.21..
3...}..27.

r51}... .
32‘ .3...

 

.... A...
.. I . ; .
LP: ...) h
..

3.)!!- a...
1...: .12 0

i
f.

. w
x? . .}K.
9.33. I‘ 1.15.!1

1.. is)!

Mr,
‘I
‘1

“HOME

...—3.. llllllllllllllllllllllllllllllllIlll
3

 

lllll

 

 

 

 

3 1293 01410043

This is to certify that the

dissertation entitled

Search for Residues Critical to Proton

Pumping in Cytochrome g Oxidase

presented by

John R . Fetter

has been accepted towards fulfillment
of the requirements for

Ph.D. -degreein Biochemistry

 

 

< 3/ o.
2 a Major professor

Date 2’ 1

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

 

 

LIBRARY
Michigan State
University

 

 

 

PLACE II RETURN BOX to remove thie checkout from your record.

TO AVOID FINES return on or betore dete due.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU le An Afflrmetive Action/Equal Opportunity lnetituion
' ulna-u

 

SEARCH FOR RESIDUES CRITICAL TO PROTON
PUMPING IN CYTOCHROME c OXIDASE

By

John R. Fetter

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Biochemistry
1995

ABSTRACT

SEARCH FOR RESIDUES CRITICAL TO PROTON
PUMPING IN CYTOCHROME c OXIDASE

By

John R. Fetter

Cytochrome c oxidase from Rb. sphaeroides is a good model for the highly
homologous mammalian oxidase. A reconstitution and assay to look for proton
pumping in the bacterial oxidase showed that it could pump protons with
similar efficiency to that of the mammalian enzyme. Site-directed
mutagenesis was used to find residues involved in a proton pumping pathway.
Mutations were made in several conserved regions in Subunit I: in the 11-111
interior loop, Asp 132Asn/Ala; in transmembrane helix VIII, Pro358Ala,
Thr359Ala; in transmembrane helix X, Tyr422Phe/Ala; and in the exterior IX-X
100p, His411Ala, Asp412Asn, Thr413Asn, and Tyr414Phe. All showed
decreased electron transfer activity indicating that they could have a role in
proton pumping. Only Asp132Asn/Ala showed inhibited electron transfer and
complete inhibition of proton pumping, while retaining normal visible
absorbance and resonance Raman characteristics. When reconstituted it
showed inhibition instead of stimulation of electron transfer with release of the
pH gradient or electrical gradient. Electron transfer activity of solubilized
Asp 132Asn showed the same response to pH as wild type. Addition of free
fatty acids with long alkyl chains such as arachidonic acid to this mutant
stimulated electron transfer activity up to six-fold, but the corresponding
alcohols had no efl‘e'ct. Analogues of arachidonic acid showed less stimulation.

Stimulation by fatty acids was not seen on wild type or other mutants. It was
concluded that the fatty acid may be replacing the missing carboxyl to improve
proton access, or allowing protons to enter the enzyme through another path.
Four more mutants were made in the DI-X loop: His411Gln, Asp412Ala,
Thr413Ala, Tyr414Ala. Previously, His411 and Asp412 were shown to
contribute to binding a manganese or magnesium. His411Gln retained native
properties indicating that the Gln may act as a replacement ligand.

Asp412Ala showed lower electron transfer activity indicating it may have lost
the metal. Thr413Ala showed wild type characteristics. 'I‘yr414Ala showed a
spectral shift of the alpha band to 604 nm providing evidence that the shift to

610 nm in Tyr414Phe is caused by an aromatic interaction with heme a.

Parts of Chapter 4 are reprinted with permission from Proc. Natl. Acad.
Sci. USA 92, 1604- 1608 (1995); © 1995, the National Academy of Sciences of
the United States of America.

Dedicated to my wife

Suzanne

ACKNOWLEDGMENTS

I wish to thank the many scientists that I have had the opportunity to
work with during my graduate training. First and foremost is Dr. Shelagh
Ferguson-Miller who has provided much guidance and creativity in worh'ng on
this project. I also would like to thank those who served on my committee Dr.
Arnold Revzin, Dr. Loran Bieber, Dr. Rawle Hollingsworth, and Dr. Gerald
Babcock. A special thanks to Jerry for all the chemical insights he has given
to us biochemists at group meetings. I would next like to thank Jon Hosler and
Thomas Nilsson for the scientific wisdom that they have passed on to me.

I am grateful to our many collaborators including Dr. Robert Gennis who
allowed me to come to his lab and learn molecular biology, and especially to J efi'
Thomas who took the time to teach me those techniques. Likewise I enjoyed
my time in the lab of Dr. Peter Nicholls working with Dr. Martyn Sharpe
learning the techniques used to measure membrane potentials. I would like to
thank Jim Shapleigh and Arturo Garcia-Horsman for making some of the
oxidase mutants. I also greatly appreciated those who ran the resonance
Raman in Jerry's lab: Einhardt Schmidt and Michelle Pressler.

I also enjoyed the time spent with my fellow graduate students: Wendy,
Taha, Jie, Yuejun, Jung, Doug and also to the undergrads past and present:
Barb, Kelly, Heather, Ann and Giancarlo.

I am glad to have met others during my stay in Michigan: Joe and his
family, Dennis, Steve, and Anita and Fr. Bill and to Suzie's family Bruce, Mike,
Liz, Justin, Jennifer, Kevin and Mom and Dad.

I am especially grateful to my parents for their patience and guidance
while I was growing up. I am glad to have two kind sisters Brenda and Joyce
along with two great brother-in-laws Doug and Keith.

A special thanks goes to my loving wife for her patience while I worked
on this thesis, and the help that she gave me, especially her photocopying
expertise.

TABLE OF CONTENTS

List of Tables ........................................................................................................... xi
List of Figures ......................................................................................................... xii
List of Abbreviations ............................................................................................. xiv
Chapter 1 - Background on cytochrome oxidase: a complex
mitochondrial enzyme and many simpler bacterial forms ............................ 1
Introduction ................................................................................................. 1
Role of the three core subunits of the oxidase ...................................... 2
Subunit I: the active site where oxygen is reduced;
likely site of the proton pump ...................................................... 2
Subunit II: the site of cytochrome c binding and
electron entry .................................................................................. 6
Subunit III: a structurally important element ....................... 8
Other subunits in the eukaryotic enzymes ........................................... 9
Overview of the oxidase reaction ............................................................. 9
Chemistry of the oxygen reaction ........................................................... 1 1
Many bacterial oxidases ........................................................................... 15
Evolutionary relationships ....................................................................... 19
Chapter 2 — Finding amino acid residues important for proton
pumping in cytochrome c oxidase ........................................................................ 2 2
Dificulties in understanding proton pumping ....................................... 22
Proton pump versus Q loop ...................................................................... 23
Theoretical aspects of proton pumping ................................................. 25
General requirements of a proton pump .................................... 25
General models for proton movement through
proteins ............................................................................................. 2 5
Continuous chains of hydrogen-bonds as
proton wires ......................................................................... 25
Proton movement through gramicidin ........................... 26
Integral injector model as a driving force ....................... 2 7
Other enzymes that transfer protons .................................................... 28
Reaction center ............................................................................... 29
F1Fo ATP synthase ...................................................................... 34
Bacteriorhodopsin .......................................................................... 35
Models which have been proposed for proton pumping in
cytochrome c oxidase ................................................................................. 42

Woodrufl' ligand shuttle model ...................................................... 43

Rousseau model requiring proximal ligand exchange

of a tyrosine ..................................................................................... 44

Chan model of ligand exchange at CuA or CuB ........................ 44

Indirect coupling of electron transfer and proton

pumping combined with a peroxide pump ................................. 45

Histidine Cycle ................................................................................ 46
Conclusions .................................................................................................. 49
Finding residues in oxidase involved in proton pumping ..................... 50

Chapter 3— Development and refinement of a reconstitution and
proton pumping assay for cytochrome c oxidase from Rhodobacter

sphaeroides ..................... _ .......................................................................................... 5 5
Introduction ................................................................................................. 5 5
Experimental procedures .......................................................................... 5 7

Materials .......................................................................................... 57
Purification of asolectin ................................................................. 5 7
Purification of cholate .................................................................... 58
Growth of bacteria and oxidase purification ............................ 58
Reconstitution ................................................................................. 59
Orientation assay ........................................................................... 6O
Proton pumping assay .................................................................. 60
Results .......................................................................................................... 6 1
Isosbestic points of phenol red and cytochrome c .................... 6 l
Controlling the pH for the proton pumping assay ................... 62
Determination of oxidase orientation after
reconstitution .................................................................................. 7 1
Respiratory control of the reconstituted oxidase ..................... 7 4
Discussion .................................................................................................... 7 5
Enzyme purity ................................................................................ 7 8
Detergent purity ............................................................................. 78
Phospholipid purity ........................................................................ 79
Accuracy of standardization ........................................................ 7 9
Presolubilization of the oxidase with cholate ............................ 80
Enzyme to lipid ratio ...................................................................... 80
pH of the HEPES/cholate bufi'er ................................................. 8 1
Concentration of the enzyme in the reconstitution ................. 8 1

Chapter 4 — Possible proton relay pathways in cytochrome c

oxidase ...................................................................................................................... 83
Introduction ................................................................................................. 83
Experimental procedures .......................................................................... 86
Results .......................................................................................................... 87

Electron transfer, spectral, and structural

properties ......................................................................................... 87

Reconstitution and respiratory control ...................................... 97

Proton pumping .............................................................................. 109
Discussion .................................................................................................... 115

General Considerations ................................................................. 1 15

Rationale for mutant selection .................................................... 1 16

HelixX. .............................................................................................. 117

IX-X Loop ......................................................................................... 1 1 7
Helix VIII ......................................................................................... 1 18
II-III loop .......................................................................................... 1 18
Conclusions ...................................................................................... 120

Chapter 5 — Measurement of membrane potential generation by

Asp132Ala and effects of free fatty acids on its activity ............................... 12 1
Introduction ................................................................................................. 1 2 1
Experimental procedures .......................................................................... 1 22

Determination of membrane potential ...................................... 122 ,

Assay of activity with fatty acids, fatty alcohols,

and analogues of arachidonic acid ............................................... 122
Results .......................................................................................................... 123

Determination of membrane potential of
reconstituted Asp132Ala, wild type, and beef heart

oxidases ............................................................................................ 123
Stimulation of Asp 132 mutants by free fatty acids ............... 123
Importance of the fatty acid carboxyl and the chain
length on stimulation of Asp 132Ala ........................................... 128
Stimulation of activity by fatty acids is unique to
the Asp132 mutants ..................................................................... 129
Stimulation of Asp132Ala by analogues of
arachidonic acid .............................................................................. 129
Restoration of reconstituted Asp 132Ala respiratory
control by addition of fatty acids ................................................. 134
Discussion .................................................................................................... 135
A nonphysiological effect of the fatty acids .............................. 140
Could there be a physiological explanation for the
stimulation of Asp132 mutants by fatty acids? ...................... 141
Reverse respiratory control ......................................................... 142
Chapter 6 — Further mutagenesis of residues in the IX-X Loop ................... 144
Introduction ................................................................................................. 144
Experimental procedures .......................................................................... 145
Materials .......................................................................................... 146
Mutagenesis and subcloning ........................................................ 146
Conjugation. ..................................................................................... 147
Small scale membrane preparations ......................................... 150
Results .......................................................................................................... 150
Discussion .................................................................................................... 156
Summary and Perspectives ................................................................................. 16 1
Appendix: publications and abstracts ............................................................... 165
Bibliography ............................................................................................................ 167

List of Tables

1.1 Variation in substrate specificity and heme type of the
bacterial oxidases ................................................................................ 1 7

4.1 Comparison of mutant and wild type oxidases ............................ 89

4.2 Reconstituted orientation assay of wild type, beef
heart, and several mutants .............................................................. 103

5.1 Fatty acid stimulation of Asp132Ala activity
increased with longer alkyl chains, while fatty alcohols
had no efl'ect ......................................................................................... 13 1

5.2 Stimulation of activity by arachidonic acid is unique to _
Asp132Ala ............................................................................................ 133

5.3 Stimulation of Asp132Ala activity with analogues of
arachidonic acid ................................................................................... 137

6.1 Visible spectral shifts and activity of the IX-X loop
mutants ................................................................................................ 158

List of Figures

1.1 Helical wheel model of the twelve predicted
transmembrane helices of subunit I from the Rb.

sphaeroides cytochrome c orn'dase .................................................... 5
1.2 Pathway for the reduction of oxygen to water in

cytochrome c oxidase.... ...................................................................... 14
2.1 Model of the QB site in reaction center .......................................... 33
2.2 Photocycle of bacteriorhodopsin ...................................................... 39

2.3 Model of the predicted hydrogen-bonded network that
develops in bacteriorhodopsin during the M4 12
intermediate ......................................................................................... 4 1

2.4 Condensed version of the histidine cycle showing the
changes that would occur at the binuclear center to

drive proton pumping ......................................................................... 48
2.5 Topological model of subunit I of the aa3-type orn'dase

from Rb. sphaeroides .......................................................................... 54
3.1 Spectral components in proton pumping assay ........................... 64

3.2 Equilibration of vesicles in proton pumping bufi‘er and
adjustment of pH to preset baseline ............................................... 6 7

3.3 Proton pumping of the wild type enzyme from
Rhodobactersphaemides .................................................................... 70

3.4 Determination of percentage of reconstituted wild type
cytochrome c orn'dase with its cytochrome c binding site
facing outward ..................................................................................... 7 3

3.5 Summary of some variables to consider during the
reconstitution of the Rhodobacter sphaeroides enzyme .............. 7 7

4.1 A model of the active site region of cytochrome c
oxidase indicating residues analyzed by mutagenesis ................. 8 5

4.2 Visible and resonance Raman spectra indicating a
native active site of purified Asp132Asn-cytochrome c
oxidase ................................................................................................... 92

4.3 Asp132Asn shows lower CO binding than wild type ................... 94

4.4 Assay of the heme A content of Asp 132Asn using
pyridine hemochromagen .................................................................. 96

4.5 Assaying Asp132Asn mutant with catalase fails to
show production of H202 .................................................................. 99

4.6 Determination of percentage of reconstituted
Asp 132Asn with its cytochrome 0 binding site facing
outward ................................................................................................. 1 O 1

4.7 Effects of valinomycin and CCCP on the oxygen
consumption of reconstituted wild type and mutant
Asp132Asn cytochrome oxidase ...................................................... 106

4.8 A comparison of the efl‘ect of pH on the activity of
purified Asp132Asn and wild type cytochrome c
orddase ................................................................................................... 108

4.9 Tyr422Phe shows normal proton pumping ................................... 1 1 1

4.10 Spectral analysis of proton pumping and cytochrome
c oxidation by reconstituted wild type and mutant
cytochrome c on’dases ........................................................................ 1 13

5.1 Reconstituted Asp132Ala shows a normal membrane
potential ................................................................................................ 125

5.2 Arachidonic acid stimulation of the Asp132Ala
mutant .................................................................................................. 127

5.3 Restoration of respiratory control by the addition of
arachidonic acid ................................................................................... 1 39

6.1 Site-directed mutagenesis and subcloning to change
residues in subunit I of Rb. sphaeroides .......................................... 149

6.2 Spectra of IX-X loop mutants in detergent-solubilized
membranes .......................................................................................... 1 53

6.3 Spectra of the purified IX-X loop mutants compared to
wild type ................................................................................................ 155

A.U.

bR
CCCP
DCCD
DEAE
disc-3,5
EPR
ESEEM
FPLC
FTIR

H+/e'

HEPES

His-tag

Im

RCR
SDS
TRIS
TMPD
ApH
AW

List of Abbreviations

absorbance units

bacteriorhodopsin

carbonyl cyanide m-chlorophenylhydrazone
N,N'-dicyclohexylcarbodiimide
diethylaminoethyl
3',3'-dipropylthiadicarbocyanine

electron paramagnetic resonance

electron spin echo envelope modulation

fast protein liquid chromatography

Fourier transform infrared spectroscopy

ratio of protons pumped per electron
transferred from cytochrome 0 into
cytochrome oxidase

4-(2-hydroxyethyl)-1-piperazine-
ethanesulfonic acid

histidines that are genetically attached to
an enzyme to allow purification by a nickel
amnity column

imidazole

ubiquinone

respiratory control ratio

sodium dodecyl sulfate
Tris(hydroxymethyl)aminomethane
N,N,N',N'-tetramethyl-p-phenylenediamine
transmembrane pH gradient
transmembrane electrical gradient

xiv

Chapter 1 - Background on cytochrome oxidase: a
complex mitochondrial enzyme and many simpler
bacterial forms

Introduction

Cytochrome oxidases participate along with other respiratory enzymes
in the conversion of energy from exergonic electron transfer events into a
transmembrane proton gradient. The respiratory enzymes occur in the inner
mitochondrial membrane of eukaryotes, or in the inner membrane of
prokaryotes. The proton gradient can be used directly to drive the transport of
ions across the membrane or for mechanical work such as driving flagellar
motion in bacteria. Another role of the proton gradient is to drive the
conversion of ADP to ATP by the F1Fo ATP synthase. When the energy is
stored in the form of ATP, it can easily difiuse throughout the cell to the site of
numerous energy requiring reactions. Cytochrome oxidase, as the final
electron acceptor, also has the unique role of catalyzing a reaction which
provides the final low potential sink for electrons, by reducing oxygen to water.
This reaction requires substrate protons that are derived from the
mitochondrial matrix, or interior of the bacterium, and thus contribute to the
formation of the gradient. At the same time oxidase uses the energy derived
from the thermodynamically favorable conversion of 02 to H20 to drive proton
pumping, further contributing to the gradient (For a textbook overview see
Nicholls and Ferguson 1992).

Much of the research on orn'dase has been done with the beef heart
enzyme which contains thirteen subunits (Kadenbach et al., 1983; Capaldi
1990b). Other oxidases such as yeast contain anywhere from 9 (Trawick et
al., 1992) to 12 (Taanman and Capaldi 1992) subunits and the plant enzyme

2
contains 8 (Peiffer et al., 1990) subunits. The discovery that bacterial

oxidases contain as few as three subunits that are highly homologous to the
largest subunits of the other enzymes, provides a simpler enzyme to work
with. Analysis of the aa3-type enzyme from Rhodobacter sphaeroides using
EPR, visible absorbance and resonance Raman shows it to have the same
metal centers as the eukaryotic enzyme. The enzyme also has high electron
transfer activity, and normal proton pumping. These results give strong
evidence that only the three subunit core is required for cytochrome c binding,
electron transfer, and oxygen reduction (Hosler et al., 1992). It has been
possible to make deletions of the genes for the subunits of bacterial oxidases,
and to reintroduce the genes on plasmids, making possible the creation of site-
directed mutants. This has been done in Rb. sphaeroides (Cao et al., 1991; Cao
et al., 1992; Shapleigh and Gennis 1992), and E. coli (Chepuri and Gennis
1990). Extensive site-directed mutagenesis and spectroscopic
characterization has been done on subunit I allowing the definition of the active
site of the enzyme in more detail than previously possible.

Only low resolution crystal structures have been published of the beef
heart enzyme, and these do not provide information about the location of
predicted transmembrane helices or the metal centers (V alpuesta et al., 1990).
However, recently detailed crystal structures have been determined of the
four subunit Paracoccus Denitrificans aa3 (Personal communication: Hartmut

Michel), and of the beef heart oxidase (Y oshikawa, personal communication).

Role of the three core subunits of the oxidase

Subunit I: the active site where oxygen is reduced; likely site of
the proton pump This subunit contains three metal centers including heme
a which accepts electrons from CuA in subunit II, and heme a3 and CuB where

3
oxygen reduction occurs. In Rb. sphaeroides it has a molecular weight of 63

kDa (Hosler et al. , 1992), with hydropathy analysis and phoA/lacZ fusions
predicting twelve transmembrane helices (Chepuri and Gennis 1990). The
placement of the metal centers has become much clearer through
mutagenesis of the proposed ligands, six totally conserved histidines, giving a
more detailed model of the subunit (See Figure 1.1). Individual mutation of any
four of the histidines 284, 333, 334, and 419 leaves heme a intact. Mutants of
Hi3102 and His421 lose the heme a, leading to the assignment of those two
histidines as its ligands. Since mutation of His333 to asparagine leaves the
heme a3 iron-his bond unaltered, it must be concluded to be a CuB ligand; the
adjacent His334 must then also be a CuB ligand, since if it were the heme a3
ligand it would put the as and Cu]; in a sterically untenable position (Lemieux
et al., 1992; Minagawa et al., 1992; Shapleigh et al., 1992). Initial studies were
unclear about whether His284 or His419 was the ligand to heme (13. However
more recent results indicate that mutation of Hi3284 to alanine leaves at least
some heme a3 iron-his bond in a native state eliminating this residue as a
heme a3 ligand and suggesting its role as a CuB ligand. In contrast, since all
mutations of His 419 disrupt the heme (13, it is proposed that this residue is
the iron ligand of heme a3 (Calhoun et al., 1993). These assignments have now
been confirmed by the crystal structure.

Other mutations have provided further information about the structure
of subunit I and will be discussed in more detail in latter chapters of this thesis.
Data show that the external loop between helices D( and X lays across the top
of the binuclear center, since mutations in that region have some effects on
both heme a and heme a3/CuB (Hosler et al., 1994a; Hosler et al. , 1994b).
Mutations in the highly conserved hydrophilic face of helix VIII greatly or
completely inhibit the electron transfer activity indicating that the helix

Figure 1.1 Helical wheel model of the twelve predicted transmembrane
helices of subunit I from the Rb. sphaeroides cytochrome c oxidase.
The indicated ligands to the metal centers are all histidines with
Hile2 and His421 acting as heme a ligands. His419 is the heme
a3 ligand, and His333, His334 and Hi8284 are shown as the Cu];

ligands (Calhoun et al., 1993; Hosler et al., 1993).

nan

Mfleo‘e 6 e a a .30
@/ We

9

o 0‘0 . ......
o o c a

 

6
lies near the binuclear center (Hosler et al., in press). A mutation of the highly

conserved aspartic acid in the loop between helices II and III inhibits electron
transfer and proton pumping but has no effect on the visible and resonance
Raman spectra indicating an important role in proton translocation but from a
position that is not adjacent to the binuclear center (Fetter et al., 1995). This
confirms the idea that subunit I may play an important role in proton pumping
implying fairly direct coupling of proton translocation to the active site
reactions.

The sum of the results of the mutational analysis leads to an ordered
placement of at least 10 of the 12 helices of subunit I surrounding the metal
centers, providing a useful semblance of a 3D model on which to base further
predictions. The distances between metal centers match reasonably well
those predicted from spectral analysis techniques (Trumpower and Gennis
1994). Thus this has become a generally accepted worln'ng model, giving
considerable impetus to mechanistic thinking and experimental design. The

accumulated results from much mutagenesis work on both the M3 oxidase
from Rb. sphaeroides and the homologous bo3 enzyme from E. coli have so far
been consistent with this model (Hosler et al., 1993).

Subunit II: the site of cytochrome c binding and electron entry
This subrmit with a molecular weight of 33 kDa in Rb. sphaeroides (Hosler et
al. , 1992) is proposed to contain 2 transmembrane helices and a soluble C-
terminal domain based on hydropathy analysis and betagalactosidase/alkaline
phosphatase (phoA/lacZ) gene fusion analysis (Chepuri and Gennis 1990).
Cytochrome c is the soluble electron carrier that delivers electrons to the
oxidase by first binding to a proposed site in the C-terminal domain. This
subunit contains several conserved carboxyl side chains which appear to
interact with conserved positively charged residues surrounding the exposed

7
heme edge of cytochrome c (Capaldi et al., 1982). Reaction of the oxidase with

carboxyl modifying reagents inhibits cytochrome c binding, while binding of
cytochrome c (Millett et al., 1983) or antibodies to subunit II protects against a
number of carboxyl modifications (Taha and Ferguson-Miller 1992). There is
also evidence for a second site of cytochrome c binding in subunit II which could
contribute to the biphasic steady state kinetics (Ferguson-Miller et al., 1978;
Taha and Ferguson-Miller 1992). An alternative explanation for the biphasic
kinetics does not require a second binding site: it is proposed that the enzyme
has two conformations that can accept electrons, only one of which can pump
protons (Brzezinski and Malmstriim 1986). More recent studies (Robinson
1995 Biochemistry, in press) lend credence to the former hypothesis.

Subunit II contains what was previously behaved to be a mononuclear
copper (Cup) in the C-terminal domain. Although the EPR signal from the
CuA center is different fi'om the blue copper proteins, the difference was
proposed to be caused by ligation of the copper by two cysteines and two
histidines (Stevens et al., 1982), instead of the single cysteine of the blue
coppers (Saraste 1990). Recent studies using S-band for EPR measurements
have found similarity of the Cup, site to a copper site in nitrous oxide
reductase, with both giving a seven-line hyperfine splitting that can be
explained by the presence of a binuclear copper (Kroneck et al., 1989;
Antholine et al., 1992). Other evidence also predicts a binuclear Cu in oxidase
including metal analysis showing 3 Cu/2 Fe (Bombelka et al., 1986; Stefi'ens et
al., 1987)and characterization of the binuclear copper in an isolated C-terminal
domain of P. denitrificans (Lappalainen et al., 1993). Again, recent
crystallographic studies confirm the 2 Cu center.

There has been much controversy over the years as to whether the

electrons from cytochrome c are transferred directly to CuA or to heme a or

8
both (Einarsdottir 1995) . Recent fast kinetic spectral studies suggest that
electron transfer occurs most rapidly between cytochrome c and Cup, with a
forward rate constant of 105/3, followed by transfer from On); to heme a with a
forward rate constant of 7,000/s (Hill 1991; Hill 1994). It has been possible to
measure values as high for electron transfer between cytochrome c and Cup,
using a photoactivatable ruthenium compound attached to cytochrome c (Pan
et al., 1993). Noting a conserved aromatic region in subunit II, possibly
between CuA and heme a, it was proposed that the aromatics might assist the
electron transfer between these centers (Stefi‘ens and Buse 1979). However,
in reaction center, aromatics do not seem to be critical for fast electron
transfer (Huber 1989). Initial site-directed mutagenesis efforts to remove one
of the copper ligands indicate that when all of the CuA is lost but the other
metal centers remain intact the enzyme retains 10% of the activity of the wild
type (Y uejun Zhen and Shelagh Ferguson-Miller: personal communication).
This result firrther supports the idea that under normal conditions the most
kinetically efficient site of entry of electrons is through CuA.

Subunit III: a structurally important element This subunit does
not contain any metal centers and in Rb. sphaeroides has a molecular weight of
30 kDa (Hosler et al. , 1993) and is predicted to contain 7 transmembrane
helices from hydropathy analysis and gene fusion experiments (Chepuri and
Gennis 1990). It had been considered for a possible role in proton pumping,
since DCCD was found to bind to a glutamate in subunit III and inhibit proton
pumping (Casey et al., 1980; Prochaska et al., 1981). However, subunit III-
depleted oxidases are still capable of pumping protons (Thompson and Gregory
1985; Prochaska and Pink 1987) and mutagenesis of the glutamate to which
DCCD binds in subunit III does not appear to inhibit proton pumping in intact
Paracoccus denitrificans (Haltia et al., 1991).

Other subunits in the eukaryotic enzymes

It has been shown that in yeast (Schatz and Mason 1974), Neurospora
crassa (Macino and Morelli 1983) and in mammals (Anderson et al., 1982), the
three-subunit core of the oxidase is encoded in the mitochondria. The other
subunits are imported into the mitochondria, being encoded by the nucleus
(Capaldi 1990a). Evidence exists that the smaller subunits are involved in
regulation of the enzyme (Kadenbach et al., 1991). It has been shown that
oxygen levels can effect the expression of subunit isoforms in Dictyostelium
discoideum (Bisson and Schiavo 1986; Bisson and Schiavo 1988) and in
yeast(Poyton et al., 1988). Comparison of the reconstituted bovine and P.
denitn'ficans enzymes showed that intraliposomal ATP could increase the Km
for cytochrome c and ADP decrease the Km for the bovine enzyme while no
efi'ect of these nucleotides could be shown on the bacterial enzyme(Hiither and
Kadenbach 1988). Phosphate (Edge and Kadenbach 1986; Malatesta et al.,
1987) and fi'ee fatty acids (Labonia et al., 1988; Thiel and Kadenbach 1989)
also have some modulatory effect on the bovine oxidase activity. Tissue
specific expression of isoforms of some of the subunits has been shown for the
mammalian oxidases and these appear to be species specific (Kadenbach et al.,
1991).

Overview of the oxidase reaction
Much of what is known about the mechanism of cytochrome c oxidase
was deciphered before the recent mutationally-derived structural model
through the use of various types of spectroscopy. The overall reaction for
oxidase is:
8H+in + 4e' + 02 —> 2H20 + 4H+out

1 0
Thus for each electron entering the enzyme, one proton is taken up from the

cytoplasm of bacteria or the matrix of mitochondria to permit reduction of
oxygen to water. There are four electrons required for the complete reaction,
and there is evidence to suggest that the last two electron transfers to the
oxygen intermediates are coupled to the proton pump, with two protons being
pumped at each of these reduction steps (Wikstrom 1989). Measured from the
inside, 2H+/e' are taken up which contributes to the protonmotive force Ap (Ap
contains an electrical component AV and a pH component ApH).

There has been some controversy about the actual number of protons
being pumped. Measurements have been made in whole mitochondria of
2H+/e' being pumped (Lehninger et al., 1981), while others made strong
arguments for 1 H+/e' from measurements in whole mitochondria and
artificially reconstituted oxidase vesicles (Wikstrtim 1977; Casey et al., 1983;
Wikstrtim 1984). Mitochondrial experiments are difficult because of the ion
transporters that can affect the apparent stoichiometry. Reconstituted
experiments are much cleaner, but it is possible that something could be lost
in purification of the enzyme, and the apparent stoichiometry can be highly
variable depending on the exact conditions including the quality of detergent
and lipids used. From experiment and thermodynamic arguments it is now
generally accepted that 1 H+/e‘ is pumped (Babcock and Wikstriim 1992). In
the case of the reconstituted Rb. sphaeroides orddase proton pumping appears
slightly lower than the beef heart enzyme, with H+/e' somewhere between 0.5
and 0.8 (Fetter et al. , 1995). (This is likely an artifact of the reconstitution
process, possibly due to a requirement for difl'erent lipids to give optimum
pumping.

There has been a question about whether the protons pumped to the

intermembrane space of mitochondria contribute to the protonmotive force

1 1
since they may diffuse into the cytoplasm. Some recent studies of

mitochondrial structure indicate that this space may be subdivided into
separate intra-cristae and inter-membrane compartments which may not be
in rapid equilibrium and thus could contribute more to formation of a pH
gradient (Mannella et al., 1994). It is also unclear whether protons that are
pumped to the periplasm of bacteria contribute to the protonmotive force of
most bacteria because small molecules from the periplasm can diffuse to the
environment through pores in the outer membrane structure. Study of the
alkalophiles may provide a clearer understanding of this question because they
can make ATP even when the external pH is as high as 12 (Krulwich et al.,
1988). This is postulated to occur through some localized transfer of protons
fi'om the oxidase to the ATP synthase in which the pumped protons fi'om the
oxidase immediately return through ATP synthase, with this cycle possibly
being compartmentalized (W esterhoff et al., 1984; Ferguson 1985; Rottenberg
1985).

Chemistry of the oxygen reaction

Several types of experiments can be done to look for the intermediates
in the oxygen reduction mechanism. Because of the speed of electron transfer
events, many of the experiments require techniques which allow the analysis of
fast kinetics on the usec or nsec time scale. An important experimental
methodology is to inhibit the enzyme by binding CO to the high spin heme a3
where 02 normally binds. CO dissociates slowly in the dark (in seconds) so it is
possible to mix the sample with 02, and then initiate a fast dissociation with a
nanosecond laser flash, allowing oxygen to react in its place (Gibson and
Greenwood 1963; Greenwood and Gibson 1967 ). The two types of spectroscopy
most often used in the analysis of the intermediates are uv-visible absorbance

12
(Oliveberg and Malmstrtim 1991) which can monitor the absorption of the
hemes and CuA, and resonance Raman (Varotsis et al., 1993), which is much
more sensitive to details of each of the heme environments. Visible
absorbance can give better kinetic resolution showing four phases of the
reaction, while the time-resolved resonance Raman provides spectral details
which, through isotope experiments and comparison with model compounds,
make clearer the true nature of the intermediates. Another method to detect
intermediates uses optical absorption on whole mitochondria and a high
protonmotive force to reverse the chemistry of the oxidase (Wikstriim 1989;
Wikstrtim and Morgan 1992). FTIR has also been helpful in resolving the
changes at the binuclear center caused by mutation. In this technique CO is
bound to the oxidase and then light is used to flash CO off of the Fe of heme a3,
after which it binds to CuB (Calhoun et al. , 1993). These methods have
provided pieces to the puzzle of the mechanism.

These data have been synthesized recently to provide a self-consistent
model of the sequence of the oxygen intermediates and the kinetics of their
interconversion (Babcock and Wikstriim 1992; Varotsis et al. , 1993;
Einarsdottir 1995). Figure 1.2 shows a summary of the chemistry at the
binuclear center. Oxygen binds only after two electrons enter the enzyme
probably by first interacting with the CuB (W oodruff et al., 1991) and then
moving to the heme 83 (2' and 2). Oxygen is then rapidly converted to the end
bound peroxy form (3-5). Entry of the third electron to the binuclear center
initiates the splitting of the oxygen to the ferryl (7). Finally the fourth electron

13

Figure 1.2 Pathway for the reduction of oxygen to water in cytochrome c
oxidase with numbering as in Babcock and Wikstriim (1992). Only
the changes occurring at the heme a3/CuB are shown, although
electron transfers are also occurring from the other centers CuA
and heme a. The question mark indicates the possibility of
additional intermediates. Only the substrate protons are indicated
in the figure. Pumped protons are predicted to be moved in going
from 3 to 7 and from 7 to 10 with two protons at each step
(Wikstriim 1989; Varotsis et al., 1993).

 

 

 

 

14

 

 

 

 

 

 

 

#2'
Fez+ CU31+(02) _>
Fe3+— (K CUBZ+ 7—»
O' -
e
#5 7
Fe3+— Ck CuB1+ 7"
0(H) H+
#9

 

 

 

Fe3+— OH' CuBZ+(0H') ->

 

 

Figure 1.2

15
converts the ferryl to a ferric hydroxide (9). Time-resolved resonance Raman

has identified modes which have been assigned for (2', 5, 7, and 9) (V arotsis et
al., 1993).

The least clear part of the mechanism is the peroxide form of the
enzyme (3-5). Previously a mode at 358 cm‘1 observed by resonance Raman
was assigned as a protonated peroxy (5) by both Babcock and Kitagawa's
groups (Ogura et al., 1991; Varotsis et al. , 1993). Similarly optical absorption
of intermediates formed during reversed electron transfer have assigned 607
nm as a peroxy and 580 nm as a ferryl (Wikstrom and Morgan 1992).
Recently, the 607 nm form was characterized using resonance Raman but
surprisingly gave a mode consistent with a ferryl intermediate (Proshlyakov et
al., 1994). Since only one group has been able to make this measurement so
far, it requires further confirmation.

Many bacterial oxidases

Bacterial oxidases contain only three or four subunits, much simpler
than those in eukaryotes, which contain up to thirteen (Capaldi 1990a; Capaldi
1990b). As discussed above, only three subunits appear to be intrinsically
involved in the activity and thus the bacterial enzyme provides a good model
system for studying the mechanism. The genes for most of these oxidases
have been sequenced and comparisons among them that identify the
conserved residues are offering insight into which residues are important. All of
the oxidases so far discovered, except the bd oxidase in E. coli. , have homology
to each other (Trumpower and Gennis 1994). Bacteria often contain multiple
oxidases that differ in oxygen afinity and are induced under different conditions

16

Table 1.1 Variation in substrate specificity and heme type of the bacterial
oxidases. The two columns categorize the oxidases by electron
donor substrate. Rows indicate similar heme type, and show that
the heme type does not relate to substrate specificity. Numbers in
parenthesis listed after each bacterial type indicate references for
that oxidase as follows: 1) (Hosler et al. , 1993; Trumpower and
Gennis 1994) 2) (de Gier et al., 1994) (van der Oost et al., 1991) 3)
(Mather et al., 1993) 4) (Sane et al., 1988) 5) (Sone et al., 1994) 6)
(Goldberg et al., 1992) 7) (Garcia-Horsman et al., 1994b) 8) (Gray
et al., 1994) 9) (de Gier et al., 1994) 10) (Preisig et al., 1993) 11)
(Lauraeus and Wikstrtim 1993) 12) (Anemiiller and Schafer 1990)
13) (Matsushita et al., 1990; Matsushita et al., 1992) 14)
(Trumpower and Gennis 1994) 15) (Ludwig 1992; de Gier et al. ,
1994) 16) (Trumpower and Gennis 1994)

Table 1. 1

Bacterial

cytochrome c reductant*

23?,

Rb. sphaeroides (1)
P. denitn'ficans (2)

c333
Thermus thermophilus (3)
Bacillus PS3 (4) (also c203) (5)

ba3
T. thermophilus (6)

Oxidases
quinol reductant

383

B. subtilis (11)
S. acidocaldarius (12)

ba3(b03)
A. aceti (13)

 

Cbb3

Rb. sphaeroides (7)
Rb. capsulatus (8)
P. denittificans (9)
B. japonicum (10)

 

b03(bb3)(003)
E. coli (1 4)

bb3 (ba3)
P. denitn'ficans (1 S)

 

bd**

E. coli (16)

*all cytochrome c oxidases contain Cu A except for the cbb3
**bd oxidase is not genetically related to the other oxidases

18

that may increase the chance of survival in variable environments (Rice and
Hempfling 197 8; Hosler et al., 1992; Garcia-Horsman et al., 1994a).

Oxidases have been found to use either cytochrome c or quinol as
substrate (See Table 1.1) A major difference even within these two categories
is the type of hemes used with heme A or heme B almost always occurring in
the low spin site, and heme A, B or O in the high spin site. The only known
exception is that under overexpression conditions an 003 with normal electron
transfer activity can be produced in E. coli (Puustinen et al., 1992). In general,
variation of the heme does not appear to relate to any other characteristics of
the oxidase. However, it has been suggested that high 02 may be required for
heme A synthesis (Garcia-Horsman et al. , 1994a). Several of the thermophilic
bacterial cytochrome c oxidases have the substrate, cytochrome c , covalently
attached to subunit II (Mather et al., 1993; Sone et al., 1994). While almost all
the cytochrome c oxidases have a CuA site, cytochrome cbb3 oxidase does
not(Garc1’a-Horsman et al., 1994b; Gray et al., 1994) . In this respect it
resembles the quinol oxidases. that lack the ligands for CuA. Thus in the quinol
oxidases the electron transfer must go directly to heme a from oxidation of
quinol. It is proposed that part of the quinone binding site occurs in subunit II
of the b03 (Garcia-Horsman et al., 1994a).

Cytochrome bd has no homology to the other oxidases. It contains two
subunits, I and 11, predicted to contain seven and eight transmembrane helices
by hydropathy and gene fusions. It uses quinol as substrate and contains
three metal centers, a low spin heme B, and two high spin sites, a second heme
B, and the heme D. Initial electron transfer is into the low spin heme B while
oxygen binding occurs to the heme D. The role of the second high spin heme D
is suggested to correspond to that of CuB (Trumpower and Gennis 1994).

19
All the bacterial oxidases so far tested have shown proton pumping

except cytochrome bd (Trumpower and Gennis 1994; van der Oost et al.,

1994). Although a model for proton pumping using ligand exchange around
CuA had previously been proposed (Gelles et al., 1987 ), it is clear from current

results with bacterial enzymes that CuA is not required (Hosler et al. , 1993).
While a different pumping mechanism might exist in the quinol oxidases that
lack CuA or even in the multisubunit eukaryotic oxidases, the high degree of
homology among these enzymes makes this unlikely (Trumpower and Gennis
1994).

Evolutionary relationships

Sequence comparisons among the oxidases has provided some
predictions about evolutionary relationships (Saraste and Castresana 1994).
A sequence alignment of the subunit I gene of cbb3 oxidase (FixN) shows 16%
identity to the cytochrome b subunit of nitric oxide reductase (NorB) (van der
Oost et al., 1994). It is possible that the cbb3 oxidase and the NO reductase
may have a common ancestor (van der Oost et al., 1994). The Paracoccus aa3
oxidase subunit I gene (CtaDII) and the E. coli cytochrome bo oxidase subunit
I gene (CyoB) show 35% identity to each other but only 11% identity to FixN
and NorB (van der Oost et al. , 1994). Thus the aa3 and b0 oxidases show a
more distant relationship to the cbb3. The subunit I gene of Rb. sphaeroides
(Coxl) that has high homology to the Paracoccus aaa and to the b0 oxidase of E.
coli shows the greatest relationship to the bovine subunit I gene with 52%
identity, providing evidence that the eukaryotic oxidases could have developed
from these closely related bacterial oxidases (Hosler et al. , 1993).

The Cup, domain in subunit II of some oxidases may have a common

ancestor with the N20 reductase since both show some similarity in the

20
region proposed to contain the Cup, ligands (Buse and Steffens 1991; van der
Oost et al. , 1991) and both have been shown to contain a binuclear copper
(Kroneck et al. , 1989; Antholine et al. , 1992).. Because of the homology
between the quinol oxidases and the cytochrome c oxidases it is proposed that
the quinol oxidases have lost the Cup, binding site (Castresana et al., 1994;
Saraste and Castresana 1994; van der Oost et al. , 1994). Further evidence of
the close relationship between these two oxidases is given by the fact that the
CuA ligands can be put into the quinol oxidase giving a binuclear copper with
similar spectral characteristics as the native CuA domain (van der Oost et al.,
1992; Kelly et al., 1993),

Mitochondria are proposed to have developed as organelles by
prokaryotes being taken into an ancestral cell, setting up a symbiotic
relationship (Gray 1989). Evidence for this comes fi'om comparison of the
sequences of ribosomal RNA genes from the nucleus, mitochondria, and
bacteria. This analysis shows a closer relationship of the prokaryotes,
especially Rb. sphaeroides, to mitochondria, than between mitochondria and
the nucleus (W oese 1987). The nuclear encoded subunits of many of the
respiratory complexes could have been added to the simpler bacterial core to
allow greater regulation of the enzyme (Kadenbach et al. , 1991).

The close sequence, function, and spectral similarity between Rb.
sphaeroides cytochrome aa3 and the mitochondrial oxidase makes worthwhile
the more time consuming processes involved in working with Rb. sphaeroides
relative to E. coli. With Rhodobacter it is necessary to do site-directed
mutagenesis in E. coli and then transfer the plasmid to Rb. sphaeroides by
conjugation. It is also necessary to grow Rb. sphaeroides on a minimal media
to prevent genetic rearrangement (personal communication: Timothy

Donohue) which takes 2-3 days, compared to a single day on a rich media.

2 1
Although working with the oxidase fi'om Rb. sphaeroides may not provide much
information about the regulation of the eukaryotic enzyme, it does make it
- easier to define the details of the structure, the oxygen reduction chemistry,
and the proton pumping reaction.

Chapter 2 - Finding amino acid residues important for
proton pumping in cytochrome c oxidase

An interesting goal remains in bioenergetics, to determine the molecular
mechanism by which cytochrome oxidases pump protons from the inside to the
outside of bacteria or mitochondria. Three of the enzymes involved in the
electron transport pathway contribute to the proton gradient, the NADH _
dehydrogenase (complex I), the bcl complex (complex III), and the cytochrome
oxidase (complex IV), while complex 11, the succinate dehydrogenase, does not.
Although it is known that the NADH dehydrogenase pumps protons, much less
is known about its mechanism than for bc1 complex and cytochrome oxidase.
While cytochrome oxidase and the bc1 have simpler enzymes in bacteria
amenable to mutagenesis studies, the dehydrogenase still contains 14
subunits, making analysis more difficult (W eidner et al., 1993; Yano et al.,
1994).

Difiiculties in understanding proton pumping

Some problems exist in trying to better understand proton pumping by
cytochrome oxidase at the molecular level. Since this enzyme is membrane
bound, it has been dificult to get high quality crystals useful for x-ray
diffraction data and the resulting structure which is critical to developing a
clearly defined model. Recently, however, the 3-D crystal structure of the
simpler bacterial oxidase fiom Paracoccus denitrificans has been obtained at 3
A resolution (Michel personal communication), and a 2.8 A resolution Structure
of the beef heart enzyme is at an early stage of development (Y oshikawa

personal communication).

22

23

A second problem is that proton interaction sites and movements are
more difficult to measure than those of other substrate molecules with
enzymes, since protons are ubiquitous and dificult to detect. However, in
bacteriorhodopsin it has been possible to monitor the protonation state of
residues using FTIR difi‘erence spectroscopy (Rothschild 1992). While these
experiments may be possible in cytochrome oxidase, they have not been done
yet, and will probably be more dificult because of the larger size of oxidase
relative to bacteriorhodopsin (125 kDa versus 27 kDa).

Proton pump versus Q loop

One way cytochrome oxidase differs from bcl is that oxidase does not contain
soluble proton carriers. Instead the redox energy fi'om electron transfer
through the metal centers of the oxidase to the oxygen/water couple are used to
drive the movement of protons through the protein and across the membrane,
a proton pump mechanism (See Chapter 1 for a detailed discussion of the
oxidase chemistry).

For the bc1 complex, the mechanism involves what is referred to as a Q
cycle in which the membrane soluble ubiquinol carries two electrons and two
protons with the protons being picked up fi'om the inside and released to the
outside. This method of moving protons is more complex than it might appear.
Bc1 contains two B hemes, an Fe/sulphur cluster, a covalently bound
cytochrome c, and a binding site for a soluble cytochrome c. Ubiquinone (Q) is
reduced by either the N ADH dehydrogenase or succinate dehydrogenase and
picks up two protons from the matrix of mitochondria (or the cytoplasm of
bacteria). Ubiquinol (QH2) moves through the membrane to bind the bc1 near
the Fe/sulphur cluster towards the outside of the membrane (center P). Two
protons are released and one electron travels through the Fe/S to the bound

24

- cytochrome cl and the other electron travels through a high potential and then
a low potential B heme where it reduces a temporarily bound ubiquinone (Q) to
ubisemiquinone (Q-‘) (center N). After the ubiquinone (Q) near the Fe/S exits,
the site, it will be one of many ubiquinones that is capable of binding near the
low potential B heme site. Now a second ubiquinol (QH2) binds near the Fe/S,
releases its protons and transfers its electrons, one to cytochrome 01 and the
other through the B hemes. Thus the ubisemiquinone at the low potential B
heme becomes fully reduced and protonated from the matrix to form ubiquinol
(QH2) and exits the site, becoming part of the ubiquinol pool which can move to
the Fe/S site for that part of the reaction. The site where ubiquinol binds near
the Fe/S is predicted to be in a hydrophobic environment, so that after electron
transfer the protons must move through the protein to the outside of the
mitochondria or bacteria (Beattie 1993). If the protons move through the
protein, this part of the mechanism may be more like that of a proton pump.
Electrons from cytochrome cl travel to the bound soluble cytochrome c which
then departs from the bcl and carries the electrons to the cytochrome oxidase.

Thus the net reaction becomes:
QH2 + 2 cytoox + 2H"'inside -> Q + chtcred + 4H+outside

This mechanism for bc1 fits well with an aspect of Mitchell's model for
the buildup of a proton gradient only using redox loops. Mitchell's chemiosmotic
hypothesis proposed that it was the buildup of a proton gradient that could be
used to drive the synthesis of ATP fi'om ADP, instead of a direct chemical
reaction that caused synthesis of ATP (Mitchell 1976). Another aspect of the
original Mitchell hypothesis was that it was through such redox loops that
electrons moved toward the negatively charged side of the membrane allowing
protons to be picked up and moved to the positive side of the membrane
(Mitchell 1976). Thus for cytochrome oxidase Mitchell proposed only the

25
uptake of substrate protons. Wikstrom showed otherwise with the discovery

that oxidase actually pumped protons across the membrane and did not just
consume them in the reduction of oxygen (Wikstrtim 1977). Much controversy
existed over this and many possible artifacts had to be ruled out before it was
completely accepted that oxidase could pump protons (Mitchell et al., 1985;
Wikstriim and Casey 1985b).

Theoretical aspects of proton pumping

General requirements of a proton pump Most of the models for
proton pumping are based on the concept that an entry path and an exit path
for protons are required along with a directional driving force and some
mechanism of gating, so that once protons are released toward the exit side a
change occurs (gate closes) to prevent the protons from traveling back down
the gradient (For an overview see (Gelles et al., 1987)).

The simplest of these models shows four states, with two conformations
and two protonation changes. A redox protein has two more states
(reduced/oxidized) giving it eight possible states (23) which is referred to as the
cubic model(Babcock and Wikstrom 1992). To have efficient pumping the
protein can only change between certain states while others are generally
disallowed, otherwise the proton pumping could become decoupled fi‘om
electron transfer. Under certain conditions decoupling may occur, allowing
slipping of the pump (Brand et al., 1994).

General models for proton movement through proteins

Continuous chains of hydrogen-bonds as proton wires One of the
early models for proton pumping was based on data from solid state chemistry
in particular studies of proton transfers in ice (N agle and Morowitz 1978). This
model proposed a proton pump using a network of hydrogen-bonded serine

26
residues in a beta sheet structure. Two important concepts about the

hydrogen-bond transfer of a proton between two residues were used. When a
hydrogen bond is short, it is easier for a proton transfer to occur (hopping), but
this makes it more dificult for the proton to reorient away from the bond
(turning). In contrast when a hydrogen bond is long, it is more difficult for a
proton transfer to occur, but it is easier to reorient the proton away from the
bond.

For the model, an energy-driven conformational change in the protein
causes a lengthening of the hydrogen bonds at the ends of the chain making it
favorable for a reorientation of the protons to occur relative to the side chains
with this propagating fi'om the proton entry side to the proton exit side. As the
protein relaxes back to the original conformation the hydrogen bonds at the
ends of the chain are shortened making it more energetically favorable for the
protons to be transferred. Thus a proton is pumped from one end of the chain
to the other by sequential displacement of the side chain protons at which
point the cycle can occur again. This theory was based on data from ice which
showed this could occur at 105 per second. It was expected to be on a similar
time scale in proteins, which with more thermal motion, would cause slower
transfer of the hydrogens, but would also allow rapider proton reorientation.
However, since a mixture of side chains in a protein instead of only serines
might introduce irregularities, it was predicted to be slowed.

Proton movement through gramicidin Gramicidin is a small
channel-forming antibiotic composed of an interwound dimer of peptides with
alternating D and L amino acids. It is selective for the transport of protons
and monovalent cations (W oolley and Wallace 1992). Proton conduction
through a network of waters in gramicidin has shown maximal proton transfer
rates of 109 per second at low pH (Akeson and Deamer 1991). This would be

27
slower at physiological pH with a lower concentration of protons. Calculations
from this data to make a comparison to the CF 1Fo ATP synthase (capable of
1200 H+/ channel'lseC'1 during ATP synthesis (Althofl' et al. 1989) indicate
that at pH 5 the rate of proton transfer in gramicidin would drop off to 7 x
103/sec (Akeson and Deamer 1991). These authors also noted that under the
conditions for the mitochondrial ATP synthase where proton uptake is from
the cytosol at pH of 7.5 the rate would drop to 20/sec. The authors suggested '
three explanations for the low rate: 1) "a much wider channel mouth"; 2) "a
much slower rate of ATP synthesis", or 3) Av driven hydrolysis of water to
release protons into the entrance of the channel (Kasianowicz et al., 1987 ).

It is interesting that Akeson and Deamer concluded that a pure water
channel like that in gramicidin could not exist in CFo. This is because the
gramicidin water channel is not selective for protons and can easily transfer
Na+ and K4“ (although at a much slower rate) while the CFO does not even
allow these ions to enter the channel (Lill et al., 1987). It was concluded that
this could explain why amino acid residues would be required during proton
transfer: to hold the waters in place so that alkali ions could not push the
waters through the channel as occurs in gramicidin (Levitt 1984), and to be
selective for proton binding (NH 2 groups have a lower afinity for alkali cations
than for protons (Williams 1988)).

Integral injector model as a driving force Other theoretical models
have been proposed with one of the simpler ones to understand being the
integral injector model (N agle and Tristram-Nagle 1984). This model has a
residue in the middle of the chain which is the driving force, or injector, for the
pump, making it easier to visualize how this might be applied to
bacteriorhodopsin, or oxidase. This residue releases a proton into a hydrogen
bonded network pushing the other protons in the network along until one exits

 

 

 

28

into the solvent. Next a conformational change occurs with the injector residue
turning 180° towards a hydrogen bonded network facing the uptake side of the
membrane, making reprotonation from the exit side unfavorable.
Reprotonation of the injector causes the protons to pass down the chain, with
reprotonation finally occurring fi'om the solvent. Immediate rotation of the
injector towards the exit path would make protonation from the uptake path
unfavorable, and allow the continuation of the cycle. This model can be
adapted so that a redox change (oxidase), or light induced conformational
changes (bacteriorhodopsin) could be the driving force of the injector's
conformational change. Studies fiom ice predict that the injection of the
proton toward the exit side would be faster than the uptake of the proton,
making it kinetically favorable to have the injector nearer the uptake side of
the membrane (N agle and Tristram-Nagle 1984). This does not appear to be
the case in bacteriorhodopsin where the retinal is in the middle of the
membrane (Henderson et al., 1990). Similarly oxidase models predict that the
binuclear center sits towards the outside of the membrane (Hosler et al. ,
1993), although exiting protons may travel through more protein on this side if
as is predicted subunit II lies on top of the membrane (Hosler et al. 1994a;
Hosler et al., 1994b). (This also appears to be the case in Michel's recent
crystal structure of oxidase.)

Other enzymes that transfer protons

To study the proton pumping mechanism, it is useful to look at what is
known about proton transfers over a distance in other enzymes where some
structural information is available. The most detailed information on proton
transfer has been determined for bacteriorhodopsin (Krebs and Khorana 1993)
and reaction center (Ermler et al. 1994b) that have crystal structures, and

29
have been characterized using site-directed mutagenesis. A crystal structure

exists for the soluble portion of the enzyme F1F0 ATP synthase, the F1 part
that contains the binding site of ATP/ADP (Abrahams et al., 1994). While no
structure has been resolved for the F0 which contains the proton pore, an
NMR structure has been determined on subunit c of the F0 (Girvin et al. 1995;
Girvin et al. 1993). Also site-directed mutagenesis results have given some
insights into the structure/function relationships in the F0 part (Fillingame
1992). ATP synthase and bacteriorhodopsin are similar to oxidase in that
they are capable of pumping protons, with the first using the ATP to ADP
conversion as the driving force for the reaction. Bacteriorhodopsin uses a light-
driven conformational change to drive the pumping. While reaction center is
not a proton pump, it drives the uptake of protons from the aqueous media
through the protein in response to the arrival of an electron at the quinone
sites. Part of the proton uptake in the oxidase reaction is similar to that of
reaction center in that the electrons reducing the oxygen drive the uptake of
protons to convert 02 to 2H20. For the pumping process, the driving force and
gating in oxidase may be more similar to that of bacteriorhodopsin and
ATPase, involving an energy driven conformational change, in this case
possibly coming from the changing redox state of the metal centers and the
oxygen intermediates.

Reaction center While reaction center is not a true proton pump in
that it does not move protons across a membrane, it does use a change in
potential at the quinone site to drive the movement of protons through the
protein. Reaction center from R. viridis was the first membrane protein for
which a high resolution x-ray crystal structure was determined (Deisenhofer et
al., 1984; Deisenhofer et al., 1985). It has also been crystallized fi'om
Rhodobacter sphaeroides (Allen et al., 1988; Chang et al., 1991). Recent

30
refinements of both structures have allowed better predictions about the

location of bound waters (Ermler et al., 1994a; Deisenhofer et al., 1995). These
structures have allowed more accurate thinking about the mechanism and
testing of hypotheses by site-directed mutagenesis. Ten cofactors exist in the
four subunit protein including four chlorophylls, two pheophytons, and two
quinones, with these eight cofactors bound symmetrically in the L and M
subunits. A carotenoid is bound asymmetrically near the chlorophylls and an
iron is located between the quinones. Excitation of two of the chlorophst (the
special pair) causes an electron to travel through one set of the symmetrically
positioned cofactors, referred to as the A branch. This electron rapidly travels
through another chlorophyll in the A branch (Kirmaier and Holten 1993; Zinth
and Kaiser 1993). The electron passes through a pheophyton, and to the
quinone (QA), and then through the iron (removal of the iron by mutagenesis of
its ligands shows only some inhibition of the activity, (Williams et al., 1991) ) to
the other quinone (Q13). A second electron follows the same path to fully reduce
QB. It is not known whether QB picks up the first proton before or after the
second electron transfer (For a review of the electron and proton transfers see
Okamura and Feher 1992)).

Analysis of the crystal structure from Rb. sphaeroides has shown two
chains of residues which could have a role in proton transfer to QB (Allen et al. ,
1988). These residues include five arginines, four aspartic acids, a glutamate,
four histidines, a serine, two threonines, and a tyrosine. Refinement of the
crystal structure from Rb. sphaeroides has resolved a chain of fourteen waters
connecting QB to the solvent outside of the protein, while the chain fi'om Rps.
viridis is significantly shortened and interrupted by a glutamic acid (Ermler et
al. , 1994a).

3 1

Site-directed mutagenesis studies have shown several residues near QB
to be required for proton transfer: G1uL212, SerL223, and AspL213 (See
Figure 2.1; L refers to the subunit) (Okamura and Feher 1992). Mutants of all
three show greater than a 90% inhibition in turnover. Studies predict SerL223
to be involved in transferring the first proton while GluL212 seems to be
necessary for the transfer of the second proton. AspL213 appears to be
necessary for transfer of the first proton, and some evidence exists for its
involvement in transfer of the second proton (Paddock et al., 1994). Figure 2.1
shows how the AspL213 could be involved in either pathway of proton transfer.
Further evidence of a role for GluL212 and AspL213 in proton transfer was
provided by experiments in which the addition of high concentrations of weak
acids increased the electron transfer rate of by presumably by improving
proton access to Q3 (Takahashi and Wraight 1991). Mutating the amino acids
implicated in the proton transfer to others with similar side chains did not
afl'ect activity: GluL212-9Asp (Paddock et al., 1990a) and SerL223—>Thr
(Paddock et al., 1990b). It is interesting that mutation of some residues in the
active site caused only a 50% inhibition in turnover, ruling out an absolute
requirement of these residues for proton transfer: ArgL217, AspL210, and
HisL190 (Okamura and Feher 1992).

The studies of reaction center seem to indicate an important role for
hydrophilic residues in proton transfer, but firrther mutagenesis studies need
to be done especially with the recent predictions of a pathway of waters in the
Rb. sphaeroides structure (Ermler et al. , 1994). This should help determine if it
is the waters that are critical in proton transfer, and if the mutants that afl'ect
the proton uptake are disrupting the path of the waters.

32

Figure 2.1 Model of the QB site in reaction center adapted fi‘om Okamura and
Feher (1992). This shows the predicted path of the two protons
transferred to Q3 driven by electron transfer to the cofactor.
Boxed residues indicate those residues predicted to participate in
the two paths based on studies of site-directed mutants at those
sites. Water molecules are also predicted to be part of the second
path.

33

3 93E

 

 

 

 

 

£51.:

 

 

om ...—hf

34

F1F0 ATP synthase The ATP synthase is composed of two
components: an F0 (composed of a363ybs subunits) that is membrane
embedded and through which protons flow towards the F1; and the F1
(a 1b2c 10) which protrudes into the interior of bacteria or matrix of the
mitochondria and contains the site for the conversion of ADP to ATP
(Fillingame 1992). Under physiological conditions the proton gradient
established by the electron transport chain drives the conversion of ADP to
ATP, but the reaction can be run in reverse so that the enzyme functions as a
proton pump. It is an important current goal in studying this enzyme to
develop a detailed mechanism of the coupling between the proton flow and its
conversion into the chemical energy of ATP. Experiments have shown there to
be a single site for the nucleotide conversion on each of the three [3 subunits,
and that at any one time each site is in a different stage of the catalysis (For a
review see Pedersen and Amzel 1993). Recently, the F1 portion has been
characterized crystallographically to atomic resolution. This has shown that
the F1 has 3 pairs of alternating or and [3 subunits organized in a circle with the
7 subunit in the middle. It appears that the contact between the 7 subunit and
the surrounding on and 3 pairs is very hydrophobic and may allow rotation of
the outer part around 7 and 5 which make up a stalk connecting to the F0
(Eytan 1982; Abrahams et al., 1994).

Site-directed mutagenesis studies have also shown that certain residues
in F0 are important for proton pumping (Senior 1988; Fillingame 1990;
Fillingame 1992). Mutagenesis has shown that in subunit a, His-245, G1u-
219, and Arg-210 are important for H+ flow. Mutation of Asp-61 in subunit c
is also critical. Recently it was shown that modification of Asp-61 in a single c
subunit by DCCD could stop the proton flow (Hermolin and Fillingame 1989).

35
An NMR structure for subunit c shows it to be made of two antiparallel helices

(Girvin et al. 1993). A previous study showed that a double mutant of
Asp61Gly/Ala24Asp retains functionality (Miller et al. 1990). A recent more
detailed NMR structure of purified subunit c agrees with this result showing
that the Asp61 and Ala24 are in opposing helices, but are adjacent to each
other (Girvin et al. 1995).

An FTIR study on the reconstituted F0 showed a continuum from about
3000 to 1800 cm-1 indicating the presence of a chain of hydrogen-bonding
residues with high proton polarizability (Bartl et al., 1995). This continuum
disappeared when the F0 was either modified with DCCD or dehydrated
showing the disappearance of the hydrogen-bonded network. Based on these
results and the mutagenesis studies of other groups and the conservation of
other residues, the following network was proposed: the carboxyl of Ala-79 c,
Tyr- 10 c, Glu-219 a, His-245 a, Asp-61 c, Lys-34 c, Arg-210 a, Arg-41 c, and
possibly intervening waters. Mutagenesis studies on any residues not
previously examined can test this hypothesis. It would also seem useful to
apply the FTIR technique to mutants believed to be blocked in proton transfer
to see if the mutation is having an effect on the hydrogen-bonded network or
through some other structural effect on the enzyme.

Bacteriorhodopsin Bacteriorhodopsin has been highly characterized
as to the mechanism of proton pumping. Its three dimensional structure has
been determined fi'om two dimensional crystalline arrays by electron
diffraction and was found to confirm previous structural predictions based on
site-directed mutagenesis in combination with biophysical studies (Henderson
et al. , 1990). More detailed information is known about its mechanism of
proton transfer than for the enzymes discussed above, and it has provided a

model for consideration during this research on cytochrome oxidase. In the

36
case of bacteriorhodopsin, light induces the conformational change of all trans

to 13-cis retinal which serves as the driving force for the pump.

Although no high resolution structure of bacteriorhodopsin existed until
recently, a low resolution structure and biochemical data predicted the relative
orientation of 7 transmembrane helices (Henderson and Unwin 1975; Khorana
1988). Khorana's group has extensively mutated amino acids which have
side chains that could be involved in transferring protons to unprotonatable
residues (Krebs and Khorana 1993). The mutants were assayed to see if light
driven proton pumping was blocked. Two residues, Asp85 and Asp96 were
absolutely required for proton pumping, while Asp212 had a significant effect
(Mogi et al., 1988). Although mutation of Tyr185 does not inhibit proton
pumping, it slows the conversion from the 0 state to the bR state and may
interact with Asp212 (Rothschild 1992). While it was concluded that mutation
of individual serines and threonines did not have any dramatic afi‘ect on proton
pumping, changing several of these residues, including Thr46 and Thr89,
caused a reduction in the efficiency of pumping (Marti et al., 1991;
Verkhovskaya et al., 1992). (These are included in a hydrogen-bonded network
for a model suggested by Rothschild. See below.) Mutation of Arg82 slows
proton pumping (Stern and Khorana 1989) and appears to be important for
exit of the pumped protons (Otto et al., 1990).

Finally, the advent of the higher resolution structure fi‘om 2D crystals
showed that many of the conclusions already made were accurate and allowed
more detailed predictions to be made (Henderson et al. , 1990). It also allowed
the definition of two channels: a smaller, hydrophobic one containing Asp96
leading to the retinal, and a larger, hydrophilic one leading away finm the
retinal. It is predicted that water molecules could fill these channels, although
the waters are not visible at this resolution. It is further predicted that the

3 7
hydrophilic residues and the waters could be involved in a hydrogen-bonded

network similar to that suggested by Nagle and Morowitz (1978).

With the crystal structure and further analysis of the mutants, a more
detailed mechanism has been pr0posed by Rothschild to account for
protonation changes seen during the FTIR difference spectra along with kinetic
measurements of proton uptake (Rothschild 1992). Figure 2.2 shows an
overview of the photocycle, and Figure 2.3 shows the position of the residues
that are involved in proton pumping. Initially in the state bR570 the
protonated positively charged Schiff base formed by the attachment of the
retinal to Lys 216 is stabilized by other negatively charged residues: Asp-85,
Arg-82 and Asp-212. Light causes the conformational change of the retinal
which goes fi'om all-trans to the cis configuration at C-13 resulting in the
change in state from bR570 to K630. This disrupts an interaction of the
positively charged Schiff base with the negatively charged residues, and
disrupts the network of residues that were dissipating the charge of the Asp-
212. In the L550 to M412 transition, a proton is transferred from the Schiff
base to Asp-85 on the exit side of the proton pathway which allows a hydrogen
bonded network to form from the proton uptake side through Asp-96, Thr-46, a
water, Thr-89, a water, Tyr- 185, Asp-212, to the Schiff base (see Figure 2.3).
Protonation of the Asp-85 releases its interaction with Arg-82, so that a water
will release a proton to the exterior leaving a hydroxyl to stabilize the Arg-82.
From M412 to N550, the proton on Asp-96 moves forward into the hydrogen-
bonded network with each proton shifting by one residue so that the final
proton reprotonates the Schiff base. From N550 to 0640, reprotonation of the
Asp-96 occurs and the retinal converts fi'om the 13-cis back to the all trans

38

Figure 2.2 Photocycle of bacteriorhodopsin. Each of the intermediates is
designated by a letter name and the subscripted absorption
maxima in nanometers. Light causes the conformational change

fiom bR563 to K625. This drives the pumping of a proton in going
fiom L550 to M412, and the uptake of a proton in going fi'om N520
to 0640-

39

excitation by light

aH-trans
/ bR568 \
13-cis
0640 K625
H+ ”520 L550
M412
'44-

Figure 2.2

40

Figure 2.3 Model of the predicted hydrogen-bonded network that develops in
bacteriorhodopsin during the M412 intermediate (Rothschfld
1992). This path indicates that the protons are in a rapid
equilibrium along this network like that shown recently using FTIR
(Olejnik et al., 1992). At the same time the negative charge of the
oxygen would be dispersed throughout the network. The large
arrow indicates the proton transfer predicted to occur fi'om
Asp212 to the Schiff base in the transition fi'om M412 to N550.
Simultaneously with this transfer, a proton at the other end of the
network from Asp96 would be removed.

 

41

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

42
configuration. Finally, from 0640 to bR570, Asp-85 releases a proton to the
hydroxyl that is interacting with Arg-82. This frees the Arg-82 to again
interact with the other residues surrounding the Schiff base.

N 0 direct structural evidence for the hydrogen-bonded network was
available when this model was proposed. The hydrogen-bonded networks were
proposed based on theory, model systems such as gramicidin (both were
discussed above), and FTIR experiments of Zundel showing the possibility of
the formation of hydrogen-bonded networks in which a proton can have high
polarizability allowing for extremely fast proton movement (picoseconds)
(Zundel 1986; Zundel 1988). Recently FTIR difference spectra were taken of
bacteriorhodopsin to compare K630, L550, and M412 to bR570 (Olejnik et al. ,
1992; Zundel 1994). This showed that in fact a continuum develops in the
FTIR with formation of the L550 intermediate indicating a hydrogen-bonded
network, and that in M412, the network is broken but that some strong
hydrogen bonds still exist. This would seem to require a modification to the
model proposed by Rothschild in which the network would form during the
M412 and Asp—96 is deprotonated during L550 (ROthschild 1992). However
Asp-96 does not deprotonate under low temperature experiments like that
used to show the FTIR continuum, but only at room temperature (Braiman et
al., 199 1). Thus the hydrogen-bonded network may form earlier, during L550,
at low temperature, but later, during M412 at room temperature.

Models which have been proposed for proton pumping in
cytochrome c oxidase

Since the discovery that cytochrome c oxidase could act as a proton
pump, many creative models for proton pumping have been envisioned. Some

of these appear less likely considering what is now known about the oxidase. A

43
model was proposed for using the change in redox state at heme a to change

the strength of a hydrogen bond to the heme formyl group (Babcock and
Callahan 1983). This model would not be possible in all oxidases considering
that it is now known that heme B, which has no formyl group, can participate
in the low spin site (Garcia-Horsman et al. , 1994a)

Woodrufl' ligand shuttle model A model for proton pumping has
recently been proposed using what is referred to as a ligand shuttle in which a
protonatable residue could switch between CuB and the distal side of heme 03,
causing displacement of the proximal histidine ligand (Casey 1986; Woodrufi‘
1993). Upon association of the distal ligand with heme a3, it would lose a
proton to be pumped. The distal ligand would then be displaced fi'om the heme
and the proximal histidine would rebind. The model then requires
reprotonation of the distal ligand by the proximal histidine, after which the
distal residue would move back to the copper, and reprotonation of the
proximal histidine would occur. The loss of the proximal ligand under some
conditions offers an interesting mechanism for control of electron transfer
rates between hemes a and a3, since detachment of the proximal histidine
would force a through space jump significantly slower than the through bond
transfer when the ligand is attached. However, the model presents some
difficulties as a proton pumping mechanism. One dificulty is imagining a
proton path through or around the heme. It also seems that the protons would
be pumped only in the binding and exchange of oxygen between CuB and heme
a3. Once the oxygen reduction begins at a3 it would be less likely for ligand
exchange to occur, although not impossible if one of the intermediates moved to
the CuB. With the use of a histidine to carry two protons at a time, a two
proton exchange occurring once during the oxygen binding and a second time
during the chemistry might account for the stoichiometry of 1H/e‘ of pumped

44
protons. This would still require an explanation for why the pumped protons
are only seen during the final intermediates of the oxygen chemistry (W ikstrtim
1989).

Rousseau model requiring proximal ligand exchange of a
tyrosine Another model recently proposed involves that of ligand exchange
between the proximal histidine and a proximal tyrosine based on the known
selectivity of the ferric form of hemoglobin for tyrosinate over histidine
(Rousseau et al., 1993). Both residues would transfer one proton each with two
exchanges of the tyrosinate for the histidine occurring at the ferric and ferryl
steps in the chemistry. This model has recently been tested by the mutation
of the Tyr422 which is in a possible location for this exchange. Mutants at this
residue were still capable of pumping protons (See chapter 4).

Chan model of ligand exchange at CuA or CuB A very complete
model for pumping was proposed based on changes in the ligands around CuA
(Gelles et al. , 1987; Chan and Li 1990). This model would not be possible in the
quinol oxidases or the cbb3 oxidase which do not contain CuA but are able to
pump protons, although a version of this model applied to the Cu]; site would
still be possible (Larsen et al., 1992). In the model for CuB, reduction of Cu2+
to Cul'l' would cause a ligand exchange, which would cause a change fi'om a
strained square planar to a tetrahedral configuration, and the movement of a
proton fi'om the incoming ligand to the departing ligand. Next, as the electron
left the Cu, the site would convert back to a Cu2+ acting as the gate, making
it electrostatically unfavorable for the proton to reverse its movement.
Finally, the proton would be released to the side of the forming gradient and the
ligands could exchange back to their original location for another cycle. Larsen
et al. proposed that their model could only pump 0.75 H+/e- and only 0.5 during
the peroxy to ferryl and ferryl to hydroxy transitions when pumping is believed

45
to occur (Wikstrom 1989), making it seem unlikely as the only pumping
mechanism, but the authors concluded that it could be part of the pump along
with CuA-

Indirect coupling of electron transfer and proton pumping
combined with a peroxide pump Other models have been proposed
involving redox Bohr effects in which the pumping at a site away from the
electron transfers could be coupled through the protein (Chance et al. 1977 ;
Brzezinski and Malmstriim 1986). These models are more complicated to
study making it more difficult for them to be ruled out or confirmed. Other
models have been proposed involving hydroxide and water as ligands at the
binuclear center which might be involved in pumping protons (Mitchell 1987;
Mitchell 1988).

A model using peroxide similar to Mitchell's model and proposing some
inclusion of redox Bohr effects has recently been proposed (Papa et al., 1994).
In this model oxygen would bind at CuB and be temporarily converted to a
peroxide, after which the protons fipm the peroxide would be pumped across
the membrane by a Bohr effect due to redox change at heme a leaving a bound
oxygen. Then this oxygen would bind to the heme a3 where it would begin going
through its characteristic reduction. A second oxygen could also bind to the
Cu}; and again be involved in pumping two more protons before it was released.
Although the chemical basis for this mechanism is questionable, the lack of
ability to monitor CuB during the oxygen chemistry makes it difficult at the
present time to completely mle out a mechanism such as this. As noted by
the author, it is interesting that there are believed to be Cu2+-peroxide
complexes that form during the chemistry of copper monooxygenases

(Klinman and Brenner 1988). One dificulty with this mechanism might be

46
making it energetically favorable for the protons to be pumped fi'om the

peroxide intermediate once it is formed (Morgan et al., 1994).

Histidine Cycle One of the most well-developed models recently
proposed uses Hi3284, a CuB ligand, as a key residue in the proton pump (See
Figure 2.4) (Morgan et al. , 1994; Wikstrom et al., 1994). Their idea to use a
histidine is based on evidence showing that the cbb3 oxidase can pump protons
(Raitio and Wikstrom 1994). If this is the case, then considering that very few
residues besides the histidines are conserved in this distantly related oxidase, it
would appear necessary to use them for pumping, assuming a conserved
pumping mechanism. The lack of conserved residues in the cbb3 also tends to
rule out mechanisms where the pump is coupled at a distance from the site of
electron transfers and oxygen chemistry, since such linkage would likely
require other conserved residues. They discuss five postulates for the model
summarized as follows : 1) Histidine can carry two protons to keep the
2H+/1e' stoichiometry (Wikstrtim 1989). 2) Oxygen intermediates have a role
but do not carry the pumped protons. 3) Imidazolium (ImH2"’) acts to
increase the redox potential of the peroxy and ferryl intermediates, thus
promoting electron transfer. 4) Substrate protons bind to oxygen during
intermediate formation, and thus promote the release of the pumped protons
fi'om the histidine. 5) The protons which bind the oxygen to form water and
those that are pumped involve separate and independent pathways.

Figure 2.4 shows a summarized version of the model. Notice the
alternating access of the proton paths allowed by the oxygen chemistry: (1)
pumped —) (2) substrate —) (3) pumped —) (4) substrate. Pumped protons bind
to the imidazolate (Im’) when oxygen binds and forms the peroxy intermediate
(1). The imidazolium (lmH2+) stabilizes and is stabilized by the oxygen

47

Figure 2.4 Condensed version of the histidine cycle showing the changes that
would occur at the binuclear center to drive proton pumping. For
the complete chemistry see Morgan et al. (1994). The pathway on
the left is for substrate protons and the one on the right is for
pumped protons. The diagram does not indicate when the protons
enter fiom outside the oxidase since presumably a hydrogen-
bonded network exists, but indicates when the protons act at the
active site. In #1, the sequence of entry into or at least near the
active site is electron, proton , electron, proton, oxygen binding,
electron, proton, with one of those protons not shown, but located
in a position to act as a substrate proton in #2 (This is done to
keep the neutrality of the site.) Two substrate protons are
required at two steps in going from #2 to #3, and again in going
from #4 to #1.

48

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2H+

Figure 2.4

49
intermediate (2), increasing the redox potential, and allowing the electrons to

convert the peroxy (2) to the ferryl (3) while simultaneously the substrate
protons are taken up causing the release of the imidazolium (ImH2+ ) protons
to the exit path. This same general mechanism works again in (3) to (4) when
the ferryl is converted to a hydroxy intermediate again stabilizing ImH2+; with
the uptake of substrate protons the final two protons are pumped. It is a
complete mechanism that thoroughly takes into account the data that is
known about the oxidase chemistry. The challenge is in trying to test this
hypothesis since modification of any of the histidines that are involved results
in an inactive oxidase (Hosler et al. , 1993). The main dificulty would seem to
be the intermediates (2) and (4) in which the doubly protonated histidine
interacts with the negatively charged perorn'de without protonating it. It would
be useful to find model compounds or to do theoretical calculations that could

provide an explanation for how this could occur.

Conclusions
The data from the various enzymes provides some information about

the flow of protons in proteins. Evidence is increasing for some type of
h)Vdr'ogen-bonded network, but is not conclusive. One problem is the idea of
haVing a network seems to imply that a break anywhere in the chain, perhaps
Caused by a single mutation would disrupt this network. That does not seem to
be the case in bacteriorhodopsin in that very few residues have a significant
e’fl‘ect on proton pumping. It has been pointed out that waters tend to order
Well in hydrophobic environments (Zundel and Brzezinski 1992). Perhaps when
some residues that are important in transferring protons are mutated, a water
is capable of replacing the residue. Of course one way to address this would be
to have higher quality crystal structures of bacteriorhodopsin. Reaction center

50
may be useful in this regard in that already various mutant crystal structures

have been determined, although they are not to a refinement that allows the
prediction of water sites (Chirino et al., 1994). If mutants could be made to the
recently seen path in the Rb. sphaeroides structure and high resolution
structures determined, it would help to answer this question. It would also be
useful to have greater application of the experimental FTIR technique
developed by Zundel to show hydrogen-bonded networks in proteins. It will be
especially interesting to see if networks visualized by this method can be
disrupted by mutations. Our group is planning to test this on some of the Rb
sphaeroides mutants.

As to the driving force and gating mechanism for the pumping enzymes,
this seems to be most clearly worked out in the case of bacteriorhodopsin. In
the case of the cytochrome orn'dases this is still a challenging question. Many
models have been proposed and some have been ruled out based on the result
that the relatively unconserved cbb3 oxidase pumps protons and the evidence
that pumping is associated with the final two electron transfers to the oxygen
intermediates. Since these two results appear to rule out many of proton
pumping models, it would be useful to have verification of these two findings by
different techniques or systems. Presently, however, the histidine ligand
exchange at CuB proposed by Wikstriim provides the best fit for the existing

data. It will be challenging to develop experiments to test this model.

Finding residues in oxidase involved in proton pumping
Similar to the research on bacteriorhodopsin, the search for proton
pumping residues in cytochrome oxidase has followed a strategy of mutating
amino acid residues with side chains that could be involved in a hydrogen-
bonded network. A big advantage, however, has been the more than 80 species

51

for which the oxidase genes have been sequenced. Thus the comparison of
these showed residues which were both highly conserved and with side chains
that could be involved in proton transfer (See Figure 2.5). While other
conserved residues could play a role by hydrogen bonding to the carbonyl or
amide groups, the involvement of these in stabilizing the helical structure
makes this less likely. Also considered were the possibility of the involvement
of prolines which, through providing flexibility to the helix or their capability of
cis/trans isomerization, might have a role in a proton relay system (Brandl and
Deber 1986). The focus is presently on subunit I since it contains the
binuclear center where the largest energy release occurs during reduction of
oxygen and where direct coupling could be most readily accomplished. Since
this reaction occurs at the heme 3.3/CuB center, most of the focus is near this
site. Subunit II could play a role through an indirect coupling mechanism or as
part of the proton exit pathway.

The focus of this study was on three of the most highly conserved
regions in subunit 1, and a tyrosine on the proximal side of heme a3. On the
proton loading side of the enzyme in a loop between helices II and III, a
conserved aspartate (Asp 132) was mutated. In the amphipathic
transmembrane helix VIII, several residues were mutated, Thr352, Pro358,
Thr359, Lys362. On the proton ern’t side of the enzyme in the loop between
helices DI and X, another set of conserved residues were mutated, His411,
Asp412, Thr413, and Tyr414. In helix X, which provides the ligands for both
heme a and a3, Tyr422 was mutated to test a model for proton pumping
recently proposed (Rousseau et al. , 1993). These mutants were then

52
characterized as to their visible absorbance spectra, carbon monoxide binding
to heme a3, steady state electron transfer activity, proton pumping activity,
and resonance Raman spectra to provide predictions about their role in the

oxidase pumping mechanism.

53

Figure 2.5 Topological model of subunit I of the aa3-type oxidase from Rb.
sphaeroides. It is based on hydropathy analysis and the phoA/lacZ
fusion analysis done in the homologous cytochrome bo fi'om E. coli
(Chepuri and Gennis 1990). Only protonatable residues that
might play a role in proton pumping through their side chain are
indicated. Those residues that are totally or almost totally
conserved (>90%) from more than 80 species of oxidase are
indicated by squares. Residues that are highly conserved (>80%),
but less so than the first set are indicated by circles. The six
histidines that act as the ligands to heme a, heme a3, and Cu]; are
indicated by hexagons. Residues that were studied for this thesis
are indicated in bold. These are located in the II-II loop, helix VIII,
the DC-X loop, and the Tyr422 in helix X.

54

 

noo— x - x—

> 5.2.
G

___>

 

as 23E

>

a8. =. - =
«2

 

Chapter 3- Development and refinement of a reconstitution
and proton pumping assay for cytochrome c oxidase from
Rhodobacter sphaeroides

Introduction

In order to characterize mutants that may be inhibited in pumping
protons, it was necessary to set up an assay to measure this. Several
methods have been previously used. The first consideration was whether to
use whole bacterial cells, or to reconstitute the purified enzyme into artificial
vesicles. Whole mitochondria from beef heart have been used to measure
proton pumping (Wikstrtim 1984; Wikstrtim and Casey 1985a) and
spheroplasts, whole bacteria with a disrupted outer membrane, are used to
assay proton pumping of cytochrome bo oxidase from E. coli (Verkhovskaya et
al. , 1992). Reconstituted cytochrome c oxidase from the beef heart enzyme
has also been used to measure proton pumping (Krab and Wikstriim 1978;
Casey et al., 1984). In the case of Rhodobacter sphaeroides, an additional
cytochrome c oxidase (cbb3) is known to be expressed along with the aa3
enzyme (Garcia-Horsman et al. , 1994b). Therefore there would be no easy way
to distinguish which oxidase was pumping protons in spheroplasts fi'om Rb.
sphaeroides. Thus it was necessary to develop a reconstitution procedure for
the purified aa3 cytochrome c oxidase from Rb. sphaeroides. The method of
cholate dialysis was used in which soybean phospholipids are sonicated to
clarity in the presence of the detergent cholate. After the addition of enzyme,
dialysis is used to remove the detergent, allowing the orn'dase to incorporate
into bilayer vesicles. This method results in an oriented oxidase insertion with
the cytochrome c binding site facing outward (Eytan 1982).

A second consideration in assaying for proton pumping is the method of
initiating the oxidase turnover that drives proton pumping. One method is to

55

56
assay the enzyme with an oxygen pulse. First the enzyme is fully reduced with

reductants such as ascorbate, TMPD and cytochrome c so that all oxygen is
removed, followed by the addition of oxygen saturated buffer that initiates the
pumping (Casey 1986). A second method, and the one developed for the
experiments in our lab, is to start with the enzyme in a solution containing
oxygen, and to initiate the reaction with the addition of cytochrome c (Casey et
al., 1979; Casey et al., 1984; Casey 1986). A major advantage is that there is
no necessity for working under anaerobic conditions, and thus it can be used
more easily when measuring pH change by spectral methods.

A final consideration is the method used to measure the pH changes,
either with a high sensitivity pH electrode, or spectrophotometrically using a
pH sensitive dye (Casey et al., 1979). The pH electrode had been previously
used in this lab (Gregory 1988), but it is susceptible to electrical interference
and rapid aging (loss of sensitivity) making it dificult to use for the small pH
changes of proton pumping. The spectral method using phenol red was chosen
for this reason and because of its additional potential for further development
for rapid kinetic measurements in the future.

It is known that many variables including preparation variation of the
beef heart oxidase can influence the reconstitution (Casey et al., 1979), so that
it was necessary to find optimum conditions with the bacterial enzyme to give
reproducibility. It had previously been published that the Rb. sphaeroides
enzyme lacking subunit III could not pump protons (Gennis et al., 1982), but
considering the high degree of homology to the mammalian enzyme it seemed
likely that this oxidase should have the same proton pumping activity. In fact,
the highly homologous two subunit enzyme fi'om Paracoccus denitn'ficans has
been shown to pump protons (Solioz et al., 1982). With the development of an
improved purification in our lab (Hosler et al., 1992), it was important to

57
establish whether this enzyme could pump protons. If it could not, it would not

serve as a good model system for studying the energy transduction mechanism
of cytochrome oxidase. After many attempts at reconstitution and fine tuning
of the proton pumping assay, I was finally able to detect proton pumping fi'om

the Rb. sphaeroides enzyme (Hosler et al. , 1992). Many further refinements of
the reconstitution and the assay were necessary before enough reproducibility

was achieved to allow the reliable comparison of the proton pumping activity

from different mutants.

Experimental procedures

Materials Horse heart cytochrome c (Sigma type VI) was purified by
carboxymethyl cellulose chromatography (Brautigan et al ., 1978), rapidly
frozen in liquid nitrogen, and stored at -80 0C. To prepare reduced cytochrome
c for proton pumping, dithionite was added to fully reduce the purified
cytochrome c, and then the reducing agent was removed by gel filtration on a
1.0 x 21-cm Sephadex G-10 column, using 1 mM HEPES-KOH, pH 7.4, 44.6
mM KCl. Stock solutions of valinomycin (2 mM) and CCCP (1 uM and 10 mM)
(both from Sigma) were made by dissolution in ethanol and stored at ~20 0C.
Lauryl B—D maltoside was purchased from Anatrace Inc., Toledo OH. Acid
standard (0.1 M HCl) was obtained fi'om Aldrich.

Purification of asolectin Asolectin from Associated Concentrates
(now out of business) was purified to remove fatty acids and other oxidative
products, or with inconsistent results, unpurified asolectin was used fi'om
Sigma (L-or-phosphatidylcholine, type II-SXUnpurified Sigma type IV-S may
also be usable but has not been tested for this oxidase). Purification of the
phospholipids was done by an adaptation of the procedure in Sone et al. (197 7 )
as follows. Asolectin (25 g) was stirred in 500 mL of dry acetone containing

58
500 mg of betahydroxytoluene (BHT); the cloudy supernatant was removed by

vacuum filtration using Type 5 Whatman Filter Paper, leaving dry
phospholipids. These were dissolved in 100 mL of dry diethylether. Then 900
mL of dry acetone was titrated into the diethylether solution at 4 °C to form a
precipitate. This was collected by filtration through Type 5 Whatman filter
paper. This was repeated starting at dissolution in the diethylether, after
which the phospholipids were dissolved in diethylether with 200 mgs of alpha I
tocopherol and centrifuged at 8000 x g for 10 min. The supernatant was
concentrated using a rotovap, and then allowed to further dry under vacuum in
a desiccator overnight. The lipids were stored under argon in small tubes at -80
oC, and each tube was only used once.

Purification of cholate Cholate from Serva was recrystallized based
on the procedure of Kagawa and Racker (1971) as follows. Cholate (200 g) was
dissolved in a minimal volume of 80% ethanol at 60 0C (approximately 700
mL). Charcoal was heated for at least 2 hr at 120 °C to activate, and then
added to the ethanol and stirred for 1/2 hr. This was filtered several times
through Type 5 Whatman filter paper. The filtrate was stored on ice to allow
recrystallization to occur. The recrystallization was repeated a second time.
Cholic acid was dissolved using KOH to make a final stock concentration of
20% which should be colorless.

Growth of bacteria and oxidase purification Rhodobacter
sphaeroides strain CY91 was grown at 30 °C with 25 ug/mL of kanamycin (the
antibiotic resistance comes from the deletion of the b562 gene which leads to
overproduction of the oxidase), and then purified following the procedure in
Hosler et al. (1992), except that the hydroxyapatite purification step was
excluded. During the early stages of development of the reconstitution and

59
proton pumping assay, it appeared that the hydroxyapatite step gave samples

that did not reconstitute as well.

Reconstitution The reconstitution was done generally as indicated in
Hosler et al. (1992). Glassware was rinsed with ethanol and then thoroughly
with water to remove any traces of dish washing detergent. Asolectin
phospholipids at 40 mg/mL, 2% cholate, and 75 mM HEPES-KOH, pH 7.2-7.4
(a lower pH seemed to give better results) were sonicated with a Heat
Systems-Ultrasonics sonicator, model W-255, using a microtip, alternating
between a power setting of 5 for 30 s and then off for 30 s at 0° C under argon
until clarity occurred. This sample was centrifuged for 15 min at 12,000 x g to
remove titanium particles and any unsuspended phospholipids. Cytochrome c
oxidase in 0.2% lauryl maltoside was diluted with 300 mM KCl, 0.2% lauryl
maltoside, after which 6.7 mg of cholate was added per nmol of enzyme to give
a final oxidase concentration of 1.8 uM (cholate was now 4%) and allowed to
incubate for 2 hr at 0 0C. (Note that for beef heart oxidase twice the
concentration of enzyme was used.) The asolectin suspension was added to the
enzyme to give a final concentration of 0.65 uM oxidase. This sample was
added to Spectrapor dialysis tubing (number 25225/204, 12- 14,000 M;- cutoff)
and the tubing was closed using a floatable plastic clamp on one end and a
weighted plastic clamp on the other. Dialysis was performed with medium
stirring speed on a Sybron Thermolyne, N uova II stir plate using the following
protocol: 6 hr in 100 volumes of 75 mM HEPES-KOH, pH 7 .4, 14 mM KCl,
0.1% cholate; 12 hr in 100 volumes of 75 mM HEPES-KOH, pH 7 .4, 14 mM
KCl; 12 hr in 100 volumes of 50 mM HEPES-KOH, pH 7 .4, 24 mM KCl, 15
mM sucrose; 12 hr in 500 volumes of 1 mM HEPES-KOH, pH 7 .4, 44.6 mM
KCl, 43.4 mM sucrose.

60
Orientation assay Reconstituted oxidase was tested to determine the

percentage that was in the outward facing orientation, which exposes the site
on oxidase where cytochrome c binds, allowing binding of this substrate and
electron transfer to the oxidase. Spectra were taken on a computerized Perkin
Elmer Lambda 4B spectrophotometer using 0.5 cm cuvettes at 24 0C. To 400
uL of 50 mM potassium phosphate at pH 7.0, 80 uL of vesicles were added.
Cytochrome c was added to a concentration of 35 uM along with ascorbate to
24 mM to reduce the oxidase with an outwardly facing cytochrome c binding
site. This was allowed to incubate for several minutes to be sure complete
reduction had occurred after which a spectrum was taken from 540 nm to 660
nm. A lipid soluble reductant, TMPD, was then added to reduce inwardly
facing oxidase, allowed to incubate for several minutes, and a second spectrum
was collected. To be sure that complete reduction had occurred, a pinch of the
reductant dithionite was added to allow complete reduction of the enzyme (in
some cases lauryl maltoside was added to 0.5% to dissolve the vesicles, with
the same results), and this spectrum was determined. The same experiment
was also performed using blank vesicles (containing no oxidase), and spectra
were collected after the same additions. These spectra were used as a
background which were subtracted from the spectra taken on the oxidase
containing vesicles, using the Perkin Elmer UVDM software package. To
determine the oxidase concentration an extinction coefficient of 27 cm-l mM- 1
was used for the peak at 605 nm with a baseline from 580 to 640 nm.

Proton pumping assay Proton pumping was measured at 22 °C at a
measuring wavelength of 556.8 with a reference wavelength of 504.7 nm and a
3 nm slit width on an Aminco DW-2 dual wavelength spectrophotometer using
the pH sensitive dye phenol red in a stirred cell. HEPES at 100 mM, 50 uM
phenol red, at pH 7.4 was used to set a baseline on the spectrophotometer.

61
The assay buffer was 50 uM NaHCO3-KOH, 45 mM KCl, 44 mM sucrose, 50

uM phenol red, pH 7 .4 at a volume of 2.35 mL. After addition of 150 uL of
reconstituted oxidase, containing 0.1 nmol of enzyme, to the assay buffer, 3.2
uM valinomycin and 0.4 nM CCCP were added to allow equilibration. Then the
pH was adjusted to 7 .4 with 0.1 M KOH using the baseline previously set. An
addition of cytochrome c was used to initiate proton pumping measured as
acidification of the external medium. After the addition of uncoupler, 5 uM
CCCP, a second addition of cytochrome c was added to allow quantitation of the
net alkalinization caused by consumption of substrate protons inside the
vesicles. Additions of 1 nmol H+ standard, before and after the addition of the 5
uM CCCP, allowed calibration of the size of the acidification and the size of the
alkalinization. To determine the H+/e' ratio, the extent of acidification was
determined by extrapolating the alkalinization rate, due to proton leakage back
into the vesicles, back to zero time-the time of cytochrome c addition. Electron.
input was based on 1 e' per cytochrome c with the cytochrome 0 concentration
being determined at 550 nM using an extinction coefficient of 28 mM'lcm‘1.

Results

Isosbestic points of phenol red and cytochrome c The proton
pumping assay developed for the Rhodobacter sphaeroides enzyme is similar to
that previously used for the reconstituted cytochrome oxidase from beef heart
(Casey et al. , 1979). However, a number of refinements were required to
obtain reproducible pumping with the bacterial enzyme. Figure 3.1 A shows
the spectra of phenol red in the buffer used for proton pumping at three pH's
7.15, 7.43, and 7.75. Part B shows the overlapping spectra of reduced and
oxidized cytochrome c at a concentration 50 times higher than what is

62
normally used in the pumping assay. It can be seen that the region around

560 nm shows the maximal change in absorbance with pH. Other
experiments, not shown, indicate that the largest change with a particular
quantity of acid or base occur near pH 7 .4. Thus this pH provides maximum
sensitivity and was used throughout the reconstitution and in the proton
pumping assay. A potential problem is that the reductant used to initiate
turnover, cytochrome c, also has an absorbance within this region. The
isosbestic points can be seen to be at 503.8, 525.3, 541.1, and 556 nm. Thus
the two isosbestic points for cytochrome c are used, one at 556 and the other
at 503.8 nm, so that as the cytochrome c goes fi'om reduced to oxidized no
change will interfere with the spectra of the phenol red . The actual values for
the isosbestic points were recalculated for the difi‘erent calibration of the
Aminco, giving 556.8 and 504.7 on that instrument. A small spectral change
will occur with the addition of cytochrome c, but this is corrected for by
adjusting the pH of the cytochrome 0 sample until no change is observed when
it is added to blank vesicles (dialyzed through the same protocol as the oxidase-
containing vesicles). This adjusts for both the initial absorbance change upon
addition of cytochrome c and for any differences in the pH of the cytochrome c
relative to the vesicles.

Controlling the pH for the proton pumping assay Since small pH
changes are being measured, almost no bufi‘ering is used in the assay. Initial
experiments were done using 50 uM HEPES at pH 7.4, with additional
buffering from the aliquot of vesicles providing another 50 uM HEPES. This
buffer is fairly unstable to pH as 002 is continuously absorbed giving a

63

Figure 3.1 Spectral components in proton pumping assay. A) Determination
of phenol red and cytochrome c isosbestic points. Spectra were
taken using the proton pumping buffer, 50 uM NaHCOB-KOH, 45
mM KCl, 44 mM sucrose, 50 uM phenol red, pH 7 .4. Additions (0.5
uL) of 0.02 M NaOH were added to give pH's at 7.0, 7.15, 7.32,
7.43, 7.75, and 8.15. Only those at 7.15, 7.43, and 7.75 are shown.
B) Oxidized cytochrome c (50 uL) was added to 3 mL of 100 mM
potassium phosphate buffer, pH 7.0 and the spectra was scanned
between 300-650 nm on a computerized Perkin Elmer Lambda
4B. Dithionite was used to reduce the cytochrome c, after which a
second scan was taken. Only the spectra from 450 to 650 are
shown. This can be used to identify the four isosbestic points at
503.8, 525.3, 541.1, and 556. This can be converted to values for
the Aminco DW2a by adding 0.8 nm.

64

 

 

- pH 7.15

Ph 7.43

 

l

 

 

 

l

T

 

 

300 350 400 450 500 550 600 650

 

wavelength (nm)

 

 

ox cytc

cytcred

V'V'V'V‘

 

 

 

 

10.1 A.U.

  

650

 

 

~QM’--.‘A--rj

 

450

Wavelength (nm)

Figure 3.1

65
baseline drift. It was found that using a 50 uM N aHC03 bufl'er stabilized the
pH to some extent during the time required for the assay.

Another development that greatly improved the reproducibility of the
assay can be seen in Figure 3.2, namely, the adjustment of the pH to 7 .4 for
each run before starting the pumping assay. Originally it was assumed that
from one assay to the next the pH would be stable at 7.4. There were,
however, gradual changes in the pH of the buffer and differences between
vesicles, causing the starting conditions to vary. There were clearly differences
between protein samples used in the reconstitution of different mutants, even
though reconstituted under identical conditions. The nature and extent of these
differences was revealed by carefully monitoring the starting pH and the
response of the vesicles to various ionophores. During the assay it is
necessary to have valinomycin present to equilibrate K+ across the membrane
and prevent build up of a membrane potential. The immediate creation of a
membrane potential could cause the protons to be pulled back into the vesicles
too quickly for a pH change to be measured (Nicholls and He 1993). A very low
concentration of CCCP was also found to be beneficial in the beginning of the
assay since it allows more rapid and complete equilibration of the vesicle
interior with the medium, so protons are not being pumped against an existing
gradient.

The presence of an en'sting gradient is revealed by the fact that after
the addition of valinomycin and the low concentration of CCCP the pH rises
significantly (approximately 0.1-0.2 pH units) and then gradually levels off.
There is variation in the final pH after equilibration, depending on which set of
vesicles are used and even with the same vesicles over time due to 002

absorption.

66

Figure 3.2 Equilibration of vesicles in proton pumping buffer and adjustment
of pH to preset baseline (The full scale shown is equal to 0.5 A. U.).
Top) Initially a concentrated buffer containing phenol red is used
to adjust the baseline of the spectrophotometer to pH 7.4. Bottom)
The reconstituted vesicles are added to the indicated proton
pumping buffer in a stirred cell. Their starting pH ranges fi'om 7.1-
7.2. To prepare the vesicles for the proton pumping assay ,
valinomycin, and a low concentration of CCCP are added and
cause the vesicles to alkalinize. N aOH (0.02 M) is added to adjust
the pH to the previously set baseline at 7 .4. The full scale shown
is equal to 0.5 AU.

67

«.m 9:6:—

: In .8. .223 .21 8
6382.. .25 3

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

n.000 . ..Ov. 2E we
23% . mwde .21 on
All 08:. :_o>Eoc__w> moozmz 5.1 cm
. \ 5 85m 8.3% wagon,“
\ ~\r - .2525
a IIO / 1
r 1 V new}. 05.
228 H M 53225205.“ I‘ll uca
9me 1 W 9:3 c_£m_>
v.» In .3. 65% .2: cm
I GEE. .mwme SE cow . .
h 3mm 0mm
» 1 x. .2520
06.8 L W A 25.5 oz...
. . H 535258;; A .. II can
came 4 W I 9:3 o3E>
.b I. I II I IIVNIQIIIW WV

 

 

 

68

This variable introduced some error into the measurements, which could
be reduced by adjusting the pH to 7 .4 as noted above and diagrammed in
Figure 3.2. A concentrated bufl‘er at pH 7 .4 containing phenol red is used to set
the baseline of the spectrophotometer. As shown in the bottom part of the
figure, the addition of the valinomycin and dilute CCCP causes the pH of the
vesicle suspension to alkalinize to different extents with different vesicles.
Then base is added to bring the pH to the baseline set on the
spectrophotometer. This two step procedure is used with each experiment to
give greater precision and reproducibility.

Proton pumping of the wild type cytochrome c oxidase from Rb.
sphaeroides Figure 3.3 shows proton pumping data for the wild type enzyme.
Addition of cytochrome c causes the protons to be pumped fi‘om inside of the
vesicles to the outside, which gives the initial rapid spectral decrease of the
phenol red indicative of acidification. Next, the direction of spectral change
reverses with an exponentially decaying in rate. This alkalinization proceeds
past the baseline due to the net consumption of protons in the conversion of Oz
to H20 in the interior of the vesicles. The size of this change can be calibrated
by the addition of an acid standard. A control experiment is done by adding
concentrated CCCP to the vesicles bringing its concentration to 5 uM. This
allows rapid equilibration of protons across the vesicle membrane. Subsequent
addition of cytochrome c causes only alkalinization. This is also calibrated
against an acid standard and can then be compared to the known amount of
cytochrome c that is added during the experiment. Knowing that one proton per
electron is consumed during this reaction provides a control for any artifactual
pH change caused by the cytochrome c alone.

69

Figure 3.3 Proton pumping of the wild type enzyme from Rhodobacter
sphaeroides. Reconstituted oxidase was assayed using phenol red
to spectrophotometrically detect pH changes at 556.8 minus
504.7 nm (See Experimental Procedures). A) Addition of 2.1 nmol
of cytochrome 0 caused a rapid acidification indicative of proton
pumping. B) After the addition of concentrated CCCP (5 uM),
addition of cytochrome 0 produced a rapid alkalinization indicative
of substrate proton consumption. Addition of 0.5 nmol of acid
standard before and after the addition of the concentrated CCCP
was used to calibrate the size of the pH changes. The assay was

normally performed on an 0.02 AU. full scale and gave AA.U.
values of 0.002 to 0.004.

70

A) Reconstituted wild type Rb. sphaeroides
oxidase shows proton pumping.

4H++4e'+02fi2H20

 

cytochrome c

10 sec
8) After the addition of uncoupler (CCCP), |__.l

only the alkalinization is seen.

4m

.+ 4e' + off—o2 H20 M

cytochrome c
\ I W

Figure 3.3

 

71

As discussed in Hosler et al. (1992), various controls were done to assure
that proton pumping was being measured. Sequential additions of cytochrome
c gave the same size of pH changes, indicating a quantitatively repeatable
reaction. Addition of azide or cyanide to the system causes elimination of the
proton pumping reaction, indicating that it is caused by the turnover of the
enzyme. Finally, when the ldnetics of the cytochrome c oxidation are
monitored, complete oxidation is observed as measured by decreased
absorbance at 550 nm with vesicles containing oxidase, but no oxidation
occurred with blank vesicles.

Determination of oxidase orientation after reconstitution
Previously it had been shown that when beef heart oxidase was reconstituted
by this type of procedure, 80% or more was oriented with the cytochrome c
binding site facing outward (Wrigglesworth and Nicholls 197 9; Gregory 1988).
It is necessary for the oxidase to be in this orientation to measure proton
pumping and thus, if mutants are found that do not appear to pump protons, it
is essential to establish that they are correctly oriented. To determine the
orientation, the reconstituted oxidase is reduced first with water-soluble
reductants cytochrome c and ascorbate. Under these conditions only orddase
with its cytochrome c binding site facing outward is reduced. A hydrophobic
reducing agent, TMPD, which can penetrate the vesicles, is then added to
reduce the inwardly facing enzyme. Figure 3.4 shows the spectra of the alpha
band at 605 nm of the reconstituted wild type cytochrome c oxidase.

72

Figure 3.4 Determination of percentage of reconstituted wild type
cytochrome c oxidase with its cytochrome c binding site facing
outward. Cytochrome c oxidase vesicles were assayed as indicated
in the Experimental Procedures by first reducing the outward
facing enzyme with cytochrome c and ascorbate. Inward facing
oxidase was assayed by the further addition of TMPD. The size of
the peak at 605 nm from the hemes of the enzyme indicates the
concentration of oxidase. By dividing the peak intensity of the first
spectrum by the peak intensity of the second spectrum, the
percentage of outward facing oxidase is determined.

73

 

 

 

 

 

 

 

— ascorbate
:I: 0.001 A.U.
...... TMPD
.IJX ,
I l l l
580 600 620 640
Wavelength (nm)

Figure 3.4

 

7 4
It can be seen that there is little change in the reduction of the oxidase in going

from the cytochrome c/ascorbate spectrum to the spectrum with the added
TMPD, except that the spectra are offset. Addition of the powerful reducing
agent dithionite and then the detergent lauryl maltoside to dissolve the vesicles
showed the same level of reduction, indicating that the ascorbate and TMPD
were effective in fully reducing the enzyme. Calculations fiom these spectra
show that 100% is in the outward orientation. The concentration of oxidase
can also be calculated from these spectra and shows that 60% of the enzyme
originally used was still present after several days of reconstitution. Thus the
reconstitution of the bacterial enzyme gives as high a percentage of correct
orientation as is seen with the thoroughly characterized beef heart enzyme.
Respiratory control of the reconstituted oxidase Another
important criterion for determining the success of a reconstitution is the
measurement of "respiratory control" in the reconstituted enzyme. If the
enzyme is properly reconstituted, stimulation of activity with excess reducing
agents (ascorbate, TMPD, and cytochrome 0) will cause a build up in the
protonmotive force across the membrane which will inhibit the oxidase.
Addition of valinomycin and CCCP will release this inhibition and allow oxygen
consumption to proceed. Dividing the released rate by the initial inhibited rate
gives a respiratory control ratio (RCR). After each reconstitution the
respiratory control is determined. A previous report compared several batches
of reconstituted beef heart enzyme and showed that by looking at the ratio of
the release of inhibition with valinomycin and CCCP to that with valinomycin
alone, a correlation was found with the efficiency of proton pumping (Wilson
and Prochaska 1990). This was not found in the studies on the Rb. sphaeroides
wild type or mutant oxidases, but other variables in this system may have

7 5
been more influential. A significant RCR is needed for pumping, but values

from as low as 3 to as high as 8 did not correlate with the efficiency of proton

pumping.

Discussion

Contrary to previous results (Gennis et al. , 1982), the oxidase from Rb.
sphaeroides is capable of pumping protons. It seems to reconstitute with a
normal orientation in the membrane and to be capable of showing good
respiratory control. Why this was not observed in the previous measurements
probably relates to the fact that the reconstitution procedure must be
independently established for different enzyme sources. Reconstitution of
membrane proteins was deveIOped by Racker (Kagawa and Racker 1971;
Hinkle et al., 1972; Eytan 1982) and has been useful to assaying many
proteins in a more native environment while still free from potential artifacts
caused by the presence of other enzymes and transporters in the native
membrane. Many of the problems that I encountered during the last several
years are only occasionally discussed in the literature. Clearly most of the
details for reconstituting a particular enzyme need to be worked out on a case
by case basis and many subtleties about any procedure remain unpublished.
In fact, few studies have been done to establish in any systematic way what
the important variables are in reconstituting membrane proteins. It seems
worth discussing potential problems of the reconstitution procedure, and thus
the variables which may be worth modifying to improve the procedure. Figure
3.5 lists the major ones that I have noticed during my experiments listed from
most to least important. Those at the tap of the list caused more obvious
changes in the quality of the reconstitution, while some of the other variables

76

Figure 3.5 Summary of some variables to consider during the reconstitution
of the Rhodobacter sphaeroides enzyme. These are listed in order
of what were considered to be most important (three check marks)
to least important (one check mark). For those listed with one or
two check marks, the effect of changing these were difficult to
quantitate because of the subtle changes that often occurred
relative to other variability in the assay.

77

Some Important Considerations
for the Reconstitution and
Proton Pumping Assay

Vo/s/ purity of enzyme

Vt/t/ purity of cholate

Vi/t/ purity of phospholipid

Vili/ type of syringe used to deliver acid

Vt/

VV

5’“

standard during assay

pretreatment of enzyme with cholate
before reconstitution

standardization of proton pumping
buffer pH for start of assay

ratio of enzyme / lipid during
reconstitution

pH of HEPES/ cholate buffer used in
reconstitution

concentration of enzyme sample during
reconstitution

Figure 3.5

78
were more difficult to quantify, but even small improvements were helpfirl.

Standardization of the proton pumping buffer pH was previously discussed in
the Results section in conjunction with Figure 3.2; the other considerations will
be discussed now.

Enzyme purity Of course one of the biggest effects on the quality of
the reconstitution was the purity of the enzyme. It might not seem that purity
of the oxidase should afi'ect the reconstitution. However, attempts to use the
oxidase after isolating inner membranes, doing a N aBr wash, and a single
DEAE anion exchange purification, but leaving out the second anion exchange,
did not give good reconstitution and proton pumping. The spectrally visible
contaminants still left at this stage are b and c cytochromes. The enzyme
may also be in an aggregated state and have higher levels of associated
bacterial lipid. During the reconstitution, these contaminants may
coreconstitute with the oxidase and make it more difficult for proper vesicles to
form. Possibly, it is a problem for multiple membrane proteins to fit into the
narrower curvature of the vesicles relative to the native membranes. Other
contaminants that are not visible spectrally might also cause a problem, such
as F0 from the ATPase which would uncouple the vesicles.

Detergent purity Similarly the purity of the cholate was critical for a
good reconstitution. Although what would appear to be highly pure cholate is
available (>99%), it is still not of a quality suitable for reconstitutions. This
can easily be seen by making an aqueous solution of the detergent which will
form a clear yellow solution. It is necessary to do a recrystallization to remove
the remaining contaminants. Problems have also been seen in this lab when
using the unpurified cholate for the preparation of beef heart cytochrome
oxidase following the procedure of Suarez et al. (1984).

7 9

Phospholipid purity Purity and composition of the phospholipids was
also important. Asolectin, which is soya bean phospholipid, is a mixture of
predominantly phosphatidylcholine, phosphatidylethanolamine, and
phosphatidylinositol. Unfortunately, these ratios are not constant between
different brands and batches. One batch of asolectin that was useful in the
reconstitutions was put through an organic extraction to remove oxidation
products and fatty acids. This may be necessary if the fatty acid content of a
particular batch is high since these will uncouple the membrane (Sharpe et al.,
1994). Another batch of type II-S lipids from Sigma was found to give good
results without being put through the purification, although a more recent
batch of the same type does not work well. One conclusion fiom these results
is that with the increased availability of individual phospholipids at lower
prices (fi'om Avante for example), development of their use in a reconstitution
of oxidase may provide more consistent results.

Accuracy of standardization One of the most critical elements of the
experiment was found to be the type of syringe that is used to deliver acid
standard during the assay. Before this was discovered, analysis of most
assays showed that the alkalinization after the CCCP had about a 20%
shortfall from the expected size of the change. This was thought to be an
artifact created by the low buffering capacity and the slowly decreasing pH
caused by 002 absorption fi'om the atmosphere. It was finally discovered that
the problem was the use of a Hamilton syringe that tended to inject slightly
larger amounts of acid than expected, which then produced an error in the
calculations. It was shown that when the syringe was emptied and rinsed
several times in the assay bufi'er, changes in pH occurred that were several
times the entire capacity of the syringe, indicating that the metal of the
syringe may be binding excess acid. A positive displacement syringe from SMI

80
using siliconized glass capillaries gave smaller injections of acid, eliminating the
shortfall from the alkalinization. Further confirmation that this was now an
accurate standard was found when a 10 uL Gilson gave the same results.

Presolubilization of the oxidase with cholate As mentioned in
Fetter et al. (1995), it was found useful to solubilize the enzyme with cholate
detergent before doing the reconstitution. The bacterial oxidase is purified in
lauryl maltoside detergent which has a lower CMC than the cholate and is thus
more difficult to remove by dialysis. The bacterial enzyme also retains
phospholipids. Although phospholipids solubilized in cholate are added to the
enzyme at the start of the reconstitution, it was found useful to saturate the
enzyme with even more cholate before mixing with the sonicated
cholate/phospholipid mixture. Addition of this higher concentration of cholate
may help displace the lauryl maltoside and the bacterial phospholipids which
could interfere with association of the oxidase with added phospholipids. Higher
cholate may also improve the dispersion of the oxidase. This procedure
resulted in improved respiratory control values and higher efficiency in proton
pumping-

Enzyme to lipid ratio It has previously been shown that varying the
enzyme to lipid ratio has an efl'ect on the quality of the reconstitution and
proton pumping (Casey et al., 1979). That group found that for reconstitution
of the beef heart enzyme there was variability in the optimum ratio between
difi‘erent preparations of the enzyme. In the case of the bacterial enzyme
better reconstitutions were found by increasing the amount of lipid relative to
the beef heart oxidase. By purifying each growth of wild type and the mutants
to the same extent, much of the variation was eliminated so that the same
ratio of enzyme to lipid could be used for each sample reconstituted.
Presolubilization of the enzyme in cholate also helped to eliminate some

8 1
variation, since the amount of lauryl maltoside and lipids bound to the enzyme

could have varied between different purifications.

pH of the HEPES/cholate bufl'er The reconstitution buffer contains
a high concentration of HEPES buffer. This becomes incorporated into the
interior of the vesicles and helps prevent a build up of the pH gradient during
turnover, with higher concentrations of internal buffer allowing a higher
apparent pumping eficiency (Proteau et al., 1983). In many of the initial
reconstitutions, a stock solution of HEPES was made up to a high
concentration and the pH adjusted to 7 .4. The pH of the cholate stock solution
is similarly adjusted. These solutions are diluted along with the phospholipids
to make up the reconstitution buffer. It was found that by not adjusting the
pH of the stock solution of HEPES the initial pH was lowered to about 7.2.
Since the first dialysis bufi’er is at pH 7.4 the pH of the reconstitution mixture
should rapidly equilibrate back to 7 .4. Yet, for some reason the reconstitution
started at lower pH seemed to give better proton pumping results.

Concentration of the enzyme in the reconstitution Another
problem is the concentration of the enzyme sample used for the reconstitution,
since each purified sample is at a different concentration as it comes ofi‘ the
column. To reduce the variable of adding different volumes of enzyme, each
sample of oxidase was diluted with the final elution buffer from the anion
exchange column to the same concentration. It was difficult to discern a clear
effect of controlling this variable, but it was one less complication to worry
about.

This discussion has covered considerations which I found to be
important in doing the reconstitution. To answer the question of how oxidase
pumps protons, a reproducible assay is needed. A better understanding of the

reconstitution process would provide information about how to better control it.

82

In general, reconstitution of membrane proteins may become a more
important issue in the future, as more receptors and channels are cloned and
purified and an understanding of their activity is sought. This knowledge would
also be useful to the pharmacologist trying to optimize drug delivery using
artificial vesicles.

Chapter 4 — Possible proton relay pathways in cytochrome
c oxidase

Introduction

Cytochrome c oxidase, a key enzyme in aerobic energy metabolism,
reduces oxygen to water, yielding substantial energy which drives the
formation of a proton gradient; however, the mechanism of coupling between
oxygen reduction and proton translocation remains obscure.

Recognition of the strong homology between mitochondrial and bacterial
enzymes (Saraste 1990; Gennis 1991) has stimulated the application of
molecular genetic tools to the analysis of the oxidase mechanism. The genes
for cytochrome c oxidase from Rhodobacter sphaeroides have been cloned,
sequenced, deleted and reintroduced into the bacterium, and sequence
comparisons reveal a high degree of homology with the three mitochondrially
encoded subunits of mammalian oxidase (Cao et al., 1991; Cao et al., 1992;
Hosler et al. , 1992; Shapleigh and Gennis 1992; Hosler et al. , 1993). Extensive
site-directed mutagenesis of the largest subunit, COX 1, has permitted the
assignment of the ligands for the three redox active metal centers, heme a,
heme a3, and CuB (Lemieux et al. , 1992; Minagawa et al. , 1992; Shapleigh et
al., 1992; Hosler et al. , 1993), suggesting that all three metal centers are
located in COXI toward the outer side of the membrane, while substrate and
pumped protons come fi'om the inside (Wikstrtim 1984). Thus some kind of
proton channel or relay system is required to convey protons to the site of
oxygen reduction, the heme a3-CuB center, and beyond. It is reasonable to
look for residues involved in proton pumping near this center, although indirect

coupling and involvement of other subunits are also possible.

83

84

Figure 4.1 A model of the active site region of cytochrome c oxidase indicating
residues analyzed by mutagenesis (filled boxes, white letters).
Tyr422 is not shown, but would be located near the two histidines
on helix X. This model shows six of the twelve predicted
transmembrane helices of subunit I. These contain five of the
proposed histidine ligands of heme a, heme a3 and CuB. Helix VI,
not shown, contains a third CuB ligand, H284, which is indicated.
The dotted arrows indicate the hypothetical, tested pathways for
protons through the protein. Two possible pathways are indicated
for substrate (Hg‘*’) and pumped (Hp+) protons as suggested in
Morgan et al. (1994). Other highly conserved residues are
indicated with bold circles.

85

   
   

9090990
lX-X loop 0 Mn2+

H411 D412

06) .
l-III loop (Doc

Figure 4.1

86
This chapter covers the results of analysis of mutant forms of Rb.

sphaeroides cytochrome c oxidase with respect to their proton and electron
transfer activity. Four regions in COX I were targeted, each of which contains
highly conserved residues capable of participating in a proton relay system
(see Figure 4.1). The results strongly suggest a critical role in proton pumping
for an aspartate in the loop between helices II and III as well as possible
involvement of helix VIII and the DI-X loop region. Tyr422 in helix X was
predicted to be involved in proton pumping through ligand exchange with
His419, the proximal ligand of heme a3 (Rousseau et al. , 1993). Mutation of
this tyrosine did not affect proton pumping ruling out a requirement for the

residue.

Experimental procedures

Site-directed mutagenesis was done using various restriction fragments
of coal as previously described (Shapleigh et al. , 1992) (see also chapter 6).
The Asp132Asn/Ala mutants were made using a 450 bp fragment coding for
the residues 68-226, and both were sequenced to check for any secondary
mutations. The mutants in helix VIII were made using a 450 bp fragment
coding for the residues 225-378.

Determination of CO spectra, the pyridine hemochrome assay,
resonance Raman spectroscopy were done as previously described (Hosler et
al. , 1992). Proton pumping was determined as indicated in chapter 3. Other
methods and experimental conditions are described in the figure legends.

87
Results

Figure 4.1 indicates the predicted locations of the residues subject to
mutagenesis in this study. (Tyr422Ala/Phe) in helix X; (His411Ala,
Asp412Asn, Thr413Asn, Tyr414Phe) in the DI-X loop; (Thr352Ala, Pro358Ala,
Thr359Ala, Lys362Met) in helix VIII; and (Asp132Asn/Ala) in the II-III loop.
All were purified as in Hosler et al. (1992) and characterized as described below.
Lys362Met was inactive and Thr352Ala had very low activity and could not be
purified sufficiently to permit proton pumping measurements.

Electron transfer, spectral, and structural properties As shown in
Table 4.1, all the purified mutants have lower activity than the 1700 s-1
turnover (molecular activity) of the wild type. Excluding Asp 132Asn/Ala,
activities range from 300 sec-1 to 1300 sec-1. The Asp132Asn mutant has a
lower turnover, only 70 sec-1, raising the possibility that this activity could be
from a contaminating oxidase, such as cytochrome cbb3 (Garcia-Horsman et
al., 1994b; Gray et al., 1994). However, visible spectra (Figure 4.2) show the
enzyme to be highly pure of the cbb3 oxidase and analysis by SDS gel (data not
shown) gives the same purity and subunit composition as wild type. This
mutant is also oxidized to the same extent as wild type in the native
membrane, as is the case for other active mutants; whereas, inactive
mutants usually show almost complete reduction. The Asp 132Ala mutant
shows essentially identical characteristics to Asp132Asn.

Most of the mutants studied here have spectral characteristics
indicative of a native binuclear center, and in so far as any alteration in the
visible and Raman spectra is observed, it does not correlate with the loss of the
ability to pump protons. Specifically, Tyr414Phe shows a 5 nm or band red
shift and His411Ala shows a 0.8

88

Table 4.1 Comparison of mutant and wild type oxidases: activity before and
after reconstitution, respiratory control ratios, and proton pumping
efficiency. Proton pumping values were calculated using data
similar to that in Figure 4.10. Activity of the purified enzyme was
measured polarographically in 50 mM KPi, pH 6.5, 0.056% lauryl
maltoside, 2 mg cholate-solubilized soybean phospholipid, at 25 °C,
with 30 uM cytochrome c kept fully reduced with 2.8 mM ascorbate
and 1.1 mM TMPD. Activity in the vesicles was measured using
the reaction conditions as in Figure 4.7., assuming no loss of the
oxidase during the reconstitution procedure. Respiratory control
ratios are calculated by dividing the activity after valinomycin plus
CCCP addition by the initial activity. H+/e‘ ratios are calculated

as described in Figure 4.10. Activities and H+le' are the average
from two or more independent reconstitutions.

89

Table 4.1

 

 

 

Enzyme Activity Activity Vesicles (S'l) RCR H+/e‘

Purified

(s-l) initial +val +val

+CCCP

b. h. 600 30 100 200 6 07-08
w. t. 1700 200 400 900 5 06-08
D132N 70 80 50 70 0.9 0
D132A 50 50 20 20 0.4 0
P358A 600 100 200 400 4 0.5-0.7
T359A 500 50 100 200 4 07-08
H411A 800 100 300 400 4 0.5-0.6
D412N 300 100 300 400 4 0.5-0.6
T413N 1300 200 400 700 3 0.5-0.6
Y414F 700 100 200 500 4 0.5-0.7
Y422F 800 90 100 300 4 0.5-0.6
Y422A 600 80 100 300 4 0.5-0.6

 

90
nm or band blue shift (Hosler et al. , 1994b), but neither is altered in proton
pumping efficiency. In contrast, the severe loss of proton pumping activity in
the Asp 132 mutants is accompanied by no alteration in the visible spectra nor
in the Raman spectra, as shown for Asp132Asn (Figure 4.2; for wild type
spectra see Hosler et al. , 1992). However, after bubbling the reduced enzyme
with 1 mL of CO which gives 100% conversion of wild type cytochrome oxidase
to a CO bound form (Hosler et al., 1992), the Asp132 mutants were only 7 0-
85% saturated (Figure 4.3). A similar slightly lower CO binding (84%) is
observed when His411 is converted to alanine (Hosler et al. , 1994b), a mutant
with high activity and normal proton pumping. Pyridine hemochromagen
analysis was done on Asp 132Asn, in which the heme A is extracted and can be
quantitated by its visible absorbance spectra (Berry and Trumpower 1987).
Figure 4.4 shows the spectra of the extracted heme A Using an extinction
coefficient of 25 mM-l cm-1, the concentration of the heme was determined to
be 16.7 uM. The actual concentration of the enzyme is 6.8 uM (at two heme
A's per enzyme), indicating no significant loss of the heme, suggesting only
minor disturbance of the heme pocket. Similarly, analysis of His41 1Ala
(Hosler et al. , 1994b) shows no loss of heme A. Whether the lower CO binding
represents an altered affinity for CO or a heterogeneous population is not clear.
Also note that recently Asp132Asn was rapidly purified using a genetically
attached histidine tail and a nickel affinity column (for the procedure see
Mitchell and Gennis, in press). This sample showed 100% CO binding, and still
had the same characteristics including a lack of measurable proton pumping.

91

Figure 4.2 Visible and resonance Raman spectra indicating a native active
site of purified Asp 132Asn-cytochrome c oxidase. A Visible
spectrum: 0.8 M in 100 mM KPi, 0.2% lauryl maltoside, pH 7 .0
reduced by dithionite. B. Resonance Raman spectra: 35 11M
oxidase in 100 mM potassium phosphate, reduced with dithionite,
excitation at 441.6 nm (Shapleigh et al., 1992). The following
modes are indicated: 214 cm'l, Fe-N his stretch of heme a3; 365
cm’l, ring bending of heme a3; 1611 cm‘l, formyl stretch of heme
a; 1624 cm' 1, vinyl stretch of heme a; 1662 cm4, formyl stretch
of heme a3.

92

a... charm

r .53 £5 Steam 3.5 532953

005—. ooww ‘ mum omw 005 com com oov
. 1 . ~ ~ . _ . . . _ . . _ . _

é

 

 

mom

7 ”-9
Argsuarul
-—>
aoueqrosqv

399L'——>

 

 

 

vzsr—>
crew-9

 

 

 

 

 

 

93

Figure 4.3 Asp132Asn shows lower CO binding than wild type. Visible
spectra were collected on 0.78 uM wild type and on 0.76 uM
Asp 132Asn in 100 mM potassium phosphate bufl‘er, 0.2% lauryl

maltoside, pH 7.0, at 22 0C, using a computerized Perkin Elmer
Lambda 4B. Samples were reduced with dithionite, and a
spectrum was taken. Then 1 mL of CO was slowly bubbled
through the solution to allow the oxidase to bind CO, and a second
spectrum was taken. Shown are the difference spectra of the CO
bound form minus the reduced form of the enzyme. The trough at
447 nm indicates the remaining reduced form relative to the CO
binding. The size of this trough for the mutant relative to the wild
type indicates the percentage of CO binding. With less CO binding,
the size of the trough will decrease as can be seen for the

Asp 132Asn. For this particular experiment the CO binding was
70% of that seen in the wild type oxidase.

94

 

0.005 A.U.

 

 

 

------ Asp132Asn

— wild type

 

 

 

 

 

 

i

 

f1 T

 

IIWTI

Wavelength (nm)

Figure4.3

.l....,... .l....Tj..
400 450 500 550 600 650

 

700

93

Figure 4.3 Asp132Asn shows lower CO binding than wild type. Visible
spectra were collected on 0.78 uM wild type and on 0.76 uM
Asp 132Asn in 100 mM potassium phosphate buffer, 0.2% lauryl
maltoside, pH 7 .0, at 22 0C, using a computerized Perkin Elmer
Lambda 43. Samples were reduced with dithionite, and a
spectrum was taken. Then 1 mL of CO was slowly bubbled
through the solution to allow the oxidase to bind CO, and a second
spectrum was taken. Shown are the difference spectra of the CO
bound form minus the reduced form of the enzyme. The trough at
447 nm indicates the remaining reduced form relative to the CO
binding. The size of this trough for the mutant relative to the wild
type indicates the percentage of CO binding. With less CO binding,
the size of the trough will decrease as can be seen for the
Asp 132Asn. For this particular experiment the CO binding was
70% of that seen in the wild type oxidase.

94

 

0.005 A.U.

 

 

 

...... Asp132Asn

— wild type

 

 

 

 

 

 

i

 

 

,,,,
400 450 500 550 600 650
Wavelength (nm)

Figure4.3

T I I I I I I I

 

700

95

Figure 4.4 Assay of the heme A content of Asp 132Asn using pyridine
hemochromagen. Protein samples were denatured and the heme
extracted by dissolving them in 100 mM NaOH, 20% pyridine.
This solution was oxidized using potassium ferricyanide, and
spectra were collected. After reducing the enzyme with dithionite,
spectra were again collected. Displayed is the reduced minus
oxidized spectra. Using an extinction coefficient of 25 mM'1 cm4,
the concentration of heme A was determined to be 16.7 uM. The
concentration of the oxidase before the extraction was 6.8 uM.

96

 

 

 

0.002 A.U.

 

 

510

l

l
560 610
Wavelength (nm)

Figure4.4

l

660

 

710

97

Another test of the stability of Asp132Asn was done by assaying the
detergent solubilized enzyme for the release of peroxide. Normally, the pocket
of the binuclear center prevents any escape of intermediates in the 02
reduction reaction (Chance et al., 197 9). Since the pumping is proposed to be
coupled to the peroxy to ferryl and ferryl to hydroxy transitions (Wikstrtim
1989), if the mutant enzyme is blocked in proton pumping, the possibility
exists that the enzyme might release hydrogen peroxide during steady state
turnover. There is some evidence that hydrogen peroxide could be released, in
that it can be added to the wild type enzyme to undergo a transition at the
active site (Proshlyakov et al. , 1994), indicating that under some conditions the
reverse might be possible. Catalase which converts hydrogen peroxide to
water and oxygen was used to assay for the formation of hydrogen peroxide
during the steady state assay (see Figure 4.5). Addition of catalase after three
minutes from the start of the assay did not generate any measurable amount
of oxygen. Control experiments showed that hydrogen peroxide was stable in
the reaction mixture. This experiment should have detected a release of
hydrogen peroxide during 20% of the turnovers of the oxidase. A lower rate of
conversion might not be detected, but the evidence suggests that the mutant is
intact enough that no peroxide is released and that the mutant completes the
catalytic cycle of converting oxygen to water greater than 80% of the time.

Reconstitution and respiratory-control The activities of the
mutants assayed before and after reconstitution are listed in Table 4.1. In
most cases activity measured for the reconstituted vesicles is significantly

lower than that of the purified enzyme due to different assay conditions:

98

Figure 4.5 Assaying Asp 132Asn mutant with catalase fails to show
production of H202. Assays were determined on an oxygen
polarograph at 25 °C in 50 mM potassium phosphate, 0.056%
lauryl maltoside pH 6.5 with 5.6 mM ascorbate and 0.28 mM
TMPD and 30 uM cytochrome c. Solubilized Asp 132Asn was
added to a stirred cell. Catalase (cat) was added (244 units) after 3
min to test for the presence of H202 (catalase converts 2 H202 to
2 H20 plus 02). No significant amount of H202 was released
fi‘om the mutant (100% conversion of 02 to H202 would yield 215
nmol). Control experiments in which 44 nmol and 88 nmol of
H202 were added initially, showed measurable levels remaining
after a 3 min incubation during turnover of the mutant.

99

ad. arm

«Ow: BE: 3+
0 30+
cm<~oram<

 

NONI .05: 31..
0 So...
cm<~mvam<

55 r

0 S94
cm<~opam<

o [ouru SH

0 So...
cm<~msam<

100

Figure 4.6 Determination of percentage of reconstituted Asp 132Asn with its
cytochrome c binding site facing outward. Cytochrome c oxidase
vesicles were assayed as indicated in the Experimental Procedures
of Chapter 3 by first reducing the outward facing enzyme with
cytochrome c and ascorbate. Inward facing oxidase was assayed
by the further addition of TMPD. Addition of dithionite after each
assay showed that complete reduction of the oxidase had already
occurred with the TMPD addition. The size of the peak at 605 nm
from the hemes of the enzyme indicates the concentration of
oxidase. By dividing the peak intensity of the first spectrum by
the peak intensity of the second spectrum, the percentage of
outward facing oxidase is determined. In this case it was
determined to be 90% (This Asp 132Asn contained a bound His-tag
which was used to purify the sample (Mitchell and Gennis, in
press)). .

101

 

 

0.002 A.U.

 

 

...... +TMPD

 

— + ascorbate

 

 

 

 

 

560

I

580

l l
600 620 640

Wavelength (nm)

Figure 4.6

l

 

660

102

Table 4.2 Reconstituted orientation assay of wild type, beef heart, and
several mutants. These samples were assayed as indicated in
Figure 4.6. Correct orientation indicates the percentage of the
oxidase with its cytochrome c binding site facing outward from the
vesicles. The percent remaining indicates the amount of oxidase
remaining after the reconstitution as compared to the lmown
amount of oxidase before the reconstitution. The concentration in
the vesicles were determined from the same data used to determine
the orientation, by using the absorbance at 605 nm after the
addition of the TMPD, and an extinction coeficient at (605-580

nm) of 27 mM'1 cm‘l.

103

 

 

 

 

 

 

 

 

 

 

Table 4.2
% correct % remaining
orientation
reconstituted
oxidase _
wild type 79 7 9
beef heart 74 94
Asp 132Asn 453 45
Asp 132Asn(His-tag) 90 59
Asp 132Ala 89 80
Tyr422Phe 92 59
Tyr422Ala '79 59

 

 

 

 

 

104

including lower TMPD levels, higher pH, and some degree of inverted
orientation of the enzyme; and also due to some loss of the enzyme during
reconstitution. In contrastuthe mutants Asp 132Asn/Ala and Asp412Asn
show no loss of activity suggesting that they are stabilized by reconstitution.
Orientation was determined for the wild type, beef heart,
Asp132Ala/Asn, and Tyr422Ala/Phe (Wrigglesworth and Nicholls 197 9; and
Chapter 3). Figure 4.6 shows the spectra for Asp 132Asn reduced with
ascorbate to determine amount of oxidase with its cytochrome c binding site
facing outward, and with TMPD to determine the oxidase with this binding site
facing the interior of the vesicles. There is little significant difference indicating
that most oxidase is in the correct orientation, and ruling this out as a possible
problem for measuring proton pumping. Table 4.2 shows the orientation
results for this mutant, the wild type, beef heart, and the Tyr422 mutant.
Beef heart only shows 7 0% in the correct orientation, but the Rb. sphaeroides
enzymes are incorporated correctly 90-100% of the time. There does appear to
be some loss of the enzyme during the reconstitution, which may be the major
cause for the reduced activity seen after the reconstitution (See Table 4.1).
Addition of cytochrome c to the reconstituted oxidase does not cause
much increase in oxygen consumption if the vesicles are intact and the oxidase
is inserted properly, because a proton gradient is immediately established
which inhibits electron transfer activity. Addition of a protonophore (CCCP)
and an ionophore (valinomycin) releases the proton and electrical gradient and
stimulates the 02 consumption rate of the native reconstituted enzyme.

105

Figure 4.7 Effects of valinomycin and CCCP on the oxygen consumption of
reconstituted wild type and mutant Asp132Asn cytochrome
oxidase. Rates of oxygen consumption were measured
polarographically at 25 00 in 10 mM HEPES, 41 mM KCl, 38 mM
sucrose, pH 7.4, with 5.6 mM ascorbate and 0.28 mM TMPD and
30 M cytochrome c. After adding 5 uL of wild type vesicles or 100
uL of Asp132Asn (D 132N) vesicles to 1.8 mL of medium in a
stirred cell, controlled activity was measured. The effect is shown
for addition of A valinomycin (4.4 uM) followed by CCCP (5.6 M)
or B: CCCP followed by valinomycin.

106

 

 

 

 

 

 

 

 

 

 

 

107

Figure 4.8 A comparison of the effect of pH on the activity of purified D132N
and wild type cytochrome c oxidase. Activities were assayed
polarographically using 0.003 “M wild type or 0.06 M D132N in
50 mM potassium phosphate for pH 6.5 to 7.5 and in 50 mM TRIS
for pH 7.5 to 9, including 0.056% lauryl maltoside, no
phospholipids, but using the concentrations of cytochrome c,
ascorbate, and TMPD as indicated in Table 4.1. Relative activities
were calculated as a percent of the highest turnover measured: for

wild type (1000 s-l) (~); for D132N (70 s-1) (0).

108

. 5

 

 

I l l J

8.5

7

6.5

 

 

O O O O
O) N In 00

(rseufiiu or annalau)
Aunnov %

0
O
F

Figure 4.8

109
Significant stimulation shows intactness of the vesicles and good incorporation

of the oxidase (Table 4.1). All of the vesicles except those of the Asp132
mutants show expected increased activity. In contrast, the Asp132 mutants
show decreased activity in response to either valinomycin or CCCP (See Figure
4.7 ) or nigericin (data not shown). Nevertheless, the fact that the
reconstituted Asp 132 mutants do respond to a change in the proton or
electrical gradient is evidence that the oxidase inserts properly and that an
intact membrane is maintained. Indeed, disruption of the vesicles with
detergent overcomes the inhibition, presumably by completely eliminating
membrane gradients. Furthermore no inhibition is observed with the purified
enzyme at similar concentrations of lipid, valinomycin, CCCP, or nigericin.

A direct effect of altered internal or external pH on the activity of the
mutant was also considered. However, no difference is observed in the pH
response of purified wild type and Asp132Asn (Figure 4.8); both activities
increase with decreasing pH as previously reported for the beef heart enzyme
(Gregory and Ferguson-Miller 1988).

Proton pumping Proton pumping of the Tyr414Phe/Ala mutants was
determined to test the hypothesis that it may be a critical residue for proton
pumping (see Figure 4.9 for the proton pumping of Tyr422Phe). Both mutants
showed normal proton pumping in which the addition of cytochrome c caused a
rapid acidification. After the addition of the uncoupler CCCP, a second addition
of cytochrome c gave only the net alkalinization showing substrate
consumption of 1 H+/e'. Proton pumping was calculated to be 0.5-0.6 H+le‘,
which was the same range as that shown by the wild type for the same set of
experiments.

110

Figure 4.9 Tyr422Phe shows normal proton pumping. Proton pumping
measurements were made as indicated in Figure 4.10 using a
spectrophotometric assay by following the pH change of phenol
red. To 2.35 mL of proton pumping buffer, 150 uL of reconstituted
vesicles were added containing Tyr422Phe (0.1 nmol).
Equilibration was done as in Figure 4.10. Addition of 2.24 nmol of
cytochrome 0 caused the decrease in absorbance indicating the
acidic pH change of proton pumping. After the addition of CCCP
only the alkalinization is seen. For this particular experiment 0.5-
0.6 H+/e' were pumped. An H+/e' of 0.5-0.6 was also obtained for
the wild type oxidase during this set of experiments.

111

3. 25E

0 088535 +1
@000

 

 

\ x

g

omm S

c @6859on

\

112

Figure 4.10 Spectral analysis of proton pumping and cytochrome c oxidation
by reconstituted wild type and mutant cytochrome c oxidases.
Both activities were assayed at 22 °C in 2.5 mL of 50 11M
NaHCO3, 45 mM KCl, 44 mM sucrose, with or without 50 11M
phenol red, pH 7 .4, with 0.08 nmol of cytochrome c oxidase
reconstituted into soybean phospholipid vesicles. Following
addition of the vesicles to a stirred cell in an Aminco DW2a
spectrophotometer, valinomycin (3.2 uM final concentration) and
dilute CCCP (0.4 nM final concentration) were used to equilibrate
the vesicles. The upper tracings show the oxidation of reduced
cytochrome c at 550 minus 505 nm in the absence of phenol red.
Extravesicular pH changes are recorded in the lower tracings, as
measured at 556.8 minus 504.7 nm by phenol red. Rapid
acidification occurred upon addition of 2.1 nmol cytochrome c
(final concentration 0.8 M) to wild type (A) or Thr413Asn (C),
but not in the case of Asp132Asn (B). In the presence of

concentrated CCCP ( 5 uM), addition of cytochrome c produced
the expected alkalization. Addition of 0.5 nmol HCl standard

before CCCP addition (H+o), and after (H+i) caused the changes
in absorbance indicated and were used to calculate the protons
extruded and the protons consumed, respectively. The H+/e'
ratio can be calculated from the protons extruded (H+), knowing
the amount of cytochrome c added (e‘).

113

 

com S 4

 

 

2m PE.

com or

 

mOO.O" mcmémm < H

 

2.85 m

 

 

 

.55

1 14
The upper portions of Figure 4.10 are recordings of the change in

oxidation state of cytochrome c after addition to the oxidase-containing
vesicles; the lower portions show the concomitant changes in pH in the
extravesicular solution. Shown are wild type, Asp 132Asn, which does not
pump protons, and Thr413Asn, another example of a mutant that has normal
proton pumping eficiency. Reduced cytochrome c is rapidly and completely
oxidized in all cases, though with a half-time about three-fold greater for
Asp132Asn than for wild type or Thr413Ala. This three-fold difference in rate
is observed at 0.8 11M cytochrome c and is in marked contrast to the >20 fold
difference in maximal velocity observed at 30 pM cytochrome c with the
purified enzyme. These results suggest that the mutation has a profound
effect on the internal rate of electron transfer that would only be limiting at
saturating cytochrome c.

The proton concentration changes shown in the lower portion of Figure
4.10A are typical for wild type enzyme. Acidification is followed by
alkalinization due to proton leakage back into the vesicles in response to
internal proton consumption. A similar result is obtained with Thr413Ala
(Figure 4.10C), but Asp132Asn (Figure 4.10B) shows no acidification, only
alkalinization. Asp132Ala showed similar results to Asp132Asn. In all three
cases shown, addition of the uncoupler CCCP allows proton equilibration
across the membrane and only net rapid alkalinization is seen. Stimulation of
the rate of alkalinization after addition of CCCP is evidence that the vesicles
were maintaining a proton gradient. Even Asp132Asn showed stimulation
under these assay conditions where valinomycin had already been added,
consistent with the oxygen consumption assay (Figure 4.7). The other
mutants (Table 4.1) showed proton pumping behavior similar to wild type and
Thr413Asn giving normal H+le‘ ratios.

115

Discussion

General Considerations Bacteriorhodopsin provides the only protein
model for a proton pumping mechanism that is backed by extensive structural
and mechanistic data. Several aspartic acids and an arginine have been
identified as part of a proton pumping pathway for which the driving force is a
light-induced conformational change of retinal (Henderson et al. , 1990; Krebs
and Khorana 1993). It is suggested that the path for protons also involves
water molecules within the protein to complete a proton relay system
(Henderson et al. , 1990) similar to that proposed by Nagle and Morowitz(Nagle
and Morowitz 1978).

To test whether the proton pumping model provided by
bacteriorhodopsin is applicable to cytochrome oxidase, it is important to
identify mutants in which proton pumping is inhibited. A complication in
interpreting any effects on proton pumping is the dual role of protons in the
oxidase reaction: four protons are consumed from the inside of the membrane
for each pair of electrons transferred, two protons in the reduction of oxygen to
water and two in the pumping reaction. Different proton paths may exist for
substrate and pumped protons, since substrate protons must directly access
the oxygen intermediates at the binuclear center while pumped protons may
not.

If there is only one path for protons to the binuclear center, then
blocking this path would be expected to inhibit electron transfer and proton
transport equally, unless protons were used preferentially for reduction rather
than for pumping. Even if two separate paths exist, inhibition of the proton
pumping pathway could slow electron transfer in so far as there is tight
coupling between the two processes. .

1 16
Although theoretical and mechanistic arguments can be made for

obligatory coupling (Brand et al. , 1994), there is evidence for decoupling or
slipping in cytochrome c oxidase (Murphy 1989). Reaction of oxidase with
DCCD, or removal of subunit III, reduces the number of protons pumped per
electron transferred (Azzi et al., 1984; Prochaska and Fink 1987; Gregory and
Ferguson-Miller 1988). Other evidence suggests that proton pumping by the
fully reduced enzyme is only 50% eflicient (Oliveberg et al., 1991; Babcock and
Wikstriim 1992). The proton to electron ratio may also vary depending on the
rate of electron input to cytochrome oxidase in whole mitochondria and
reconstituted vesicles (Capitanio et al., 1991; Papa et al., 1991).

Rationale for mutant selection Site-directed mutagenesis in
cytochrome aa3 was focused on residues that could participate in a proton
transfer network and were highly conserved. Likely candidates are amino acids
with side chains having available protons. While any amino acid except proline
might play some role in a network through its backbone carbonyl or amide,
these groups are normally occupied in hydrogen bonds to stabilize the or-helix.
In the case of proline, its amide nitrogen is involved in forming the cyclic
structure of the side chain. Thus the carbonyl of the amino acid preceding the
proline by four residues in the chain, will be free to accept another hydrogen-
bond (Piela et al., 1987). It has also been proposed that prolines in
transmembrane helices could be important for regulating ion transport by
increasing the flexibility of the helix through cis/trans isomerization (Brandl
and Deber 1986). While in the lac permease these residues do not appear to be
critical (Consler et al., 1991), in bacteriorhodopsin mutation of the prolines has
an effect on the rates of deprotonation and reprotonation of the schifl' base
(Zhang et al., 1993).

1 1 7
According to the bacteriorhodopsin model, acidic residues should be

prime candidates for a proton relay system. Only one conserved acidic residue,
E286, is predicted to crust in the membrane domain of cytochrome oxidases.
While the E286N mutant of cytochrome 003 is inactive, the same mutation in
the homologous cytochrome be is active and pumps protons (Thomas et al.,
1993); therefore a critical role in proton pumping is unlikely.

Helix X Tyrosine 422 is predicted to be in transmembrane helix X
which also contains one of the heme a ligands, His421, and the single heme a3
ligand, His419. Experiments have indicated that the His419 ligand dissociates
during photolysis of CO (W oodrufl' et al., 1991). This led to the development of
a detailed mechanistic model of proton pumping in which the Tyr422 and the
His419 could carry protons, and through an alternating exchange of the
ligands, drive proton pumping (Rousseau et al., 1993) (also see Chapter 2). If
oxidase follows this mechanism, then mutation of the tyrosine should have a
severe efi‘ect on the activity of the enzyme. However, the mutation of the
residue to phenylalanine or alanine still leaves significant electron transfer and
does not inhibit proton pumping. Thus it is not required for proton pumping in
the way that is suggested. Since electron transfer is inhibited there is still the
possibility of it having some role in the proton pumping mechanism, but it is
not critical for the pump.

IX-X Loop Among the mutants analyzed in this extra-membrane
region, Thr413Asn is most like wild type with only slightly lower electron
transfer activity and normal pumping, arguing against an important
mechanistic role. Modification of Asp412 to Asn, on the other hand, causes
strong inhibition of electron transfer, but minimal alteration of spectral

properties (Hosler et al. , 1994b) and no change in protons per electron pumped.
Tyr414Phe and His411Ala have shifted or bands (Hosler et al. , 1994b) and

1 1 8
inhibited electron transfer but again unaltered pumping efficiency. Thus there

is no positive evidence to implicate residues 411, 412, and 414 in proton
pumping, but a role in this process cannot be ruled out since their overall
activity is diminished. It should be noted that studies indicate that residues
411 and 412 participate in a Mn/Mg binding site (Espe et al., 1995; Hosler et
al., 1995); loss of the metal after mutation of these residues could account for
the loss of function (Hosler et al., 1995). Indeed, a role for this non-redox active
metal in proton pumping can be envisaged.

Helix VIII Because Lys362Met was inactive and Thr352Ala was
difficult to purify and had low activity, neither could be tested for proton
pumping. Both these residues have been proposed for a role in proton transfer
to the binuclear center (Hallen and Nilsson 1992; Hosler et al. , 1993). Their
activity characteristics and altered spectral properties support the idea of
helix VIII being proximal to the binuclear center, but do not establish its role.
Mutants Thr359Ala and Pro358Ala show reduced activity, about 30% of wild
type, but no significant change in spectral properties or proton pumping
efficiency, as measured by H+/e'. Since the native spectral characteristics of
these mutants indicate no general disruption of protein structure, the observed
loss of activity could be due to a limitation of proton access to the active site.
Indeed, if two pathways for substrate and pumped protons exist, as indicated
in Figure 4.1, inhibition of a substrate proton pathway might be expected to
give these results. Similarly, inhibition of a tightly coupled pathway for
pumped protons would equally limit electron transfer. Thus these results are
consistent with a central role for helix VIII in proton transfer.

II-III loop Conversion of Asp132 in the II-III loop to Asn or Ala caused
a dramatic inhibition of electron transfer and these were the only mutations
that eliminated proton pumping while retaining the ability to reduce oxygen at

1 1 9
a significant rate. Visible absorbance spectra show that neither mutation has
a significant efi'ect on hemes a or a3. The absence of changes in the
vibrational modes of either heme a or heme a3 in resonance Raman analysis of
Asp132Asn is further strong evidence of native heme environments. Only the
CO binding appears to be slightly altered. Heterogeneity of the sample or a
minor structural change in the binuclear center (possibly in the vicinity of
spectrally silent CuB) may have occurred such that the CO binding equilibrium
is altered. Although CO binding would suggest a quantitatively similar
structural perturbation in His411Ala (Hosler et al. , 1994b), it remains
competent in proton pumping.

Lack of evidence of a significant structural change at the binuclear
center favors the idea that electron transfer activity in the Asp 132 mutants is
limited by either availability of substrate protons to the active site (single
pathway model) or by the obligatory coupling of proton pumping to electron
transfer (two pathway model). Either scenario, along with the predicted
location of the II-III loop, would implicate Asp132 as a critical residue in a
proton relay system from the intracellular side of the membrane. Similar
conclusions have been reached fi'om studies of mutants in the homologous
residue in cytochrome bo of E. coli (Thomas et al. , 1993). However, in that
study, proton pumping was carried out in spheroplasts where the detailed
response to pH and electrical gradients could not be easily examined.

When incorporated into phospholipid vesicles, mutants of Asp132 show
an unusual response to addition of ionophores or protonophores: their activity
is inhibited rather than stimulated by release of either the proton or electrical
gradient. A possible explanation for this response is that the mutation has
changed the pH sensitivity of the mutant. Although the pH inside and outside
of the vesicles should be the same after reconstitution, when valinomycin is

120
added before beginning the proton pumping assay, external alkalinization is

observed. This implies a potassium gradient (higher internal [K+]), but is also
consistent with the observation of Madden and Redelmeier (1994) that there is
a higher pH inside than outside of the vesicles (greater than 8.6, versus 7.4). If
Asp132Asn were less active than wild type at low pH, then addition of
valinomycin (or CCCP or nigericin) might inhibit the enzyme activity.
However, the pH dependence of the activity of the purified D132N is
essentially identical to wild type, arguing against a simple pH effect. Thus it
seems more likely that release of a gradient of pH or charge is involved in the
inhibition. If D132 were normally involved in a salt bridge holding the II-III
loop in a stable conformation, loss of its negative charge would make this
region more conformationally flern'ble and could alter its sensitivity to changes
in the electrical/pH gradient. Alternatively, the loss of D132 as an entry point
for protons could bring into play a new, less efficient pathway that might
respond differently to an electrochemical potential.

Conclusions None of the residues examined by mutagenesis in helix X,
the DK-X extracellular loop or helix VIII preferentially inhibit proton pumping,
but since a number retain native spectral properties and have lower or no
activity, their involvement in proton transfer remains a real possibility. More
definitively, loss of proton pumping with retention of low but significant
electron transfer activity in the case of mutants at Asp132 is strongly
suggestive of a critical role in proton pumping. It remains to be determined
whether Asp132 is a direct participant in a proton relay system or an
important determinant of a structural element that is essential for this

process.

Chapter 5 — Measurement of membrane potential
generation by Asp132Ala and effects of free fatty acids on
its activity.
Introduction

In Chapter 4 the inhibitory effect of ionophores was shown on the
reconstituted Asp 132 mutants. A further understanding of this response may
help explain why the Asp132 mutants do not pump protons and provide
mechanistic information about the pumping reaction. While some evidence
was given in the previous chapter that the ionophores are not inhibiting the
enzyme by a direct effect but through changing the membrane potential or pH
gradient, further confirmation of this would be useful. Experiments in this
chapter help to confirm this result and show another interesting feature of the
mutation, the addition of free fatty acids can partially overcome the defect.
This 'chemical rescue' by fatty acids was unique to this mutant. The
fluorescent dye diSC-3,5 was used to measure the membrane potential: its
fluorescence decreases when a potential is developed, due to its movement into
the hydrophobic membrane environment. This technique of measuring
membrane potentials has been well characterized (Bashford and Smith 1979)
and used previously on reconstituted beef heart enzyme (Singh and Nicholls
1985). It is difficult to determine exactly how much membrane potential is
produced using this technique. By comparing the result between Asp132Ala
and controls reconstituted simultaneously using beef heart oxidase that is
known to have a normal membrane potential and wild type oxidase that would
be expected to have a normal membrane potential, it is possible to establish
whether the responses are qualitatively similar. The stimulation of activity by
fatty acids was studied extensively by looking at the effects of changing chain
length, the alcohol versus the acid, the efl‘ect on other mutants, and whether
analogues of the fatty acid could also stimulate.

121

122

Experimental procedures

Reconstitution was done as indicated in Chapter 3.

Determination of membrane potential Fluorescence intensity was
measured on a Perkin Elmer fluorescence spectrophotometer with excitation
set to 622 nm, emission set to 688 nm (10 nm slit), and using 0.45 uM diSC 3,5
as the membrane potential probe at 30 0C. Reconstituted vesicles (50 uL),
containing either wild type or Asp132Ala were added to the buffer: 10 mM K+
HEPES, 20 mM K2804, 50 mM sucrose, pH 7 .4. To reduce the enzyme
cytochrome c was added to a final concentration of 1.1 uM. Excess reducing
equivalents were added, ascorbate to 5.6 mM and TMPD to 280 uM, to
continuously turnover the enzyme allowing generation of a membrane
potential. Nigericin was added to 0.5 uM to release the pH gradient; this
caused only a small change in the diSC-3,5 fluorescence. Valinomycin was
then added to 0.33 uM, releasing the membrane potential and causing an
increase in fluorescence to the level seen before the addition of ascorbate and
TMPD.

Assay of activity with fatty acids, fatty alcohols, and analogues of
arachidonic acid Steady state turnover rates were measured as indicated in
Chapter 4 with the specific details as listed in the figure legends. The fatty
acids, fatty alcohols and analogues of arachidonic acid were purchased from
Sigma. Arachidonic acid purchased from either Sigma or Cayman Chemical
gave similar results. These reagents, except for indomethacin, were made up
as 250 mM stock solutions in ethanol and stored at -20 °C. Indomethacin was
only soluble in ethanol to 50 mM. Stock solutions of 50 mM and 5 mM were
prepared. The same amount from each stock could then be added to the stirred
polarograph cell during the steady state turnover assay to give final

123
concentrations of 5 uM, 50 uM, and 250 uM. Wild-type and mutant enzymes

were purified as described in Chapter 4.

Results

Determination of membrane potential of reconstituted
Asp132Ala, wild type, and beef heart oxidases As can be seen in Figure
5.1, a similar response is observed in the fluorescence changes of diSC-3,5 for
wild type and the Asp 132Ala. Beef heart oxidase (not shown) also has the
same type of response. For an unknown reason, the addition of oxidized
cytochrome c causes some decrease in fluorescence. Then ascorbate and
TMPD are added to continuously reduce the cytochrome c and maintain a
steady state turnover of the orddase causing the build up of a membrane
potential. This causes the diSC-3,5 to be pulled into the vesicle membrane
resulting in a decreased fluorescence. Nigericin induces an electrically neutral
exchange of protons for potassium across the vesicle membrane, resulting in
loss of the pH gradient without loss of the membrane potential. Valinomycin
allows K+ to equilibrate across the membrane and thus releases the
membrane potential causing the fluorescence to increase to the intensity
before the addition of ascorbate/TMPD. Reconstituted Asp132Ala shows the
production of a membrane potential like that of the wild type and the beef
heart enzymes.

Stimulation of Asp132 mutants by free fatty acids Figure 5.2
shows the polarographic trace of Asp 132Ala stimulated by cytochrome c,
ascorbate and TMPD. Addition of arachidonic acid ( 5 uM, 50 uM, and 250 uM)
caused significant stimulation of the enzyme with the amount of stimulation

124

Figure 5.1 Reconstituted Asp 132Ala shows a normal membrane potential.
Addition of 1.1 uM orddized cytochrome c as the substrate caused
an initial decrease in fluorescence unrelated to turnover. To
generate steady state turnover 5.6 mM ascorbate and 280 uM
TMPD were added. This led to a decrease m the fluorescence of the
probe, indicating the formation of a membrane potential. N rgencrn
was added to 0. 5 uM to release the pH gradient. Valinomycin was
then added to 0. 33 uM which releases the membrane potential
causing the fluorescence to increase to the same level as before
generation of a membrane potential.

Fluorescence Intensity

125

 

 

cytc l

 

 

asc/TMPD

t

nigericin I

 

 

 

 

 

Asp132Ala

wild type

I valinomycin

 

 

0

Time (s)

Figure 5.1

200

300

126

Figure 5.2 Arachidonic acid stimulation of the Asp 132Ala mutant. Activity
was assayed as indicated in Table 5.1. A. shows stimulation of
activity by multiple additions of arachidonic acid and inhibition by
2 mM cyanide. B. shows a single addition of arachidonic acid with
inhibition by 9 mM azide. Activity per second is listed along the
trace.

127

arachidonic acid

  
 
    

 
  

   

50 0M
70
/ 250 uM
ochrome c
M 20 /
B_ 180

cyanide

/

250 pM
arachidonic
acid

 

 

T ' / F

 

60 nmol azide
12
i -i
2 min

Figure 5.2

128

increasing along with the increasing concentration of the fatty acid, bringing
the turnover from 30 sec‘1 to near 200 sec’l. Inhibition of all cytochrome c
orddase activity by cyanide (part A) or azide (part B), inhibitors that bind at
the active site of the oxidase, provides evidence that the fatty acid is
stimulating the oxidase activity and is not just causing an artifactual
secondary reaction with other reagents. Further evidence of this is given by ‘
tests of the fatty acid under the same assay conditions excluding oxidase that
showed no significant changes in the background rate of oxidation. Both

Asp 132Ala and Asp 132Asn showed stimulation by fatty acids to a similar
degree.

Importance of the fatty acid carboxyl and the chain length on
stimulation of Asp132Ala This fatty acid effect was more thoroughly
characterized to determine whether the carboxyl group on the fatty acid was
important for the stimulation and also to determine the effect of the chain
length (See Table 5.1). The measurements were made with consecutive
additions of low," medium, and high concentrations of each fatty acid or fatty
alcohol. It showed that no stimulation occurred with acetate or propionate,
indicating that more than just the acidic group was required Medium chain
fatty acids such as octanoic acid and lauric acid showed some stimulation,
while the fatty alcohols of the same chain length did not show an efl'ect. The
long chain fatty acids fi'om 16-22 carbons showed the most significant effects
with the greatest stimulation fi'om arachidonic acid and docosahexanoic acid.
Again, alcohols of these two fatty acids did not show any stimulation of the
activity. Clearly, the stimulation is caused by more than just a hydrophobic
interaction with the protein.

129

Stimulation of activity by fatty acids is unique to the Asp132
mutants To determine if the fatty acid stimulation was specific for the
Asp132 mutants, the same tests with arachidonic acid and its alcohol were
performed on wild type, beef heart, and other mutant oxidases (See Table 5.2).
Since wild type and many of the mutant oxidases are stimulated by soya bean
phospholipids (a mixture of mainly phosphatidylcholine,
phosphatidylethanolamine, and phosphatidylinositol), these enzymes were
tested to see if fatty acid stimulation might relate to the stimulation by
phospholipid. Asolectin had a stimulatory effect on wild type, beef heart,
Thr359Ala, Asp412Asn, and Phe391Gln. N 0 effect was observed on the
Asp132 mutants or on Lys362Met, Met262Leu, or His259Asn. In contrast,
there was no significant stimulation by fatty acid or fatty alcohol seen with the
wild type and beef heart enzymes. Similarly no effect was observed with
several of the mutants including Asp412Asn, Lys362Met, and Hi8259Asn.
Thr359Ala and Met262Leu showed some stimulation by fatty acid but a
similar effect was seen with the alcohol. Phe391Gln showed stimulation by the
fatty alcohol, but not the fatty acid. None showed nearly the level of
stimulation as seen with the Asp132 mutants. There was little correlation
between the stimulatory effect of the phospholipids and that of the fatty acids,
indicating the unique role the fatty acid has in stimulating the Asp132
mutants.

Stimulation of Asp132Ala by analogues of arachidonic acid
Arachidonic acid is a substrate for prostaglandin synthase, and various
inhibitors that bind in the active site of that enzyme, where arachidonic acid
normally binds, have been characterized (Laneuville et al., 1994). Several of
these were tested

130

Table 5.1 Fatty acid stimulation of Asp 132Ala activity increased with longer
alkyl chains, while fatty alcohols had no efi‘ect. Activity was
assayed using a Gilson oxygraph that measured oxygen
consumption of the 13 nM Asp 132Ala mutant at 30 uM
cytochrome c, 1.1 mM TMPD, and 2.8 mM ascorbate in 50 mM
potassium phosphate, pH 6.5 at 25 0C. All fatty acids or alcohols
were dissolved in ethanol. For each fatty acid or alcohol
consecutive additions were made to a final concentration of 5 uM,
50 uM, and 250 uM to the enzyme during turnover. The numbers
in the table are relative to 100 which represents a turnover 50 seC'
1 for the Asp 132Ala mutant.

131

 

Table 5.1
% activity
with sequential addition of
WW

5 50 250
Fatty Acid or Fatty Alcohol
Na acetate 90 80 70
propionate 80 7 0 7 0
propanol 90 90 90
octanoic acid 90 100 120
octanol 100 80 80
lauric acid 120 120 150
lauryl alcohol 1 10 100 70
palmitic acid 140 2 10 150
hexadecanol 120 90 80
oleic acid (18:1A9) 120 160 140
linoleic acid (182139.12) , 190 260 310
linolenic acid (18:3A9,12,15) 130 170 150
arachidonic acid (20:A5,8,11,l4) 180 300 410
arachidonoyl alcohol 100 90 100

docosahexanoic acid (22:A4,7,10, 13,16,19) 260 380 330
docosahexanol 90 90 130

 

132

Table 5.2 Stimulation of activity by arachidonic acid is unique to Asp 132Ala.
Other mutants were tested by the protocol in Table 5.1 for
stimulation by arachidonic acid or arachidonyl alcohol. They were
also tested with 2 mg of cholate solubilized phospholipid. Two
mutations were in subunit II (noted in parenthesis), while the
others are all in subunit 1.

133

Table 5.2

maximum % stimulation of activity
with additign gt

 

initial arachidonic arachidonyl
activity acid alcohol asolectin

oxidase

wild type 1400 0 0 20
beef heart 450 0 0 10
Asp 132Ala 50 300 0 0
Thr359Ala 350 20 30 30
Asp4 12Asn 370 0 0 40
Phe391Gln 30 0 70 70
Lys362Met 20 0 0 0
Met262(II)Leu 130 40 20 0

Hi8259(II)Asn 10 0 0 0

 

134

on the Asp 132Ala mutant to see if they might have a stimulatory effect on the
activity. Table 5.3 indicates the results of testing these. Benzoic acid although
not a characterized inhibitor of the prostaglandin synthase is similar in
structure to aspirin. Aspirin may not be the best analogue since it blocks the
activity through a covalent modification, whereas the other two tested,
flurbiprofen and indomethacin are competitive inhibitors . Benzoic acid showed
the least stimulatory effect, indicating that more than just a hydrophobic
group and a carboxyl are needed for stimulation. Flurbiprofin showed a small
amount of stimulation at the highest concentration. Indomethacin showed
more significant levels of stimulation at the lowest concentrations and a
moderate level of stimulation at the highest concentration. In each experiment
after the addition of the highest concentration of fatty acid, arachidonic acid
was added. In all cases this showed further stimulation, and did not show more
than the normal amount of stimulation after indomethacin, indicating that the
effect was not cumulative. Thus it seems likely that indomethacin is
stimulating by the same mechanism as arachidonic acid.

Restoration of reconstituted Asp 132Ala respiratory control by
addition. of fatty acids The unusual inhibitory effect on the reconstituted
activity of Asp 132Ala by ionophores was discussed in Chapter 4 (see also
Figure 5.3 B). It was found that with the addition of fatty acid to the
reconstituted enzyme before the addition of the valinomycin and CCCP, the
inhibitory effect could be prevented (Figure 5.3 A). Since the fatty acid is
capable of acting as an uncoupler (Sharpe et al. , 1994), the concentration was
kept lower than that required for complete uncoupling. While it appears that
this may have restored the respiratory control, the effect is complicated
because of the dual role of uncoupling and of stimulation of the activity. As

135
shown in the previous chapter after CCCP is added to the Asp132Ala vesicles

it causes inhibition. Ifthis is followed with the addition of valinomycin a
stimulation in activity occurs. It seems possible then that the stimulation
with valinomycin as seen in Figure 5.3 A could be the same effect. It is still not
clear how the removal of one gradient (pH or membrane potential) can cause
inhibition of activity, then if followed by removal of the other gradient, a
stimulation occurs.

Experiments were also done to see if proton pumping might be restored.
Higher concentrations such as that used for the effect seen in Figure 5.3 A
started to cause uncoupling, seen as a faster rate of alkalinization. While it is
possible that some proton pumping could have been restored, it was not

measurable because of this uncoupling.

Discussion

The result that the reconstituted Asp132Ala generates a normal
membrane potential under steady state turnover conditions indicates that the
enzyme is inserted correctly in the membrane and protons are taken up
normally from the interior of the vesicles. This is the first direct evidence that
protons are being taken up from inside of the vesicles during steady state
turnover.

Fatty acids were shown to significantly increase the electron transfer
activity of the Asp 13 2Ala mutant. Longer chain length and the carboxyl of the
fatty acid are required for the stimulation with the largest effect from

arachidonic acid and docosahexanoic acid. This effect was only seen

136

Table 5.3 Stimulation of Asp 132Ala activity with analogues of arachidonic
acid. Both indomethacin and flurbiprofen are prostaglandin
synthase inhibitors. Benzoic acid is similar to aspirin that can
covalently modify the active site of prostaglandin synthase. Three
consecutive additions of fatty acid were made to the assay with a
final addition of arachidonic acid. Conditions for the assay are
indicated in Table 5.1.

137

 

Table 5.3
Flurbiprofen Indomethacin
Cl
I °“’
. ..
O ‘
COOH C - O
o . .,
CHzCOOH
caao
Arachidonic Acid
@335
%activity
with sequential addition of
WW followed by
50 umol
5 50 250 arachidonate

Analogue
benzoic acid 1 10 110 120 260
flin'biprofin 1 10 1 10 140 370

indomethacin 120 130 200 290

 

138

Figure 5.3 Restoration of respiratory control by the addition of arachidonic
acid. Respiratory control ratios were determined on an oxygen

polarograph at 25 0C in 1.8 mL of 10 mM HEPES, 41 mM KCl, 38
mM sucrose pH 7 .4. Ascorbate (10 uM final concentration) and
N,N,N’,N’-tetramethyl—p-phenylenediamine (TMPD) (0.5 uM final
concentration) were used to keep cytochrome c fully reduced. The
substrate cytochrome c at 30 uM final concentration was added to
stimulate turnover. To get maximal activity 2 nmol of
valinomycin and 10 nmol of carbonyl cyanide m-
chlorophenylhydrazone (CCCP) are added. A the restoration of
respiratory control with the addition of arachidonic acid to 20 uL of
vesicles giving a final concentration of 7 nM oxidase. B. the
inhibition of activity with the addition of valinomycin when fatty
acid is not added, using 50 uL of vesicles giving a final
concentration of 18 nM oxidase.

    
  

 

D1 32A
vesicles

 

139

arachidonic acid

valinomycin

 

Figure 5.3

140

with the Asp132 mutants, not with other mutants, nor with the wild type or
beef heart enzymes, indicating a unique interaction with the Asp132 mutants.
Since the Asp 132 mutants are not stimulated by phospholipids but instead by
fatty acids, while wild type and the other mutants show stimulation by
phospholipid but not by fatty acid, it can be concluded that the fatty acid and
phospholipid stimulatory efi‘ects occur by different mechanisms.

Although two analogues of arachidonic acid, benzoic acid and
flurbiprofin, showed very little stimulatory effect, indomethacin was able to
stimulate up to two fold, indicating that the binding site of the fatty acid may
have some similarity to the prostaglandin synthase active site. This provides
some evidence that the fatty acid may be in a folded conformation. Since the
stimulation was not as great as with the fatty acids, the greater flexibility of
the fatty acid may be required for the maximal effect.

A nonphysiological effect of the fatty acids A possible explanation
for the restoratory effect of the fatty acid is that it is binding in a position so it
can supply the missing carboxyl of the Asp132 mutants. Alternatively, it
could be interacting at another site on the enzyme to improve proton access to
the active site. This mechanism is similar to that observed in
bacteriorhodopsin (Otto et al., 1989; Tittor et al., 1989) and reaction center
(Takahashi and Wraight 1991) where the activity of mutants with deleted
acidic residues is restored by addition of soluble acids. Such acids have not
been tested on the Asp 132 mutants because most are inhibitory to the
oxidase.

If the fatty acids are improving activity by restoring proton access
through a different pathway the fatty acid might be expected to stimulate
activity of other mutants which are limited by proton access. Since none of

14 1
the other mutants tested showed any significant stimulation with fatty acids,

we may conclude that either these mutants are not limited by proton access or
if access is limited the fatty acids are not capable of binding at the mutation
site but bind only at the Asp132 mutation site. Further experiments will be
required to clarify the exact role of the fatty acids in the stimulatory activity
on the Asp132 mutants.

Could there be a physiological explanation for the stimulation of
Asp132 mutants by fatty acids? In considering the possibility of a
physiological effect, evidence exists that addition of fatty acids can stimulate
the activity of the beef heart cytochrome oxidase (Labonia et al. , 1988; Thiel
and Kadenbach 1989). This has not been shown to occur with the bacterial
wild type enzyme. However, these data show that noncovalently attached
fatty acids are capable of stimulating activity of the Asp 132 mutants. If one
or more fatty acids are important to the activity of the oxidase then mutation
of the Asp132 residue might decrease the affinity for fatty acids, so that higher
concentrations are required for stimulation. This mechanism can be compared
to the uncoupling protein in brown fat that passively transports protons
(Jezek et al., 1994). For proton transport to occur fatty acids must be
associated with the uncoupling protein, but are not covalently attached (Rial et
al., 1983). Theoretical models have been proposed in which the fatty acid is
bound in a stable conformation, with its carboxyl being used to move protons
(Klingenberg 1993). Other models propose that the fatty acid itself is moved
by the protein and in the process carries a proton across the membrane
(Skulachev 1991). More recent experiments by Garlid and coworkers indicate
that the fatty acid anion is transported by the uncoupling protein (Garlid,
personal communication). Whether this type of fatty acid role can relate to
oxidase activity or regulation remains to be determined.

142
One last possibility to consider is whether the fatty acid could play a

physiological role through a covalent attachment to the oxidase. Since fatty
acids have been found covalently attached through the N-terminal amino
group of glycine (Casey 1994; Milligan et al., 1995), it might be possible that a
fatty acid could attach to the oxidase through one of the two conserved
asparagines that have been shown to interact with the Asp 132 (Garcia?
Horsman et al., 1995). Covalent binding of fatty acids has also been shown to~
cysteines in membrane incorporated G proteins (Milligan et al., 1995), and in
the case of rhodopsin the alkyl chain appears to act as an anchor in the
membrane (Moench et al. , 1994). This type of mechanism could not be easily
reconciled with the results in the Asp 132 mutants because of the evidence
that the carboxyl group of the fatty acid, which would be unavailable if
covalently attached, is necessary for the stimulatory effect.

Reverse respiratory control The studies in this chapter have not
provided a much clearer explanation of the unusual respiratory control of the
Asp132 mutants. However, a possible explanation for the unusual effect of
ionophores on the reconstituted Asp 132Ala/Asn was proposed through a
personal communication with Marten Wikstrtim, and is worth some discussion
at this point. During turnover of the reconstituted enzyme the mutant may
still allow some proton uptake from the vesicle interior as in wild type,
generating a membrane potential and pH gradient. Formation of the
membrane potential may make it easier for the mutant to draw protons into
the active site from the outside of the membrane allowing oxygen reduction to
go faster. When either the pH or the electrical gradient is removed, proton
uptake from the outside would be less favorable, causing a decrease in the

activity of the enzyme.

143

This dual access model is harder to reconcile with the observation that
the rate shows a decrease with the removal of one gradient (Aw or ApH) and
then when this is followed by the removal of the other gradient, an increase in
rate occurs. However, the amount of inhibition and stimulation varies with the
preparation of enzyme from a very small stimulation (Figure 5.3B) to almost
total restoration of the rate to that observed before addition of ionophores
(Figure 4.7). It is therefore likely that two opposing efi‘ects are involved: first,
a normal inhibition of the enzyme activity by the gradients, and second, a
stimulation of the activity caused by the gradients promoting easier proton
uptake from the outside. Thus release of both gradients removes the normal
inhibitory efl'ect and allows the rate to increase to the extent that proton
access routes will allow. Further studies are needed to characterize this effect
which could provide new information about the mechanism of proton pumping
and respiratory control in cytochrome oxidase.

Chapter 6 — Further mutagenesis of residues in the IX-
X Loop

Introduction

As discussed in Chapter 4, the mutation of three of the four highly
conserved residues in the DK-X extracellular loop caused some inhibition of
electron transfer, but did not inhibit proton pumping. The lowered activity
could be indicative of a role in proton pumping, if the normal proton pathway
could be circumvented by another pathway or the mutated residue replaced by
a water. Other studies have shown that the loop overlays the active site in
subunit 1, with His411 and Asp412 being near the heme a3/CuB site and the
Thr413 and Tyr414 close to heme a (Hosler et al., 1994a; Hosler et al. , 1994b).
It was also shown that mutation of His411 to Ala, or Asp412 to Asn, causes
the loss of a nonredox active manganese and inhibits the electron transfer
activity by about 50% (Hosler et al. , 1995). Studies of this manganese in Rb.
sphaeroides show that it can be replaced by a magnesium (Hosler et al., 1995).
Low levels of manganese (Fee et al., 1986; Numata et al., 1989; Lauraeus et
al., 1991; Haltia 1992) or stoichiometric levels of magnesium (Buse and
Stefl‘ens 1991) are found in other bacterial oxidases, and all eukaryotic
enzymes so far examined, suggesting some general role of this non-redox active
metal site. Because the soluble portion of subunit 11 containing the binuclear
CuA must closely approach the binuclear center, it was proposed that the
manganese/magnesium might be involved in stabilizing the interaction of
subunits I and II to facilitate electron transfer between the binuclear Cu A and
heme a(Hosler et al. , 1995). Using X-band and Q-band EPR, and ESEEM the
manganese has been shown to be a six-coordinate metal (Espe et al. , 1995).
This study also showed that the redox state of the enzyme had only a small
efi‘ect on bond lengths or angles at this metal center. Since it appears to be

144

145
located near the active site, the Mn might be expected to be sensitive to

conformational changes that occur during reduction of the enzyme. Thus it
was concluded that the protein does not undergo major conformational
rearrangements between the two states in the area of the active site. A model
of the Mn site based on the crystal structure of a manganese site with similar
EPR characteristics in lectins (Espe et al. , 1995), predicts that His411 is a
direct ligand to the metal, while Asp412 may act as the ligand to a water that
is a direct ligand to the metal.

Mutation of Thr413 and Tyr414 did not have an effect on manganese
binding. But while mutation of Thr413 to Asn showed insignificant spectral
effects, mutation of the Tyr414 to Phe caused a shift of the alpha band from
605.4 to 610 nm. This agrees with data fi'om the oxidase in Tetrahymena
pyriformis that has a phenylalanine in place of the tyrosine and an alpha band
at 617 nm (Ziaie and Suyama 1987).

This IX-X loop is located where residues may be involved in transferring
protons away from the binuclear center. Thus the possibility exists that the
Mn/Mg could play some role in this process. To further investigate the
importance of these residues, four more mutations were made including,
His411 to Gln and Asp412, Thr413, and Tyr414 to alanine. A preliminary
characterization is reported in this chapter.

Experimental procedures

Any basic molecular biology details not listed below followed those in
Sambrook et al. (1989). Collection of visible absorbance spectra and
determination of steady state electron transfer activity were done as indicated
in Chapter 4 with any other details listed in the figure legends.

146
Materials Restriction enzymes and ligase were obtained from either

Boehringer Mannheim Biochemicals or Gibco BRL. The reagents for the
mutagenesis procedure were as follows: T4 DNA polymerase and Exonuclease
111 (USB), 5Me-dCTP (Pharmacia), T4 Kinase and T4 Ligase (NEB or
Amersham), MspI (BRL), Hha I (NEB), Nucleotides (Pharmacia), ATP
(Sigma). Sequencing was done using the Sequenase version 2.0 kit fi'om USB
Large scale purification of oxidase was done as indicated in Hosler et al. (1992),
except that no N aBr wash was done, and only a single DEAE anion exchange
chromatography on an FPLC was used for His411Gln, and Tyr414Ala. For
Thr413Ala and Asp412Ala two anion exchange chromatography steps were
used.

Mutagenesis and subcloning The Batt Method of mutagenesis was
used to make the mutants (Vandeyar et al., 1988), with some modifications to
the procedure (Lemieux et al. , 1992). Primers 18-21 bp in length were made for
each of the mutants. For the Thr413Ala mutant a primer was made that
introduced the MluI restriction site, to allow easier detection of this mutant.
The procedure required the use of several previously used plasmids (Shapleigh
and Gennis 1992). Single stranded DNA was prepared from M13 containing a
220 bp Sal I/Sph I piece from the subunit I cytochrome oxidase gene from Rb.
sphaeroides. T4 DNA polymerase was used to extend the primer using dATP,
dGTP, d'ITP, and 5—methyl-dCTP as indicated in (Lemieux et al. , 1992). The
methylation served to protect this strand from MspI and HhaI which could
then be used to nick the non-mutant strand. Exonuclease III was used for
complete cleavage and removal of this strand. This was transformed into E.
coli and allowed to propagate. Single stranded DNA was prepared fi'om
multiple plaques. Sequencing of this DNA was done using the Sequenase 2.0 kit
fi'om USB to detect those colonies which contained the mutant of interest.

147
Next the Sal I/Sph I piece was subcloned into a larger piece of pND38

(see Figure 6.1). This was required as an intermediate step because the
complete subunit I sequence contains 2 Sal I sites, preventing a direct ligation
into that piece. From pND38 a Bgl II/Hind III segment was removed and
ligated into pJ S3. Next the entire subunit I piece was put into pRK415-1 using
HindIII/EcoRI, and screened using XGAL and IPTG (Sambrook et al. , 1989).
This plasmid is stable in E. coli and Rb. sphaeroides, allowing it to be a shuttle
to the Rb. sphaeroides. A Sal I/Hind III piece was moved into M13 and
sequenced using the Sequenase 2.0 kit, to confirm that the mutation was still
present. Then the pRK415-1 was transformed into 817-1, a strain of E. coli
capable of conjugating with Rb. sphaeroides. After conjugation with J S 100, the
subunit I deletion strain of Rb. sphaeroides, the mutant DNA fiom Rb.
sphaeroides was transformed into DH5a'. Again the Sal I/Hind III piece was
moved into M13 to allow a final confirmatory sequencing.

Conjugation These strains were conjugated based on the procedure in
Donahue et al. (1988) as follows. E. coli Sl7-1 containing the mutant in
pRK415-1 was grown overnight in 3 mL LB, 25 ug/mL tetracycline at 37 0C.
Rb. sphaeroides (JSlOO) was grown in 50 mL of Sistrom's media at 30 0C
without antibiotics. On the second day 500 uL of the overnight culture of Sl7-'
1 was added to 50 mL of LB, 25 ug/mL tetracycline and grown for 1.5 hours.
From the E. coli and the Rb. sphaeroides cultures, 1 mL of each was spun down
at 5000 x g for 1 min in a microcentrifuge. To each pellet, 500 uL of LB was
added, and then both samples were mixed together. Following centrifugation
again at 5000 x g for 1 min, the pellets were resuspended in 100 uL of LB. For
the actual conjugation, Millipore HA filters (0.45 11) were placed

148

Figure 6.1 Site-directed mutagenesis and subcloning to change residues in
subunit I of Rb. sphaeroides. Site-directed mutagenesis was done
in M13 using the Batt Method. Subcloning through pND38
containing a piece of the subunit I gene was necessary because
pJ S3 contains two Sal I sites. The entire subunit I gene is
contained in pJ S3. Vector pRK415-1 can propagate in E. coli and
Rb. sphaeroides, allowing it to be used in the conjugation. E. coli
strain Sl7-1 can conjugate with the Rb. sphaeroides to transfer
the subunit 1 gene.

149

  
     
    
   
   

 

 

 

 

 

 

Batt Method
Site Directed 4M»
Mutagenesis W033 pND38
in M13:
Sequence to
Find Mutant
Bgl u
8.“ SP“ Sail Sphl I-llndlll
syn/Hindu: ”use
EcoRl EcoRl
Blue/White
8cm" '0', Hind Ill I Ecol! l
Insertion . PRK‘" 5'1 «isms-1
into \ Ml
pRK41 5-1 \
3; ll
sell 8PM "‘"d'" Sall Sp,” mm
ltranatormation
Sequence to III
check that
mutation is [ 817.1 I .j + C M aa3 V]
in pRK415-1
5- 60” Rb. sphaeroides
subunit I deletion
conjugation i
ll
. .
Rb. sphaeroides

Figure 6.1

150
on an LB plate, to which the cells were added and allowed to grow at 30 0C for
4 hours. The filter was put in a tube with 1 mL of Sistrom's media, that was
shaken for 10 minutes to resuspend the mixture of E. coli and Rb. sphaeroides.
The solution was centrifuged at 1500 x g for 5 min, and resuspended in 1 mL of
Sistrom's. From this sample 25 uL were plated on Sistrom's plates containing
1 ug/mL tetracycline, 50 ug/mL streptomycin, and 50 ug/mL spectinomycin,
and grown for 3 days at 30 00.

Small scale membrane preparations Two colonies from each
mutant conjugation were grown in 100 mL Sistrom's in 500 mL Erlenmeyer
flasks with shaking at 300 rpm at 30 0C for 2 days. Preparation of
membranes for spectra followed the protocol in Hosler et al. (1992) except that
the cells were broken by a single pass through a French press without the
addition of protease inhibitors. After the low speed spin to remove cell debris at
20,000 x g for 20 min, the membranes in the supernatant were pelleted at
200,000 x g for 1.5 h. The pellet was solubilized in 50 mM potassium
phosphate bufi‘er, pH 7.2 and 2% lauryl maltoside. This was spun in a
microcentrifuge at 16,000 x g for 10 minutes and the supernatant was used for
the spectra. Dithionite reduced-ferricyanide oxidized spectra were taken to
confirm the presence of cytochrome oxidase. An extinction coefficient of 24
mM"1cm‘1 at 606-650 was used to determine the concentration.

Results

Thr413Ala was the first mutant to be completed, since it had the
introduced MluI restriction site, which allowed screening for the mutation after
each step of subcloning and ligation. Once in Rb. sphaeroides, a large scale
preparation of this enzyme was done, and it showed almost identical
characteristics to that of the wild type enzyme, like that observed for

15 1
Thr413Asn (Hosler et al., 1994b). This mutant was reconstituted and showed

normal proton pumping (data not shown).

The other three mutants were somewhat more difficult to work with
since in moving the first piece into pND38 and then into pJ S3, the wild type
insert was already present in the vector, maln'ng it impossible to distinguish
the wild type from the mutant until the sequencing was done.

Figure 6.2 shows the reduced-oxidized difference spectra of the mutants
in the detergent-solubilized membranes. It appears that the spectrum for the
His411Gln is shifted to 607 run, while the other mutants look normal.
However, it was discovered in looking at previous membrane preparations,
that the slope of the baseline could cause spectra to shift, making it difficult to
obtain an accurate peak wavelength. Thus spectra at this stage of
preparation can only be used to indicate the presence of an intact oxidase, not
to assess spectral properties.

Large scale purifications were done to isolate each enzyme. Only a
single purification over a DEAE anion exchange was used for His411G1n and
Asp412Ala. Thr413Ala and Tyr414Ala were purified twice with the DEAE
column. The spectra of the purified samples are shown in Figure 6.3. The wild
type sample was purified from an overexpression strain, omitting the sucrose
gradient and the hydroxyapatite chromatography steps. This gave a more
active enzyme than what is normally isolated. This sample looks spectrally
pure as can be seen by the lack of contaminating C heme at 416 nm in the
Soret. This 0 heme is seen in the other spectra, and comes from either the bc1

complex or from the cbb3 oxidase or membrane bound c that is associated

152

Figure 6.2 Spectra of IX-X loop mutants in detergent-solubilized membranes.
Thr413Ala is from a large scale purification (12 L) after the
sucrose gradient, while the others are from small scale (100 mL)
purifications. Only 1% of the large scale preparation was used for
the spectra but the entire sample from the small scale
preparations were used. All were solubilized in 50 mM potassium
phosphate, pH 7 .2 and 2% lauryl maltoside detergent. After
centrifugation, the supernatant was divided into two cuvettes.
Dithionite was used to reduce one sample and ferricyanide was
used to oxidize the other. Spectra are displayed as reduced-
oxidized. The shape on the blue side of the Thr413Ala peak is
likely due to a scattering artifact. Its peak position is normal, and
its purified spectrum in Figure 6.3 looks normal.

153

 

Y41 4A

T413A

D412A
H41 1 .

 

 

 

 

500 605 700
Wavelength (nm)

Figure 6.2

154

Figure 6.3 Spectra of the purified Di-X loop mutants compared to wild type.
Samples were solubilized in 100 mM potassium phosphate buffer,
pH 7 .2, 0.2% lauryl maltoside, giving a range of final
concentrations from 0.309 uM. Samples were allowed to reduce
to completion after the addition of dithionite. Spectra were
normalized using DeltaGraph software so that each was on the
same scale. Most of the spectral contribution at the 605 nm peak
is fi'om heme a, while the peak at 445 nm has an equal
contribution fiom heme a and heme 03. Small blue shifts at 605
nm can be seen for the Asp and Thr mutants; the His mutant
shows no shift at 605 nm; and the Tyr mutant shows the largest
shift to 604 nm.

Absorbance

155

 

 

 

 

 

 

 
  

 

Tyr414Ala

Thr413Ala

Asp412Ala

His411Gln

wild type

 
 
   
 
 

 

 

 
 

 

 

400

T

450

I i
500 550

T
600

Wavelength (nm)

Figure 6.3

156
with the aa3 oxidase. This contaminant makes it dificult to reconstitute the

enzyme and assay for proton pumping. Steady state electron transfer activity
can still be determined.

Electron transfer activity for the mutants and visible spectral
characteristics are shown in Table 6.1. Results are also given for the
previously characterized mutants in this loop, along with the results for wild
type enzyme purified by the complete procedure that gives somewhat lower A
activity. All of the newly made mutants showed significant electron transfer
activity. Asp412Ala showed the greatest inhibition. Although Tyr414Ala
showed high electron transfer activity, it had the largest spectral shift, to 604
nm. This, however, is a shift in the opposite direction to the previously
characterized Tyr414Phe mutant which shifts to 610 nm (Hosler et al. ,
1994b). Doing the same assay with the addition of a cholate/phospholipid
mixture before addition of the enzyme gave higher activity in each case.
Previously characterized activities were only done only with the mixture of
cholate/phospholipids added.

Discussion

The mutants were cloned out of Rb. sphaeroides into E. coli. and then
into M13 one last time to do sequencing. None of the mutants previously
studied by this lab had been resequenced after being put into Rb. sphaeroides,
but it was done this time as an extra precaution. Thr413Ala was not checked
since it showed normal characteristics. It was also decided to check the
previously studied His411Asn because of its complete lack of activity even
though Asn is considered to be a good replacement for His. The mutants
His411Gln, Asp412Ala, and Tyr414Ala were shown to have the

157

Table 6.1 Visible spectral shifts and activity of the IX-X loop mutants. The
upper table contains data from the newly made mutants. In the
bottom is data from previous mutants with the activity from
Fetter et al. (1995) and the spectral characteristics from Hosler et
al. (1994b). Peaks for the alpha and Soret region were determined
from the spectra in Figure 6.3. Steady state electron transfer
activity was assayed as indicated in Table 4.1 except that an
assay was also done without the addition of phospholipids.

158

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 6.1
(nm) (sec'l) plus lipids
(sec '1)
newly
characterized
syncs
wild type 605.4 444.4 1500 2200
His 411Gln 605.4 443.8 1200 1500
Asp412Ala 604.8 443.5 400 600
Thr413Ala 604.8 443.1 1200 1400
Tyr414Ala 604 443.3 1200 1500
max (nm) plus lipids
(sec'l)
previously
characterized
,flzjmes
wild type 605.4 444.4 1700
His411Ala 604.8 443.8 800
His 411Tyr 604.2 441.1 0
Asp412Asn 605.4 443.8 300
Thr413Asn 605.4 443.8 1300
Tyr414Phe 610 445.6 700

 

159
expected sequence. His411Asn was found to actually have the sequence for

Tyr at the position of 411. This better explains the results of its loss of
activity, its visible spectral shifts, and its loss of the Fe-His stretch in the
resonance Raman spectra (Hosler et al. , 1994b). It seems likely that the
larger Tyr (Chothia 1975) with aromatic character would interact with the
protein environment in a different way than a His, or Asn and be less likely to
be a substitute ligand for Mn or Mg.

His411Gln shows only a slight shift in the Soret spectra and high
activity. Since the previously characterized mutant at this site, His411Asn,
.had major spectral alterations and had lost all activity, Gln might have been
expected to be similar, but with the information from resequencing that the
mutant was actually a Tyr, the result is more interpretable. It is interesting
that the His411Gln retains such high electron transfer activity (1500 sec’l),
higher than the previously characterized His411Ala (800 sec‘l). Since the
His41 lAla had lost the nonredox active manganese, it would seem that the Gln
may still be providing a ligand for that metal site, keeping the enzyme almost
as stable and active as the wild type enzyme.

Asp412Ala shows some small visible absorbance spectral shifts and
significantly lower electron transfer activity (600 sec’l), similar to that of the
previously characterized Asp412Asn (300 sec'l). Since Asp412Asn was
shown to lose the non-redox active manganese (Hosler et al., 1995), it seems
likely that this is also occurring with the Asp412Ala, and that this introduces
some structural instability that is causing the lower electron transfer activity.
These results might be compared to those found with reaction center, in which
mutation of the ligands to the Fe that lies between QA and QB in the electron
transfer path slows down the steady state activity, but does not completely
inhibit it (Williams et al., 1991).

160
It is very interesting that Tyr414Ala shows only a small shift in the

alpha band (to 604 nm) and in the opposite direction to the large shift of
Tyr414Phe (to 610 nm). We previously hypothesized that either loss of the
hydroxyl and a possible hydrogen bond, or repositioning of the aromatic moiety
close to heme a, caused the shift in the heme a spectrum. Since Tyr414Ala
shows a much smaller effect on the alpha band and would also have lost a
potential hydrogen bond, the shift to 610 nm caused by Tyr414Phe seems
likely to be the result of a new interaction of the phenylalanine with heme a as
was previously proposed (Hosler et al. , 1994b).

These results support the previous data for the residues in the Df-X loop
that showed that this loop lies over heme a, heme a3/CuB, and are consistent
with the model for the active site that has been proposed (Hosler et al., 1994b).
Further analysis of the mutants will provide new information on their functional
or structural importance.

Summary and Perspectives

A reconstitution and proton pumping assay were developed as indicated
in Chapter 3. Many variables in the reconstitution make it difiicult to get
reproducible and accurate results in the proton pumping assay. Of the factors
found to be potential problems, the most important are the purity of the
enzyme, cholate, and the phospholipids. One potential for improvement would
be to switch to purified lipids instead of the variable mixtures of soya bean
phospholipids that are now used. Other factors were found for which it is more
diflicult to determine an optimum such as the details of the pretreatment of
the enzyme with cholate, the ratio of enzyme to lipid used, and the
concentration of enzyme used during the reconstitution. Since the main goal of
this project was to be able to measure proton pumping and not to understand
reconstitutions, time constraints did not allow a complete characterization of
these variables.

Much of the project focused on purifying and characterizing various
mutant oxidases that might show some inhibition of proton pumping. Until
interesting mutants are found this is a repetitive search through many
mutants that may not show a significant effect on proton pumping. This was
the approach used on bacteriorhodopsin, where the focus of the mutagenesis
was not limited by looln'ng at conserved residues as has been done in oxidase.
Hundreds of residues were mutated to find several that showed the strongest
effects on proton pumping (Krebs and Khorana 1993). This may say
something about the resiliency of proton pumping enzymes to mutation, and
may indicate the possibility for either multiple proton paths through the
protein, or that water molecules may be able to replace lost residues in the
path. Many of the mutants in Rb. sphaeroides oxidase show some inhibition in
electron transfer, so that they could not be completely ruled out of an

161

162

involvement in proton pumping. Of course with the present data, it also can
not be ruled out that the mutations have altered the structure in a way to slow
electron transfer without inhibiting proton pumping. Better techniques for
determining the effects of the mutations need to be developed to determine the
cause of the lower electron transfer activities.

One possibility is to develop a stopped-flow assay to measure proton
pumping on a faster time scale, allowing a more quantitative measurement in
changes in the rate of proton pumping. Unfortunately to increase the time
resolution simultaneously requires a greater optical absorption signal. Thus
more enzyme is needed to make larger batches of vesicles to generate larger
changes in proton pumping. This will require better overexpression of the
oxidase as is currently being developed by others in the lab. A possible way
around the requirement may be to use a steady state proton pumping assay.
While this can cause uncoupling of proton pumping fi'om electron transfer
(Capitanio et al. 1991; Papa et al. 1991), it may still be possible to determine
differences in the proton pumping activity of the mutants.

It may also be of interest to apply to the aa3 oxidase any of those
techniques used to measure fast electron transfer, either by optical absorption
or by resonance Raman. These should be ideal to determine how specific
mutations are affecting the enzyme. Presently, studies will need to focus on
those techniques such as rapid electron transfer from ruthenated-cytochrome
c which require smaller amounts of sample (Pan et al. 1993). Techniques such
as the information-rich flow flash resonance Raman spectroscopy will require
much larger amounts of enzyme than are now available. Hopefully,
improvements in overexpression of the Rb. sphaeroides oxidase will also make
this feasible.

163

It is exciting when an obvious effect on proton pumping is found as is the
case with the loss of proton pumping with mutation of Asp132. Studies of a
single mutant can provide many insights into the proton pumping mechanism
in cytochrome oxidase. Study of this mutant is only now in the beginning
stages, but has already shown something interesting with the discovery that it
has a unique stimulatory response to fatty acids not seen with the other
proton pumping-inhibited mutants. Experiments showed that a longer alkyl
chain length gave the greatest stimulation of the Asp 132 mutant, and that the
carboxyl of the fatty acid was required for this effect. Further experiments can
be done to provide more detail about this interaction. Where exactly on the
enzyme is the carboxyl of the fatty acid interacting? Where is the hydrophobic
tail interacting? What is the conformation of the tail? The studies in Chapter
5 showing that indomethacin can have a two-fold stimulatory effect indicate
that the fatty acid may be folded. Yet, the fatty acid has a larger effect
implying that some flexibility may be important. Determination of the crystal
structure by Michel opens up new possibilities for study. Perhaps the crystals
could be soaked with fatty acids, or with analogs of the fatty acids to see if they
bind in a specific region of the oxidase.

Finally, four more mutants were made in the loop between helices D(
and X. This was important training in allowing me to gain experience in doing
site-directed mutagenesis, DNA sequencing, and other molecular biology
techniques. It has also provided new information about the loop. The low
activity of the Asp412Ala indicates that it may also have lost a manganese
(magnesium), and that probably because of a structural effect is slowing
electron transfer between Cup, and heme a. A mutant in which His411 is
converted to Gln retains high activity, and likely still retains the manganese

164
(magnesium). It will be of interest to see how changing the ligand from a

histidine to a glutamine has affected the EPR properties of this metal center.
All together the mutations have provided some information about the
regions that are required for proton pumping. The II-III loop appears to be
important as the proton entry point into the oxidase. On the other side of the
membrane the IX-X loop appears to have an interaction with all of the metal
centers of the enzyme, to have ligands to a manganese (magnesium) ion, and
to lie between the CuA and the active site. Clearly protons that are exiting on
this side of the membrane must be nearby, and further studies of this loop, and
of conserved residues in other loops on this side of the membrane with
comparison to the crystal structure may clarify where the protons are exiting.
One interesting loop that has not received much attention yet is the XI-XII
loop that contains two conserved arginines, and a conserved tyrosine, proline
and aspartic acid. In the middle of the membrane, mutation of residues such
as Thr352 and Lys362 cause complete inhibition of electron transfer. They
may very well play a role in proton transfer, but it will require more subtle
methods to determine their role. Sitting between the two hemes Tyr422 has
been shown not to be critical for proton pumping as proposed by Rousseau et
al. 1993, narrowing down the search for those residues that are. With the
crystal structure of oxidase, the ability to make mutations, and the advances
in the various spectroscopic techniques used to study the enzyme, many
interesting discoveries about the mechanism of this enzyme should be made in

the ensuing years.

APPENDIX
Publications and Abstracts

APPENDIX

PUBLICATIONS

D.S. Wagner, A. Salari, D.A Gage, J. Leykam, J. Fetter, R. Hollingsworth,
and J .T. Watson (1991) “Derivatization of Peptides to Enhance Ionization
Efficiency and Control Fragmentation During Analysis by Fast Atom
Bombardment Tandem Mass Spectrometry” Biological Mass Spectrometry 20,
419-425.

J .P. Hosler, J. Fetter, M.M.J. Tecklenburg, M. Espe, C. Lerma, and S.
Ferguson-Miller (1992) “Cytochrome aa3 of Rhodobacter Sphaeroides as a Model
for Mitochondrial Cytochrome c Oxidase: Purification, Kinetics, Proton Pumping
and Spectral Analysis” Journal of Biological Chemistry 267, 24264-24272.

J .P. Hosler, S. Ferguson-Miller, M.W. Calhoun, J .W. Thomas, L. Lemieux,
J. Ma, C. Georgiou, J. Fetter, J. Shapleigh, R.B. Gennis (1993) “Insight into the
Active-Site Structure and Function of Cytochrome Oxidase by Analysis of Site-
Directed Mutants of Bacterial Cytochrome aa3 and Cytochrome bo” Journal of
Bioenergetics and Biomembranes 25, 121-136.

J .P. Hosler, J .P. Shapleigh, M.M.J., Tecklenburg, J .W. Thomas, Y. Kim, M.
Espe, J. Fetter, G.T. Babcock, J .O. Alben, R.B. Gennis, S. Ferguson-Miller (1994)
“A Loop between Transmembrane Helices IX and X of Subunit I of Cytochrome c
Oxidase Caps the Heme a-Heme a3-CuB center” Biochemistry 33, 1194-1201.

J .R. Fetter, J. Shapleigh, J .W. Thomas, A. Garcia-Horsman, E. Schmidt, J.
Hosler, G.T. Babcock, R.B. Gennis, S. Ferguson-Miller. “Possible Proton Relay
Pathways in Cytochrome c Oxidase” Proceedings of the National Academy of
Science, 92, pp. 1604-1608.

ABSTRACTS

J. Hosler, J. Fetter, J. Shapleigh, M. Espe, J. Thomas, Y. Kim, R. Gennis,
G. Babcock, and S. Ferguson-Miller (1992) “Native and engineered forms of
Rhodobacter sphaeroides cytochrome c oxidase. A model for the mitochondrial
enzyme.” EBEC Short Reports 7, II-45.

J. Fetter, J .P.. Hosler, J .P. Shapleigh, M.M.J. Tecklenburg, R.B. Gennis,
G.T. Babcock and S. Ferguson-Miller (1992) “Analysis of Proton Pumping By
Site-Directed Mutagenesis of Rhodobacter Sphaeroides Cytochrome c Oxidase”
Biophysical Journal 61, No.2, Part 2, A203.

J. Fetter, J .P. Hosler, J .P. Shapleigh, J .W. Thomas, M.M.J. Tecklenburg, Y.
Kim, R.B. Gennis, G.T. Babcock, and S. Ferguson-Miller (1993) “Effects of Site-
Directed Mutagenesis on Vectorial Proton Translocation in Rhodobacter
Sphaeroides Cytochrome C Oxidase” Biophysical Journal 64, No. 2, A103.

165

166

J .P. Hosler, M. Espe, J. Fetter, J .P. Shapleigh, J. Thomas, M.M.J.
Tecklenburg, Y. Kim, Y. Zhen, R.B. Gennis, G.T. Babcock, and S. Ferguson-Miller
(1993) “Metal Center Ligation and Proton Pumping in Cytochrome c Oxidase”
Workshop on Structure-Function of Ion-Translocating Complexes, Freiburg,
Germany

J. Fetter, J .W. Thomas, Y. Kim, R.B. Gennis, G.T. Babcock, S. Ferguson-
Miller (1994) “Identification of a Carboxylate Residue Critical for Proton Pumping
on the Inner Side of Subunit I of Cytochrome c Oxidase” Biophysical Journal 66,
No. 2, A366.

J. Hosler, J. Fetter, D. Mitchell, M. Pressler, J. Alben, R. Gennis, G.
Babcock, & S. Ferguson-Miller (1995) “An Experimental Test of a Ligand
Switching Mechanism for Cytochrome c Oxidase” Biophysical Journal 68

J .R. Fetter, M.A. Sharpe, J. Qian, P. Nicholls, & S. Ferguson-Miller
(1995) “Free Fatty Acids Stimulate the Activity of a Proton Pumping Mutant
of Cytochrome aa3” FASEB Journal 9, No. 6, A552.

BIBLIOGRAPHY

Bibliography

Abrahams, J. P., Leslie, A. G. W., Lutter, R. & Walker, J. E., (1994) Structure
at 2.8 A Resolution of F l-ATPase from Bovine Heart Mitochondria.

Nature 370, 621-628.

Akeson, M. & Deamer, D. W., (1991) Proton Conductance by the Gramicidin
Water Wire. Biophys. J. 60, 101-109.

Allen, J. P., Feher, G., Yeates, T. 0., Komiya, H. & Rees, D. C., (1988)
Structure of the Reaction Center from Rhodobacter Sphaeroides R-26:
Protein-Cofactor (Quinones and Fe2+) Interactions. Proc. Natl. Acad.
Sci. USA 85, 8487-8491.

Althoff, G., Lill, H. & Junge, W., (1989) Proton Channel of the Chloroplast ATP
Synthase. CFO: Its Time-Averaged Single-Channel Conductance as
Function of pH, Temperature, Isotopic and Ionic Medium Composition.
J. Membr. Biol. 108, 263-271.

Anderson, S., De Bruijn, M. H. L., Coulson, A. R., Eperon, I. C., Sanger, F. &
Young, I. G., (1982) Complete Sequence of Bovine Mitochondrial DNA:
Conserved Features of the Mammalian Mitochondrial Genome. J. Mol.
Biol. 156, 683-717.

Anemtiller, S. & Schafer, G., (1990) Cytochrome aa3 from Sulfolobus
acidocaldan'us. A Single-subunit, Quinol-Oxidizing Archaebacterial
Terminal Oxidase. Eur. J. Biochem. 191, 297-305.

Antholine, W. E., Kastrau, D. H. W., Stefl'ens, G. C. M., Buse, G., Zumft, W. G.
& Kroneck, P. M. H., (1992) A Comptirative EPR Investigation of the
Multicopper Proteins Nitrous-Oxide Reductase and Cytochrome c
Oxidase. Eur. J. Biochem. 209, 875-881.

Azzi, A., Casey, R. P. & N alecz, M. J ., (1984) The Effect of N',N'-
Dicyclohexylcarbodiimide on Enzymes of Bioenergetic Relevance.
Biochim. Biophys. Acta 768, 209-226.

Babcock, G. T. & Callahan, P. M., (1983) Redox-Linked Hydrogen Bond
Strength Changes in Cytochrome 0: Implications for a Cytochrome
Oxidase Proton Pump. Biochem. 22, 2314-2319.

Babcock, G. T. & Wikstriim, M., (1992) Oxygen activation and the
conservation of energy in cell respiration. Nature 356, 301-309.

167

168

Bartl, F., Deckers-Hebestreit, G., Altendorf, K. & Zundel, G., (1995) The F0

Complex of the ATP Synthase of Escherichia coli Contains a Proton
Pathway with Large Proton Polarizability Caused by Collective Proton
Fluctuation. Biophys. J. 68, 104-110.

Bashford, C. L. & Smith, J. C., (1979) The Use of Optical Probes to
Monitor Membrane Potential. Methods Enzymol. 55, 569-586. ’

Beattie, D. S., (1993) A Proposed Pathway of Proton Translocation through the
be Complexes of Mitochondria and Chlomplasts. J. Bioenerg. Biomembr.
25, 233-44.

Berry, E. A. & Trumpower, B. L., (1987) Simultaneous Determination of
Hemes a, b, and c from Pyridine Hemochrome Spectra. Anal. Biochem.
161, 1-15.

Bisson, R. & Schiavo, G., (1986) Two Different Forms of Cytochrome c Oxidase
Can be Purified fipm the slime Mold Dictyostelium discoideum. J. Biol.
Chem. 261, 4373-4376.

Bisson, R. & Schiavo, G., (1988) Slime Mold Cytochrome c Oxidase. Ann. NY.
Acad. Sci 550, 325-336.

Bombelka, E., Richter, F.-W., Stroh, A. & Kadenbach, B., (1986) Analysis of
the Cu, Fe, and Zn Contents in Cytochrome c Oxidases from Different
Species and Tissues by Proton-Induced X-ray Emission (PIXE).
Biochem. Biophys. Res. Commun. 140, 1007-1014.

Braiman, M. S., Bousché, O. & Rothschild, K. J ., (1991) Protein Dynamics in
the Bacteriorhodopsin Photocycle: Submillisecond Fourier Transform
Infrared Spectra of the L, M, and N Photointermediates. Proc. Natl.
Acad. Sci. USA 88, 2388-2392.

Brand, M. D., Chien, L.-F. & Diolez, P., (1994) Experimental Discrimination
Between Proton Leak and Redox Slip During Mitochondrial Electron
Transport. Biochem J. 297, 27-29.

Brandl, C. J. & Deber, C. M., (1986) Hypothesis About the Function of
Membrane-Buried Praline Residues in Transport Proteins. Proc. Natl.
Acad. Sci. 88, 917-921.

Brautigan, D. L., Ferguson-Miller, S. & Margoliash, E., (1978) Mitochondrial
Cytochrome c: Preparation and Activity of Native and Chemically
Modified Cytochromes c. Methods Enzymol. 53, 128-164.

169

Brzezinski, P. & Malmstrom, B. G., (1986) Electron-Transport—Driven Proton
Pumps Display Nonhyperbolic Kinetics in the Oxidation of
Ferricytochrome c catalyzed by Cytochrome c Oxidase. Proc. Natl. Acad.
Sci. 83, 4282-4286.

Biige, U. & Kadenbach, B., (1986) Influence of Bufi‘er Composition, Membrane
Lipids and Proteases on the Kinetics of Reconstituted Cytochrome c
Oxidase from Bovine Liver and Heart. Eur. J. Biochem. 161, 383-390.

Buse, G. & Stefl‘ens, G. C. M., (1991) Cytochrome c Oxidase in Paracoccus
denitrificans. Protein, Chemical, Structural, and Evolutionary Aspects.
J. Bioenerg. Biomembr. 23, 269-289.

Calhoun, M. W., Thomas, J. W., Hill, J. J., Hosler, J. P., Shapleigh, J. P.,
Tecklenburg, M. M. J ., Ferguson-Miller, S., Babcock, G. T., Alben, J. O. &
Gennis, R. B., (1993) Identity of the Axial Ligand of the High-Spin Heme
in Cytochrome Oxidase: Spectroscopic Characterization of Mutants in
the bo-type Om'dase of Escherichia coli and the aa3-type Oxidase of

Rhodobacter sphaeroides. Biochemistry 32, 10905- 10911.

Cao, J ., Shapleigh, J ., Gennis, R., Revzin, A. & Ferguson-Miller, S., (1991) The
Gene Encoding Cytochrome c Oxidase Subunit II from Rhodobacter
sphaeroides; Comparison of the Deduoed Amino Acid Sequence with
Sequences of Corresponding Peptides from Other Species. Gene 101,
133-137 .

Cao, J ., Shapleigh, J ., Gennis, R., Revzin, A & Ferguson-Miller, S., (1992)
Cytochrome aa 3 of Rhodobacter sphaeroides as a Model for
Mitochondrial Cytochrome c Oxidase. J. Biol. Chem. 267, 24273-24278.

Capaldi, R. A, (1990a) Structure and Assembly of Cytochrome c Oxidase.
Arch. Biochem. Biophys. 280, 252-262.

Capaldi, R. A, (1990b) Structure and Function of Cytochrome c Oxidase.
Annu. Rev. Biochem. 59, 569-596.

Capaldi, R. A, Darley-Usmar, V., Fuller, S. & Millett, F., (1982) Structural and
Functional Features of the Interaction of Cytochrome c With Complex
III and Cytochrome c On'dase. FEBS Lett. 138, 1-7.

Capitanio, N., Capitanio, G., De Nitto, E., Villani, G. & Papa, S., (1991) H+/e'
Stoichiometry of Mitochondrial Cytochrome Complexes Reconstituted in
Liposomes. FEBS Lett. 288, 179-182.

Casey, P. J ., (1994) Lipid Modifications of G Proteins. Current Opin. Cell Biol.
6, 2 19-225.

170

Casey, R. P., (1986) Measurement of the H+ Pumping Activity of
Reconstituted Cytochrome Oxidase. Methods in Enzymology 126, 13-21.

 

Casey, R. P., Chappell, J. B. & Azzi, A, (197 9) Limited-Turnover Studies on
Proton Translocation in Reconstituted Cytochrome c Oxidase-
Containing Vesicles. Biochem. J. 182, 149-156.

Casey, R. P., O'Shea, P. S., Chappell, J. B. & Azzi, A, (1983) A Quantitative

Characterization of H+ Translocation by Cytochrome Oxidase Vesicles.
Biochim. Biophys. Acta 765, 30-37.

Casey, R. P., O'Shea, P. S., Chappell, J. B. & Azzi, A, ( 1984) A Quantitative

Characterization of H+ Translocation by Cytochrome c Oxidase
Vesicles. Biochim. Biophys. Acta 765, 30-37.

Casey, R. P., Thelen, M. & Azzi, A, (1980) Dicyclohexylcarbodiimide Binds
Specifically and Covalently to Cytochrome c Oxidase While Inhibiting

Its H+-Translocating Activity. J. Biol. Chem. 255, 3994-4000.

Castresana, J ., Liibben, M., Saraste, M. & Higgins, D. G., (1994) Evolution of
Cytochrome Oxidase, An Enzyme Older Than Atmospheric Oxygen.
EMBO J. 13, 2516-2525.

Chan, S. I. & Li, P. M., (1990) Cytochrome c Oxidase: Understanding Natures's
Design of a Proton Pump. Biochemistry 29, 1-12.

Chance, B., Leigh, J. S. J. & Waring, A, in Structure and Function of Energy-
Transducing Membranes van Dam, K. & van Gelder, B. F., Eds.
(Elsevier/North-Holland, Amsterdam, 197 7) pp. 1- 10.

Chance, B., Sies, H. & Boveris, A, (197 9) Hydroperoxide Metabolism in
Mammalian Organs. Physiol. Rev. 59, 527-605.

Chang, C.-H., El-Kabbani, O., Tiede, D., Norris, J. & Schifl‘er, M., (1991)
Structure of the Membrane-bound Protein Photosynthetic Reaction
Center from Rhodobacter sphaeroides. Biochemistry 30, 5352-5360.

Chepuri, V. & Gennis, R. B., (1990) The Use of Gene Fusions to Determine the
Topology of All of the Subunits of the Cytochrome 0 Terminal Oxidase
Complex of Escherichia coli. J. Biol. Chem. 265, 12978-12986.

Chirino, A J ., Lous, E. J ., Huber, M., Allen, J. P., Schenck, C. C., Paddock, M.
L., Feher, G. & Rees, D. C., (1994) Crystallographic Analyses of Site-
Directed Mutants of the Photosynthetic Reaction Center from
Rhodobacter sphaeroides. Biochemistry 33, 4584-4593.

Chothia, C., (1975) Structural Invarients in Protein Folding. Nature 254, 304-
308.

171

Consler, T. G., Tsolas, O. & Kaback, H. R., (1991) Role of Proline Residues in
the Structure and Function of a Membrane Transport Protein.
Biochemistry 30, 1291-1298.

de Gier, J .-W., Lfibben, M., Reijnders, W. N. M., Tipker, C. A, Slotboom, D.-J.,
van Spanning, R. J. M., Stouthamer, A H. & van der Oost, J ., (1994)
The Terminal Oxidases of Paracoccus denitrificans. Mol. Microbiol. 13,
183- 196.

Deisenhofer, J ., Epp, 0., Miki, K., Huber, R. & Michel, H., (1984) X-Ray
Structure Analysis of a Membrane Protein Complex: Electron Density
Map at 3 A Resolution and a Model of the Chromophores of the
Photosynthetic Reaction Center from Rhodopseudomonas viridis. J.
Mol. Biol. 180, 385-398.

Deisenhofer, J ., Epp, 0., Miki, K., Huber, R. & Michel, H., (1985) Structure of
the Protein Subunits in the Photosynthetic Reaction Centre of
Rhodopseudomonas viridis at 3A Resolution. Nature 318, 618-624.

Deisenhofer, J ., Epp, O., Sinning, I. & Michel, H., (1995) Crystallographic
Refinement at 2.3 A Resolution and Refined Model of the Photosynthetic
Reaction Centre from Rhodopseudomonas viridis. J. Mol. Biol. 246, 429-
457 .

Donohue, T. J ., McEwan, A G., Van Doren, S., Crofts, A R. & Kaplan, S.,
(1988) Phenotypic and Genetic Characterization of cytochrome c2
Deficient Mutants of Rhodobacter sphaeroides. Biochemistry 27, 1918-
1925.

Einarsdéttir, O., (1995) Fast Reactions of Cytochrome Oxidase. Biochim.
Biophys. Acta 1229, 129-302.

Ermler, U., Fritzsch, G., Buchanan, S. K & Michel, H., (1994a) Structure of
the Photosynthetic Reaction Centre from Rhodobacter sphaeroides at
2.65 A Resolution: Cofactors and Protein-Cofactor Interactions.
Structure 2, 925-936.

Ermler, U., Michel, H. & Schifi‘er, M., (1994b) Structure and thction of the
Photosynthetic Reaction Center from Rhodobacter sphaeroides. J.
Bioenerg. Biomembr. 26, 5- 15.

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

172

Eytan, G. D., (1982) Use of Liposomes for Reconstitution of Biological
Functions. Biochim. Biophys. Acta 694, 185-202.

Fee, J. A, Kuila, D., Mather, M. W. & Yoshida, T., (1986) Respiratory Proteins
from Extremely Thermophilic, Aerobic Bacteria. Biochim. Biophys. Acta
853, 153-185.

Ferguson, S. J ., (1985) Fully Delocalized Chemiosmotic or Localized Proton
Flow Pathways in Energy Coupling? A Scrutiny of Experimental
Evidence. Biochim. Biophys. Acta 811, 47-95.

Ferguson-Miller, S., Brautigan, D. L. & Margoliash, E., (1978) Definition of
Cytochrome c Binding Domains by Chemical Modification. J. Biol.
Chem. 253, 149-159.

Fetter, J. R., Qian, J ., Shapleigh, J ., Thomas, J. W., Garcia-Horsman, Schmidt,
E., Hosler, J ., Babcock, G. T., Gennis, R. B. & Ferguson-Miller, S., (1995)
Possible Proton Relay Pathways in Cytochrome c Oxidase. Proc. Natl.
Acad. Sci USA 92, 1604-1608.

Fillingame, R. H., in The Bacteria: A Treatise on Structure and Function
Krulwich, T., Eds. (Academic Press, Inc., New York, 1990), vol. XII,.

Fillingame, R. H., (1992) H+ Transport and Coupling by the F0 Sector of the
ATP Synthase: Insights into the Molecular Mechanism of Function. J.
Bioenerg. Biomembr. 24, 485-491. '

Garcia-Horsman, J ., Barquera, B., Rumbley, J ., Ma, J. & Gennis, R. B.,
(1994a) The Superfamily of Heme-Copper Respiratory Oxidases. J.
Bacteriol. 176, 5587-5600.

Garcia-Horsman, J. A, Berry, E., Shapleigh, J. P., Alben, J. O. & Gennis, R. B.,
(1994b) A Novel Cytochrome c Oxidase from Rhodobacter Sphaeroides
that Lacks Cu A. Biochemistry 33, 3113-3119.

Garcia-Horsman, J. A, Puustinen, A, Gennis, R. B. & Wikstrtim, M., (1995)
Proton Transfer in Cytochrome b03 Ubiquinol Oxidase of Escherichia

coli: Second-Site Mutations in Subunit I That Restore Proton Pumping
in the Mutant Asp 135—)Asn. Biochemistry 34, 4428-4433.

Gelles, J ., Blair, D. F. & Chan, S. I., (1987) The Proton-Pumping Site of
Cytochrome c Oxidase: A Model of its Structure and Mechanism.
Biochim. Biophys. Acta 853, 205-236.

Gennis, R. B., (1991) Some Recent Advances Relating to Prokaryotic
Cytochrome c Reductases and Cytochrome c Oxidases. Biochim.
Biophys. Acta 1058, 21-24.

173

Gennis, R. E, Casey, R. P., Azzi, A. & Ludwig, B., (1982) Purification and
Characterization of the Cytochrome c Oxidase from Rhodopseudomonas
sphaeroides. Eur. J. Biochem. 125, 189-195.

Gibson, Q. & Greenwood, C., (1963) Reactions of Cytochrome Oxidase with
Oxygen and Carbon Monoxide. Biochem. J. 86, 541-554.

Girvin, M. E. & Fillingame, R., (1993) Helical Structure and Folding of Subunit
c of FIFO ATP Synthase: 1H NMR Resonance Assignments and NOE

Analysis. Biochemistry 32, 12167-1217 7.

Girvin, M. E. & Fillingame, R. H., (1995) Determination of Local Protein
Structure by Spin Label Difference 2D NMR: The Region Neighboring
Asp61 of Subunit c of the FIFO ATP Synthase. Biochemistry 34, 1635-
1645.

Goldberg, R. A, Einarsdéttir,'O., Dawes, T. D., O'Connor, D. B., Surerus, K. K.,
Fee, J. A. & Kliger, D. S., (1992) Magnetic Circular Dichroism Study of
Cytochrome ba 3 from Thermus thermophilus: Spectral Contributions
from Cytochromes b and a 3 and Nanosecond Spectroscopy of CO

Photodissociation Intermediates. Biochemistry 31, 9376-9387.

Gray, K., Grooms, M., Myllykallio, H., Moomaw, C., Slaughter, C. & Daldal, F.,
(1994) Rhodobacter Capsulatus Contains a Novel cb-Type Cytochrome c
Oxidase without a Cu A Center. Biochemistry 33, 3120-3127.

Gray, M. W., (1989) The Evolutionary Origins of Organelles. Trends Gen. 5,
294-299.

Greenwood, C. & Gibson, Q. H., (1967) The Reaction of Reduced Cytochrome c
Oxidase with Oxygen. J. Biol. Chem. 242, 1782-1787.

Gregory, L. C., Dissertation, Michigan State University (1988).

Gregory, L. C. & Ferguson-Miller, S., inAdvances in Membrane Biochemistry
and Bioenergetics Kim, c. H., Tedeschi, H. & Diwan, J. J., Eds. (Plenum,
New York, 1988).

Hallen, S. & N ilsson, T., (1992) Proton Transfer During the Reaction Between
Fully Reduced Cytochrome Oxidase and Dioxygen: pH and Deuterium
Isotope Efi‘ects. Biochemistry 31, 11853-11859.

Haltia, T., (1992) Reduction of Cu A Induces a Conformational Change in

Cytochrome c Oxidase from Paracoccus denitrificans. Biochim. Biophys.
Acta 1098, 343-350.

174

Haltia, T., Saraste, M. & Wikstrom, M., (1991) Subunit III of Cytochrome c
Oxidase is not Involved in Proton Translocation: A Site-Directed
Mutagenesis Study. EMBO J. 10, 2015-2021.

Henderson, R., Baldwin, J. M., Ceska, T. A, Zemlin, F., Beckmann, E. &
Downing, K. H., (1990) Model for the Structure of Bacteriorhodopsin
Based on High-resolution Electron Cryo-microscopy. J. Mol. Biol. 213,
899-929.

Henderson, R. & Unwin, P. N. T., (197 5) Three-Dimensional Model of Purple
Membrane Obtained by Electron Microscopy. Nature 257, 28-32.

Hermolin, J. & Fillingame, R. H., (1989) H+-ATPase Activity of Escherichia
coli FIFO Is Blocked After Reaction of Dicyclohexylcarbodiimide with a

Single Proteolipid (Subunit c) of the F0 Complex. J. Biol. Chem. 264,
3896-3903.

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

Hill, B. C., (1994) Modeling the Sequence of Electron Transfer Reactions in the
Single Turnover of Reduced, Mammalian Cyctochrome c Oxidase. J.
Biol. Chem. 269, 2219-2226.

Hinkle, P. C., Kim, J. J. & Racker, E., (1972) Ion Transport and Respiratory
Control in Vesicles Formed From Cytochrome Oxidase and
Phospholipids. J. Biol. Chem. 247, 1338-1339.

Hosler, J. P., Espe, M. P., Zhen, Y., Babcock, G. T. & Ferguson-Miller, S., (1995)
Analysis of Site-Directed Mutants Locates a N on-Redox-Active Metal
near the Active Site of Cytochrome c Oxidase of Rhodobacter
sphaeroides. Biochemistry 34, 7586-7 592.

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., (1993) 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. 25, 121-136.

Hosler, J. P., Fetter, J ., Tecklenburg, M. M. J ., Espe, M., Lerma, C. &
Ferguson-Miller, S., (1992) Cytochrome aa 3 of Rhodobacter sphaeroides

as a Model for Mitochondrial Cytochrome c Om'dase. J. Biol. Chem. 267 ,
24264-24272.

Hosler, J. P., Kim, Y., Shapleigh, J ., Gennis, R., Alben, J ., Ferguson-Miller, S. &
Babcock, G., (1994a) Vibrational Characteristics of Mutant and Wild-
Type Carbon-Monoxy Cytochrome c Oxidase: Evidence for a Linear

175

Arrangement of Heme a, a3, and CuB. J. Am. Chem. Soc. 116, 5515-
5516.

Hosler, J. P., Shapleigh, J. P., Kim, Y., Pressler, M., Georgiou, C., Babcock, G.
T., Alben, J. O., Ferguson-Miller, S. & Gennis, R. B., (in press) Polar
Residues in Helix VIII of Subunit I of Cytochrome c Oxidase Influence
the Activity and the Structure of the Active Site.

Hosler, J. P., Shapleigh, J. P., Tecklenburg, M. M. J ., Thomas, J. W., Kim, Y.,
Espe, M., Fetter, J ., Babcock, G. T., Alben, J. O., Gennis, R. B. &
Ferguson-Miller, S., (1994b) A Loop between Transmembrane Helices
D( and X of Subunit I of Cytochrome c Oxidase Caps the Heme a-Heme
a3-CuB Center. Biochemistry 33, 1194- 1201.

Huber, R., (1989) Nobel Lecture: A Structural Basis of Light Energy and
Electron Transfer in Biology. EMBO J. 8, 2125-2147.

Hiither, F.-J. & Kadenbach, B., (1988) Intraliposomal Nucleotides Change the
Kinetics of Reconstituted Cytochrome c Orn'dase fi'om Bovine Heart but
Not fi'om Paracoccus Denitrificans. Biochem. Biophys. Res. Commun.
153, 525-534.

Jezek, P., Orosz, D. E., Modriansky, M. & Garlid, K. D., (1994) Transport of
Anions and Protons by the Mitochondrial Unooupling Protein and Its
Regulation by Nucleotides and Fatty Acids. J. Biol. Chem. 269, 26184-
26190.

Kadenbach, B., Jarausch, J ., Hartrnann, R. & Merle, P., (1983) Separation of
Mammalian Cytochrome c Oxidase into 13 Polypeptides by a Sodium
Dodecyl Sulfate-Gel Electrophoretic Procedure. Anal. Biochem. 129,

5 1 7 -52 1.

Kadenbach, B., Shroh, A, Hiither, F.-J., Reimann, A & Steverding, D., (1991)
Evolutionary Aspects of Cytochrome c Oxidase. J. Bioenerg. Biomembr.
23, 321-334.

Kagawa, Y. & Backer, E., (1971) Partial Resolution of the Enzymes Catalyzing
Oxidative Phosphorylation: XXV Reconstitution of Vesicles Catalyzing

32Pi-Adenosine Triphosphate Exchange. J. Biol. Chem. 246, 5477-5487.

Kasianowicz, J ., Benz, R. & McLaughlin, S., (1987) How Do Protons Cross the
Membrane Solution Interface? Kinetic Studies on Bilayer Membranes
Exposed to the Protonophore S-13 (5-chloro-3-tert-butyl-2'-chloro-4'-
nitrosalicylanilide). J. Membrane Biol. 95, 7 3-89.

Kelly, M., Lappalainen, P., Talbo, G., Haltia, T., van der Oost, J. & Saraste, M.,
(1993) Two Cysteines, Two Histidines, and one Methionine are Ligands
of a Binuclear Purple Copper Center. J. Biol. Chem. 268, 16781-16787.

176

Khorana, H. G., (1988) Bacteriorhodopsin, a Membrane Protein That Uses
Light to Translocate Protons. J. Biol. Chem. 263, 7439-7442.

Kirmaier, C. & Holten, D., in The Photosynthetic Reaction Center Deisenhofer, J.
& Norris, J. R., Eds. (Academic Press, San Diego, 1993), vol. II, pp. 49-
70.

Klingenberg, M., (1993) Dialectics in Carrier Research: The ADP/ATP Carrier
and the Uncoupling Protein. J. Bioenerg. Biomembr. 25, 447-457.

Klinman, J. P. & Brenner, M., in Oxidases and Related Redox Systems King, T.
E., Mason, H. S. & Morrison, M., Eds. (Alan Liss, New York, 1988) pp.
227-246.

Krab, K & Wikstrom, M., (1978) Proton-Translocating Cytochrome c Oxidase
in Artificial Phospholipid Vesicles. Biochim. Biophys. Acta 504, 200-214.

Krebs, M. P. & Khorana, H. G., (1993) Mechanism of Light-Dependent Proton
Translocation by Bacteriorhodopsin. J. Bacteriol. 175, 1555- 1560.

Kroneck, P. M. H., Antholine, W. A, Riester, J. & Zumfi, W. G., (1989) The
Nature of the Cupric Site in Nitrous Oxide Reductase and of Cu A in

Cytochrome c Oxidase. FEBS Lett. 248, 212-213.

Krulwich, T. A, Hicks, D. B., Seto-Young, D. & Gufl'anti, A A, (1988) The
Bioenergetics of Alkalophilic Bacilli. CRC Crit. Rev. Microbiol. 16, 15-36.

Labonia, N., Muller, M. & Azzi, A, (1988) The Effect of N on-esterified Fatty
Acids on the Proton-Pumping Cytochrome c Oxidase Reconstituted into
Liposomes. Biochem. J. 254, 139-145.

Laneuville, O., Breuer, D. K., DeWitt, D. L., Hla, T., Funk, C. D. 8: Smith, W. L.,
(1994) Differential Inhibition of Human Prostaglandin Endoperoxide H
Synthases-l and -2 by Nonsteroidal Anti-Inflamatory Drugs. J.
Pharmacol. Exp. Ther. 271, 927-934.

Lappalainen, P., Aasa, R., Malmstrtim, B. G. & Saraste, M., (1993) Soluble
Cu A-binding Domain from the Paracoccus Cytochrome c Oxidase. J.

Biol. Chem. 268, 26416-26421.

Larsen, R. W., Pam, L.-P., Musser, S. M., Li, Z. & Chan, S. I., (1992) Could CuB

be the Site of Redox Linkage in Cytochrome c Oxidase? Proc. Natl. Acad.
Sci. USA 89, 723-727.

Lauraeus, M., Haltia, T., Saraste, M. & Wikstrom, M., (1991) Bacillus subtilis
Expresses Two Kinds of Haem-A-Containing Terminal Oxidases. Eur. J.
Biochem. 197, 699-705.

177

Lauraeus, M. & Wikstrom, M., (1993) The Terminal Quinol Oxidases of
Bacillus Subtilis Have Different Energy Conservation Properties. J.
Biol. Chem. 268, 11470-11473.

Lehninger, A. L., Reynafarje, B., Davies, P., Alexandre, A, Villalobo, A &
Beavis, A, Mitochondria and Microsomes. Lee, C. P., Schatz, G. &
Dallner, G., Eds., (Addison-Wesley Publishing Co., 1981).

Lemieux, L. J ., Calhoun, M. W., Thomas, J. W., Ingledew, W. J. & Gennis, R. B.,
(1992) Determination of the Ligands of the Low-Spin Heme of the
Cytochrome o Ubiquinol Oxidase Complex Using Site-Directed
Mutagenesis. J. Biol. Chem. 267, 2105-2113.

Levitt, D. G., (1984) Kinetics of Movement in Narrow Pores. Curr. Top. Membr.
Transp. 21, 181-197.

Lill, H., Althofl', G. & Junge, W., (1987) Analysis of Ionic Channels by a Flash
Spectrophotometric Technique Applicable to Thylakoid Membranes:
CFO the Proton Channel of the Chloroplast ATP Synthase, and for

Comparison, Gramicidin. J. Membrane Biol. 98, 69-78.

Ludwig, B., (1992) Terminal Oxidases in Paracoccus denitrificans. Biochim.
Biophys. Acta 1101, 195-197 .

Macino, G. & Morelli, G., (1983) Cytochrome Oxidase Subunit 2 Gene in
Neurospora crassa Mitochondria. J. Biol. Chem. 258, 13230-13235.

Madden, T. D. & Redelmeier, T. E., (1994) Transmembrane Distribution of
Lipophilic Cations in Response to an Electrochemical Potential in
Reconstituted Cytochrome c Oxidase Vesicles and in Vesicles Exhibiting
a Potassium Ion Diffusion Potential. J. Bioenerg. Biomembr. 26, 221-
230.

Malatesta, F., Antonini, G., Sarti, P. & Brunori, M., (1987) Modulation of
Cytochrome Oxidase Activity by Inorganic and Organic Phosphate.
Biochem. J. 248, 161-165.

Mannella, C. A, Marko, M., Penczek, _P., Barnard, D. & Frank, J ., (1994) The
Internal Compartrnentation of Rat-Liver Mitochondria: Tomographic
Study Using the High-Voltage Transmission Electron Microscope.
Microsc. Res. Tech. 27, 278-283.

Marti, T., Otto, H., Mogi, T., Rosselet, S. J ., Heyn, M. P. & Khorana, H. G.,
(1991) Bacteriorhodopsin Mutants Containing Single Substitutions of
Serine or Threonine Residues Are All Active in Proton Translocation. J.
Biol. Chem. 266, 6919-6927. '

Mather, M., Springer, P., Hensel, S., Buse, G. & Fee, J. A, (1993) Cytochrome
Oxidase Genes fiom Thermus Thermophilus: Nucleotide Sequence Of

178

the Fused Gene and Analysis of the Deduced Primary Structures for
Subunits I and III of Cytochrome caa3. J. Biol. Chem. 268, 5395-5408.

Matsushita, K., Ebisuya, H., Ameyama, M. & Adachi, O., (1992) Change of the
Terminal Oxidase from Cytochrome al in Shaking Cultures to
Cytochrome o in Static Cultures of Acetobacter aceti. J. Bact. 174, 122-
129.

Matsushita, K., Shinagawa, E., Adachi, O. & Ameyama, M., (1990)
Cytochrome “1 of Acetobacter Aceti is a Cytochrome ba Functioning as

Ubiquinol Orddase. Proc. Natl. Acad. Sci. USA 87, 9863-9867.

Miller, M. J ., Oldenburg, M. & Fillingame, R. H., (1990) The Essential Carboxyl
Group in Subunit c of the FlFo ATP Synthase Can Be Moved and H+-

translocating Function Retained. Proc. Natl. Acad. Sci. USA 87, 4900-
4904.

Millett, F., de J ong, C., Paulson, L. & Capaldi, R. A, (1983) Identification of
Specific Carboxylate Groups on Cytochrome c Oxidase That are
Involved in Binding Cytochrome c. Biochemistry 22, 546-552.

Milligan, G., Parenti, M. & Magee, A I., (1995) The Dynamic Role of
Palmitoylation in Signal Transduction. TIBS 20, 181- 186.

Minagawa, J ., Mogi, T., Gennis, R. B. & Anraku, Y., (1992) Identification of
Heme Ligands in Subunit I of the Cytochrome bo Complex in
Escherichia coli. J. Biol. Chem. 267, 2096-2104.

Mitchell, D. M. & Gennis, R. B., (in press) Rapid Purification of Wildtype and

Mutant Cytochrome c Oxidase from Rhodobacter sphaeroides by Ni2+-
NTA Affinity Chromatography. FEBS Lett.

Mitchell, P., (1976) Possible Molecular Mechanisms of the Protonmotive
Function of Cytochrome Systems. J. Theor. Biol. 62, 327 -367.

Mitchell, P., (1987) A New Redox Loop Formality Involving Metal-Catalysed
Hydrorn'de-ion Translocation. FEBS. Lett. 222, 235-245.

Mitchell, P., (1988) Possible Protonmotive Osmochemistry in Cytochrome
Oxidase. Ann. NY. Acad. Sci. 550, 185- 198.

Mitchell, R, Mitchell, R., Moody, A. J ., West, I. C., Baum, H. & Wrigglesworth,
J ., (1985) Chemiosmotic Coupling in Cytochrome Oxidase: Possible
Protonmotive 0 Loop and 0 Cycle Mechanisms. FEBS Lett. 188, 1-7.

Moench, S. J ., Moreland, J ., Stewart, D. H. & Dewey, T. G., (1994)
Fluorescence Studies of the Location and Membrane Accessibility of the
Palmitoylation Sites of Rhodopsin. Biochemistry 33, 5791-5796.

179

Mogi, T., Stern, L. H., Marti, T., Chao, B. H. & Khorana, H. G., (1988) Aspartic
Acid Substitutions Afi‘ect Proton Translocation by Bacteriorhodopsin.
Proc. Natl. Acad. Sci. USA 85, 4148-4152.

Morgan, J. E., Verkhovsky, M. I. & Wikstrom, M., (1994) The Histidine Cycle:
A New Model for Proton Translocation in the Respiratory Heme-Copper
Oxidases. J. Bioenerg. Biomembr. 26, 599-608.

Murphy, M. P., (1989) Slip and Leak in Mitochondrial Oxidative
Phosphorylation. Biochim. Biophys. Acta 977, 123-141.

Nagle, J. F. & Morowitz, H. J ., (197 8) Molecular Mechanisms for Proton
Transport in Membranes. Proc. Natl. Acad. Sci. USA 75, 298-302.

Nagle, J. F. & Tristram-Nagle, S., in Information and Energy Transduction in
Biological Membranes(Alan R. Liss, Inc., New York, 1984) pp. 103-111.

Nicholls, D. G. & Ferguson, S. J ., Bioenergetics 2 (Academic Press, 1992).

Nicholls, P. & He, J ., (1993) Direct and Indirect Effects of Valinomycin upon
Cytochrome c orddase. Arch. Biochem. and Biophys. 301, 305-310.

Numata, M., Yamazaki, T., Fukumori, Y. & Yamanaka, T., (1989) Some
Properties of Nitrosomonas europea Cytochrome c Oxidase (aa3-Type)

Which Lacks Cu A' J. Biochem. 105, 245-248.

Ogura, T., Takahashi, S., Shinzawa-Itoh, K, Yoshikawa, S. & Kitagawa, T.,
(1991) Time-Resolved Resonance Raman Investigation of Cytochrome

Oxidase Catalysis: Observation of a New Oxygen-Isotope Sensitive
Raman Band. Bull. Chem. Soc. Jpn. 64, 2901-2907.

Okamura, M. Y. & Feher, G., (1992) Proton Transfer in Reaction Centers from
Photosynthetic Bacteria. Annu. Rev. Biochem. 1992, 861-896.

Olejnik, J ., Brzezinski, B. & Zundel, G., (1992) A Proton Pathway with Large
Proton Polarizability and the Proton Pumping Mechanism in
Bacteriorhodopsin-Fourier Transform Infrared Difference Spectra of
Photoproducts of Bacteriorhodopsin and of its Pentadimethyl Analogue.
J. Mol. Struct. 271, 157-173.

Oliveberg, M., Hallén, S. & Nilsson, T., (1991) Uptake and Release of Protons
During the Reaction Between Cytochrome c Oxidase and Molecular
Oxygen: A Flow-Flash Investigation. Biochemistry 30, 436-440.

Oliveberg, M. & Malmstrom, B. G., (1991) Internal Electron Transfer in
Cytochrome c Oxidase: Evidence for a Rapid Equilibrium between
Cytochrome a and the Bimetallic Site. Biochemistry 30, 7053-7057 .

180

Otto, H., Marti, T., Holz, M., Mogi, T., Lindau, M., Khorana, H. G. & Heyn, M.
P., (1989) Aspartic Acid-96 is the Internal Proton Donor in the
Reprotonation of the Schiff Base of Bacteriorhodopsin. Proc. Natl. Acad.
Sci. USA 86, 9228-9232.

Otto, H., Marti, T., Holz, M., Mogi, T., Lindau, M., Khorana, H. G. & Heyn, M.
P., (1990) Substitution of Amino Acids Asp-85, Asp-212, and Arg-82 in
BacteriorhodOpsin Affects the Proton Release Phase of the Pump and
the pK of the Schiff Base. Proc. Natl. Acad. Sci. USA 87 , 1018-1022.

Paddock, M. L., Feher, G. & Okamura, M. Y., (1990a) pH Dependence of
Charge Recombination in RCs from Rb. sphaeroides in which Glu-L212
is Replaced with Asp. Biophys. J. 57, 569a.

Paddock, M. L., McPherson, P. H., Feher, G. & Okamura, M. Y., (1990b)
Pathway of Proton Transfer in Bacterial Reaction Centers:
Replacement of Serine-L-223 by Alanine Inhibits Electron and Proton
Transfers Associated with Reduction of Quinone to Dihydroquinone.
Proc. Natl. Acad. Sci. USA 87, 6803-6807.

Paddock, M. L., Rongey, S. H., McPherson, P. H., Juth, A, Feher, G. &
Okamura, M. Y., (1994) Pathway of Proton Transfer in Bacterial
Reaction Centers: Role of Aspartate-LZ 13 in Proton Transfers
Associated with Reduction of Quinone to Dihydroquinone. Biochemistry
33, 734-745.

Pan, L. P., Hibdon, S., Liu, R.-Q., Durham, B. & Millet, F., (1993) Intracomplex
Electron Transfer Between Ruthenium-Cytochrome c Derivatives and
Cytochrome c Oxidase. Biochemistry 32, 8492-8498.

Papa, S., Capitanio, N., Capitanio, G., De Nitto, E. & Minuto, M., (1991) The
Cytochrome Chain of Mitochondria Exhibits Variable H+/e'
Stoichiometry. FEBS Lett. 288, 183-186.

Papa, S., Lorusso, M. &-Capitanio, N., (1994) Mechanistic and
Phenomenological Features of Proton Pumps in the Respiratory Chain
of Mitochondria. J. Bioenerg. Biomembr. 26, 609-618.

Pedersen, P. L. & Amzel, L. M., (1993) ATP Synthases: Structure, Reaction
Center, Mechanism and Regulation of One of N ature's Most Unique
Machines. J. Biol. Chem. 268, 9937-9940.

Peiffer, W. E., Ingle, R. T. & Ferguson-Miller, S., (1990) Structurally Unique
Plant Cytochrome c Oxidase Isolated from Wheat Germ, a Rich Source
of Plant Mitochondrial Enzymes. Biochem. 29, 8698-9701.

Piela, L., Némethy, G. & Scheraga, H. A, (1987) Proline-Induced Constraints
in tat-Helices. Biopolymers 26, 1587-1600.

181

Poyton, R. 0., Trueblood, C. E., Wright, R. M. & Farrell, L. E., (1988)
Expression and Function of Cytochrome c Oxidase Subunit Isologues.
Ann. NY. Acad. Sci. 550, 289-307.

Preisig, O., Anthamatten, D. & Hennecke, H., (1993) Genes for a
Microaerobically Induced Orddase Complex in Bradyrhizobium
Japonicum are Essential for a Nitrogen-Fixing Endosymbiosis. Proc.
Natl. Acad. Sci. USA 90, 3309-3313.

Prochaska, L. J ., Bisson, R., Capaldi, R. A, Steffens, G. C. M. & Buse, G.,
(1981) Inhibition of Cytochrome c Oxidase Function by
Dicyclohexylcarbodiimide. Biochim. Biophys. Acta 637, 360-373.

Prochaska, L. J. & Fink, P. S., (1987) On the Role of Subunit III in Proton
Translocation in Cytochrome c Oxidase. J. Bioenerg. Biomem. 19, 143-
164.

Proshlyakov, D. A, Ogura, T., Shinzawa-Itoh, K., Yoshikawa, S., Appelman, E.
H. & Kitagawa, T., (1994) Selective Resonance Raman Observation of
the "607 nm" Form Generated in the Reaction of Oxidized Cytochrome c
Oxidase with Hydrogen Peroxide. J. Biol. Chem. 269, 29385-29388.

Proteau, G., Wrigglesworth, J. M. & Nicholls, P., (1983) Protonmotive
Functions of Cytochrome c Oxidase in Reconstituted Vesicles. Biochem.
J. 210, 199-205.

Puustinen, A, Morgan, J. E., Verkhovsky, M., Thomas, J. W., Gennis, R. B. ]. &
Wikstrom, M., (1992) The Low Spin Heme Site of Cytochrome o from E.
coli is promiscuous with Respect to Heme Type. Biochemistry 31,
10363-10369.

Raitio, M. & Wikstrtim, M., (1994) An Alternative Cytochrome Oxidase of
Paracoccus denitrificans Functions as a Proton Pump. Biochim. Biophys.
Acta 1186, 100-106.

Rial, E., Poustie, A & Nicholls, D. G., (1983) Brown-Adipose-Tissue
Mitochondria: the Regulation of the 32 000-Mr Uncoupling Protein by

Fatty Acids and Purine Nucleotides. Eur. J. Biochem. 137, 197 -203.

Rice, C. W. & Hempfling, W. P., (197 8) Oxygen-Limited Continuous Culture
and Respiratory Energy Conservation in Escherichia coli. J. Bacteriol.
134, 115-124.

Rothschild, K. J ., (1992) FTIR Difference Spectroscopy of Bacteriorhodopsin:
Toward a Molecular Model. J. Bioenerg. Biomembr. 24, 147-167.

Rottenberg, H., (1985) Proton-Coupled Energy Conservation: Chemiosmotic
and Intramembrane Coupling. Mod. Cell Biol. 4, 47-83.

182

Rousseau, D. L., Ching, Y.-c. & Wang, J ., (1993) Proton Translocation in
Cytochrome c Oxidase: Redox Linkage through Proximal Ligand
Exchange on Cytochrome a3. J. Bioenerg. Biomembr. 25, 165-17 6.

Sambrook, J ., Fritsch, E. F. & T., M., Molecular Cloning: A Laboratory Manual
(Cold Spring Harbor Laboratory, Cold Spring Harbor, 1989).

Saraste, M., (1990) Structural Features of Cytochrome Oxidase. Q. Rev.
Biophys. 23, 331-366.

Saraste, M. & Castresana, J ., (1994) Cytochrome Oxidase Evolved by
Tinkering with Denitrification Enzymes. FEBS Lett. 341, 1-4.

Schatz, G. & Mason, T. L., (1974) The Biosynthesis of Mitochondrial Proteins.
Annu. Rev. Biochem. 43, 51-87.

Senior, A. E., (1988) ATP Synthesis by Oxidative Phosphorylation. Physiol.
Rev. 68, 177-231.

Shapleigh, J ., Hosler, J. P., Tecklenburg, M. J ., Ferguson-Miller, S., Babcock, G.
T. & Gennis, R. B., (1992) Identification of the Heme Axial Ligands for
the Cytochrome a Component of Cytochrome c Oxidase. Proc. Natl.
Acad. Sci. USA 89, 4786-4790.

Shapleigh, J. P. & Gennis, R. B., (1992) Cloning, Sequencing, and Deletion fi'om
the Chromosome of the Gene Encoding Subunit I of the aa3-type

cytochrome c oxidase of Rhodobacter sphaeroides. Mol. Microbiol. 6, 635-
642.

Sharpe, M. A, Cooper, C. E. & Wrigglesworth, J. M., (1994) Transport of K+
and Other Cations across Phospholipid Membranes by Nonesterified
Fatty Acids. J. Membrane Biol. , .

Singh, A. P. & Nicholls, P., (1985) Cyanine and Safi'anine Dyes as Membrane
Potential Probes in Cytochrome c Oxidase Reconstituted
Proteoliposomes. J. Biochem. Biophys. Meth. 11, 95-108.

Skulachev, V. P., (1991) Fatty Acid Circuit as a Physiological Mechanism of
Uncoupling of Oxidative Phosphorylation. FEBS Lett. 294, 158- 162.

Solioz, M., Carafoli, E. & Ludwig, B., (1982) The Cytochrome c Oxidase of
Paracoccus denitriflcans Pumps Protons in a Reconstituted System. J.
Biol. Chem. 257, 1579-1582.

Sone, N., Ogura, T., Noguchi, S. & Kitagawa, T., (1994) Proton Pumping
Activity and Visible Absorption and Resonance Raman Spectra of a
cao-type Cytochrome c Oxidase Isolated from the Thermophilic
Bacterium Bacillus PS3. Biochemistry 33, 849-855.

183

Sone, N., Yokoi, F., Fu, T., Ohta, S., Metso, T., Raitio, M. & Saraste, M., (1988)
Nucleotide Sequence of the Gene Coding for Cytochrome Oxidase
Subunit I from the Thermophilic Bacterium PSB. J. Biochem. 103, 606-
610.

Sone, N ., Yoshida, M., Hirata, H. & Kagawa, Y., (1977) Reconstitution of
Vesicles Capable of Energy Transformation from Phospholipids and
Adenosine Triphosphatase of a Thermophilic Bacterium. J. Biochem. 81,
5 19-528.

Stefl'ens, G. C. M., Biewald, R. & Buse, G., (1987) Cytochrome c Oxidase is a
Three-Copper, Two-heme-A protein. Eur. J. Biochem. 164, 295-300.

Stefi'ens, G. J. & Buse, G., in Cytochrome Oxidase King, T. E., Orii, Y., Chance,
B. & Okunuki, K., Eds. (Elsevier, 1979) pp. 79-90.

Stern, L. J. & Khorana, H. G., (1989) Structure-Function Studies on
Bacteriorhodopsin. X. Individual Substitutions of Arginine Residues by
Glutamine Affect Chromophore Formation, Photocycle, and Proton
Translocation. J. Biol. Chem. 264, 14202-14208.

Stevens, T. H., Martin, C. T., Wang, H., Brudvig, G. W., Scholes, C. P. & Chan,
S. I., (1982) The Nature of Cu A in Cytochrome c Oxidase. J. Biol. Chem.

257 , 12106-12113.

Suarez, M. D., Revzin, A, Narlock, R., Kempner, E. S., Thompson, D. A. &
Ferguson-Miller, S., (1984) The Functional and Physical Form of
Mammalian Cytochrome c Oxidase Determined by Gel Filtration,
Radiation Inactivation, and Sedimentation Equilibrium Analysis. J. Biol.
Chem. 259, 13791-13799.

Taanman, J .-W. & Capaldi, R. A, (1992) Purification of Yeast Cytochrome c
Oxidase with a Subunit Composition Resembling the Mammalian
Enzyme. J. Biol. Chem. 267, 22481-22485.

Taha, T. S. M. & Ferguson-Miller, S., (1992) Interaction of Cytochrome c with
Cytochrome c Oxidase Studied by Monoclonal Antibodies and a Protein
Modifying Reagent. 31, 9090-9097. '

Takahashi, E. & Wraight, C. A, (1991) Small Weak Acids Stimulate Proton
Transfer Events in Site-Directed Mutants of the Two Ionizable

Residues, Gaul-212 and AspL213, in the QB-binding Site othodobacter
Sphaeroides Reaction Center. FEBS 283, 140-144.

Thiel, C. & Kadenbach, B., (1989) Influence of Non-esterified Fatty Acids on
Respiratory Control of Reconstituted Cytochrome c Orddase. FEBS Lett.
251, 270-274.

184

Thomas, J. W., Puustinen, A, Alben, J. O., Gennis, R. B. & Wikstrom, M.,
(1993) Substitution of Asparagine for Aspartate- 135 in Subunit I of the
Cytochrome bo Ubiquinol Oxidase of Escherichia coli Eliminates Proton-
Pumping Activity. Biochemistry 32, 10923- 10928.

Thompson, D. A. & Gregory, L., (1985) Cytochrome c Oxidase Depleted of
Subunit III: Proton-Pumping, Respiratory Control, and pH Dependence
of the Midpoint Potential of Cytochrome a. J. I norg. Biochem. 23, 357-
364.

Tittor, J ., Soell, C., Oesterhelt, Butt, H.-J. & Bamberg, E., (1989) A Defective

Proton Pump, Point-Mutated Bacteriorhodopsin Asp 96 —> Asn is Fully
Reactivated by Azide. EMBO. J. 8, 3477-3482.

Trawick, J. D., Kraut, N., Simon, F. R. & Poyton, R. O., (1992) Regulation of
Yeast COX6 by the General Transcription Factor ABFl and Separate
HAP2- and Heme-Responsive Elements. Molecular and Cellular Biology
12, 2302-2314.

Trumpower, B. L. & Gennis, R. B., (1994) Energy Transduction by Cytochrome
Complexes in Mitochondrial and Bacterial Respiration: The Enzymology
of Coupling Electron Transfer Reactions to Transmembrane Proton
Translocation. Annu. Rev. Biochem. 63, 675-716.

Valpuesta, J. M., Henderson, R. & Frey, T. G., (1990) Electron Cryo-
microscopic Analysis of Crystalline Cytochrome Oxidase. J. Mol. Biol.
214, 237-251.

van der Oost, J ., de Boer, A P. N., de Gier, J .-W. L., Zumft, W. G., Stouthamer,
A H. & van Spanning, R. J. M., (1994) The Heme-Copper Oxidase
Family Consists of Three Distinct Types of Terminal Oxidases and is
Related to Nitric Ordde Reductase. FEMS Microbial. Lett. 121, 1- 10.

van der Oost, J ., Haltia, T., Raitio, M. & Saraste, M., (1991) Genes Coding for
Cytochrome c Oxidase in Paracoccus denitrificans. J. Bioenerg.
Biomembr. 23, 257-267.

van der Oost, J ., Lappalainen, P., Musacchio, A, Warne, A, Lemieux, L. J .,
Rumbley, J ., Gennis, R. B., Aasa, R., Pascher, T., Malmstrom, B. G. &
Saraste, M., (1992) Restoration of a Lost Metal-Binding Site:
Construction of Two Different Copper Sites Into a Subunit of the E. coli
Quinol Oxidase Complex. EMBO J. 11, 3209-3217.

Vandeyar, M. A, Weiner, M. P., Hutton, C. J. & Batt, C. A, (1988) A Simple
and Rapid Method for the Selection of Oligodeoxynucleotide-Directed
Mutants. Gene 65, 129-133.

185

Varotsis, C., Zhang, Y., Appelman, E. H. & Babcock, G. T., (1993) Resolution of
the Reaction Sequence During the Reduction of 02 by Cytochrome

Oxidase. Proc. Natl. Acad. Sci. USA 90, 237-241.

Verkhovskaya, M., Verkhovsky, M. & Wikstrom, M., (1992) pH Dependence of
Proton Translocation by Escherichia coli. J. Biol. Chem. 267 , 14559-
14562.

Weidner, U., Geier, S., Ptock, A, Friedrich, T., Leif, H. & Weiss, H., (1993) The
Gene Locus of the Proton-Translocating NADH: Ubiquinone
Oxidoreductase in Escherichia coli. Organization of the 14 Genes and
Relationship Between the Derived Proteins and Subunits of
Mitochondrial Complex I. J. Mol. Biol. 233, 109-122.

Westerhofl', H. V., Melandri, B. A, Venturoli, G., Azzone, G. F. & Kell, D. B.,
(1984) A Minimal Hypothesis for Membrane-Linked Free Energy
Transductiom the Role of Independent, Small Coupling Units. Biochim.
Biophys. Acta. 768, 257-292.

Wikstrom, M., (197 7) Proton Pump Coupled to Cytochrome c Oxidase in
Mitochondria. Nature 266, 27 1-27 3.

Wikstrtim, M., (1984) Pumping of Protons from the Mitochondrial Matrix by
Cytochrome Oxidase. Nature 308, 558-560.

Wikstrcm, M., (1989) Identification ‘of the Electron Transfers in Cytochrome
Oxidase that are Coupled to Proton-Pumping. Nature 338, 776-778.

Wikstrom, M., Bogachev, A, Finel, M., Morgan, J. E., Puustinen, A, Raitio, M.,
Verkhovskaya, M. L. & Verkhovsky, M. I., (1994) Mechanism of Proton
Translocation by the Respiratory Oxidases. The Histidine Cycle.
Biochim. Biophys. Acta 1187, 106-111.

Wikstrom, M. & Casey, R., (1985a) The Oxidation of Exogenous Cytochrome c
by Mitochondria. FEBS 183, 293-298.

Wikstrtim, M. & Casey, R., (1985b) The Oxidation of Exogenous Cytochrome c
by Mitochondria: Resolution of a Long-Standing Controversy. FEBS
Lett. 183, 293-298.

Wikstrom, M. & Morgan, J. E., (1992) The Dioxygen Cycle: Spectral, Kinetic
and Thermodynamic Characteristics of Ferry] and Peroxy
Intermediates Observed by Reversal of the Cytochrome Oxidase
Reaction. J. Biol. Chem. 267, 10266-10273.

Williams, J. C., Paddock, M. L., Feher, G. & Allen, J. P., (1991) Efl‘ects of Iron
Ligand Substitutions in Reaction Centers from Rhodobacter sphaeroides.
Biophys. J. 59, 142a.

186

Williams, R. J. P., (1988) Proton Circuits in Biological Energy Interconversions.
Annu. Rev. Biophys. Biophys. Chem. 17, 71-97.

Wilson, K S. & Prochaska, L. J ., (1990) Phospholipid Vesicles Containing
Bovine Heart Mitochondrial Cytochrome c Oxidase and Subunit III-
Deficient Enzyme: Analysis of Respiratory Control and Proton
Translocating Activities. Arch. Biochem. Biophys. 282, 413-420.

Woese, C. R., (1987) Bacterial Evolution. Microbial. Rev. 51, 221-271.

Woodruff, W. H., (1993) Coordination Dynamics of Heme-Copper Oxidases.
The Ligand Shuttle and the Control and Coupling of Electron Transfer
and Proton Translocation. , .

Woodruff, W. H., Einarsdottir, O., Dyer, R. B., Bagley, K. A, Palmer, G.,
Atherton, S. J ., Goldbeck, R. A, Dawes, T. D. & Kliger, D. S., (1991)
Nature and Functional Implications of the Cytochrome a3 Transients

after Photodissociation of CO-Cytochrome Oxidase. Proc. Natl. Acad.
Sci. USA 88, 2588-2592.

Woolley, G. A & Wallace, B. A, (1992) Model Ion Channels: Gramicidin and
Alamethicin. J. Membrane Biol. 129, 109-136.

Wrigglesworth, J. M. & Nicholls, P., (1979) Turnover and Vectorial Properties
of Cytochrome c Oxidase in Reconstituted Vesicles. Biochim. Biophys.
Acta 547, 36-46.

Yano, T., Sled, V. D., Ohnishi, T. & Yagi, T., (1994) Expression of the 25-
Kilodalton Iron-Sulfur Subunit of the Energy-Transducing NADH-
Ubiquinone Oxidoreductase of Paracoccus denitrificans. Biochemistry 33,
494-499.

Zhang, Y. N., El-Sayed, M. A, Stern, L. J ., Marti, T., Magi, T. & Khorana, H. G.,
(1993) Effects of Mutagenetic Substitution of Prolines on the Rate of
Deprotonation and Reprotonation of the Schiff Base During the
Photocycle of Bacteriorhodopsin. Phytochem. Phytobial. 57, 1027-1031.

Ziaie, Z. & Suyama, Y., (1987) The Cytochrome Oxidase Subunit I Gene of
Tetrahymena: a 57 Amino Acid NHZ-Terminal Extension and a 108

Amino Acid Insert. Curr. Genet. 12, 357-368.

Zinth, W. & Kaiser, W., in The Photosynthetic Reaction Center Deisenhofer, J. &
Norris, J. R., Eds. (Academic Press, San Diego, 1993), vol. II, pp. 71-88.

Zundel, G., (1986) Proton Polarizability of Hydrogen Bonds: Infrared Methods,
Relevance to Electrochemical and Biological Systems. Methods
Enzymol. 127, 439-455.

187

Zundel, G., (1988) Proton Transfer in and Proton Polarizability of Hydrogen
Bonds: IR and Theoretical Studies Regarding Mechanisms in Biological
Systems. J. Mol. Struct. 177, 43-68.

Zundel, G., (1994) Hydrogen-bonded Chains with Large Proton Polarizability as
Charge Conductors in Proteins Bacteriorhodopsin and the F0 Subunit of

E. coli. J. Mol. Struct. 322, 33-42.

Zundel, G. & Brzezinsln', B., in Proton Transfer in Hydrogen-Bonded Systems
Bountis, T., Eds. (Plenum Press, New York, 1992) pp. 153-166.

MICHIan STnTE UNIV. LIBRRRIES
l“WWlllmlW“||||WWWIIIIIIWIHWHI
31293014100436