u
FENDOR=100%
i ii iii iV

Figurel.l9: A) Davies ENDOR pulse sequence, B) Evolution of spin state population
during the Davies ENDOR pulse sequence. Electron (11.) and nuclear (an) spin
transitions excited by selective microwave and RF pulses, respectively. State populations
are shown using colored bars. The ﬁgure is a modiﬁcation of a ﬁgure from Tyryshkin, A.
et a1.

49

4. Conclusion:

Progress has been made in the study of C00 for the past three decades, but many
questions still have to be addressed. Proton transfer and electron transfer during protein
turnover and how these processes are coupled, are still not well understood. In this thesis,

the author attempts to understand the enzyme from the following perspectives:

Previous work indicates that electron transfer ﬁom heme a to the binuclear center
depends on a protonation event and the reaction is limited by the proton transfer 6568. It
was also found that electron transfer from Cup, to heme a is also coupled to proton
transfer69‘73. Studies on the dependence of electron transfer ﬁ'om Cu to heme a on proton
transfer has led to conflicting results in the literature. In the second chapter, a series of
mutants at position 44, close to heme a, were studied by EPR, fast kinetics, and

electrochemistry methods in a systematic way with the hope of shedding some new light

on proton coupled electron transfer reaction associated with this site.

Proton movement pathways in CcO are of great importance in trying to understand
the mechanisms of proton translocation across the membrane. Crystal structures of C60
have enhanced our understanding of likely proton uptake and exit pathways. However,
the dynamics, or motions, of the protein cannot easily be observed in a static crystal
structure. Additionally, the resolution of the crystal structure is usually not good enough
to determine the placement of protons, but EPR is a complementary technique in this
regard. Previous studies, using the Mn-substituted enzyme and rapid ﬁeeze quench

coupled EPR, provided some intriguing results on possible proton/water exit(s)8’ 14. In

50

chapters three and four, ENDOR and ESEEM studies of Mn-substituted CcO are
described in detail, following up on the previous work to better deﬁne the dynamics of

water and other ligands in this critical region.

High ﬁeld cw EPR was also performed to study the enzyme at different oxidation
states in the presence of deuterium or l7O-labeled water, utilizing the strong Mn2+ signal.
Some unexpected disappointing results were explained and several factors explaining the
ﬁndings were also found. In the future, high-ﬁeld EPR could be very useful as it has the

potential to facilitate our understanding of CcO.

51

REFERENCES:

l. Bushnell, G. W.; Louie, G. V.; Brayer, G. D., High-resolution three-dirnensional
structure of horse heart cytochrome c. J Mol Biol 1990, 214, (2), 585-95.

2. Iverson, T. M.; Luna-Chavez, C.; Cecchini, G.; Rees, D. C., Structure of the
Escherichia coli fumarate reductase respiratory complex. Science 1999, 284, (5422),
1961-6.

3. Iwata, S.; Ostermeier, C.; Ludwig, B.; Michel, H., Structure at 2.8 A resolution of
cytochrome c oxidase ﬁom Paracoccus denitrzﬁcans. Nature 1995, 376, 660-669.

4. Svensson-Ek, M.; Abrarnson, J.; Larsson, G.; Tomoth, S.; Brzezinski, P.; Iwata,
S., The X-ray crystal structures of wild-type and EQ(I—286) mutant cytochrome c
oxidases from Rhodobacter sphaeroides. J. Mol. Biol. 2002, 321, 329-339.

5. MacMunn, C. A., Researches on myohaematin and the histohaematins.
Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences
1886, 177, 267-298.

6. Keilin, D.; Hartree, E. F., Cytochrome and cytochrome oxidase. Proceedings of
the Royal Society of London. Series B: Biological Sciences 1939, 127, 167-191.

7. Ferguson-Miller, S.; Babcock, G., Heme/Copper Terminal Oxidases. Chemical
Reviews 1996, 96, (7), 2889-2907.

8. Schmidt, B.; McCracken, J .; Ferguson-Miller, S., A discrete water exit pathway in
the membrane protein cytochrome c oxidase. Proc Natl Acad Sci U S A 2003, 100, (26),
15539-42.

9. Tsukihara, T.; Aoyama, H.; Yamashita, E.; Tomizaki, T.; Yamaguchi, H.;
Shinzawa-itoh, K.; Nakashirna, R.; Yaono, R.; Yoshikawa, 8., Structure of Metal Sites of
Oxidized Bovine Heart cytochrome c Oxidase at 2.8 A. Science 1995, 269, 1069-1074.

10. Qin, L.; Hiser, C.; Mulichak, A.; Garavito, R. M.; Ferguson-Miller, 8.,
Identiﬁcation of conserved lipid/detergent-binding sites in a high-resolution structure of
the membrane protein cytochrome c oxidase. Proc Natl Acad Sci U S A 2006, 103, (44),
16117-22.

ll. Hosler, J. P.; Fetter, J.; Tecklenburg, M. M. J.; Espe, M.; Lerma, C.; Ferguson-
Miller, S., Cytochrome am of Rhodobacter sphaeroides as a Model for Mitochondrial
Cytochrome c Oxidase. J. Biol. Chem. 1992, 267, (34), 24264-24272.

12. Hosler, J. P.; Ferguson-Miller, 8.; Mills, D. A., Energy Transduction: Proton
Transfer Through the Respiratory Complexes. Annu Rev Biochem 2006.

52

13. Tsukihara, T.; Aoyama, H.; Yamashita, E.; Tomizaki, T.; Yamaguchi, H.;
Shinzawa-Itoh, K; Nakashima, R.; Yaono, R.; Yoshikawa, S., The Whole Structure of
the 13-Subunit Oxidized Cytochrome c Oxidase at 2.8 A. Science 1996, 272, 1136-1144.

14. Florens, L.; Schmidt, B.; McCracken, J.; Ferguson-Miller, S., Fast deuterium
access to the buried magnesium/manganese site in cytochrome c oxidase. Biochemistry
2001, 40, (25), 7491-7497.

15. Marcus, R. A.; Sutin, N., Electron transfers in chemistry and biology. Biochim.
Biophys. Acta 1985, 811, 265-322.

16. Sirnonson, T., Gaussian ﬂuctuations and linear response in an electron transfer
protein. Proceedings of the National Academy of Sciences of the United States of America
2002, 99, (10), 6544-6549.

17. Gray, H. B.; Malrnstrorn, B. G.; Williams, R. J. P., Copper coordination in blue
proteins. Journal of Biological Inorganic Chemistry 2000, 5, (5), 551-559.

18. Winkler, J. R.; WittungStafshede, P.; Leckner, J.; Malrnstrom, B. G.; Gray, H. B.,
Effects of folding on metalloprotein active sites. Proceedings of the National Academy of
Sciences of the United States of America 1997, 94, (9), 4246-4249.

19. Hopﬁeld, J. J ., Electron-Transfer between Biological Molecules by Thermally
Activated Tunneling. Proceedings of the National Academy of Sciences of the United
States of America 1974, 71, (9), 3640-3644.

20. Mcconnell, H., Intramolecular Charge Transfer in Aromatic Free Radicals.
Journal of Chemical Physics 1961, 35, (2), 508-&.

21. Davis, W. B.; Svec, W. A.; Ratner, M. A.; Wasielewski, M. R., Molecular-wire
behaviour in p-phenylenevinylene oligomers. Nature 1998, 396, (6706), 60-63.

22. Witt, H.; Zickermann, V.; Ludwig, B., Site-directed mutagenesis of cytochrome c
oxidase reveals two acidic residues involved in the binding of cytochrome c. Biochim
Biophys Acta 1995, 1230, (1-2), 74-6.

23. Geren, L. M.; Beasley, J. R.; Fine, B. R.; Saunders, A. J.; Hrbdon, S.; Pielak, G.
J.; Durham, B.; Millett, F., Design of a Ruthenium-Cytochrome c Derivative to Measure
Electron Transfer to the Initial Acceptor in Cytochrome c oxidase. J. Biol. Chem. 1995,

270, 2466 - 2662.
24. Ramirez, B. E.; Mahnstrorn, B. G.; Winkler, J. R.; Gray, H. B., The currents of

life: The terminal electron-transfer complex of respiration. Proc. Natl. Acad. Sci. USA
1995, 92,11949 - 11951.

53

25. Regan, J. J.; Ramirez, B. E.; Winkler, J. R.; Gray, H. B.; Mahnstrom, B. G.,
Pathways for electron tunneling in cytochrome c oxidase. Journal of Bioenergetics and
Biomembranes 1998, 30, (1), 35-39.

26. Garnelin, D. R.; Randall, D. W.; Hay, M. T.; Houser, R. P.; Mulder, T. C.;
Canters, G. W.; de Vries, S.; Tolman, W. B.; Lu, Y.; Solomon, E. I., Spectroscopy of

mixed-valence CuA-type centers: Ligand-ﬁeld control of ground-state properties related
to electron transfer. J. Am. Chem. Soc. 1998, 120, 5246-5263.

27. George, S. D.; Metz, M.; Szilagyi, R. K.; Wang, H. X.; Cramer, S. P.; Lu, Y.;
Tohnan, W. B.; Hedman, B.; Hodgson, K. 0.; Solomon, E. I., A quantitative description
of the ground-state wave function of Cu-A by X-ray absorption spectroscopy:

Comparison to plastocyanin and relevance to electron transfer. Journal of the American
Chemical Society 2001, 123, (24), 5757-5767.

28. Medvedev, D. M.; Daizadeh, 1.; Stuchebrukhov, A. A., Electron transfer tunneling
pathways in bovine heart cytochrome c oxidase. Journal of the American Chemical
Society 2000, 122, (28), 6571-6582.

29. Liu, W. T.; Thorp, H. H., Photoinduced Proton-Transfer between
Transdioxorhenium(V) and Metal-Hydrides - a Convenient Method for Determining
Kinetic Acidities. Abstracts of Papers of the American Chemical Society 1995, 210, 168-
INOR.

30. Cukier, R. I., Mechanism for Proton-Coupled Electron-Transfer Reactions.
Journal of Physical Chemistry 1994, 98, (9), 2377-2381.

31. Cukier, R. I., Proton-Coupled Electron-Transfer through an Asymmetric
Hydrogen-Bended Interface. Journal of Physical Chemistry 1995, 99, (43), 16101-16115.

32. Fang, J. Y.; Hammes-Schiffer, S., Proton-coupled electron transfer reactions in
solution: Molecular dynamics with quantum transitions for model systems. Journal of
Chemical Physics 1997, 106, (20), 8442-8454.

33. Christen, G.; Renger, G., The role of hydrogen bonds for the multiphasic
P680(+center dot) reduction by Y-z in photosystem II with intact oxygen evolution

capacity. Analysis of kinetic H/D isotope exchange effects. Biochemistry 1999, 38, (7),
2068-2077.

34. de Grotthuss, C. J. T., Sur la decomposition de l'eau et des corps qu'elle tient en
dissolution a l'aide de l'electricite galvanique. Annales de Chimie (Paris) 1806, LVIII, 54-
74.

35. Onsager, L., Ion Passages in Lipid Bilayers. Science 1967, 156, (3774), 541-&.

54

36. Nagle, J. F.; Morowitz, H. J., Molecular Mechanisms for Proton Transport in
Membranes. Proceedings of the National Academy of Sciences of the United States of
America 1978, 75, 298-302.

37. Pomes, R.; Roux, B., Molecular mechanism of H+ conduction in the single-ﬁle
water chain of the gramicidin channel. Biophys J 2002, 82, (5), 2304-16.

38. Trurnpower, B., A concerted, alternating sites mechanism of ubiquinol oxidation
by the dirneric cytochrome bcl complex. Biochim. Biophys. Acta 2002, 1555, 166-173.

39. Lanyi, J. K.; Luecke, H., Bacteriorhodopsin. Curr Opin Struct Biol 2001, 11, (4),
415-9.

40. Garczarek, F.; Gerwert, K, Functional waters in intraprotein proton transfer
monitored by FTIR difference spectroscopy. Nature 2006, 439, (7072), 109-112.

41. Hosler, J. P.; Ferguson-Miller, S.; Calhoun, M. W.; Thomas, J. W.; Hill, J.;
Lemieux, L.; Ma, J.; Georgiou, C.; Fetter, J.; Shapleigh, J. P.; Tecklenburg, M. M. J.;
Babcock, G. T.; Gennis, R. B., Insight into the Active-Site Structure and Function of
Cytochrome Oxidase by Analysis of Site-Directed Mutants of Bacterial Cytochrome M3
and Cytochrome bo. J. Bioenerg. Biomembr. 1993, 25, 121-136.

42. Hosler, J. P.; Shapleigh, J. R; Mitchell, D. M.; Kim, Y.; Pressler, M.; Georgiou,
C.; Babcock, G. T.; Alben, J. 0.; Ferguson-Miller, S.; Gennis, R. B., Polar Residues in
Helix VIII of Subunit I of Cytochrome c Oxidase Inﬂuence the Activity and the Structure
of the Active Site. Biochemistry 1996, 35, 10776-10783.

43. Thomas, J. W.; Lemieux, L. J.; Alben, J. 0.; Gennis, R. B., Site-directed
mutagenesis of highly conserved residues in helix VIII of subunit I of the cytochrome bo
ubiquinol oxidase ﬁom Escherichia co 1i: an amphipathic transmembrane helix that may

be important in conveying protons to the binuclear center. Biochemistry 1993, 32, (41),
1 1 173-80.

44. Garcia-Horsman, J. A.; Puustinen, A.; Gennis, R. B.; Wikstrom, M., Proton
transfer in cytochrome bo3 ubiquinol oxidase of Escherichia coli: second-site mutations
in subunit I that restore proton pumping in the mutant Asp135-->Asn. Biochemistry 1995,
34, (13), 4428-33.

45. Fetter, J. R.; Qian, J.; Shapleigh, J.; Thomas, J. W.; Garcia-Horsman, J. A.;
Schmidt, E.; Hosler, J.; Babcock, G. T.; Gennis, R. B.; Ferguson-Miller, 8., Possible
Proton Relay Pathways in Cytochrome c Oxidase. Proc. Nat. Acad. Sci. U. SA. 1995,
92, 1604-1608.

46. Yoshikawa, S.; Shinzawa-Itoh, K.; Nakashima, R.; Yaono, R.; Yamashita, E.;
Inoue, N.; Yao, M.; Fei, M. J.; Libeu, C. P.; Mizushima, T.; Yamaguchi, H.; Tomizaki,

55

T.; Tsukihara, T., Redox-coupled crystal structural changes in bovine heart cytochrome c
oxidase. Science 1998, 280, (5370), 1723-9.

47. Mills, D. A.; Florens, L.; Hiser, C.; Qian, J.; Ferguson-Miller, S., Where is
"outside" in cytochrome c oxidase and how and when do protons get there? Biochim.
Biophys. Acta 2000, 1458, 180-187.

48. Mills, D. A.; Ferguson-Miller, 8., Understanding the mechanism of proton
movement linked to oxygen reduction in cytochrome c oxidase: lessons from other
proteins. Febs Letters 2003, 545, (1), 47-51.

49. Qian, J.; Mills, D. A.; Geren, L.; Wang, K. F.; Hoganson, C. W.; Schmidt, 3.;
Hiser, C.; Babcock, G. T.; Durham, B.; Millett, F.; Ferguson-Miller, 8., Role of the
conserved arginine pair in proton and electron transfer in cytochrome c oxidase.
Biochemistry 2004, 43, (19), 5748-5756.

50. Popovic, D. M.; Stuchebrukhov, A. A., Proton exit channels in bovine
cytochrome c oxidase. Journal of Physical Chemistry B 2005, 109, (5), 1999-2006.

51. Gottschalk, M.; Dencher, N. A.; Halle, B., Microsecond exchange of internal

water molecules in bacteriorhodopsin. Journal of Molecular Biology 2001, 311, (3), 605-
621.

52. Preston, G. M.; Carroll, T. P.; Guggino, W. B.; Agre, P., Appearance of Water
Channels in Xenopus Oocytes Expressing Red-Cell Chip28 Protein. Science 1992, 256,
(5055), 385-387.

53. Fu, D.; Libson, A.; Miercke, L. J.; Weitzman, C.; Nollert, P.; Krucinski, J .;
Stroud, R. M., Structure of a glycerol-conducting channel and the basis for its selectivity.
Science 2000, 290, (5491), 481-6.

54. Sui, H. X.; Han, B. G.; Lee, J. K; Walian, P.; Jap, B. K, Structural basis of
water-speciﬁc transport through the AQPl water channel. Nature 2001, 414, (6866), 872-
878.

55. Walz, T. G. a. T., the structure of aquaporins. quarterly Reviews of Biophysics
2006, 39, (4), 361-396.

56. Loida, P. J.; Sligar, S. G., Molecular Recognition in Cytochrome-P-450 -
Mechanism for the Control of Uncoupling Reactions. Biochemistry 1993, 32, (43),
11530-11538.

57. Oprea, T. I.; Hummer, G.; Garcia, A. E., Identiﬁcation of a functional water
channel in cytochrome P450 enzymes. Proc Natl Acad Sci U S A 1997, 94, (6), 2133-8.

56

58. Tsvetkov, S. A. D. a. Y. D., electron spin echo envelope modulation (ESEEM)
spectroscopy. 1992.

59. Mims, W. B., Envelope Modulation in Spin-Echo Experiments. Physical Review
B 1972, 5, (7), 2409-&.

60. Mims, W. B., Pulsed Endor Experiments. Proceedings of the Royal Society of
London Series a-Mathematical and Physical Sciences 1965, 283, (1395), 452-&.

61. Davies, E. R., New Pulse Endor Technique. Physics Letters A 1974, A 47, (1), 1-
2.

62. Deligiannakis, Y.; Louloudi, M.; Hadjiliadis, N., Electron spin echo envelope
modulation (ESEEM) spectroscopy as a tool to investigate the coordination environment
of metal centers. Coordination Chemistry Reviews 2000, 204, 1-112.

63. Tan, X. L.; Bernardo, M.; Thomann, H.; Scholes, C. P., Pulsed and Continuous
Wave Electron Nuclear Double-Resonance Patterns of Aquo Protons Coordinated in
Frozen Solution to High-Spin Mn2+. Journal of Chemical Physics 1993, 98, (7), 5147-
5157.

64. Tyryshkin, A. M.; Morton, J. J. L.; Ardavan, A.; Lyon, S. A., Davies electron-
nuclear double resonance revisited: Enhanced sensitivity and nuclear spin relaxation.
Journal of Chemical Physics 2006, 124, (23), -.

65. Branden, M.; Sigurdson, H.; Namslauer, A.; Gennis, R.; Adelroth, P.; Brzezinski,
P., On the role of the K-proton transfer pathway in cytochrome c oxidase. Proceedings of
the National Academy of Sciences of the United States of America 2001, 98, 5013-5018.

66. Karpefors, M.; Adelroth, P.; Zhen, Y.; Ferguson-Miller, S.; Brzezinski, P., Proton
uptake controls electron transfer in cytochrome c oxidase. Proc Natl Acad Sci U S A
1998, 95, (23), 13606-11.

67. Pecoraro, C.; Gennis, R. B.; Vygodina, T. V.; Konstantinov, A. A., Role of the K-
channel in the pH-dependence of the reaction of cytochrome c oxidase with hydrogen
peroxide. Biochemistry 2001, 40, (32), 9695-708.

68. Moody, A. J.; Rich, P. R., The Effect of pH on Redox Titrations of Haem a in
Cyanide-Liganded Cytochrome-c Oxidase: Experimental and Modelling Studies.
Biochim. Biophys. Acta 1990, 1015, 205-215.

69. Babcock, G. T.; Callahan, P. M., Redox-Linked Hydrogen Bond Strength

Changes in Cytochrome a: Implications for a Cytochrome Oxidase Proton Pump.
Biochemistry 1983, 22, 2314-2319.

57

70. Capitanio, N.; Capitanio, G.; Demarinis, D. A.; De Nitto, E.; Massari, 8.; Papa,
8., Factors affecting the H+/e- stoichiometry in mitochondrial cytochrome c oxidase:
inﬂuence of the rate of electron ﬂow and transmembrane delta pH. Biochemistry 1996,
35, (33), 10800-6.

71. Capitanio, N.; Capitanio, G.; Boffoli, D.; Papa, S., The proton/electron coupling
ratio at heme a and Cu(A) in bovine heart cytochrome c oxidase. Biochemistry 2000, 39,
(50), 15454-61.

72. Capitanio, N.; Capitanio, G.; Minuto, M.; De Nitto, E.; Palese, L. L.; Nicholls, P.;
Papa, S., Coupling of electron transfer with proton transfer at heme a and Cu(A) (redox
Bohr effects) in cytochrome c oxidase. Studies with the carbon monoxide inhibited
enzyme. Biochemistry 2000, 39, (21), 6373-9.

73. Papa, S.; Capitanio, N.; Villani, G., A cooperative model for protonmotive heme-
copper oxidases. The role of heme a in the proton pump of cytochrome c oxidase. FEBS
Lett 1998, 439, (1-2), 1-8.

58

CHAPTER 2
CW EPR INVESTIGATIONS OF THE HEME A CENTER IN
SEVERAL S44 MUTANTS

59

1.Introduction

Cytochrome c oxidase (CcO) is the terminal electron acceptor in the respiratory
chain in the inner mitochondrial membrane in eukaryotes and in the bacterial cytoplasmic
membrane. CcO has four redox-active metal centers: heme a, heme a3, the dinuclear
copper center (Cut), and a type 11 copper center (CuB) which is magnetically coupled
with heme a 3. Time-resolved spectroscopies have shown that electrons are transferred
from a reduced cytochrome c to CuA1'3, ﬁ‘om Cu, to heme a, then to the heme a3/CuB
binuclear center, where molecular oxygen binds and is reduced to water (Fig.2.1). In
Rhodobacter sphaeroides cytochrome c oxidase (Rs.CcO), His 102 and His 421 are the
two axial ligands for heme a. Although heme a is not involved in the reduction of
molecular oxygen to water, it is proposed that heme a reduction is linked to redox
changes at the binuclear center4 and probably to the movement of protons. Bacterial CcO
has a much simpler structure than its mammalian counterpart, but shows a high degree of
sequence homology and functional equivalency 5'7, making it a good model system for
the more complex mammalian enzyme. Despite the similarities in the bacterial and
mammalian oxidase structures, there are some differences. The Rs. and Paracoccus
denitriﬁcans (Pd) bacterial CcOs have steady-state activities of ~1500s’1 which are
higher than bovine CcO does (~4005'1). A red shift in the visrble spectrum was also
observed in bacterial CcO compared to bovine CcOs. In addition, the redox potential of
bovine heme a is measured to be more positive, by ~ +13 mV, relative to the Rs. heme a,

assuming similar redox potentials of CuA9.

60

 

Figure 2.1: The Rs. CcO crystal structure (PDB ID: 1M56). Subunit I (gray) with the
two hemes in light green, Ca and Mg (green spheres) and key residues (stick) of S44,
H93, H102 (a ligand of heme a), D132 at the entrance of the proton uptake path and helix
I (bright yellow) that comprises the residue S44. Subunit H (deep red) is shown with the
dinuclear CuA (orange spheres). The ﬁgure is the courtesy ﬁgure from Dr.Mills

The g values of the EPR spectrum of heme a in bacterial oxidases are also shifted
relative to bovine oxidase, particularly, the gZ peak is 2.82 in Rs. CcO compared to 3.03 in
bovineg. It was proposed that the shifted EPR g values in Rs.CcO were due to an
increased anionic character of one or both of the heme a histidine axial ligands, possibly
due to altered hydrogen bonding of the histidines to nearby residues“). Examination of the
crystal structure of bovine CcO5 revealed that the peptide backbone oxygen of a
conserved glycine in subunit 1 is hydrogen bonded to a histidine ligand of heme a in all
mammalian aa3-type oxidases and in many other species including yeast, wheat and
T hermus thermophilus ba3 oxidaseg‘ 1‘. However, in the bacterial aa3 oxidases, Pd. and
Rs., the glycine residue has been replaced with a serine (Fig. 2.2) with the hydroxyl
group of the serine (S44) hydrogen bonded to the histidine (H102 in Rs. CcO). The other
histidine ligand of heme a (H421) is surrounded by highly conserved
hydrophobic/aromatic amino acids which coincide in an overlay of the Rs and bovine
structures; thus, it is unlikely to be the source of the altered heme a properties between

the mammalian and bacterial oxidases.

To test the possible effect of hydrogen bonding to the histidine ligand of heme a,
the serine 44 residue in Rs.CcO was mutated to a glycine to resemble the bovine enzyme,
creating a mutant S44G, and the serine was also replaced by an aspartic acid, an
glutamate, and a glutamine, creating mutants S44D, S44E, and S44N respectively. With
the serine to glycine mutation, a shift of the EPR spectrum towards bovine was observed
as predicted; from serine to aspartic acid, and glutamine, the prediction of a shift of their

EPR spectra away from bovine was also met, but to a different extent. Further electron

62

 

Bovine CcO Rs CcO

I’H102 H421
. ,3:

S44

 

 

 

 

 

 

Figure 2.2: The ligands to heme a are shown for bovine (1V54 structure) and Rs
(1M56) C60. The residues that are closest to the heme a ligands are shown in stick
structure. Bovine CcO shows a glycine (G30) with the carbonyl backbone oxygen
hydrogen-bonding to His61. Rs CcO shows a serine (S44) with its hydroxyl hydrogen-
bonding to His 102. The ﬁgure is the courtesy ﬁgure from Dr.Mills

63

transfer and fast kinetics measurements showed that the S44G mutant has a wild-type—
like kinetic behavior while the electron transfer kinetics of the S44D mutant is
signiﬁcantly altered, leading to unique insights into the control of heme a redox behavior,

and a potential tool for understanding the electron and proton movements.

2. Experimental Methods

2.1 Protein expression and puriﬁcation

All the mutants were created by Dr.Hiser as described in”. The CcO protein was
puriﬁed by NiZ+-NTA affinity chromatography” and further puriﬁed by two tandem
DEAE ion-exchange column (Tosohaas DEAE-SPW 10p particle size, 8 mm x 7.5 cm)
using a Pharmacia AKTA FPLC as previously described13 to remove excess subunit I and
the remaining contaminants. The enzyme was then concentrated and washed into the
desired buffer using a Millipore centrifugal ﬁlter with a MW cut-off 100K The amount
of Mn2+ incorporated into the protein depends on the [Mg] to [Mn] ratio in the grth
medium"). The YZ-100 strain of R. sphaeroides, which over-expresses the wild-type CcO
with a histidine-tag added to the C-terminus of subunit I, was grown on “high Mn”
medium (with MnSO.t and MgSO4 at ﬁnal concentrations of 700pM and SOpM,
respectively). The substitution with Mn2+ was necessary to make the native Mg2+ center
into an EPR visible site. The visible spectral characteristics of Mg- and Mn-substituted

wild-type enzymes were identical and oxygen consumption activities were not altered.

2.2 Reconstitution. of cytochrome c oxidase into phospholipid vesicles (COVS)

64

All glassware was rinsed with ethanol. Asolectin was suspended to 40mg/ml by
sonication in 2% cholate, 75 mM HEPES-KOH, pH 7.4, at 0 °C under argon. A Heat
Systems-Ultrasonics sonicator (Model W-225), equipped with a microtip, was used at a
power setting of 5 for intervals of 30 S on, 30 3 off, until clarity was reached. The
suspension was centrifuged for 15 min at 12,000 X g to remove titanium particles.
Puriﬁed cytochrome c oxidase in 0.2% lauryl maltoside was preincubated with cholate
detergent for 6.7mg per nmol enzyme for two hours before adding to the phospholipid
suspension to a ﬁnal concentration of 2 nmol/ml and then dialyzed with rapid stirring at 4
°C in Spectrapor dialysis tubing (number 25225/204, 12-14,000 M, cutoff) by the
following protocol: 6 h in 100 volumes of 75 mM HEPES-KOH, pH 7.4, 14 mM KCl,
0.1% cholate; 12 h in 100 volumes of 75 mM HEPES-KOH, pH 7.4,14 mM KCl; 12 h in
100 volumes of 50 mM HEPES-KOH, pH 7.4, 24 mM KC1,15 mM sucrose; 12 h in 500

volumes of] mM HEPES-KOH, pH 7.4, 44.6 mM KCl, 43.4 mM sucrose” ‘5.

2.3 Activity of COVs under steady-state and stopped-ﬂow proton pumping measurements
CcO was ﬁrst reconstituted into asolectin lipid vesicles (COVS). Measurements of
proton pumping were made on an OLIS-rsm stopped-ﬂow with 100 1.1M phenol red, 50
11M HEPES-KOH pH 7.4, 45 mM KCI on the outside of the COVs, and 75 mM HEPES-
KOH pH7.4 on the inside. Activities were obtained from averaging at least three
measurements of cytochrome c oxidation from scanning a range of 155 mm (x mm -y nm)
using double monochromators with a 500 blaze wavelength and a groove density of 600
lines. The rapid seaming of the wavelength range was at the rate of 1 scan/msec. The

OLIS-rsm SVD Global Works software package was used to analyse the

65

spectra.Additionally, exponential ﬁtting of the cytochrome c changes monitored at 550
nm using Microcal OriginTM, in the absence of phenol red, was performed. The activities
of the COVs were measured in the absence of ionophores, forming both a membrane
potential and a pH gradient during turnover (controlled state). The addition of
valinomycin to COVs removes the NP (membrane potential), and this is the condition in
which proton pumping is normally observed as a net acidiﬁcation (decrease in
absorbance) of the phenol red on the outside. The addition of FCCP removes the pH
gradient (uncontrolled state) by allowing protons to equihbrate across the membrane,
resulting in alkalinization on the outside (increase in absorbance), due to the net

consumption of protons for the formation of H20 at the active site.

2.4 UV-visible Spectroscopy

UV-visrble spectra of fully reduced wild-type, S44G, S44D, S44N, and S44E
proteins by addition of dithionite were recorded on a Perkin-Ehner Lamda 40P
spectrophotometer scanning wavelengths from 400nm to 750nm after appropriate

dilution. into pH7.4 100mM HEPES, 0.1% lauryl maltoside.

2.5 Electron Paramagnetic Resonance Spectroscopy (EPR)

Continuous-wave X-band EPR spectra of puriﬁed Rs.CcO WT, S44G, S44D,
S44E, and S44N were measured on a Bruker ESP300 equipped with a TE102 cavity
resonator. Temperature was maintained at 4.5K using an Oxford ESR900 helium cryostat
assembly. The g values were determined by direct measurements of the magnetic ﬁeld

strength and microwave frequency. Spectrometer conditions: microwave frequency

66

9.458GHz; microwave power SOpW; modulation amplitude was 20.0Gauss. Other

conditions are described in ﬁgure legends.

Electron spin echo-detected EPR and three-pulse ESEEM spectra were recorded on
a Bruker E-680X spectrometer operating at X-band and equipped with a model ER4118-
MD-X-S-Wl probe that employs a 5mm dielectric resonator. The temperature was
maintained at 4.5 K using an Oxford liquid helium ﬂow system equipped with a CF—935
cryostat and an ITC-503 temperature controller. ESEEM data were collected using a
three-pulse, stimulated echo sequence (90°-1t-90°-T-90°) with 90° microwave pulse
widths of 16ns. A four-step phase cycling sequence, together with the appropriate
addition and subtraction of the integrated spin-echo intensities served to remove the

unwanted echoes and baseline offsets from the data.

3. Results

3.1 Optical spectra and activity

UV-visible spectra of two mutants S44G and S44D in Rs. CcO are similar to that of
wild type enzyme with a Soret peak at 445nm and or peak at 606 nm, suggesting that
there are no major changes in the environment of the hemes (Fig.2.3). For S44G, steady-
sate studies (Mills, DA) Show that the maximal activity across the pH range from pH 6 to
9 is not altered, thus the mutation does not change the rate-limiting step in the turnover.
Good proton pumping with a usual H+/e' ratio of ~1 from reconstituted COV (made by

Dr.Mills, see methods) also indicates that S44G behaves just like wild type enzyme. For

67

 

 

Absorbance

 

 

 

 

 

0.0 ' T ' l 1 l ' l ' l '
450 500 550 600 650

Wavelength (nm)

Figure 2.3: The visible spectra of reduced wild type, S44D, and S44G CcOs. The
ﬁgure is the courtesy ﬁgure from Dr.Mills

68

the S44D mutant, the activity is reduced to less than 30% of the wild-type activity at pH
6.5, under conditions of maximal steady-state activity (data not Shown). S44D
reconstituted into COVs also Showed a lowered respiratory control ratio (RCR:
uncontrolled rate/controlled rate). The controlled rate is normally inhibited by both a
A‘I’ (membrane potential) and ApH (pH gradient) and is usually the slowest rate. The
proton pumping stoichiometry (H+/e) for S44D is signiﬁcantly lower than wild-type and

S44G (Table 1).

S44D has the same pH dependence of its steady-state activity as the wild-type does.
However, in the stopped-ﬂow analysis, S44D shows a different pH proﬁle with an
apparent pKa less than 6.0 (Fig. 2.4), due to different rate-limiting steps in the two assay

systems.

3.2 X-band cw EPR

Mutants of serine 44 in Rs. CcO were created to test the proposal that the position 44
may be important in determining the differences in heme a EPR spectra. In Fig.2.5, the
low-spin heme a spectra for wild-type Rs. CcO and mutants S44G and S44D are shown.
The wild-type g values are 2.81, 2.31, and1.63, while bovine oxidase has different g
values of 3.03, 2.21, and 1.45 for low spin heme a (green arrows), the EPR spectral
features of Cu in bovine CcO are only slightly different from those of the Rs.CcO (gZ

=2.18 and g,,,=2.00).

69

 

 

 

 

 

 

 

 

Activity ( e /sec/aa3)

COVs pH Cont. Uncon. RCR

6.5 130312 950124 7.3
WT 7.4 303:1 1050:23 13

6.5 9031 297115 3.3
5441) 7.4 4711 185:1:5 3.9

6.5
S44G

7-4 102112 1020153 10

 

 

 

 

 

 

 

Table 1: The steady-state activity of S446, S44D and wild-type at different pH. The
steady-state activity measurements were made in a Gilson oxygraph with 30pM horse
cytochrome c, 3mM ascorbate, lmM TMPD, 2 mg of asolectin lipids, and 0.05% lauryl
maltoside. Contr.= controlled proton-pumping rate measured in the absence of any
ionophores. Uncon.= uncontrolled proton-pumping rate measured in the presence of
uncoupler (211M valinomycin + 511M CCCP). RCR= respiratory control ration, is the ratio
of the uncontrolled rate over the controlled rate.

70

 

 

 

 

 

 

 

20001 g: 400
o Wild-type
E 1600- (a -80
6; S44G @ 60
.5 1200. -
5
° 8800. -40
‘7’ g 3440
3 400. g -20
o
q) \\
$30ng a at
v q q I ‘ ‘
a E (”K
3.- . .80
a. g 6004 XWild-type ¢\D\q Wild-type
g 2 . j - -60
o. 400 ‘
Q 0 O .40
o \
’5 zoo 330%. . .2.
0' 0 “=1 $440 9 o
e . 0 0 0e . 9 e
s 7 s 9 6 7 a 9
pH pH

Figure 2.4: The steady-state activity assay of wild type, S44G, and S44D CcOs. The
top panel shows the steady-state activity of S44G, S44D and wild type at different pH.
The left panel shows the activity (e'/sec/aa3) and the right panel shows the percentage
activity. The bottom panel shows the activity measured under stopped-ﬂow conditions
where the rate of electron transfer is rate limiting. The ﬁgure is the courtesy ﬁgure from

Dr.Mills.

71

Percentage Activity

The mutation of S44 in Rs. CcO to a glycine residue to resemble bovine did produce
a shifted EPR spectrum of heme a with g values of 2.86, 2.30, and 1.60. AH of these
values are shifted towards those of bovine CcO, though the shifts are not as dramatic as
would be expected if the residue were the only contributor to the difference between Rs.

and bovine CcO.

The S44D mutant, however, shows spectral characteristics distinct from those of
wild-type and S44G. In the gz region, two peaks for gz = 2.72/2.78, and gy and gx values
of 2.30 and 1.74, are considerably shifted in the opposite direction to those of bovine and
of the S44G. When the EPR spectra are measured at different pH values ranging from 5.5
to 8.0, a big shift in the relative intensities of the peaks at 2.72/2.78 is observed: peak
2.72 had the highest intensity at pH 8.0 while peak at 2.78 had increased intensity at
pH5.5, suggesting that the EPR spectrum is inﬂuenced by a protonation event (Fig.2.6).
Spin quantiﬁcation by double integration of two peaks shows that the total area under the
peaks is comparable in magnitude at different pH values, indicating an interconversion
between them. The results agree well with the proposal that 2.72 and 2.78 respectively
represent the deprotonated and the protonated form of Asp 44 (Fig. 2.6). When the areas
under the two peaks are plotted as a function of pH (Fig. 2.7), the pKa of the
interconversion is estimated to be less than 5.5, the lowest pH that can be measured
without precipitation. This value is also consistent with the results of kinetic

measurements of the intrinsic electron transfer rate from. CuA to heme a9.

72

 

 

 

 

 

 

 

 

_‘ Cu Ag:
2.18
‘ 2.31
2.82 Cu x,
.. , 2.35 I‘ 2'72 _ .r 2.01
:- j . J. ‘ ‘1‘ .. ' ”A W
J (”I ‘ heme a 92 ' 1.74 1.631.60
. V ' I heme a g
‘ Bovine CcOI 3.03 2.21 1-45
': heme a gy
‘. ——Wild-type CcO
- — $440
a
I ﬁ I ' T ' l ' l ' l ' I F |
1 000 1 500 2000 2500 3000 3500 4000 4500

Magnetic field (Gauss)

Figure 2.5: Overview of the EPR spectral changes with wild-type Rs CcO (black),
8446 (red) and S44D (blue). Heme a g values of Rs. wild type, S44G and S44D are
labeled on top of each corresponding peak, and bovine g values are listed beneath in
green. Samples were made in buffer with 20mM HEPES-KOH, pH 8.0,14mM KC], and
0.1%lauryl maltoside with the ﬁnal concentrations of 8211M, 88uM, and 61uM
respectively. The EPR experimental conditions are: microwave ﬁ'equency: 9.458GHz;
microwave power SOuW; modulation frequency: IOOkHz; modulation amplitude, 20.0G;
conversion time: 327msec; temperature: 4.2K.

73

 

. — pH5.5
2.72 —pH6.5

— pH7.4
, --—— pH8.0

 

 

 

 

r

2300 ' 24'00 ' 25'00
Magnetic Field (Gauss)

2600

Figure 2.6: Effect of pH on heme a gz EPR spectra of S44D CcO. pH 5.5 and 6.5
buffer contained 50mM Mes-KOH, and 0.1%lauryl maltoside. pH 7.4 and 8.0 buffers
contained 50mM Hepes-KOH, and 0.1%1auryl maltoside. Ionic strength was controlled
by adjusting the conductivity using small amount of NaCl. pH 5.5, black; pH 6.5, red; pH
7.4, blue; pH8.0, cyan. EPR conditions are the same as described in ﬁgure 2.5.

74

     

 

 

14000 a
. + 2.78 peak
+
12000 A 2.72 peak
10000 -i
o
ID 8000 -‘
3
g 'I
'5 6000 4
E .
<
4000 ..
2000 -
0 —1
ﬂ ' T ' l ' I ' I ' I
5.5 6.0 6.5 7.0 7.5 8.0
buffer pH

Figure 2.7: Plot of amplitude of 2.72/2.78 double peak for heme a versus buffer pH
for S44D. Amplitudes were obtained by double integration of areas at two peaks
respectively at different pH values. The sample and spectrometer conditions are the same
as depicted in ﬁgure 2.5.

75

EPR spectra of samples at different ionic strength were also studied. Higher ionic
strength had a similar effect on S44D as lowering pH, making the peak at 2.78 somewhat
more intense, although not as dramatic an effect as when the pH is changed. Cytochrome
c binding also caused a small Shift in the relative intensities of the 2.78/2.72; in this case,
lowering the intensity of 2.78 peak upon cytochrome c binding (Fig. 2.8). These results
were also consistent with changes observed in kinetic measurements of the intrinsic

12
electron transfer rates .

As a control for the effect of charge/protonation state in the S44D mutation, the S44
was mutated to a neutral residue, asparagine (S44N). The resultant EPR spectrum shifted
in the same direction as S44D, but only had a single peak at gZ = 2.80, close to that of the
apparent protonated form of S44D (Fig. 2.9). A further interesting mutant, S44E, also
Showed a shifted EPR peaks at 2.79, 2.29, and 1.66. To get a better understanding of the
changes in EPR characteristics and electron transfer kinetics by these mutations, more

EPR and kinetic measurements need to be done.

In the EPR spectrum of the S44D mutant of Rs.CcO, a relatively intense signal
appears at g2=5.87. This “g = 6” signal is normally attributed to high spin heme a3. In
wild-type CcO, due to antiferromagnetic coupling between high spin heme a3 and CuB,
very little EPR signal is usually visible in this region. From case to case, if the binuclear
center is not perfectly coupled, a weak signal from high-spin heme a 3 is observed in this

region in bacterial CcO as well as bovine enzyme. The presence of the larger signal in the

76

—— S44D in low salt +cyt.c

. — S44D in high salt +cyt.c 278
-——- 8440 in low salt 2-72
——S44D in high salt

    

 

 

 

 

l ' T ' 1
2000 2200 2400 2600
Magnetic Field (Gauss)

Figure 2.8: Effect of ionic strength and cytochrome c binding on heme a gz EPR
spectra of S44D. Samples were made in buffers with 50mM Hepes-KOH, pH 7.4, and
0.1%lauryl maltoside for low salt and 50mM Hepes-KOH, pH 7.4, 150mM NaCl, and
0.1%lauryl maltoside for high salt. Cytochrome c and protein have same ﬁnal
concentrations. EPR conditions are the same as described in ﬁgure 2.5.

77

— WT COO

J --S44G

2.82 __ S44D
—— £441:

— S44N

 

l
l
t‘

 

1:73;“
I T I ' I ' I
2300 2400 2500 2600
Magnetic Field (Gauss)

Figure 2.9: EPR spectra overlay of WT, S44G, S44D, S44E, and S44N in the gz
region. Wild type CcO: black; S44G, red; S44D, blue; and S44E, cyan; S44N, dark gray.
EPR conditions are the same as described in ﬁgure 2.5.

78

S44D mutant suggests a decoupling that could be the result of some reduction of Cut; due

to the very low redox potential of heme a 9.

Above the heme a and heme a 3/CuB sites, there is one non-redox-active Mg metal
site. It is centrally located at the interface of subunits I and II (Fig.2.1). Mg can be
replaced by an EPR-visible metal, Mn, without any effect on enzyme ﬁrnction, and can
be used as a probe to track conformational changes in the region”). EPR data indicated
that the Mn spectrum in S44D is essentially the same as that of the wild-type enzyme.
Unlike heme a, the Mn site is unaffected by buffer pH (Fig.2.10), ionic strength, and
cytochrome c binding (data not shown), indicating that no general conformational change

is involved in the EPR spectral changes observed in heme a.

As shown in Fig.2.11, a new EPR feature at g=2.01 region appeared in mutant
S44N. This same signal was also observed in S44D samples from one preparation. The
appearance of this signal was accompanied by a high spin heme a 3 signal in the g=6.0
region of the spectrum Double integration revealed that the amount of EPR-visible CuA
did not change very much (less than 30%). The integration error was estimated to be less
than 20%; this error arises from the protein concentration estimate, the baseline
correction routine, and the fact that there are small heme signals in the g=2 region at such
a low temperature. Since a lowered redox potential of heme a in the deprotonated
oxidized form of S44D was indicated by the rapid kinetic analysis of Cu», to heme a
electron transfer, it is possible that a new electron equilibrium caused this new feature.

Various experiments were performed to determine where this feature came from. There

79

 

 

 

 

— pH8.0
—— pH7.4
—— pH6.5
— pH5.5

 

T

I r I
2800 3000

Figure 2.10: The pH profile of Mn EPR spectra of S44D. The buffer conditions and
EPR conditions are the same as described in ﬁgure2.5

80

3200 ' 3400
Magnetic Field (Gauss)

U I I l U U
3600 3800 4000

 

1 W

 

 

 

 

 

' l ' I W l ' I I ' l ' f
1000 1500 2000 2500 3000 3500 4000 4500
Magnetic Field (Gauss)

T

Figure 2.11: The cw EPR spectrum of mutant S44N. The red * indicates the new
feature at g=2.0 region. The buffer conditions and EPR conditions are the same as
described in Figure2.5

81

are many possibilities in terms of its origin. This signal could be from a disturbed CuA, or
from a decoupled CuB, or even ﬁ'om a relatively stable radical. In power saturation
experiments, microwave power from 0.199pW to 12.6mW was used, and the power
saturation pattern for this new EPR feature was different ﬁom those of the heme a, and
Curt (data not shown). The temperature dependence of its amplitudes was also studied
(Fig.2.12). The disappearance of the signal at 110 K ruled out the possibility that the
signal came from a radical since the species survived the puriﬁcation procedure at room
temperature. The heme a 3 signal at “g=6” region indicated that the heme a 3/Cu3 binuclear
center might be decoupled to some degree upon mutation. As a result of this decoupling,
the normally invisible CuB center may be observed in g=2.00 region, explaining the new
sharp peak in the CcO spectrum. To test this speculation, echo-detected EPR absorption
and three-pulse ESEEM experiments were carried out. The spectra at different shot
repetition times (srt) had no signiﬁcant difference. The three-pulse ESEEM spectrum of
S44D was essentially the same as that of the wild-type CcO partly because the resolution

is not high enough to recognize the difference if any.

4. Discussion

4.1 Spectroscopic Measurements

In the S44 mutants of Rs. CcO, there are distinctive alterations in the paramagnetic
spectra. The mutation of S44G shifts the heme a peak from gZ = 2.82 to gZ = 2.86, in the
direction of the heme a EPR signal of bovine oxidase which has a peak at g = 3.03.

Additionally, the S44G mutation causes a shift in the low spin heme a gy from 1.63 to

82

 

 

 

I ' l ' I ' I ' I ' I ' I ' I
1000 1500 2000 2500 3000 3500 4000 4500
Magnetic Field (Gauss)

Figure 2.12: The temperature-dependence EPR spectra of S44D. The buffer
conditions and EPR conditions are the same as described in ﬁgure2.5.

83

1.60, g)( ﬁom 2.31 to 2.30. These changes are not reﬂected in the visible spectrum of the

S44G, which appears identical to wild type.

Despite the shift of the heme a gZ signal in S44G to a value closer to that of the
bovine CcO, the shift by no means completely explains the major difference between the
EPR spectra of Rs and bovine CcO. The remaining difference could be due to other
residues that surround the histidine ligand on the same side as the glycine, which are not
hydrogen bonded to the histidine. These less conserved residues are mostly hydrophobic
and aromatic, and could affect the electronic conﬁguration, augmenting the difference

between the serine and the glycine hydrogen bondsg.

For many years, chemists and bioinorganic chemists have been studying low spin
heme complexes16 and hemeproteinsn'zo. Their ﬁndings provided valuable information
for understanding spin state, coordination number, the identity of axial ligands, and
spectroscopic and magnetic properties of hemes. Among these researchers, Blumberg and
Peisach21 developed a method in which three g values obtained from EPR spectroscopic
methods were reduced into two crystal ﬁeld parameters (“tetragonalitYﬁ AD»,
representative of the electron density at the iron, and “rhombicitf’, V/A, a pure geometric
term), and then plotted as various sets individual points. Depending on the nature of the
sixth ligand of heme proteins, ﬁve regions are arbitrarily drawn where parameters for the
ﬁve different compounds may be expected to reside. The C group consists of low-spin
minority species having methionine as the sixth ligand. The B and H groups consist of

complexes of heme in which the sixth ligand is a histidine. The difference between

84

compounds lying in B and H groups lies in a difference in proton content. The irnidazole
of histidine in B type has a proton bound to it, while in the H complexes this proton is
absent. The 0 region represents hydroxide forms of the or and B chains of hemes. The last
group, P, represents a series of low-spin heme compounds in which the ligand of the
heme is a alkyl mercaptide. The members of this group have similar EPR features to the
cytochrome P450 family. With the help of “truth diagram”, one can predict the electron
density at the heme iron in target heme protein and further relate this to speciﬁc

properties of heme and its axial ligands.

The tetragonality and the rhombicity parameters of bovine wild-type CcO, R.s wild
type, and S44 mutant CcOs were calculated using Taylor’s formalism”. It is interesting
to observe that, except for the parameters of bovine CcO, the parameters of all the others
fall into the H region, a region that contains hemes known to be ligated by an anionic
histidine ligand. Points corresponding to S44G, S44N, and S44D are located in order
from left to right of the H region, indicating an increase of charge density on the anionic
histidine (Fig.2.13). For the model in which heme a is ligated by two neutral histidines,
the tetragonality (A0.) is 2.88; when the model is ligated by two deprotonated histidines,
the tetragonality (NA) is 4.08. All of the calculated tetragonality values are in the middle
of these two extremes. With a N2 of 3.77, the apparently deprotonated form of S44D
leans toward the extreme where two histidines are mostly deprotonated. Wild type bovine
CcO (A/lt=3.15), on the other hand, is toward the other extreme with two neutral

histidines; while both the Rs. wild type, and S44G CcOs have higher tetragonality values

85

 

 

12 t— 0 HEMOGLOBINA
11 a ALPHA CHAINS
' ”\ /,_\ /,/"'~\a BETACHAINS ﬂ“
1.0 "" I I, } I ‘ r”’ ‘
l ‘ I, I I, I I’I” ‘
.9 - I t, I/ , ‘
’ Ir . ’r 2 N3 / / ° 5..
A I [I I, . / I ‘ ‘
s .3 l— 1 . u . 7; I ’ A A /
a. i‘ ’1’ ° 0" " I" ‘ x”
E '7 Litmus '1' It ' /’ ‘v' H” , x”
,2 _. RHOMBIC \ /\ I 9’ /’ cf..-
.6 / t 4 s
D V ‘3’ ' x t I e ‘
c x \
g -5 "' . Rs wr CcO \‘tu/ ti“ ’ ,” as” \
.r: ' H . ‘ \
o: ,4 l- . Rs. S446 \ e 0 1|
3 .— ‘ RS. S44N P\\ ¢.. . i
' v Rs. 3445 \ 05131-1],
.2 r—- ‘ RS. 84400-1) \‘.._....—’
> Rs. S44D(-)
'1 "PURE 9 Bovine WT CcO
Axm J A 1 A A
o 1 2

 

Figure 2.13: Crystal field parameters of bovine wild type CcO, Rs. wild type 0:0,
and S44 mutants in the frame of the Blumberg-Peisach truth diagram. C region
represent the heme species with a methionine as the sixth ligand, B region represent the
heme species with a neutral histidine as the sixth ligand, H region represent the heme
species with a anionic histidine as the sixth ligand, 0 region region represent the heme
species with hydroxide as the sixth ligand, and P region represent the heme species with
an alkyl mercaptide as the sixth ligand. Rs.S44D (H) refers to protonated form of S44D
where the calculations of tetragonal ﬁeld and rhombicity are performed using g values of
2.78, 2.31, and 1.68. Rs. S44D (-) refers to deprotonated form of S44D where the
calculations of tetragonal ﬁeld and rhombicity are performed using g values of 2.72, 2.31,

and 1.74.

3 4 s
Tetragonal Field (Am

86

 

(NAT-"3.33, 3.34 respectively), which means that there is an increase in electron donation

by the axial histidines (Table 2).

Unlike the shifts in the heme a EPR Spectrum of S44G, the shifts in S44D are in the
opposite direction, moving further away ﬁom bovine CcO. EPR spectra of S44D at
different pHs are consistent with data from kinetic analysis and electron transfer studies.
All these results indicate that the carboxyl group of Asp at position 44 could exist as two
differently protonated forms. At ﬁrst glance, the UV-visible spectrum of the S44D mutant
is identical to the wild-type enzyme. Close examination, however, reveals subtle changes
in the apparent extinction coefﬁcient at 606nm, a lowering by 20% to 30% compared to
the extinction coefﬁcient at 444nm. This change also shows some ionic strength- and pH-
dependence, with the lowest extinction coefﬁcient at higher pH. This data again suggests
that Asp 44 could associate with the histidine ligand (H102) of the heme a in two forms,
depending on the environmental pH conditions. The altered kinetics of electron transfer
can also be explained by the protonation of Asp at low pH, which would reduce the

negative charge in the region of the heme a ligand, H102.

Close examination of the local geometry of heme a and the nearby residue which is
expected to form a hydrogen-bonding with the axial histidine ligand of heme a in crystal
structures of bovine wild-type, Rs. wild type, S44D(H) and simulated structures of Rs.
S44N, S44E, S44G revealed a good correlation between hydrogen-bonding distance and
heme a gz value. Of these six different variants of CcO, bovine has the longest distance

of 3.22A and highest gz=3.03, followed by Rs. S44G with a distance of 3.03A and

87

92 9y
Rs. wr 2.82 2.31 1.63 2.26 4.46 3.34 0.68
Rs. S44G 2.86 2.3 1.6 2.14 4.4 3.33 0.64
Rs.S44N 2.8 2.31 1.65 2.31 4.58 3.42 0.68
Rs.S44E 2.79 2.29 1.66 2.54 4.94 3.67, 0.69
Rs. S44D(H) 2.78 2.31 1.68 2.56 5.22 3.53 0.73,
Rs. S44D(-) 2.72 2.31 1.74 2.7 5.12 3.77 0.72
BovineWT 3.03 2.21 1.45 1.68 3.99 3.15 0.53

gx VIA=A B M. =B-A/2 VIA

Table 2: EPR parameters for heme a in cytochrome c oxidase. The A and B can be

 

 

 

A: g, +g, B: g, +g,
gy+gz gz‘gr’ gy+gz gy-gx

calculated as follows:

88

gz=2.86, the S44D has the shortest distance of 2.71A and gz=2.78, and all other mutants
have distances between 3.22A and 2.71A and g values between 3.03 and 2.78 (Fig. 2.14).
The trend in distance change within all of six variants reﬂects the change in strength of
hydrogen-bonding between axial histidine ligand and its partner, which is consistent with
the result of Peisach-blumberg truth diagram analysis. Axial histidine in bovine wild type
CcO has the least anionic character because of the weakest hydrogen bond between
His6l and G30. Mutation of Rs. wild-type CcO to S44G made it more like bovine,
resulting in a longer distance between HileZ and S44G and decreased anionic character
on HileZ. Further mutations to S44N, S44E, and S44D resulted in shortened distances
and increased anionic character on Hi8102. It is interesting to observe such a high degree
of correlation and consistency within various methods of analysis, which suggests that
EPR could be a powerﬁil and independent method to investigate hydrogen-bonding
properties of histidine of heme a in CcO. Without some crystal structures available, EPR
data helped us visualize changes caused by mutations on heme a local hydrogen-bonding
with good conﬁdence. As EPR data and modeling indicated, a glycine residue could form
a hydrogen bond in Rs. CcO with a slight movement of the peptide backbone, which was
observed in crystal structure of Rs. S44D: similarly asparagine and glutamate seem to

manage to accommodate themselves to the position in a manner Similar to aspartic acid.

The origin of the new signal at g=2.01 in S44N and S44D is difﬁcult to interpret.
Previously, the observation of 3 Cut; signal in the EPR spectrum was reported in CcO
when heme a 3 was decoupled, in which EPR signals did not look like the observed

signal”. Because CuB has different ligation environment than CuA does, CuB should have

89

 

Bovine CcO

 

Rs CcO

2.8:Am

I

[y/ 102
$44

H421 '1

Rs 8440

 

 

E44

 

 

3.22 A .
2’ H61 H378
630
Rs 8445 Rs S44N
H102 H421
2.72A.-' : ‘S "

 

H102

‘70
d 2.78

   
  

 
  
 

044
Rs 8446
"‘21 H102
S /o‘3.03
e44

 

 

H421

H421

 

Figure 2.14: The ligands to heme a in crystal structures and simulated structures of
CcO. The residues that are closest to the heme a ligands are shown in stick structure. Top
ﬁgures are from crystal structures of bovine (1V54 structure) Rs wild-type CcO (1M56),
and S44D (unpublished data). Bottom ﬁgures are simulated ﬁgures generated in Amber
program using the crystal structure of Rs. wild type CcO as template.
program, local geometry minimization is employed to mimic the most possible

orientation of the mutated residue at 44.

90

In simulation

 

different ESEEM data modulation pattern. However, three-pulse ESEEM data and echo-
detected EPR absorption did not show signiﬁcant alteration in modulation pattern to
support the speculation that this new signal was ﬁom Guys. The formation of a radical is
another possible explanation for the new signal in the g=2.0 region. However, the
temperature dependence tests showed that the peak disappeared, along with other signals,
when the temperature was increased to 110 K, which ruled out the possibility that it came
from a radical considering the fact that this species survived the puriﬁcation procedure at
room temperature. The fact that there are so many overlapping signals in this region
makes it a formidable task to ﬁnd out the origin of this new feature even with higher-

resolution techniques such as ESEEM or ENDOR.

Crystal structures of both the oxidized and reduced forms of the S44D mutant in
(unpublished data, Qin,) at 2.8K resolution show that the carboxyl group in D44 is
disordered in the oxidized form, but is well resolved in the reduced state with one oxygen
atom 2.7 A from the HileZ nitrogen. This ﬁnding is consistent with the results of kinetic
analysis and the EPR spectra at different pH values, suggesting that the reduced state of
heme a is stabilized by the protonation of S44D, and the protonated form has a strong

hydrogen bond with the histidine nitrogen.

4.2 Effect of mutation on electron transfer and redox potential

For mutant S44G, little change of redox potential in heme a is observed, which is

probably due to subtle changes in the region around heme a. These changes in structure

91

are not sufﬁcient to cause any signiﬁcant change in its electron and proton transfer

behaviors in steady—state, stopped—ﬂow, and photo-induced rapid kinetic measurements”.

The mutation of S44 to an aspartate, on the other hand, caused a dramatic change in
the kinetics of electron transfer from Cup, to heme a. The electron transfer rate is very
sensitive to pH change: at higher pH, the electron transfer is greatly slowed by a factor of
1000; at lower pH, the intrinsic electron transfer rate is biphasic, with a fast rate, followed
by a slower rate similar to that at high pH. These observations, along with pH-dependent
EPR changes, are most easily rationalized as being due to the ionization state of the
aspartate carboxyl group. At higher pH, D44 is likely in its deprotonated state, resulting
in a large decrease in the redox potential of heme a due to the proximity of the negative
charge, and a slowed electron transfer ﬁom CuA. When the pH is lowered, D44 becomes
partially protonated, leading to an increased redox potential and favoring heme a
reduction. At high pH, the slow electron transfer rate from Cu to heme a suggests that
the rate is dependent on slow access of protons to the S44D site, an example of proton
coupled electron transfer (PCET). The observation that the rate of electron transfer at
high pH is slow but the reaction goes to completion, indicates that D44 is eventually
protonated and a normal redox potential is achieved. As shown in Scheme 1, at the
beginning point, D44 is deprotonated and heme a is oxidized (a3+ COO) with a lowered
redox potential; at the end, D44 becomes protonated and heme a is reduced (a2+ COOH)
with a normal redox potential. At pH 6.5 and 5.5, D44 has a mixture of
protonated/deprotonated forms, corresponding to the double peaks at 2.78/2.72 in EPR

spectra, and this also explains the observed biphasic kinetics. The fraction that is

92

protonated (peak at 2.78) results in the rapid electron transfer from CuA to heme a, as in
wild-type CcO. The fraction corresponding to deprotonated D44 (peak at 2.72) results in
a very slow rate of electron transfer from Cup, to heme a, which is rate-limited by the rate

of protonation of D44.

This raises an interesting question as to where the protons come from to protonate
D44, and to what extent the rate of proton transfer controls the intrinsic rate of electron
transfer. It is noteworthy that the S44D mutant is less efficient at proton pumping,
suggesting the possibility that the proton required for heme a reduction (D44 protonation)

is stolen from the pumping path and then leaked when heme a is re-oxidized.

An internal rather than external source of protons is suggested by a low deuterium
isotope effect12 and by the fact that low pH does not increase the rate of the slow phase,
only the proportion of fast versus slow phase in single electron transfer measurements.
That Zn inhibition is not accentuated in the mutant also supports the internal proton
source hypothesis. Further crystallographic and kinetic studies on the S44E and S44N
mutants could give more insight into the altered EPR characteristics and provide useful

tools for studying the role of heme a in the proton pumping mechanism.

93

e

a3+coo’ .:.L. a2+coo‘
Eai

WHPK: H” 1? 9'9

a3*COOH 4—' 32*COOH

Scheme]

94

REFERENCES:

1. Pan, L. P.; Hibdon, S.; Liu, R.—Q.; Durham, 3.; Millett, F., Intracomplex Electron
Transfer Between Ruthenium-Cytochrome c Derivatives and Cytochrome c Oxidase.
Biochemistry 1993, 32, 8492-8498.

2. Pan, L. R; Li, Z.; Larsen, R.; Chan, S. I., The Nature of Cux in Cytochrome c
Oxidase. J. Biol. Chem. 1991, 266, (3), 1367-1379.

3. Hill, B. C., The Reaction of the Electrostatic Cytochrome c-Cytochrome Oxidase
Complex with Oxygen. J. Biol. Chem. 1991, 266, (4), 2219-2226.

4. Rich, P.; Meunier, 3.; Mitchell, K; Moody, R., Coupling of Charge and Proton
Movement in Cytochrome c Oxidase. Biochim. Biophys. Acta 1996, 1275, 91-95.

5. Tsukihara, T.; Aoyama, H.; Yamashita, E.; Tomizaki, T.; Yamaguchi, H.;
Shinzawa-itoh, K.; Nakashima, R.; Yaono, R.; Yoshikawa, 8., Structure of Metal Sites of
Oxidized Bovine Heart cytochrome c Oxidase at 2.8 A. Science 1995, 269, 1069-1074.

6. Svensson-Ek, M.; Abramson, J.; Larsson, G.; Tomoth, S.; Brzezinski, P.; Iwata,
S., The X-ray crystal structures of wild-type and EQ(I-286) mutant cytochrome c
oxidases from Rhodobacter sphaeroides. J. Mol. Biol. 2002, 321, 329-339.

7. Qin, L.; Hiser, C.; Mulichak, A.; Garavito, R. M.; Ferguson-Miller, 8.,
Identiﬁcation of conserved lipid/detergent-binding sites in a high-resolution structure of
the membrane protein cytochrome c oxidase. Proc Natl Acad Sci U S A 2006, 103, (44),
16117-22.

8. Hosler, J. P.; Fetter, J.; Tecklenburg, M. M. J.; Espe, M.; Lerma, C.; Ferguson-
Miller, S., Cytochrome aa3 of Rhodobacter sphaeroides as a Model for Mitochondrial
Cytochrome c Oxidase. J. Biol. Chem. 1992, 267, (34), 24264-24272.

9. Millett, F.; Miller, M. A.; Geren, L.; Durham, B., Electron transfer between
cytochrome c and cytochrome c peroxidase. J Bioenerg Biomembr 1995, 27, (3), 341-51.

10. Hosler, J. P.; Espe, M. P.; Zhen, Y.; Babcock, G. T.; Ferguson-Miller, 8.,
Analysis of Site-Directed Mutants Locates a Non-Redox-Active Metal near the Active
Site of Cytochrome c Oxidase of Rhodobacter sphaeroides. Biochemistry 1995, 34, 7586-
7592.

11. Williams, P. A.; Blackburn, N. J.; Sanders, D.; Bellamy, H.; Stura, E. A.; Fee, J.
A.; McRee, D. E., The CuA domain of Thermus thermophilus ba3-type cytochrome c
oxidase at 1.6 A resolution. Nat. Struct. Biol. 1999, 6, (6), 509- 516.

12. Mills, D. A. S. X., Lois Geren, Carrie Hiser, Ling Qin, Martyn A.; Sharpe, J. M.,
Bill Durham, Francis Millett“, Shelagh Ferguson-Miller*. Proton-Dependent Electron

95

Transfer from CuA to Heme a and an Altered EPR Spectrum in Mutants Close to Heme a
of Cytochrome Oxidase Biochemistry 2008.

13. Hiser, C.; Mills, D. A.; Schall, M.; Ferguson-Miller, S., C-terrninal truncation and
histidine-tagging of cytochrome c oxidase subunit II reveals the native processing site,
shows involvemet of the C-terminus in cytochrome 0 binding, and imporves the assay for
proton pumping. Biochemistry 2001, 40, 1606-1615.

14. Fetter, J. R.; Qian, J.; Shapleigh, J.; Thomas, J. W.; Garcia-Horsman, J. A.;
Schmidt, E.; Hosler, J.; Babcock, G. T.; Gennis, R. B.; Ferguson-Miller, 8., Possible
Proton Relay Pathways in Cytochrome c Oxidase. Proc. Nat. Acad. Sci. U. SA. 1995,
92, 1604-1608.

15. Mills, D. A.; Ferguson-Miller, 8., Inﬂuence of structure, pH and membrane

potential on proton movement in cytochrome c oxidase. Biochimica et Biophysica Acta
2002, 1555, 96-100.

16. Teschner, T.; Trautwein, A. X.; Schunemann, V.; Yatsunyk, L. A.; Walker, F. A.,
Low-spin ferriheme models of the cytochromes: Correlation of molecular structure with
EPR and Mossbauer spectral parameters. Hyperfine Interactions 2004, 156, (1), 285-291.

17. Gadsby, P. M. A.; Thomson, A. J., Assignment of the Axial Ligands of Ferric Ion
in Low-Spin Hemoproteins by near-Inﬁared Magnetic Circular-Dichroism and Electron-
Paramagnetic Resonance Spectroscopy. Journal of the American Chemical Society 1990,
112, (13), 5003-5011.

18. Yonetani, T.; Schleyer, H., Electromagnetic Properties of Hemoproteins .2. Effect
of Physical States on Electron Paramagnetic Resonance Parameters of Hemoproteins.
Journal of Biological Chemistry 1967, 242, (17), 3926-&.

19. Yonetani, T.; Schleyer, H., Electromagnetic Properties of Hemoproteins .1.
Electron Paramagnetic Resonance Absorptions of Single Crystals of Ferrimyoglobin and
Cytochrome C Peroxidase. Journal of Biological Chemistry 1967, 242, (17), 3919-&.

20. Yonetani, T.; Yamamoto, H.; Erman, J. E.; Leight, J. S. J.; Reed, G. H.,
Electromagnetic Properties of Hemoproteins. V. Optical and Electron Paramagnetic

Resonance Characteristics of Nitric Oxide Derivatives of Metalloporin-Apohemoprotein
Complexes. J. Biol. Chem. 1972, 247, (8), 2447-2455.

21. Peisach, J.; Blumberg, W. E.; Adler, A., Electron-Paramagnetic Resonance
Studies of Iron Porphin and Chlorin Systems. Annals of the New York Academy of
Sciences 1973, 206, (Oct22), 310-327.

22. Taylor, C. P. S., Epr of Low-Spin Heme Complexes - Relation of T2G Hole

Model to Directional Properties of G-Tensor, and a New Method for Calculating Ligand-
Field Parameters. Biochimica Et Biophysica Acta 1977, 491, (1), 137-149.

96

23. Pezeshk, A.; Torres, J.; Wilson, M. T.; Symons, M. C. R., The EPR spectrum for
Cu-B in cytochrome c oxidase. Journal of Inorganic Biochemistry 2001, 83, (2-3), 115-
119.

97

CHAPTER 3
PULSED ELECTRON NUCLEAR DOUBLE RESONANCE

STUDY OF AQUA PROTONS COORDINATED TO Mn (II)
IN CYTOCHROME C OXIDASE

98

1 .Introduction

An important bioenergetic function of cytochrome c oxidase is to pump protons
across the membrane. Thus, it is important to be able to understand proton movements
and the mechanism of proton pumping. In crystal structures of CcO, the non-redox active
Mg site is located in an aqueous region1(Fig.3.l), on the outer side of the membrane,
which hints that it may ﬁmction as a regulator of proton/water exit. The role of the Mg
site in CcO is difﬁcult to study because of its diamagnetic and non-redox active
properties. However, the ﬁnding that Mg can be substituted by Mn without any
ﬁmctional effects opens a door for studying this site by EPR methodsz. Previous studies
in the Ferguson-Miller lab have shown some intriguing results3’ 4. EPR and rapid freeze
quench data using szO indicate that the water ligands of the an‘L/Mg2+ site, or the
protons of these waters, can exchange with bulk solvent at a rate consistent with a role for
this region in product release during turnover. Further experiments using H2170 and 1702
showed that water itself was able to reach the Mn2+fMg2+ site from the outside bulk water
and from the interior active site. In these studies, the authors established a standard curve
of Mn2+(HzO)6 in different levels of 2H20 to estimate the number of exchanged water or
protons. Although the estimated number is quite close to the result revealed by X-ray
crystal structure assuming ﬁll] substitution of liganded waters, this indirect method was a
compromise using the low resolution of the EPR machine, at that time. However, the
EPR machine has been upgraded, and a better quality, higher resolution, and more direct
measurement of echo modulation of Mn2+ by its water ligands, can be achieved. In

ESEEM, the modulation depth is proportional to n/r6 where n is the number of nuclei

99

 

 

)Q j \ W l" _
l Subunit I/II interface in CcO , OutSIde
\ V s. . '

-

1“ , H20/H“ channel ?

   
  

Figure 3.1: Proposed water channel in bovine CcO. Blue ribbons on the top are from
subunit II which houses dinuclear CuA site (blue spheres). The green ribbons on the far
right represent subunit III, which does not have any metal centers and its function is
unclear. The purple ribbons represent subunit I which contains two hemes (red) and CuB
site (light green sphere). The proposed water channel is shown as the gap between
subunits I and 11, beginning near heme a3 propionates and the Mg/Mn site (brown
sphere), with residues lining the channel in ball-and stick form. The three small blue
spheres represent crystallographically visible water molecules. The ﬁgure is from
Florens, L. et al3

100

coupled to the paramagnetic center and r is the dipolar distance. With either parameter
unknown, it is hardly possible to ﬁgure out the other parameter without guessing.
However, using ENDOR, we can measure r directly by measuring the hyperﬁne coupling
constants. Then we can gain knowledge of the number of coupled nuclei. NMR has been
the major method to study the hyperﬁne coupling interactions between Mn2+ and its
nearby nuclei. However, most of the hyperﬁne information was obtained indirectly,
which is by measuring the relaxation properties. ENDOR, on the other hand, is

straightforward in determining the hyperﬁne coupling constants.

The electron spin states of Mn2+ have relatively long lifetimes. As a result, the EPR
signals have narrower line-widths. Therefore, it is usually possible to obtain hyperﬁne-
coupling information from the resolved EPR features and obtain information on the
ligand environment. The signature feature of an Mn EPR spectrum is the sextet splittings
which arise from the hyperﬁne coupling between Mn unpaired electrons and its nucleus
(I=5/2). However, these hyperﬁne splittings are not as useﬁll as expected because the
sensitivity of the EPR measurement is somewhat diminished by the complicated
appearance of EPR spectrum resulting ﬁ‘om overlapping of the hyperﬁne splittings,
which are due to each of the ﬁve allowed EPR transitions (Fig.3.2). It is important to sort
out these couplings in the ENDOR pattern, and obtain intrinsic dipolar and contact
contributions. Scholes and coworkerss, using a simple system of manganese hexaquo as
the model, did an excellent job in addressing this issue and provided us with a good

reference to start with. In this chapter, the same method was applied to a far more

101

ZFS ZEMAN EFECT

 

 

 

 

 

 

A ”He . n+5/2
S=i3/2 : +3/2
S-i1/2 : “,2
I 5 i— -1/2
3:15/2 , ~\ '3/2

 

 

 

454%-

mI
+512
+312
+112
-1f2
~31?
-5/2
-5l2
-3l2
~10
+112

l
+3t2
+512

{ 6 lines in an

) electronic

transitions
Figure 3.2: Diagram of Mn(II) energy level and EPR transitions. ZFS represents
zero-ﬁeld-splitting. The ﬁgure is a modiﬁcation of a ﬁgure from Reed, G. et a1 6

B

 

 

+1l2
re"

-1I2

A l
l i 1‘
__// K/r‘ m / :/

 

 

 

 

 

l
/
r

 

 

 

102

complex system, CcO, using Davies ENDOR to study the hyperﬁne coupling interactions
between Mn and its water ligands. The pulse sequence of Davies ENDOR is depicted in a
ﬁgure (Fig.3.3). According to the principles established by previous researchers, ENDOR
amplitudes from nuclei with small hyperﬁne couplings (A<2MHz) are enhanced by
longer microwave pulse widths, and ENDOR amplitudes from nuclei with larger
hyperﬁne couplings (A>2MHz) are enhanced by shorter microwave pulse widths with
higher magnetic ﬁeld strengths. In my experiments, different widths of the microwave
pulse were used to selectively observe nuclei either strongly or weakly coupled to the
Mn2+ site. The inner sphere water ligands of the Mn2+ site are the focus of my study,
therefore, a shorter width of pulse could enhance the ENDOR amplitude, and was used in

these experiments.

2. Experimental Methods

2.1 Protein production and puriﬁcation

The amount of Mn2+ incorporated into the protein depends on the [Mg] to [Mn]
ratio in the growth mediumz. The YZ-100 strain of R. sphaeroides, which over-expresses
wild-type CcO with a histidine-tag added to the C-terminus of subunit I, was grown on
“high Mn” medium (with MnSO4 and MgSO4 at ﬁnal concentrations of 700uM and
50uM, respectively). The protein was puriﬁed by Ni2+-NTA afﬁnity chromatography and
further puriﬁed by DEAE ion-exchange chromatography7 to remove excess subunit I and
the remaining contaminants. The procedure involves washing the protein with buffer

containing lmM EDTA to remove the adventitious free extrinsic Mn ions. The enzyme

103

 

 

 

T

 

 

 

 

 

 

M2 1: 1t 1: is

 

 

1E

 

 

 

Preparation Mixing

Detection

Figure 3.3: Pulse sequence for Davies ENDOR experiment. The ﬁgure is a
modiﬁcation of the ﬁgure from Prisner, T. et al.8

104

was then concentrated and washed into the desired buffer using a Millipore centriﬁigal
ﬁlter with a MW (Molecular Weight) cut-off lOOKDa. The visrble spectral characteristics
of Mg- and Mn-substituted wild-type enzymes were identical and oxygen consumption
activities were not alteredz. The visible spectrum of the resting enzyme also indicated that

the puriﬁed enzyme is in its oxidized form (Fig 3.4).

2.2 Sample preparation

The oxidized sample was made by mixing wild-type Mn (II) substituted CcO in
20mM HEPES-KOH buffer, 15mMKCl, 0.1% lauryl maltoside, at pH7.4, resulting in a
sample of the protein and glycerol having a ﬁnal concentration of 148uM and 40%(v/v)
respectively. The solution was then degassed by purging with argon gas for more than 30
minutes and then quickly transferred to an EPR-ENDOR tube and ﬁ'ozen in liquid
nitrogen. The reduced sample in 100 mM HEPES-KOH buffer, 0.1% lauryl maltoside, at
pH7 .4 was made by ﬁrst reducing the enzyme with dithionite with a ﬁnal concentration
of 25mM. Then the same procedure for preparing the oxidized ENDOR sample was

followed.

2.3 X-band pulsed electron nuclear double resonance Spectroscopy (ENDOR)

X-band (microwave ﬁequency 9.72GHz) pulsed-ENDOR spectra were measured on
a Bruker E-680X spectrometer equipped with an EN 4118X-MD-4-Wl resonator. The
temperature was maintained at 4.5 K using an Oxford Instruments liquid helium ﬂow
system equipped with a CF-935 cryostat and an ITC-503 temperature controller. ENDOR

spectra were recorded using a Davies pulsed-ENDOR sequence. The relative ENDOR

105

0.134
0.144 / V l

0.10. . 3
Abs 1 \ 1l

0.06 “\r

0.02 . \

 

 

h -- Y:____‘ ._ ,_ '__ I. {ﬁx “Mk“
\__ _._, _fF' ._ m
E W- .
v r r Y r t v r I V T ‘— i _ ”M
400 440 480 520 560 6C” 640 690

wavelength (nm)

Figure 3.4: The visible spectrum of Mn-substituted 0.0 at resting (blue) and
dithionite-reduced (red) states.

106

amplitude can be enhanced based on the magnitude of the hyperﬁne-coupling
interactions. The choice of pulse widths is therefore an important factor in improving
ENDOR sensitivity. The ENDOR conditions used in this study are: microwave 11: pulse of
96ns and a 7t/2 pulse for 48ns.The radio frequency 7: pulse was 8000ns and was started at
1000ns after the ﬁrst microwave 1t pulse. The second microwave n/2 pulse followed at

1000ns at the end of the radio frequency pulse. The delay time between two pulses of the

detection Hahn-echo sequence was 400ns.

2.4 Simulation of EPR and ENDOR spectra

As a hard metal, Mn2+ has an afﬁnity for hard oxygen or nitrogen-containing
ligands such as water, carboxylate, and histidine. In CcO, the Mn2+ ion was observed to
be ligated by six ligands: three water molecules, two carboxylates, and one histidine].
The distorted octahedral coordination of Mn2+ generates zero-ﬁeld splittings in the frozen
solution. Such zero-ﬁeld splittings cause a broadening of the EPR spectrum on either side
of the central g=2.00 feature, and shift the EPR intensity away from g=2.00 to contain
contributions from Ms= i3/2 and i5/2 electron spin states, although not as signiﬁcantly

as seen in the manganese hexaquo.

EPR spectra simulations were accomplished by using software Xsophe which is the
standard soﬁware on a Bruker E-680X machine. The electronic spin Hamiltonian for
Mn2+ can be written as: He =geﬂeHOS +D [Szz-l/3S(S+l)] +AI-S where the ﬁrst term is
the electron Zeeman interaction. The second term is the zero-ﬁeld splitting term, and the

$6 ’9

z axis is the direction of the distortion ﬁ'om octahedral symmetry. Due to nearly

107

spherical distribution of electron spins about the Mn nucleus, the Mn hyperﬁne coupling
is treated as isotropic and the third term represents the isotropic electron-nuclear
hyperﬁne coupling. In the simulation, the zero-ﬁeld splitting parameters of D =145G, and

A =96.5G were usedg.

The simulation strategy for ENDOR is based on the ENDOR work done by Kwiram

110

et a . The ENDOR frequencies of the protons coupled to the Mn (11) can be expressed

as:

3.1

V(63¢) : {Z [[12 (69 ¢)(A1Ms — V0)2 ] }5

Where 0 and 4) are the spherical polar angles relating the applied magnetic ﬁeld to the
principal axes of the hyperﬁne tensor. v0 is deﬁned by ganHo/h. 11 = sin0cos¢, lz=
sinGsind), and 13: c050. The A1, A2, and A3 are principal hyperﬁne coupling constants,

and Ms represents the electron spin quantum number.

The simulation procedure contains four steps: ﬁrst, the powder pattern for each
electron spin transition was obtained according to equation 3.1, then same procedure was
applied to the next transition, until all ﬁve electron spin transitions were completed.
Second, was to use the Simpson’s rule to approximate the integral of the v(0,¢) ﬁmction
over all angles. The next step was to multiply the overall ENDOR powder pattern by the
proper weightings (underlying EPR intensities obtained from echo-detected EPR
spectrum). In the last step, the resultant histogram was smoothed by the Gaussian

convolution function in Matlab program. At 2K or less there is a thermal population

108

difference between Mn2+ with different Ms quantum number, which cannot be ignored.
This Boltzmann population factor was not included in this simulation because the
experiments were performed at a relatively high temperature, 4.5K. As a result, there is

no difference in intensity for transitions |-3/2> <—> |-l/2> and |+l/2> <—> |+3/2>. For the

same reason, |-5/2> <—-> |-3/2> and |+3/2> <—> |+5/2> have the same intensity.

3 Results and Discussion

3.] Electron spin-echo-detected absorption EPR spectroscopy

As a starting point for the later theoretical simulations, the EPR absorption
spectrum was monitored by the spin-echo amplitude at 4.5K (Fig.3.5). The spectrum
contained the EPR transitions due to the Mn (11) S=5/2 spin sextet. The six relatively
narrow lines are due to the central transition of the S=5/2, split by the hyperﬁne coupling
interactions with Mn (II)’s I=5/2 nuclear spin. The broad unresolved background is due to
other electron spin transitions that contribute intensity to the central region of the
spctrum. The relative contributions from individual transition were estimated using the
Xsophe program equipped on Bruker E680 machine. The zero-ﬁeld splitting parameters
used in the simulation are from the reference). The relative weights of these transitions
are: 0.36 for |-l/2> <——) |+1/2>, 0.20 for |-3/2> <—> [-1/2>, 0.20 for |+1/2> <——) |+3/2>, 0.12
for |-5/2> <—> |-3/2>, and 0.12 for |+3/2> (—> |+5/2>. These weightings are an approximate

estimation, but represent fairly well the contributions from different electron spin states.

3.2 1H ENDOR spectroscopy

109

 

Amplitude

 

 

 

2500 W00 3500 4000 4500
Magnetic Field (Gauss)

Figure 3.5: Experimental (black) and simulated (red) electron spin-echo-detected
absorption EPR spectra of Mn-substituted cytochrome c oxidase. Experimental
conditions for the measurement: ve =9.712GHz; central ﬁeld is 3.5KG, mw pulse widths
48ns and 96ns for rt/2 and 7t pulses, respectively; temperature, 4.5K. Simulation
conditions are: D=145G; E/D=0.193; A=96.5G; Weightings from the individual
transitions are 0.36 for |—1/2> (—> |+1/2>, 0.24 for |-3/2> <—-> |-l/2>, 0.22 for |+1/2> <—>
|+3/2>, 0.08 for |-5/2> (-—) |-3/2>, and 0.10 for |+3/2> (—> |+5/2>.

110

Davies ENDOR was recorded at a central ﬁeld of 3700G within the central
transition region, where the microwave ﬁ'equency is about 9.72GHz. As Fig.3.6 shows,
the 1H ENDOR signals lie approximately symmetrical with respect to the proton Larmor
frequency and are split by hyperﬁne coupling interactions. The spectra of oxidized and
dithionite-reduced proteins were obtained under conditions that optimize the coupling
through bond features between 9 and 22MHz. The features are centered at the free proton
Larmor ﬁequency (vp) and have splittings away ﬁom vp of i1.6, i 4.7, and i 7.4MHz.
These splittings have a ratio of approximately 1:3:5, indicating that these features were
mostly from the same type of proton nucleus, where it was respectively coupled to
electron spin states i1/2, 1- 3/2, and :t 5/2. The spectra of oxidized and dithionite-reduced
ENDOR samples are similar to each other except that the spectrum of the reduced sample
is more symmetrical in terms of the splittings centered at vp. Most of the shoulders in the
spectrum vanished in the deuterated protein sample, suggesting that the protons are
exchangeable (Fig.3.7). The residual shoulder features of the ENDOR spectrum are most
likely ﬁom incomplete deuteration because exchanging the protein into 2H20 cannot
achieve 100% efﬁciency. Another reason is that the presence of the non-exchangeable

protons ﬁom non-deuterated glycerol (added as cryoprotectant) and the protein.

Although the CcO ENDOR spectrum has the same basic shape as the Mn (HzO)6
complex, the former has a far more complicated spectrum. Peaks and shoulders from M5

= i3/2 and i5/2 were hidden in the baseline and are difﬁcult to recognize.

lll

Amplitude

l
a 1', ,l

’ .
d ‘V‘it/ i' ‘wbl" ' "i" H“-

 

 

F l T r
0 25

15 2
Frequency (MHz)

Figure 3.6: Davies ENDOR spectra of the oxidized (red) and the reduced (black)
cytochrome c oxidase at central field 3.7kG. Oxidized protein sample is in 20mM
HEPES-KOH, pH 7.4, 14mMKCl, and 0.1%laury1 maltoside buffer. The reduced protein
sample in 100 mM HEPES-KOH buffer, 0.1% lauryl maltoside, at pH7.4 was made
reducing the enzyme with dithionite with a ﬁnal concentration of 25mM. ENDOR
conditions: mw ﬁ‘equency, ve=9.72GHz; mw power, 20uW; n/2 and 1t mw pulse widths
are 48ns and 96ns respectively, RF pulse width is 8000ns; t= 400ns; T=40ns; srt=2000ns;
signal averaging number 100; scan number n=25; temperature, 4.5K.

112

Amplitude

h

l

.J/ [Wuvlf J 1' ["ﬂ* l '. ll; ‘ ll

 

 

l l r I l

5 10 15 20 25
Frequency (MHz)

Figure 3.7 : Davies ENDOR spectra of reduced cytochrome c oxidase in pH 7.4 water
buffer (red) and 75% 2H20 buffer (black). Protein sample in HzO is made in 100mM
HEPES-KOH, pH 7.4, and 0.1%lauryl maltoside buffer and reduced b dithionite with a
ﬁnal concentration of 25mM. Protein sample in szO is washed with HzO buffer of the
same components as the water buffer several times to ensure the szO level to at least
90%. ENDOR conditions: mw frequency, ve=9.72GHz; mw power, 20uW; n/2 and 7t
mw pulse widths are 48ns and 96ns respectively; RF pulse width is 8000ns; 1:= 400ns;
T=40ns; srt=2000ns; signal averaging number 100; scan number n=25; temperature,
4.5K.

ll3

3.3 ENDOR simulation

To better understand the ENDOR spectra of the Mn in the CcO enzyme, it was
US$111] to simulate ENDOR powder patterns with the purpose of predicting the positions
of peaks and shoulders in the ENDOR spectrum. A series of trial simulations of ENDOR
spectrum were performed using a Matlab-written program (Krzyaniak Matt.) with the
procedures described in the method. The simulated powder pattern ENDOR spectrum

matches well with the experimental data (Fig.3.8).

The protons around the Mn” site in CcO yielded the values of A“ and A; of
7.6MHz and —2.2MHz, respectively. From equations A” = 2Ad + A, and A i = A5- Ad, an
isotropic coupling As =1.06MI-Iz, and a dipolar coupling Ad = 3.26MHz were obtained.
The A“ and A r values represent the averaged hyperﬁne coupling interactions from

different protons within a certain distance range, ca. 5 A.

4. Conclusion

The Davies ENDOR proves to be a powerful tool to provide hyperﬁne-coupling
constants. The simulation of the overall ENDOR powder pattern yielded an isotropic
coupling As=l .07MHz and a dipolar coupling Ad=3.27MHz, where AH=2Ad+AS (7.6MHz)
and A i=As-Ad. (-2.2MHz). The major contribution to the hyperﬁne coupling is the
through-space dipolar coupling to protons that are about 2.89A from the Mn site. The
estimated distance between Mn and protons is in the range of the distances between Mn
and its inner-sphere protons (2.17~ 2.86 A) revealed in 2.0A X-ray crystal structure11 and

is close to the value of 2.8-2.9 A observed using lH ENDOR in Mn hexaaquo complex”.

114

The agreement between the estimated distance using simulated hyperﬁne coupling
constants and the distance from crystal structure validated our simulation. The hyperﬁne
coupling constants therefore would be good enough to be used in ESEEM simulation to

quantify the number of protons bound to Mn site, which is the focus of the next chapter.

115

vp=15.8MHz

Ms= -1 l2

 

Ms=+l /2

Amplitude

Ms=+3l2

 

Ms=+5l2
_ Ms=-5l2
vWVV'w'mr' ""'V'V
l T l I
20 25

 

10 15
Frequency (MHz)

Figure 3.8: Experimental (black) and simulated (red) ENDOR powder pattern
spectra at central ﬁeld H=3.7kG. The simulation used AH=7.6MHz and A i=-2.2MI-Iz

116

REFERENCES:

1. Tsukihara, T.; Aoyama, H.; Yamashita, E.; Tomizaki, T.; Yamaguchi, H.;
Shinzawa-Itoh, K.; Nakashima, R.; Yaono, R.; Yoshikawa, S., The Whole Structure of
the 13-Subunit Oxidized Cytochrome c Oxidase at 2.8 A. Science 1996, 272, 1136-1144.

2. Hosler, J. P.; Espe, M. P.; Zhen, Y.; Babcock, G. T.; Ferguson-Miller, 8.,
Analysis of Site-Directed Mutants Locates a Non-Redox-Active Metal near the Active
Site of Cytochrome c Oxidase of Rhodobacter sphaeroides. Biochemistry 1995, 34, 7586-
7592.

3. Florens, L.; Schmidt, B.; McCracken, J.; Ferguson-Miller, 8., Fast deuterium

access to the buried magnesium/manganese site in cytochrome c oxidase. Biochemistry
2001, 40, (25), 7491-7497.

4. Schmidt, 3.; McCracken, J .; F erguson-Miller, S., A discrete water exit pathway in
the membrane protein cytochrome c oxidase. Proc Natl Acad Sci U S A 2003, 100, (26),
15539-42.

5. Tan, X. L.; Bernardo, M.; Thomann, H.; Scholes, C. P., Pulsed and Continuous
Wave Electron Nuclear Double-Resonance Patterns of Aquo Protons Coordinated in
Frozen Solution to High-Spin Mn2+. Journal of Chemical Physics 1993, 98, (7), 5147-
5157.

6. Reed, G. H. a. M., G. D., EPR of Mn(II) complexes with enzymes and other
proteins. Biological Magnetic Resonance 1984, 6, 73-142.

7. Hiser, C.; Mills, D. A.; Schall, M.; Ferguson-Miller, S., C-terminal truncation and
histidine-tagging of cytochrome c oxidase subunit II reveals the native processing site,
shows involvemet of the C-terminus in cytochrome c binding, and imporves the assay for
proton pumping. Biochemistry 2001, 40, 1606-1615.

8. Prisner, T.; Rohrer, M.; MacMillan, F., Pulsed EPR spectroscopy: Biological
applications. Annual Review of Physical Chemistry 2001, 52, 279-313.

9. Kass, H.; MacMillan, F.; Ludwig, B.; Prisner, T. F., Investigation of the Mn
binding site in cytochrome c oxidase from Paracoccus denitriﬁcans by high-frequency
EPR. Journal of Physical Chemistry B 2000, 104, (22), 5362-5371.

10. Kwiram, A. L., Endor Studies in Molecular Crystals .1. Glutaric Acid at 4.2
Degrees K. Journal of Chemical Physics 1971, 55, (5), 2484-&.

11. Qin, L.; Hiser, C.; Mulichak, A.; Garavito, R. M.; Ferguson-Miller, 8.,
Identiﬁcation of conserved lipid/detergent-binding sites in a high-resolution structure of
the membrane protein cytochrome c oxidase. Proc Natl Acad Sci U S A 2006, 103, (44),
16117-22.

117

12. Hoogstraten, C. G.; Britt, R. D., Water counting: Quantitating the hydration level
of paramagnetic metal ions bound to nucleotides and nucleic acids. Rna-a Publication of
the Rna Society 2002, 8, (2), 252-260.

118

Chapter 4
ELECTRON SPIN ECHO ENVELOPE MODULATION
(ESEEM) STUDY OF WATER LIGANDS BOUND TO
Mg/Mn SITE IN CYTOCHROME C OXIDASE

119

1. Introduction

Electron spin echo envelope modulation (ESEEM) spectroscopy is a very versatile
technique. ESEEM spectra contain information on the strength of hyperﬁne coupling, the
identity of the coupled nucleus, and the number of nuclei coupled to the electron spin.
Because of this potential advantage, ESEEM has found many applications to
paramagnetic centers having an electron spin S=l/2. However, for systems with $21,
more attention has been paid to the analysis of spectral line positions and shapes than on

analysis of modulation amplitude to quantify the number of coupled nuclei.

As one of the more abundant heavy metals in the earth’s crust, Mn2+ can be found
in many metalloenzymes as part of their active sites such as the photosynthetic complex,
pyruvate kinase, superoxide dismutase. The ESEEM study of the Mn2+ sites in biological
systems interests many spectroscopists. In CcO, the Mg site can be substituted by Mn2+
and studied by ESEEM method. ESEEM studies of Mn-CcO1 indicated that a histidine
was bound to the Mg/Mn site, which was conﬁrmed later by the crystal structure. Further
studies of fast deuterium access to the buried Mn2+/Mg2+ site in CcO, combining ESEEM
with rapid freeze quench (RFQ)2, indicated that the water ligands of the Mn2+/Mg2+ site,
and the protons of these waters, can exchange with bulk solvent at a rate consistent with a
role for this metal in product release during turnover. Because of the complexity of S=5/2
system, simulations of Mn(II) ESEEM spectra are rarely done. A relatively convenient
way to quantify the number of waters coupled to the Mn site is to establish a standard

curve for quantiﬁcation. In the deuterium exchange study, Florens et al. used Mn2+(HzO)(,

120

as a standard to determine the number of exchanged water or protons. In their
experiment, the standard curve was established by ﬁrst obtaining the echo modulation of
Mn2+ by one water ligand by taking the ﬁfth root of the spectrum of Mn2+(HzO)6 in 80%
2H20 and then raising the resultant spectrum to increase the power up to ﬁve. The single-
dataset-generated standard curve introduced one problem that partly caused the
estimation of two water ligands exchanged with deuterium at the Mn site, which was
argued to be an underestimate. A better way to count the number of waters that are able
to undergo deuterium exchange at the Mn site relies on both higher signal-to-niose
ESEEM data and a more precise simulation program. Recently, Astashkin and
Raitsimring3 developed an ESEEM theory for a spin system with S>1/2 like Mn2+ that
avoids some of the shortcomings of previous work 4’ 5. Using their simulation strategy
together with information from X-ray crystal structures, we will use ESEEM

spectroscopy to quantify ligand displacement chemistry at the Mn site.

Since the release of high-resolution crystal structures of CcO, a focus on the role of
water in the facilitation of proton movement has prompted us to pay more attention to the
Mg/Mn site, which bridges subunits I and II, shares one ligand with dinuclear CuA site,
and is located in the vicinity of the aqueous region. Critical examination of bovine CcO
crystal structures in oxidized and reduced states revealed two key water molecules,
W1 Mg, which is in the sixth coordination position on Mg, and Wlﬁlgg, which is in the exit

channel and appears to undergo protonation upon CuA reduction. In the oxidized enzyme
both WIMg and WI 5198“ are neutral water molecules (Fig. 4.1A). Upon CuA reduction, a

proton movement is implied by the generation of H+-WIE193H, and a (”OH ligated to

121

Mg2+ at the WlMg position (Fig. 4.13). To test the crystallographic observations and the
implied hypothesis that the proton movement is one step in a redox-linked export of a
pumped proton from the binuclear center into the exit pathway, cw EPR and ESEEM
techniques were utilized to monitor the change in Mn cw EPR spectra and the change in
number of protons coupled to the Mn site at different oxidation states, with or without the

presence of extrinsic ligand, cyanide.

2. Experimental methods

2.1 Protein production and puriﬁcation

The amount of Mn2+ incorporated into the protein depends on the [Mg] to [Mn]
ratio in the grth medium". The YZ-100 strain of R. sphaeroides, which over-expresses
wild-type CcO with a histidine—tag added to the C-terminus of subunit 1, was grown on
“high Mn” medium (with MnSO4 and MgSO4 at ﬁnal concentrations of 700uM and
5011M, respectively). The protein was puriﬁed by Ni2+-NTA affmity chromatography and
further puriﬁed by DEAE ion-exchange chromatography7 to remove excess subunit I and
the remaining contaminants. The procedure involves washing the protein with buffer
containing lmM EDTA to remove the adventitious free extrinsic Mn ions. The enzyme
was then concentrated and washed into the desired buffer using a Millipore centrifugal
ﬁlter with a MW (Molecular Weight) cut-off 100KDa. The visible spectral characteristics
of Mg— and Mn-substituted wild-type enzymes were identical and oxygen consumption

activities were not altered". The visible spectrum of the puriﬁed enzyme showed that it

122

 

 

M “A Cw” E173“
ll .
E198 H +-W1E198"

  
 
  

 
   

     
 

   

Ti ‘ E198“
.3111. *- , ' 5’
0 on szg Fonic Bond‘ '10-)
Wrm - I“ ‘ HO-WIMg
T1271 [Wriz'r zsAl
WH291 a, .- - 1

1" .

J.,...—

 

WC

B. Reduced 2E1

 

 

‘l'Ieme “3
A. Oxidized (2DYR)

 

Figure 4.1: The formation of a hydroxide/hydronium pair upon CuA reduction;
modeled on bovine oxidase oxidized and reduced structures. A comparison of the
oxidized (A) and reduced (B) structures of bovine oxidase indicates a change in the
protonation status and bonding of two water molecules: W1519311 is HOH289 in the
bovine 2DYR (Ox) and HOH291 in the bovine 2EIJ (Red) structures. The ﬁgure is a
modiﬁcation of a courtesy ﬁgure from Dr. Martyn Sharpe.

123

was in its ﬁilly oxidized form (see chapter 3 for the spectrum).

2.2 Sample preparation

Puriﬁed protein was mixed with buffers containing 20mM HEPES-KOH,
lSmMKCl, 0.1% lauryl maltoside, at pH7.4 with varying levels of 2H20, which resulted
in ﬁnal percentages of szO ranging from 0% to 75% (v/v). Protein samples in different
levels of 2&0 had the same ﬁnal concentration of 3611M and were frozen in liquid
nitrogen. A parallel series of protein samples with 0.4M sucrose as a glassing agent were
also made. Reduced samples were made by adding 10 mM dithionite to samples which
had been de-aerated using three vacuum/argon cycles. The buffer of the reduced samples
was modiﬁed to contain100mM HEPES-KOH, 0.1% lauryl maltoside, at pH7.4 to buffer

the acidiﬁcation caused by adding the dithionite.

The model compound Mn2+-DTPA was prepared with a slight excess of chelator in
H20 and different level of 2H20, which also contained 0.4M sucrose. The Mn2+(HzO)6
was prepared by dissolving MnSO4 in H20 and different level of szO with 0.4M

SUCI'O $6.

2.3 Electron Spin Echo Envelope Modulation (ESEEM) spectroscopy and Data Analysis
Pulsed measurements were made on a Bruker E-680X spectrometer operating at X

band and equipped with a model ER 4118X-MD-X5-W1 probe that employs a 5mm

dielectric resonator. The temperature was maintained at 4.5K using an Oxford

instruments liquid helium ﬂow system equipped with a CF-935 cryostat and an ITC-503

124

temperature controller. ESEEM data were collected using three-pulse (stimulated echo)
with the sequence of 90°- r—90°-T-90° and a 90° microwave pulse widths of 16 ns. 1.‘ was
204ns in all ESEEM measurements. An integration window of 24ns was used to acquire

spin echo amplitude, and data set lengths were 512 points.

To determine the hydration level, three-pulse ESEEM was performed under such
conditions that signals from protons are suppressed. The modulations in the ESEEM
spectrum are ﬁom three sources: the magnetic nuclei from nearby residues, largely ‘4 N
from histidine imidazoles; inner-sphere water deuterons; outer-sphere water deuterons
and deuterons of the protein. Since the experimental modulation ﬁmction is the product
of all contributions, the contribution ﬁom 2H nuclei can be isolated by using ratio
method8 in which the experimental data is divided by the data of control containing the
unwanted contributions from sources such as the coordinating histidine. Individual
ESEEM spectra were normalized by dividing each data set by its maximum amplitude.
The normalized 2H ESEEM data were then divided by the correspondingly normalized 1H
ESEEM data, resulting in ESEEM data dominated by the 2H contribution to the ESEEM.
These data were then tapered using a Hamming window ﬁmction and subjected to a
cosine Fourier transformation. The resulting ESEEM spectra were obtained by taking the

absolute value of the transforms.

2.4 ESEEM simulation
For Mn(II) system, an approximate ESEEM theory was developed by Coffino and

Peisach4, and by Larson and Singels, However, in all their publications, authors were

125

focused on the spectral analysis without attempt to utilize , or compare the experimental
intensities of spectral lines with theoretical evaluations. Recently, Astashkin and
Raitsimring3 revised and corrected the previous theory and derived new expressions for
the ESEEM from high spin system Mn2+. The simulation program used in this work is
based on the original work done by Mims9 with some modiﬁcations suggested by
Astashkin and Raitsimring3 in their paper. The program was written in Matlab by

Matthew Krzyaniak (Michigan State University, Department of Chemistry).

3. Results
3.1 The use of cryoprotectant

Under cryogenic temperature (4.5K), a glassing agent or cryoprotectant is critical
to preventing solute from segregation and other adverse effects. However, many
cryoptotectants contain alcohol group that can occupy coordination site on metal ions,
and therefore introduce errors in determining the number of waters bound to metal center.

110 found in their studies that glycerol, a commonly used

For instance, Halkides et a
cryoprotectant, ligates directly to the Mn2+ bound in the active site of P21-ras by
displacing a water ligand. So the choice of cryoprotectant must be considered.
Hoogstraten et a1H demonstrated in their study of hydration level of the Mn2+ bound to
nucleic acid and nucleotide that glycerol and ethylene glycol will occupy some of the

inner-sphere coordination sites on the Mn2+ in aqueous solution, while sucrose or similar

saccharide-based cryoprotectant are relatively safe choice for water counting purpose. In

126

1.1 -
1.0 a

1 — CcO wlo cryoprotectant
0.9 - — Geo W! cryoprotectant

0.8 -

0.7 e

0.6 A
l

0.5 -

 

 

Amplitude

 

 

0.4 -1

 

 

r ‘ I

. 1 . . l
0 2000 4000 6000 8000

T+ 1(ns)

Figure 4.2: Effect of the cryoprotectant on CcO three-pulse ESEEM data. The
spectra are normalized and divided by 0%2H20 control sample. Protein samples are
made in 75% 2HzO level buffer with 20mM HEPES-KOH, pH 8.0,14mM KCl, and
0.1%lauryl maltoside with a ﬁnal concentration of 3611M. Spectrometer conditions are:
mw ﬁ'equency, 9.72GHz; central ﬁeld, 3.7kG; microwave power, 50mW; 1t/2 pulse
width, l6ns; r, 204ns; integration window, 24ns.

127

this study, 0.4M sucrose was added in Mn2+(HzO)6 and Mn-DTPA samples. For the CcO
samples, two parallel samples were made respectively with and without sucrose; ESEEM
data showed no signiﬁcant difference in modulation depth (Fig.4.2). This ﬁnding can be
explained by the large size of the CcO protein sorrounding the Mn in which electron

spins of the Mn2+ site are well segregated. But for better protection of the EPR sample

tubes during freezing, a cryoprotectant was still used.

3.2 The deuterium standard curve for Mn2+(HzO)6 and Mn-DTPA

Mn2+(HzO)(, is octahedrally coordinated, having six inner sphere water molecules
and a great number of outer-sphere water molecules. The deuterium modulation depth
consists of contributions from both inner and outer-sphere waters. While in the Mn-
DTPA complex, inner sphere binding sites are occupied by DTPA, resulting in only
outer-sphere modulation depth in ESEEM spectrum. ESEEM modulation depth from
pure inner-sphere water contribution can be determined by dividing Mn-DTPA

modulation from Mn2+(HzO)(, modulation.

The isolation of 2H ESEEM from inner-sphere water ligands is demonstrated in
Fig.4.3A which shows raw ESEEM data of Mn(HzO)(, in H20 and 2H20, along with the
szO /HzO ratio. 2Hs in the solvent give rise to strong modulations within about 500ns.
Fig.4.3B shows the corresponding ESEEM curves for the Mn-DTPA complex, which
does not have inner-sphere water ligands. Fig.4.3C shows the ﬁnal processed modulation

curve, representing pure inner-sphere bound water ESEEM in Mn(HzO)6 complex. There

128

Amplitude

Amplitude

Amplitude

1.25

 

 

 

 

 

 

1.00
0.75~
0.50«
0.25:}
0.00-
0 Y 2000 f 4000 ' 6000 Y 8000
T+t(ns)
1.25—

 

 

 

 

 

 

 

 

1.004
0.75 ‘
0.50-
0 ' 20'00 ' 500 T 6000 ' 80'00
T+t(ns)
1.00-
0.75. c
0.504
l
0.00 . . . . . . a .
0 2000 4000 6000 8000
T+1:(ns)

Figure 4.3: Three-pulse ESEEM time-domain data of Mn(H20)6(A), Mn-DTPA (B),
and Mn(H20)6/ Mn-DTPA ratio (C). Samples in H20 (blue line), D20 (black line), and
D/I-I ratio (red line). All the spectrometer conditions are the same as depicted in ﬁgure

129

are two features worth mentioning: ﬁrst, the comparison of the data for Mn(HzO)6 with
the corresponding data for Mn-DTPA indicates that inner-sphere waters contribute about
70% ESEEM modulation, which is in good agreement with the literature”; second, the
damping of the deuterium ESEEM ﬁom the difference spectrum, which contains only
inner-sphere deuteron contribution, is very rapid which is expected because of the shorter

distance of inner-sphere deuterons from the electron spinlz‘ 13 .

3.3 ESEEM spectra of oxidized CcO in deuterated solvents with cryoprotectant

Similar studies were performed on Mn2+-substituted CcO at 3.7kG. Ratios of
modulation data remove contributions from 14N from the histidine ligand to Mn (His411).
Tirne-domain spectra of these samples showed pronounced modulations from deuteron
hyperﬁne interactions and an increase in modulation depth as percentage of 2H20 was
increased, indicating increased number of coupled deuteron nuclei (Fig.4.4A). The fairly
rapid damping of the deuterium ESEEM suggests that the inner-sphere deuterons
contribute predominantly to the ESEEM spectrum. The Fourier transformation of time-
domain spectra results in the frequency domain spectra of all samples (Fig.4.4B). The
frequency modulation at 2.42MHz is from coupled deuterons. Due to the use of proton
suppression time, the proton modulation amplitude at 15.8MHz is minimized. The
intensity of the 2.4MI-Iz deuterium signal increased as the concentration of szO was
increased. A nearly linear relationship was observed between the echo modulation and

the concentration of szO. The data was ﬁt with a linear ﬁmction (Fig.4.5).

130

1.05

 

 

 

 

0.90 ‘ _ . .
m i. t r. ‘ 1‘ ‘ . 1‘ “l_ ~11 . ‘
U _' '/ V‘M‘ﬂm'ﬁ‘ﬁ‘u‘w‘1'9"*.l'\4-«+M».',~.;«-,. .
:1 ,_ _
5 0.75 11
Q .. ’ ‘ . ,.
E . l v "
< 0.60-
0.45-
0 2000 4000 6000 8000
T + 1(ns)
7
B
6
q, 5
"g
E 4
Q.
3
E
< 2
1
O
0.000 0.005 0.010 0.015 0.020 0.025 0.030
Frequency (GHz)

Figure 4.4: Three-pulse ESEEM time-domain spectra (A), frequency-domain
spectra (B) of Rs. CcO in varying levels of 2H20 buffers ranging from 25% to 75%
in the absence of cryoprotectant. Colors represent: black, 25% D20; red, 30% D20;
cyan, 35% D20; purple, 40% D20; pink, 50% D20; green, 75%D20. Sample preparations
and the ESEEM conditions are the same as depicted in ﬁgure 4.2.

131

(I)
J

- CcO with 0.4M sucrose
—— linear fit

‘1
.l _L l

O)
J

01
n I 1

OD
s4 1

N
a l

Amplitude @2.4MHz

 

0 ...,...,.,.,.....,..
0 10 20 30 40 50 50 70 80 90100

020%

 

Figure 4.5: Fit of Fourier-transformed three-pulse ESEEM spectra of oxidized CcO
in different concentration of 2H20 buffers with cryoprotectant with a linear
function. The samples are made in the same way as described in ﬁgure 4.2. And the
spectrometer conditions are the same as depicted in ﬁgure 4.2.

132

3.4 cw EPR and ESEEM studies of the oxidized and reduced CcO plus or minus cyanide
To test the hypothesis that WlMg becomes a hydroxide in the reduced CcO, the
enzyme was reduced and studied by cw EPR, Figure 4.6 shows a comparison of the cw
EPR spectra of reduced and oxidized Rs. CcO in the absence and presence of the
externally added ligand, cyanide. Addition of cyanide to the oxidized enzyme had no
effect on the line shape of the Mn EPR signal but did broaden the Mn lineshape in the
reduced enzyme sample. This line shape change indicates that cyanide is perturbing the

electronic structure of the Mn center.

ESEEM experiments were also carried out on the oxidized and reduced CcO plus or
minus cyanide at 3.7kG. The time-domain spectra of the reduced CcO are similar to those
of the oxidized enzyme. When the time-domain spectrum of the oxidized CcO plus
cyanide is divided by the corresponding spectrum of the CcO without cyanide and then
Fourier transformed, it is clear that there is no signals observed, which was consistent
with the cw EPR data (Figure 4.7A). However, the same procedure for the reduced
oxidase plus cyanide and reduced oxidase resulted in the observation of the modulation at
1.5MI-Iz, 3.4MHz, and 52le, which can be assigned in part to perturbation of the

nitrogen from the liganded histidine (Figure 4.7B).

The reduced CcO with and without cyanide in different percentages of 2H20 buffer
were also studied by ESEEM. Plotting the peak amplitude at 2.42MHz versus 2H20
concentration and ﬁtting with a linear function resulted in the similar relationship

between echo amplitude and 2H20 concentration. When two plots of the reduced CcO

133

(A)
Oxidized

Oxidized + CN

 

Amplitude

Reduced J (B)

 

Reduced + CN

 

Figure 4.6: cw EPR spectra of oxidized and reduced CcO in the presence or absence
of cyanide. The protein samples are in 20 mM HEPES-KOH, 4 mM KCl, pH 7.0 buffer
with cyanide ﬁnal concentration of 5 mM, and dithionite of 10 mM. The spectrometer
conditions are: mw frequency, 9.46 GHz; mw power, 50 uW; modulation ﬁ'equency, 100

kHz; modulation amplitude, 13.0 G; conversion time, 327ms.

134

Oxidized enzyme. c‘4N H20/H20

 

     

Reduced enzyme. 25% D20/1420

 

1.5 3.4 5.2

Reduced enzyme. CMN H20n-l20

3,7 51 Reduced enzyme. C15N H20/H20

Figure 4.7 Effect of different radionucleotide of cyanide on the three-pulse ESEEM

spectra of the Mn center of Rs. oxidase. The ﬁgure is a modiﬁcation of the courtesy
ﬁgure ﬁ'om Dr. Martyn Sharpe.

135

and the reduced CcO plus cyanide are put on the same axes for comparison, it is
interesting to observe that the difference in slopes of two lines in Figure 4.8 is
approximately 9%. This difference could give us some insight into the change in deuteron

number associated with the Mn2+ upon cyanide addition in the reduced enzyme.

3.5 ESEEM simulation

The ESEEM spectrum for CcO is simulated using the parameters obtained from
simulation of the echo-detected EPR spectrum (see chapter 3 for the details): the relative
weights of various electron spin transitions are 0.36 for |-l/2> <—) |+1/2>, 0.20 for |-3/2>
<—) [-l/2>, 0.20 for |+l/2> <—> |+3/2>, 0.12 for |-5/2> <—> |-3/2>, and 0.12 for l+3/2> <—>
|+5/2>; the super hyperﬁne coupling parameters are A“=7.6MHz A]: —2.2MI-Iz for
proton, A“=l.16MHz, A i= —0.34MHz for deuteron. Because high spin Mn2+ has a nearly
zero orbital angular momentum, the g value is usually treated as isotropic and has a value
close to ﬂee electron g value. Due to the nearly spherical distribution of unpaired
electrons about the 55Mn nucleus, the 55Mn hyperﬁne coupling is virtually isotropic

except in highly distorted complexes.

The resultant simulated spectra assuming there are four, ﬁve or six deuterons in the
inner sphere of the Mn2+ site are shown in Figure 4.9. Comparing the simulated spectrum
with the experimental data, we can observe very good match in spectra between the
simulated and the experimental in terms of the lineshape. The amplitude at deuteron

Larmor frequency (2.42MHz) in the simulated spectrum, however, does not match that in

136

0.75 ‘

0.5 1

Amp
025‘

mp :,
l I 0'
..'. 0 25 50 75
_ g, 01.020
,-‘ '2‘ Reduced CcO+CN+DZO/Reduced CcO+CN+H20
1‘ .

_. ; 1.. Reduced CcO+D20/Reduced CcO+H20
! “ \f'kr.‘

*mrw-w M.,—f‘ Manx—«an. ,

 

 

 

1 1b 15
Frequency (MHz)

20

Figure 4.8: Plot of the amplitude at 2.42MHz versus 2H20 concentration for the
reduced CcO ratio and the reduced CcO plus cyanide ratio. The ﬁgure is a
modiﬁcation of a courtesy ﬁgure from Dr. Martyn Sharpe.

137

the experimental spectrum, assuming there are three water molecules bound to the Mn

site (Figure 4.10).

4 Discussion

Previous ESEEM studies of CcO were more focused on the analysis of the
modulation frequency. In an ESEEM study, CcO was incubated in buffered 2H20 under
resting or turnover conditions for 90 minutes and then frozen for experiments”. The
observation of the deuteron Larmor frequency in ESEEM data provided an indication that
the deuterium was introduced into the CuA environment. A further ESEEM study of the
Mn site1 in CcO provided direct evidence for a nitrogen ligand bound to the Mn site and
assigned this nitrogen to a histidine ligand, which was later conﬁrmed by crystal
structures15 . One attempt to use modulation amplitude was made by Schmidt et al. 16 who
studied the deuterium exchange at the Mn2+ site in CcO and estimated the number of
bound water ligands based on the analysis of deuteron modulation amplitude. This
attempt proved to be a pretty successﬁrl trial but can be improved with a higher

sensitivity instrument and better simulation pro gram.

Examination of recent crystallographic data”’ 18 revealed that the reduced CcO
structure appeared to have a hydronium ion close to the CuA-Mn pair W198" and that one
of the water ligands of Mn, WlMg, was deprotonated. This suggested that a proton could
be cycling between W198” and WIMg as a function of the redox state of the enzyme,
playing a role in proton pumping. Experiments with cw EPR showed a perturbed Mn

spectrum when cyanide was added to the reduced CcO, but had no effect on the oxidized

138

 

 

 

 

 

4.5 -
~5.3%
4.0 -
—4 deuterons
- }-9.5%
3.5 - — 5 deuterons
o . —— 6 deuterons
u 3.0 4
a .
= 2 5 -
Q I
E 2.0 -
< .
1.5 -
1.0
0.5
0-0 I r l ' I ' l ' I f I ' I
0 5 10 15 20 25 30

F requency(MH2)

Figure 4.9: The ESEEM simulated spectra of CcO assuming different numbers of
deuterons bound to the Mn site. The ESEEM simulation conditions are: central ﬁeld,
3700G; mw frequency, 9.72GHz; starting T, 40ns; tau, 204ns; T increment, l6ns; Axx,
Ayy for deuteron, -0.34MHz, and Azz, 1.16MHz. The relative weights of these
transitions: 0.36 for |-l/2> e |+1/2>, 0.20 for |-3/2> <—) |-l/2>, 0.20 for |+1/2> e) |+3/2>,
0.12 for |-5/2> <—) |-3/2>, and 0.12 for |+3/2> e |+5/2>.

139

enzyme as predicted if cyanide could substitute for hydroxide. Because of its sensitivity
to the change in ligand environment, ESEEM is used to monitor the change in ligands.
Both cw EPR and ESEEM data showed that the extrinsic ligand cyanide could bind to the
Mn site when CuA was reduced. In the reduced crystal structures, one of three water
ligands of the Mn site appears to be deprotonated, giving a hydroxide ion ligand. This
predicts a negatively charged cyanide ligand could replace it, as observed. In the oxidized
enzyme, no cyanide binding was detected by ESEEM. Modulation amplitude analysis of
the reduced and the reduced CcO plus cyanide indicated that a drop of about 9% in
intensity at deuteron modulation was caused by the binding of the cyanide. With the help
of the crystal structures, it is calculated that if one of three water ligands is a hydroxide,
then ﬁve inner sphere deuterons are expected. If the hydroxide is replaced by a cyanide, a
decrease of one out of ﬁve deuterons is expected, corresponding to a 9.5% decrease in
signal intensity. The closeness between the experimental data and the theoretical
calculation make it reasonable to conclude that cyanide replaces a hydroxide ligand to

Mn in the reduced enzyme, but cannot displace the aqueous ligand present in the oxidized

enzyme.

ESEEM data indicates that the difference in modulation amplitude is due to the
displacement of a hydroxide ion by a cyanide ion, and supports the hypothesis that the
reduction of the Cu caused one of the water ligands of the Mn to become a hydroxide
ion. The cycling between water and hydroxide ion of Mg water ligand could play an
important role in proton transfer upon change in electron density on Cut. A role for the

CuA/Mg site in proton movements could represent an important part of the proton exit

140

mechanism. Further experiments are needed to better explore this interesting possibility,
proposed by Dr. Sharpe, M. et al. (paper submitted)

Finally, the simulation we used to help determine the number of water, turned out to
be more complicated than we thought. The simulated spectrum successﬁrlly matches in
lineshape with the experimental data but did not exactly match the intensity of the
modulation (Figure 4.10). This discrepancy in amplitude could be explained in several
reasons: ﬁrst, the way we ratio the data may cause the loss of some spectroscopic
information although it effectively simpliﬁes the spectrum; second, according to our
study, the six inner sphere deuterons in the simulated spectrum only represents about
70% of the echo amplitude and the contributions ﬁ'om out sphere deuterons were not
considered in the simulation; third, to simulate such complicated spin systems as Mn2+ is
not a trivia task: assumptions and simpliﬁcations in the simulation procedure make it

inevitable that the simulation is only an approximation.

Interestingly, the simulation program we used to monitor the change in modulation
amplitude upon the change in coupled nuclei provided us with good results. It is found
that when the number of coupled deuterons changed from ﬁve to four, the modulation
amplitude decreased by 9.5%, which agrees well with experimental data, in which the
decrease is 9.0%. When the number of deuterons changed from six to ﬁve, the
modulation amplitude decreased by 5.3%. A total change in modulation amplitude was
observed to be 15.3% for deuterons change from six to four. The agreement of simulated

data with experimental data proved to be useful in estimating the change in relative

141

numbers of coupled nuclei during the reaction and may be a good tool to help us

understand better the proton movements.

Note: Previous work done by Kass et al.19 indicated the CuA in the oxidized enzyme
dipolar coupled with the Mn site and affected the EPR characteristics of Mn site. Because
of this spin-spin interaction, the ligand ﬁeld of Mn is somewhat altered. This interaction
may affect the deuterium modulation in such a complicated manner that it is not very
useful to compare the deuterium echo amplitude in oxidized and reduced enzyme in
ESEEM experiments (see Figure 4.11). This explained why we carry out more
complicated cyanide experiments in the reduced enzyme instead of just comparing the

oxidized and reduced enzyme to test the hypothesis.

142

 

 

 

 

 

 

 

 

 

6. l
. ——4 deuterons
— 5 deuterons
5 _
q, — 6 deuterons
U l — C00 in 75% 020
3 4 '1
:2:
a. ‘ A
E 3 "
< ,
2 4
1 _
A Avith "NAAh-W L‘AAAL_ A---A—A.AM
O f I ' I ' T l I 'T'l V7 _" V I: 1' : ' II
0 5 1o 15 20 25 30

Frequency(MHz)

Figure 4.10: Comparison of the simulated and the experimental data. The
experimental data is the CcO in 75%(v/v) szO. The sample and spectrometer conditions
are depicted in ﬁgure 4.2

143

I FQEH§LHJBCICDCKD [h2(3VF12C)

1°? - Reduced CcO + CN 020/1120
1 A Oxidized CcO 020/1120
3 .l
s ..
:3
1‘: i A
a. 4 _ '
E
<
2-1
A
.l

 

 

ﬂ I j l V l ' 1 ' I 1 l ' l

I I r I
0 10 20 30 40 50 60 70 80

D20%

Figure 4.11: Plots of the amplitude at 2.42MHz versus 2H20 concentration for the
oxidized CcO ratio, reduced CcO ratio and the reduced CcO plus cyanide ratio

144

REFERENCES:

1.Espe, M. P.; Hosler, J. P.; Ferguson-Miller, S.; Babcock, G. T.; McCracken, J., A
Continuous Wave and Pulsed EPR Characterization of the Mn2+ Binding Site in
Rhodobacter sphaeroides Cytochrome c Oxidase. Biochemistry 1995, 34, 7593-7602.

2.Florens, L.; Schmidt, 3.; McCracken, J.; Ferguson-Miller, 8., Fast deuterium access to
the buried magnesium/manganese site in cytochrome c oxidase. Biochemistry 2001, 40,
(25), 7491-7497.

3.Astashkin, A. V.; Raitsirnring, A. M., Electron spin echo envelope modulation theory
for high electron spin systems in weak crystal ﬁeld. Journal of Chemical Physics 2002,
117, (13), 6121-6132.

4.Coffmo, A. R.; Peisach, J ., Nuclear Modulation Effects in High-Spin Electron-Systems
with Small Zero-Field Splittings. Journal of Chemical Physics 1992, 97, (5), 3072-3091.

5.Larsen, R. G.; Halkides, C. J.; Singel, D. J., A Geometric Representation of Nuclear
Modulation Effects - the Effects of High Electron-Spin Multiplicity on the Electron-Spin
Echo Envelope Modulation Spectra of Mn-2+ Complexes of N-Ras P21. Journal of
Chemical Physics 1993, 98, (9), 6704-6721.

6.Hosler, J. P.; Espe, M. P.; Zhen, Y.; Babcock, G. T.; Ferguson-Miller, 8., Analysis of
Site-Directed Mutants Locates a Non-Redox-Active Metal near the Active Site of
Cytochrome c Oxidase of Rhodobacter sphaeroides. Biochemistry 1995, 34, 7586-7592.

7.Hiser, C.; Mills, D. A.; Schall, M.; Ferguson-Miller, S., C-terminal truncation and
histidine-tagging of cytochrome c oxidase subunit II reveals the native processing site,
shows involvemet of the C-terminus in cytochrome c binding, and imporves the assay for
proton pumping. Biochemistry 2001, 40, 1606-1615.

8.Mims, W. B.; Davis, J. L.; Peisach, J., The Exchange of Hydrogen-Ions and of Water-
Molecules near the Active-Site of Cytochrome-C. Journal of Magnetic Resonance 1990,
86, (2), 273-292.

9.Mims, W. B., Envelope Modulation in Spin-Echo Experiments. Physical Review B
1972, 5, (7), 2409-&.

10.Halkides, C. J.; Farrar, C. T.; Singel, D. J., The effects of cryoprotection on the
structure and activity of p21 ras: Implications for electron spin-echo envelope modulation
spectroscopy. Journal of Magnetic Resonance 1998, 134, (1), 142-153.

11.Hoogstraten, C. G.; Britt, R. D., Water counting: Quantitating the hydration level of

paramagnetic metal ions bound to nucleotides and nucleic acids. Rna-a Publication of the
Rna Society 2002, 8, (2), 252-260.

145

12.Mims, W. B.; Davis, J. L.; Peisach, J., The accessibility of type I copper(II) centers in
laccase, azurin, and stellacyanin to exchangeable hydrogen and ambient water.
Biophysical Journal 1984, 45, 755-766.

13.Mccracken, J.; Peisach, J.; Dooley, D. M., Cu(Ii) Coordination Chemistry of Amine
Oxidases - Pulsed Electron-Paramagnetic-Res Studies of Histidine Imidazole, Water, and

Exogenous Ligand Coordination. Journal of the American Chemical Society 1987, 109,
(13), 4064-4072.

14.Hansen, A. P.; Britt, R. D.; Klein, M. P.; Bender, C. J.; Babcock, G. T., Endor and
Eseem Studies of Cytochrome-C-Oxidase - Evidence for Exchangeable Protons at the
Cu(a) Site. Biochemistry 1993, 32, (49), 13718-13724.

15.Tsukihara, T.; Aoyama, H.; Yamashita, E.; Tomizaki, T.; Yamaguchi, H.; Shinzawa-
Itoh, K.; Nakashima, R.; Yaono, R.; Yoshikawa, S., The Whole Structure of the 13-
Subunit Oxidized Cytochrome c Oxidase at 2.8 A. Science 1996, 272, 1136-1144.

16.Schmidt, B.; Hillier, W.; McCracken, J.; Ferguson-Miller, S., The use of stable
isotopes and spectroscopy to investigate the energy transducing protein cytochrome c
oxidase. Biochimica et Biophysica Acta 2003, In press.

l7.Shinzawa-Itoh, K.; Aoyama, H.; Muramoto, K; Terada, H.; Kurauchi, T.; Tadehara,
Y.; Yamasaki, A.; Sugimura, T.; Kurono, S.; Tsujimoto, K.; Mizushima, T.; Yamashita,
E.; T sukihara, T.; Yoshikawa, 8., Structures and physiological roles of 13 integral lipids
of bovine heart cytochrome c oxidase. Embo Journal 2007, 26, (6), 1713-1725.

18.Muramoto, K; Hirata, K.; Shinzawa-Itoh, K.; Yoko-O, S.; Yamashita, E.; Aoyama,
H.; Tsukihara, T.; Yoshikawa, S., A histidine residue acting as a controlling site for
dioxygen reduction and proton pumping by cytochrome c oxidase. Proceedings of the
National Academy of Sciences of the United States of America 2007, 104, (19), 7881-
7886.

l9.Kass, H.; MacMillan, F.; Ludwig, B.; Prisner, T. F., Investigation of the Mn binding

site in cytochrome c oxidase from Paracoccus denitriﬁcans by high-frequency EPR.
Journal of Physical Chemistry B 2000, 104, (22), 5362-5371.

146

CHAPTER 5

HIGH FIELD/FREQUENCY cw EPR STUDY OF Mn
(11) IN CYTOCHROME C OXIDASE

147

1.Introduction

High spin mononuclear Mn2+ sites are found in many active sites of
metalloenzymes. The most commonly used method to study Mn complexes is by EPR
spectroscopy, since the EPR properties of Mn2+ are governed by the shape of the
magnetization in a given coordination environment. It is known that the shape of the
magnetization is determined by the distribution of unpaired electrons around the Mn
nucleus, and this distribution is related to the types of ligands present and their
geometrical arrangement about the ion. The interpretation of the structural message

encoded in the EPR spectrum is a major part of this study.

High spin Mn ion (d5) is characterized by an electron spin S=5/2 and nuclear spin
I=5/2. Its electronic properties can be described by the following spin Hamiltonian:
He 1.3.1149 +A1as + D [sf-1356+ 1) ] + E (5.2-5,?) 5.1
where the ﬁrst two terms represent the electron Zeeman and the isotropic electron-nuclear
hyperﬁne interaction, respectively, whereas the last two terms are the zero-ﬁeld splitting
(zfs) interaction. The zfs parameters D, and E, provide insight into the symmetry of
Mn2+coordination. Larger D and E parameters means lower symmetry. For Mn (S=5/2),
there are ﬁve ﬁne structure transitions arising from transitions between Ms= iS/Z, i3/2,
and il/2. Each of these transitions is split into six lines ﬁom hyperﬁne coupling to the
Mn nucleus (I=5/2). The total allowed transitions is thirty. The energies of Ms= i5/2,
i3/2 are dependent on D and E to the ﬁrst order and are angle-dependent; therefore, in an

orientationally-disordered system, the intensities of the EPR transitions from Ms= :5/2,

148

i3/2 are weak and hard to observe. By contrast, the energies of Ms= i1/2 have a very
small angular dependence, thus it has more intense EPR signals. In CcO, most of the EPR
signals are from Ms= il/Z transitions. In addition to allowed EPR transitions, there are
some forbidden (Am; = i l) EPR transitions. These forbidden transitions have less
intensity and depend inversely on the magnetic ﬁeld. The majority of the EPR work on
Mn complexes and enzymes has been done with the X-band (9.5GHz) or Q-band
(35GHz) frequencies. But using EPR with a higher magnetic ﬁeld strength will greatly

simplify the EPR spectrum and improve the resolution.

A non-redox-active Mg site is located at an aqueous region, observed in the crystal
structure], which suggests that it may be involved in proton/water exit. The substitution
of Mg by Mn does not affect its function but opens a door for the investigation of this site
by EPRZ. Previously, studies of the Mn site in CcO provided some intriguing results. In
the deuterium-exchange experiments3, the results indicated that the ligands of the Mg/Mn
site in CcO could exchange with the bulk solvent at a fast rate (>3000s'1). In a H2170
exchange experiment, Mn cwEPR spectra were broadened upon water exchange.
Additional evidence of 170 presence near the Mn site is given by the observation of the
17O Larmour frequency in the ESEEM spectrum, during the catalytic reaction, with
limited turnover in the rapid ﬁeeze quench apparatus, with 1702 as the substrate instead of
regular oxygen gas. The observed line-broadening of the Mn EPR spectrum indicated that
the product waters were bound to the Mn site“. It was noticed that the oxidation state of
CuA was not indicated in the turnover experiment, although it is most likely in its

oxidized form. EPR experiments showed that the oxidation state of Cu», slightly affects

149

the Mn spectra in the literatures. The difference in spectra was interpreted as a
conformational change which occurred in CcO near the Mg site upon a change in the
oxidation state of CuA6. Investigations of the Mn/Mg site in Rs.CcO by cw EPR and
ESEEM concluded that the difference in the Mn EPR spectra was so small that the
geometry of Mn was barely altered by the change in oxidation states of Cqu’ 7. In 2000,
Prisner and coworkers reinvestigated the Mn site in Pd. CcO using high-ﬁeld EPR (W-
band, 95GHz) and concluded that the splitting of the Mn sextet spectrum was due to a
dipolar coupling interaction present between CuA and Mn and this interaction vanished
when the Cut site was reduced. The difference in the Mn EPR spectra was attributed to
the dipolar coupling interaction between two spin centerss. Now, with the high ﬁeld (HF)
EPR machine, the Mn site in oxidized and reduced CcO is reinvestigated to see if similar
difference in spectra could also be observed in Rs. CcO. The possible cause behind the
difference in the Mn spectra is also discussed. Besides exploring the origin of the
observed difference in CcO under oxidized and reduced conditions, it is possrble that it
could be applied to isotope-labeled water-exchange experiments, using HF-EPR to study

the ligation environment of Mn in CcO.

The hyperﬁne coupling ﬁom [70 (1= 5/2) has been used in EPR investigations of
many metalloenzymes to identify ligands for the paramagnetic metal center9. Reed and
coworkers‘o’ 1' have developed a useﬁil scheme to determine the number of inner sphere
waters of the Mn2+ ion by an EPR technique. The method relies on the detection of
hyperﬁne coupling between the unpaired electron spins of Mn2+ and the magnetically

active nuclei of ligand atoms. The magnitude of the hyperﬁne coupling between Mn2+

150

and 17O is smaller compared to the intrinsic line widths caused by other sources of
broadening—primarily unresolved 1H hyperﬁne coupling interactions and dipolar
coupling from CuA in the oxidized enzyme8 and second-order ﬁne structure broadening
due to zero-ﬁeld splitting. As a result, the broadening from 17O hyperﬁne coupling to the
EPR signals is difﬁcult to resolve, especially at lower ﬁeld such as X-band. Since the
second-order ﬁne structure broadening is inversely proportional to the ﬁeld strength, Bo,
while hyperﬁne broadening is independent of B0, it is obviously more advantageous to do

l7O-labeling experiments at higher ﬁeld.

Advances in high-ﬁeld EPR enable the investigation of paramagnetic metal ions
with S 21 and I 21, such as Mn”, for the following reasons: ﬁrst, a simpliﬁed spectrum
can be achieved due to the reduction of contrrbutions ﬁom second-order ﬁne structure
broadening; second, the intensities of forbidden transitions are reduced in high ﬁeld;
third, orientation selectivity can be improved in disordered systems; and fourth, more-
precise and more complete information about the system under study can be obtained

because of dramatic decrease in cross-relaxation of the paramagnetic center.

Here, in this chapter, the EPR spectra for CcO are used to determine the hydration
number of Mn2+ by 17O hyperﬁne coupling interactions. Protein samples including:
oxidized state, and dithionite-reduced state, in different solvents with 2H and ”O-

enrichment were made and studied under W-band (95 GHz) EPR.

2. Experimental methods

151

2.1 Protein production and puriﬁcation

The amount of Mn2+ incorporated into the protein depends on the [Mg] to [Mn]
ratio in the growth mediumz. The YZ-100 strain of R. sphaeroides, which over-expresses
wild-type CcO with a histidine-tag added to the C- terminus of subunit 1, was grown on
“high Mn” medium (with MnSO4 and MgSO4 at ﬁnal concentrations of 700uM and
SOuM, respectively). The protein was puriﬁed by Ni2+-NTA aﬂinity chromatography and
ﬁirther puriﬁed by DEAE ion-exchange chromatography” to remove excess subunit I
and the remaining contaminants. The procedure involves washing the protein with buffer
containing lmM EDTA to remove the adventitious free extrinsic Mn ions. The enzyme
was then concentrated and washed into the desired buﬁ‘er using a Millipore centrifugal
ﬁlter with a MW cut-off 100KDa. The visible spectral characteristics of Mg- and Mn-
substituted wild-type enzymes were identical and oxygen consumption activities were not

altered.

2.2 UV-visible spectroscopy

UV-visible spectra of as-puriﬁed and dithionite-reduced proteins were recorded on
a Perkin-Elmer Lamda 40P spectrophotometer scanning wavelengths from 400nm to
750nm after appropriate dilution into pH7.4 100mM HEPES, 0.1% lauryl maltoside.
From the visible spectrum, the puriﬁed oxidase is in its fully oxidized state (See spectrum

in chapter 3).

2.3 Sample preparation

152

Approximately 5011M, 200nL each of the oxidized, and the dithionite-reduced wild-
type Mn (11) substituted CcO was transferred to silica-ﬁrsed capillary sample tubes with
id. 0.40mm and od. 0.55mm (Bruker), which were sealed with silicone high vacuum
grease (Bruker). Another two aliquots of reduced enzyme were mixed with 100% 2H-
enriched and 40% 17O-enriched water (Cambridge Isotope Laboratories), resulting in
80% 2H and 30% 17O enrichment in the ﬁnal samples respectively. The samples were

then sealed again.

2.4 W-band EPR spectroscopy

W—band cw-EPR spectra were recorded using a commercial W-band EPR
spectrometer (Bruker Elexsys E680) equipped with a helium ﬂow Oxford CF93S
cryostat, a cylindrical Bruker Teraﬂex TE110 cavity, and a 6T Magnex superconducting
magnet. The microwave frequency was measured with the internal Bruker counter. Both
oxidized and reduced samples were investigated in a temperature range from 4.2 to 120K.
The EPR parameters were as follows: microwave frequency 94.0 GHz; microwave power
15.8uW; ﬁeld modulation amplitude 1G; modulation frequency lOKHz; time constant

327ms.

2.5 Strategy for determination of the number of inner sphere waters

Since the 17O has a spin of 5/2, the hyperﬁne coupling between H2170 ligand(s) and
the Mn2+ center would split each EPR transition into six approximately equally spaced
components, resulting in a broadening of the EPR spectrum. In order to extract the

coordination information out of the EPR spectrum, Reed and coworkers developed the

153

subtraction method‘o’ ”. The highest l7O-enriched water available is 40%. Therefore, the
signals Observed for a sample containing l7O-enriched ligand are a superposition of
spectra for enzyme with Mn coordinated to 17O and to other oxygen isotopes. The
experimental spectrum, Se, is a weighted sum of n+1 subspectra (assume there are n
spectroscopically equivalent H2170 ligands on the Mn2+ site). And the difference
spectrum, Sd, has the form of 8.3 = Se- FnSu (Su represents the spectrum without 170
enrichment and Fn is the adjustable factor between 0 and 1). By trying different Fn, a
series of difference spectra are obtained. Only the difference spectrum that does not have

extraneous features is useﬁil in providing the hydration number.

If the water ligand is exchanged to szO buffer, a narrowing spectrum will be
expected, due to the very small g. of deuterium. The narrowing effects can be potentially

used to determine the hydration level of Mn2+ too.

3. Results and discussion

3.1 EPR spectra of CcO in HzO

Fig.5.] shows the W-band spectra of oxidized and reduced CcO at 15K. The
reduced CcO exhibited a typical Mn2+ sextet hyperﬁne pattern centered at the g=2.0
region with an averaged splitting of 94.12G. For the oxidized CcO, each line of the sextet
was split up by more than 15G. Temperature-dependence experiments from 15K to 120K
were also performed. The reduced CcO did not show a temperature-dependent behavior,

but the oxidized CcO did show a dependent behavior (Fig.5.2). Below 40K, the EPR

154

60—

— OXIdIzed CcO
—— Reduced CcO

 

dX"/dH (an)

 

 

 

 

 

 

 

 

 

'60 I I I I 3
33.4 33.6 33.8 34.0x10
Magnetic Field (Gauss)

 

Figure 5.1: The HF-EPR spectra of the oxidized and the reduced wild type CcO at
15K. The spectrum of the oxidized CcO is shown in red and the spectrum of the reduced
CcO is shown in blue. The protein sample is in 20mM HEPES-KOH, 14mM KCl, pH
7.4, 0.1% lauryl maltoside with concentration of SOuM. The EPR conditions are:
microwave ﬁequency 94.16GHz; microwave power 15.8uW; ﬁeld modulation amplitude
1G; modulation frequency lOKHz; time constant 327ms.

155

60—

 

 

 

 

 

 

 

 

 

 

33.4 33.5 33.3 34.0x103
Magnetic Field (Gauss)

Figure 5.2: The temperature-dependence HF-EPR spectra of the oxidized wild type
CcO. The red line represents the spectrum of CcO at temperature 15K and the blue
represents the spectrum of CcO at temperature 120K. The sample and EPR conditions are
the same as depicted in Fig. 5.1.

156

spectrum showed the same splittings as observed at 15K. Between 50K and 80K, 3
change ﬁ'om a splitting doublet to a normal single sextet was observed. Above 90K, the

EPR spectrum tended to resemble the spectrum of the reduced CcO.

3.2 EPR spectra of the oxidized CcO in H20 and 75% szO

The oxidized CcO in H20 and 75% 2H20 were studied by HF-EPR at 15K. Both
EPR spectra showed the splitting of each peak of the sextet. If carefully examined, the
spectrum of the sample in 75% 2H20 buffer is slightly narrower, making the dipolar
coupling peak more pronounced. When the line widths of the two spectra were measured,
it was found that a 2Gauss narrowing in the line width of the 2H20 spectrum, which was
what we expected to observe if some of the coupled protons were exchanged into

deuterons (Fig. 5.3).

3.3 EPR spectra of the reduced CcO in 75% szO and 30% H2170

Fig.5.4 Shows spectra of the reduced CcO in water and in 30% H2170 buﬁ‘er at
15K. The two spectra look almost identical with a normal Mn2+ hyperﬁne pattern. Fig.5.5
depicts the spectra of reduced CcO in water and 2H20 at 15K. There was no change in the

EPR spectrum for Mn2+ upon szO exchange.

3.4 Discussion

157

40 -« — Oxrdrzed CcO In H2O

— OXIdIzed CcO In 020
20 n
.. 1‘
0 _..

-20 q

dX"/dH (a.u)

 

 

 

 

-40 -1

 

 

I

33.4 33.6 33.3 34.0x103
Manetic Field (Gauss)

Figure 5.3: The HF-EPR spectra of the oxidized CcO in H20 and 2H20 buffers at
15K. The samples and spectrometer conditions are the same as depicted in ﬁgure 5.1.

158

 

 

 

 

 

 

 

 

 

 

 

 

300 ‘ —-— Reduced CcO m H20
— Reduced CcO In 30% H2170
200 ~
’5 100 z
3
i‘i
:3 0 E l 1 1 If
S< l I‘ l I
U -100 - I
-200 a
'300 I I I I 3
33.4 33.6 33.8 34.0x10

Magnetic Field (Gauss)

FigureS.4: The HF-EPR spectra of the reduced CcO in H20 and H2170 buffers at
15K. The red line represents the spectrum of the enzyme in H20 and the blue represents
the spectrum of the enzyme in H21 0. The protein in 100mM HEPES-KOH, 14mM KC],
pH 7.4, 0.1% lauryl maltoside was ﬁrst reduced by 25mM dithionite and then diluted by
40% enriched HznO, resulting a protein sample with a ﬁnal concentration of 5011M and a
H2170 enrichment of 30% (v/v).The EPR conditions are the same as depicted in Fig. 5.1

159

150~

 

 

 

 

 

 

 

 

 

 

 

 

 

100 e — Reduced CcO In H2O
-— Reduced CcO In 020
A. 50 -
3.
3
I -_ ,
.2 0
5<
U -50 -
-1OO -
-150

33.4 33.6 33.3 34.0x103
Magnetic Field (Gauss)

Figure5.5: The HF-EPR spectra of the reduced CcO in H20 and 21120 buffers at
15K. The red line represents the spectrum of the enzyme in H20 and the blue represents
the spectrum of the enzyme in H20. The protein in 100mM HEPES-KOH, 14mM KCl,
pH 7.4, 0.1% lauryl maltoside was ﬁrst reduced by 25mM dithionite and then diluted by
2H20, resulting a protein sample with a ﬁnal concentration of 5011M and a H2170
enrichment of 80% (v/v). The EPR conditions are the same as depicted in Fig. 5.1.

160

The difference in the EPR spectra of Mn2+ for oxidized and reduced Pd. CcO was
ﬁrst reported by Seelig et al. in 19815. The change in the Mn2+ EPR signal that is
associated with the oxidation state of Cut, was studied by EPR. The changes were
interpreted as a rearrangement in the rhombic octahedral coordination environment of the
central Mn2+ atom, which is indicative of a redox-linked conformational transition in the
enzyme". Later studies of Rs. CcO, by cw and pulse EPR, indicated that the difference in
spectra between oxidized and reduced CcO, and therefore the change in geometry of the
Mn center upon reduction or oxidation was small]. The X-ray crystal structure of bovine
CcO revealed that CuA and Mg shared one ligand, G1u198" 13. This discovery led to an
investigation of the Mn site in Pd. CcO by cw X-, Q-, and W-band EPR methods in
20008. The difference in the Mn EPR spectra was interpreted as the dipolar coupling
interaction occurring between the CuA site and the Mn2+ site. Based on the evaluated
dipolar coupling constant, an average distance between these two metal sites was also

calculated which was in good agreement with the data from the crystal structure.

The W-band studies of Rs. CcO in oxidized and reduced states, gave similar results
to those observed in Pd. CcOs. The temperature-dependence behavior of the oxidized Rs.
CcO was also similar to that of the Pd. CcO. Based on all our observations, a conclusion
can be drawn that the splittings in the oxidized CcO are due to a dipolar coupling
interactions between the Cut center and the Mn center. Such a high sensitivity of the Mn
site to the change in oxidation state of CuA can be used as a probe in tracking charge

change during the electron transfer process.

161

The narrowing effect caused by 2H20 was observed in the oxidized CcO sample as
expected. The modest narrowing effect (about 2Gauss) was promising and prompted us
to ﬁirther experiment with reduced CcO which has narrower and simpler spectrum,

therefore, easier to observe these effects.

The 17O broadening of EPR spectrum of Mn”, due to the hyperﬁne coupling, was
used previously to identify ligands for the paramagnetic center and to characterize the
metal ligand bonding ”’ 14‘ ‘5. In CcO, previous data indicated that the Mg/Mn site might
play an important role as a water exit. In one experiment, 48%-enriched H2170 water was
rapidly mixed in a 1:1 ratio with the oxidized enzyme and then subjected to the X-band
cw EPR study. The observed broadening provided the evidence that the H2170 was in a
close position to Mn site. Experiments with the reduced CcO samples in a 30% 17O-
enrichment and 75% 2H-enrichment buffers respectively were performed in high ﬁeld
EPR in the hopes of getting hydration information about the Mn2+ site. Unfortunately, the
EPR data of CcO in 17O and 2H-enriched water buffers failed to show the predicted
broadening and narrowing effects, respectively. The unsuccessful trials could be due to
higher than we thought temperature, which ruins any solvent effects, let alone the subtle
hyperﬁne coupling effects between the Mn and the ligands. In this respect, temperature
control played a crucial role in this experiment and should be given more attention to in

the ﬁiture work.

162

REFERENCES:

l. Tsukihara, T.; Aoyama, H.; Yamashita, E.; Tomizaki, T.; Yamaguchi, H.;
Shinzawa-Itoh, K; Nakashima, R.; Yaono, R.; Yoshikawa, S., The Whole Structure of
the 13-Subunit Oxidized Cytochrome c Oxidase at 2.8 A. Science 1996, 272, 1136-1144.

2. Hosler, J. P.; Espe, M. P.; Zhen, Y.; Babcock, G. T.; Ferguson-Miller, 8.,
Analysis of Site-Directed Mutants Locates a Non-Redox-Active Metal near the Active
Site of Cytochrome c Oxidase of Rhodobacter sphaeroides. Biochemistry 1995, 34, 7586-
7592.

3. Florens, L.; Schmidt, B.; McCracken, J .; Ferguson-Miller, 8., Fast deuterium
access to the buried magnesium/manganese site in cytochrome c oxidase. Biochemistry
2001, 40, (25), 7491-7497.

4. Schmidt, B.; McCracken, J .; Ferguson-Miller, S., A discrete water exit pathway in
the membrane protein cytochrome c oxidase. Proc Natl Acad Sci U S A 2003, 100, (26),
15539-42.

5. Seelig, A.; Ludwig, B.; Seelig, J.; Schatz, G., Copper and Manganese Electron-
Spin Resonance Studies of Cytochrome-C Oxidase ﬁom Paracoccus-Denitriﬁcans.
Biochimica Et Biophysica Acta 1981, 636, (2), 162-167.

6. Haltia, T., Reduction of Cu Induces a Conformational Change in Cytochrome c
Oxidase from Paracoccus denitriﬁcans. Biochimica et Biophysica Acta 1992, 1098, 343-
350.

7. Espe, M. P.; Hosler, J. P.; Ferguson-Miller, S.; Babcock, G. T.; McCracken, J., A
Continuous Wave and Pulsed EPR Characterization of the Mn2+ Binding Site in
Rhodobacter sphaeroides Cytochrome c Oxidase. Biochemistry 1995, 34, 7593-7602.

8. Kass, H.; MacMillan, F.; Ludwig, B.; Prisner, T. F., Investigation of the Mn
binding site in cytochrome c oxidase from Paracoccus denitriﬁcans by high-frequency
EPR. Journal of Physical Chemistry B 2000, 104, (22), 5362-5371.

9. Gupta, R. K; Oesterling, R. M.; Mildvan, A. 8., Dual Divalent-Cation
Requirement for Activation of Pyruvate-Kinase - Essential Roles of Both Enzyme-Bound
and Nucleotide-Bound Metal-Ions. Biochemistry 1976, 15, (13), 2881-2887.

10. Lodato, D. T.; Reed, G. H., Structure of the Oxalate-Atp Complex with Pyruvate-
Kinase - Atp as a Bridging Ligand for the 2 Divalent-Cations. Biochemistry 1987, 26, (8),
2243-2250.

11. Reed, G. H.; Leyh, T. 8., Identiﬁcation of the 6 Ligands to Manganese(li) in

Transition-State-Analogue Complexes of Creatine-Kinase - O-17 Superhyperﬁne
Coupling from Selectively Labeled Ligands. Biochemistry 1980, 19, (24), 5472-5480.

164

12. Hiser, C.; Mills, D. A.; Schall, M.; Ferguson—Miller, S., C-terminal truncation and
histidine-tagging of cytochrome c oxidase subunit II reveals the native processing site,
shows involvemet of the C-terminus in cytochrome 0 binding, and imporves the assay for
proton pumping. Biochemistry 2001, 40, 1606-1615.

13. Tsukihara, T.; Aoyama, H.; Yamashita, E.; Tomizaki, T.; Yamaguchi, H.;
Shinzawa—itoh, K; Nakashima, R.; Yaono, R.; Yoshikawa, 8., Structure of Metal Sites of
Oxidized Bovine Heart cytochrome c Oxidase at 2.8 A. Science 1995, 269, 1069-1074.

14. Bellew, B. F.; Halkides, C. J.; Gerfen, G. J.; Grifﬁn, R. G.; Singel, D. J., High
frequency (139.5 GHz) electron paramagnetic resonance characterization of Mn(II)-
(H20)-O-17 interactions in GDP and GTP forms of p21 ras. Biochemistry 1996, 35, (37),
12186-12193.

15. Rohrer, M.; Prisner, T. F.; Brugmami, O.; Kass, H.; Spoerner, M.; Wittinghofer,
A.; Kalbitzer, H. R., Structure of the metal-water complex in Ras center dot GDP studied
by high-ﬁeld EPR spectroscopy and P-31 NMR spectroscopy. Biochemistry 2001, 40,
(7), 1884-1889.

165

   

lllililllﬂllllﬂlljljllllliljlllll