remove this checkout from
your record. FINES wil]
——* be charged if book is

returned after the date

MSU BELURNING MATERIALS;
P1ace in book drop to
LIBRARIES
”—

 

stamped be1ow.

___.—

 

 

 

 

 

 

 

 

 

 

 

 

 

CLARIFICATION OF THE ROLES OF SUBUNIT III AND
PHOSPHOLIPID IN CYTOCHROME c OXIDASE ACTIVITY

By

Linda C. Gregory

A DISSERTATION

Submitted to
Michigan State University
in partial ful llment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Biochemistry
1988

ABSTRACT

CLARIFICATION OF THE ROLES OF SUBUNIT III AND
PHOSPHOLIPID IN CYTOCHROME c OXIDASE ACTIVITY

By

Linda C. Gregory

Mammalian cytochrome c oxidase, the terminal component of the
respiratory chain, is a 13 subunit hemeprotein located in the inner
mitochondrial membrane. The roles of the subunits and of phospholipids
associated with the purified enzyme are not yet defined. When isolated from
rat liver using cytochrome c affinity chromatography in lauryl maltoside,
the enzyme is highly delipidated and depleted of subunits III, Vb, VIa, VIb
and VIIa. Although the enzyme has been reported to have a specific
requirement for 2-3 molecules of diphosphatidylglycerol (DPG) for optimal
activity, oxidase containing sub-stoichiometric amounts of DPG (0.3 mol/mol
enzyme), prepared by a variation of this method, retained high activity and
normal biphasic kinetics for the reaction with cytochrome c. Possible roles
for subunit III in (1) proton translocation, (2) regulation of activity by pH,
(3) regulation of activity by electrochemical gradients, and (4) oxygen
binding, were investigated by examining the effect of subunit III removal.
(1) Subunit III-depleted oxidase retains 50% of the proton translocating
activity of the complete enzyme, indicating that the peptide is not essential
for, but could increase the efficiency of this process. (2) The electron transfer
activity of cytochrome oxidase is highly pH dependent. The pKa values
describing this dependence and the extent of stimulation at low pH are

similar with or without subunit III demonstrating that it is not the location
of the protonatable groups that mediate the regulation of activity. (3) In
mitochondria or in reconstituted phospholipid vesicles, cytochrome oxidase
is inhibited by the transmembrane electrochemical gradient generated by
enzyme activity. Response to the pH or electrical components of the gradient
was found to be independent of the presence or absence of subunit III.
Further, studies on the steady-state reduction of heme a in response to
independent variation of the pH and electrical gradients indicate a dual
mechanism of respiratory control in which the electrical gradient regulates
electron transfer from cytochrome c to heme a, while the pH gradient
controls electron transfer from the heme a to the heme a3 redox center. (4)
Binding of CO, 02 and HCN at the oxygen binding site was not altered upon

removal of subunit III.

Parts of Chapters 3 and 4 are reprinted with permission from Biochemistry
27, 6307-6314 (1988);© 1988, the American Chemical Society. Part of
Chapter 4 is reprinted with permission from Biochemistry, publication in

press.

To my family

ACKNOWLEDGMENTS

It is with pleasure that I take this opportunity to thank Dr. Shelagh
Ferguson-Miller for providing me with the intellectual, moral, and financial
support I needed to complete my graduate studies. As a scientist and as a
person, Shelagh served as a valuable role model. I only hope that I have
learned one quarter of what she offered me.

I must also mention my graditude to an earlier role model - Ms. Pat
Lucas who provided me with such a wonderful introduction to chemistry
and biology.

Thanks are also due to Dr. Peter Nicholls (Brock University, Ontario,
Canada) and Dr. Britton Chance (University of Pennsylvania) who shared
their expertise and their laboratories so that I could study membrane
potential generation and oxygen and carbon monoxide binding by
cytochrome oxidase.

To labmates past (Bob, Bridgette, Krishna, Maria, Debra and Jerome)
and present (Taha, Jon, Wendy, Jian Li and Eileen), thanks for the
discussions, gripe sessions and afternoon parties. Thanks also to Millie,
Tracy and Joan - good friends and fellow "sufferers."

I am especially grateful to Mark and to my parents for sharing their
love and support.

TABLE OF CONTENTS

Page
LIST OF TABLES ................................ x
LIST OF FIGURES ............................... xi
LIST OF ABBREVIATIONS ......................... xiii
INTRODUCTION ................................ 1
The Role of Cytochrome c Oxidase in Energy
Conservation in Mitochondria ...................... 2
Electron Transfer Activity ......................... 3
Proton Translocating Activity ...................... 4
Structural Basis of Cytochrome c Oxidase Activity ......... 4
Polypeptide Composition .......................... 5
Three-Dimensional Structure ...................... 8
Interactions with Membrane Lipids .................. 9
Prosthetic Groups .............................. 10
Functions of Protein Subunits ...................... 12
Subunit III .................................... 14
Effects of Reagents that Bind Subunit III .............. 16
Subunit III-Depleted Cytochrome c Oxidase ............. 18

CHAPTER 1. PURIFICATION OF LIPID- AND
SUBUNIT III- DEPLETED CYTOCHROME c OXIDASE:
STRUCTURAL AND FUNCTIONAL CONSEQUENCES . . . . 23

Introduction ................................... 23
Experimental Procedures .......................... 26
Materials .................................... 26
C chrome Oxidase Preparations ................... 26
S S-Polyacrylamide Gel Electrophoresis ............... 28
[14C]-DCCD Labelling of Cytochrome Oxidase ........... 28
Western Blot Analysis ........................... 29
Sfie ectral Assays ................................ 29
Posphorus Assays .............................. 30
Electron Transfer Activity ......................... 30
Results ....................................... 31
Subunit Composition of "Complete" Cytochrome c
Oxidase .................................... 31
Subunit Composition of Subunit III-Depleted
Cytochrome c Oxidase ........................... 31
Phospholipid Content ............................ 41
Effect of Phospholipid Content on Electron
Transfer Activity .............................. 41
Discussion ..................................... 48

Page

CHAPTER 2. EFFECT OF SUBUNIT III REMOVAL
ON THE PROTON TRANSLOCATING ACTIVITY OF

CYTOCHROME c OXIDASE ........................ 51
Introduction ................................... 51
Experimental Procedures .......................... 55

Materials .................................... 55
Cytochrome Oxidase Preparations ................... 55
Reconstitution ................................. 55
Respiratory Control Assays ........................ 56
Proton Translocation Assays ....................... 56
Results ....................................... 58
Respiratory Control ............................. 58
Proton Translocating Activity ...................... 61
Discussion ..................................... 70

CHAPTER 3. THE EFFECT OF SUBUNIT III REMOVAL
ON CONTROL OF CYTOCHROME c OXIDASE

ACTIVITY BY pH ............................... 73
Introduction ................................... 73
Experimental Procedures .......................... 75
Materials .................................... 75
Cytochrome Oxidase Preparations ................... 75
SDS-Polyacrylamide Gel Electrophoresis ............... 75
Reconstitution ................................. 75
pH Dependence of Cytochrome Oxidase Activity .......... 76
Results ....................................... 77

pH Dependence of Kinetic Parameters of

Reconstituted Cytochrome Oxidase With and

Without Subunit III ............................ 77
Effect of pH on the Spectral Properties of

Cytochrome Oxidase With and Without Subunit

III ........................................ 86
Comparison of Beef Heart and Rat Liver
C chrome Oxidases Depleted of Subunit III ........... 87

E ect of DCCD on the Kinetic Parameters

of Cytochrome Oxidase With and Without
Subunit III .................................. 94
Discussion ..................................... 96

CHAPTER 4. CONTROL OF RESPIRATION IN
CYTOCHROME c OXIDASE VESICLES: INDEPENDENT
REGULATION BY H AND ELECTRICAL GRADIENTS

AND EFFECT OF UBUNIT III DEPLETION ........... 103
Introduction ................................... 103
Experimental Procedures .......................... 107

Materials .................................... 107
Cytochrome Oiddase Preparations ................... 107
Reconstitution ................................. 108
Assays of Electron Transfer Activity .................. 108

Page

CHAPTER 4. (cont)
Steady-State Reduction Studies ..................... 108
Results ....................................... 110
Independent Elimination of Aw and A H ............... 110

Changes in Steady-State Reduction 0 Heme a

and Cytochrome c in Response to Alterations

in the Membrane Potential ....................... 113
Effect of Aw and A H on Steady-State Kinetics

of the Reaction 0 Cytochrome c with

C chrome Oxidase ............................ 122
E ect of Subunit III Content on the Control

of Steady-State Reduction Levels of

Heme a ..................................... 125
Correlation of Heme (1 Reduction Levels and
the Respiratory Control Ratio ..................... 128
Discussion ..................................... 134
A Dual Mechanism for Respirato Control ............. 134
Effect of Subunit III Removal on e Control
of Steady-State Reduction of Heme a ................. 142

CHAPTER 5. THE EFFECT OF SUBUNIT III REMOVAL
ON THE BINDING OF CARBON MONOXIDE, OXYGEN

AND CYANIDE TO CYTOCHROME c OXIDASE .......... 144
Introduction ................................... 144
Ex erimental Procedures .......................... 147
ytochrome Oxidase Preparations ................... 147
Carbon Monoxide-Difference Spectra ................. 147
Carbon Monoxide-Binding Assays ................... 148
Oxygen-Binding Assays .......................... 148
Cyanide-Binding Assays .......................... 149
Results ....................................... 150
Carbon Monoxide-Difference Spectra ................. 150
Rates of Carbon Monoxide Binding ................... 153
Rates of Oxy en Binding .......................... 153
Cyanide Bin ' g ............................... 155
Discussion ..................................... 159
SUMMARY ..................................... 161
APPENDIX ..................................... 163
BIBLIOGRAPHY ................................. 165

ix

LIST OF TABLES
T 1e Page

1 Properties of Subunit III-Depleted Cytochrome
c Oxidases Prepared by Different Methods ...... 19

2 Cytochrome c Oxidase Preparations: Subunit III
Content, Electron Transfer Activity, and
Phospholipid Content ........................ 42

3 Comgarison of Steady-State Kinetic Parameters

 

Oi tained bKIS pectrophotometric and

Polarograp c Assays ........................ 78
4 Effect of Varying Subunit III Content on the

Kinetic Parameters of Cytochrome Oxidase ...... 93

5 Effect of DCCD on Kinetic Parameters of
Reconstituted Cytochrome Oiddase With and
Without Subunit III ........................ 95

6 Effect of Subunit III Depletion on the Rates
of Combination of Cytochrome Oxidase with
Carbon Monoxide and Oxygen ................ 154

LIST OF FIGURES

Figu_re Page

1 Subunit composition of mammalian cytochrome

c oxidase .................................. 7
2 Subunit composition of cytochrome c oxidase

from beef heart and rat liver .................... 33
3 [14C]-DCCD binding to purified cytochrome c

oxidase ................................... 36
4 Subunit composition of subunit III-depleted

cytochrome c oxidase isolated from rat liver .......... 38
5 SDS-PAGE and Western blot analysis of

cytochrome c oxidase isolated from rat liver .......... 40
6 Eadie-Hofstee plots of the kinetics of

cytochrome c oxidase with cytochrome c ............ 44

7 Eadie-Hofstee plots of the kinetics of
delipidated cytochrome oxidase with
cytochrome c ............................... 47

8 Demonstration of respiratory control
in cytochrome oxidase vesicles ................... 60

9 Cytochrome c-induced proton translocation
by cytochrome c oxidase vesicles .................. 65

10 Dependence of H+/e- ratios on (A) charge
compensation and (B) turnovers per assay ........... 68

11 Stability of reconstituted cytochrome c
oxidase versus incubation pH .................... 81

12 Double reciprocal plots showing steady-
state uncoupled activit of cytochrome c
oxidase as a function 0 cytochrome c
concentrations at various pH values ............... 83

13 The pH dependence of (A) maximal activity
(TNmax) and (B) Km for reconstituted
cytochrome oxidase .......................... 85

14 Subunit III-depleted beef heart cytochrome
c oxidase .................................. 90

Figs

15

16

17

18

19

20

21

22

23

24

25

26

27

Double reciprocal plots for reconstituted
beef heart cytochrome oxidase of va 'ng
subunit III content at (A) pH 5.0 an (B)

pH 7.4 ................................

pH dependence of: (A) log (TNm) and (B)

10g (TNmax/Km) ...........................

Stimulation of activity in cytochrome oxidase

vesicles versus nigencin concentration ..........

Effect of independent elimination of Au; and
ApH on oxidation of ferrocytochrome c by

cytochrome oxidase vesicles .................

Effect of independent elimination of Aw and
ApH on the reduction of heme a and cytochrome
c: difference spectra (steady-state minus

oxidized) of cytochrome oxidase vesicles .........

Cha es in levels of reduction of heme a and
cytoc rome c upon independent equilibration

of A1]; and ApH ...........................

Eadie-Hofstee plots of activity of
reconstituted cytochrome oxidase under

coupled and uncoupled conditions .............

Changes in levels of reduction of heme a
and cytochrome 0 upon independent elimination
of Aw and ApH in the presence and absence of

subunit III .............................

Effect of DCCD treatment on the steady-state
level of reduced heme a in the presence and

absence of ionophores ......................

The relationship of RCR and the level of

steady-state reduction of heme a ..............

Hypothetical scheme for the dual control of
respiration in cytochrome oxidase vesicles

by the electrochemical gradient ...............

Carbon monoxide difference spectra of
cytochrome c oxidase with and without

subunit III .............................

The kinetics of cyanide bindi to
cytochrome c oxidase with an without

subunit III .............................

xii

...92

...99

...112

....115

...118

.. .121

....124

...127

...130

...133

...137

...152

...157

CCCP
[MCI-DCCD
DCCD
DEAE

DPG

Em

FPLC
H+/e'

Hepes

LDAO
MES

M.

RCR

SDS
SDS-PAGE

TBS

TMPD

Tris-Cl

A pH
AV
Aux-1+

LIST OF ABBREVIATIONS

Carbonyl cyanide m-chlorophenylhydrazone
N,N’-dicyclohexy1 [14C]-carbodiimide
N,N’-dicyclohexylcarbodiimide
Diethylaminoethyl

Diphosphatidylglycerol

Midpoint oxidation-reduction potential
Fast protein liquid chromatography

Ratio of protons translocated er

electrons transferred by cytoc rome c

oxidase

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

Lauryldimethylamine N-oxide
2-(N-morpholino)ethanesulfonic acid
Molecular weight

Respiratory control ratio

Sodium dodecyl sulfate

Sodium dodecyl sulfate-polyacrylamide
gel electrophoresis

Tris buffered saline
N,N,N’,N’-tetramethyl-p-pheny1enediamine
Turnover number

Tris(hydroxymethyl)aminomethane
hydrochloride

Transmembrane pH gradient
Transmembrane electrical gradient

Transmembrane electrochemical gradient

INTRODUCTION

Oxidative phosphorylation is a complex series of reactions in aerobic
cells that convert the reducing power from metabolism of carbohydrates,
proteins, and lipids into high energy phosphate bonds of ATP needed to
drive cellular functions. The mitochondrial respiratory chain functions as
the energy transducer in this process by accepting reducing equivalents
abstracted from organic substrates, transferring the electrons through a
series of redox centers to mediate a stepwise loss in potential energy, and
capturing the energy released at particular steps in the form of charge and
proton concentration gradients across the inner mitochondrial membrane.
The chemiosmotic hypothesis (Mitchell, 1961) represented the bulk phase
electrochemical potential gradient of protons (Ahm), created during
respiration, as the obligatory intermediate required to drive ATP synthesis.
However, the molecular mechanism for generating and utilizing the
gradient to drive ATP synthesis remains undefined and controversial.
Furthermore, in some recent hypotheses it is postulated that localized
protons (associated with the membrane surface or membrane interior) and
altered protein conformations are essential intermediates, and bulk phase
Aum is regarded as only a byproduct or alternative mode of the energy
conversion process (see recent review by Slater, 1987). These possibilities
emphasize the potential importance of organization of electron transfer
complexes in the membrane and of protein conformation changes that may

accompany electron transfer events.

To achieve a better understanding of the principles involved in the
energy transduction process, more knowledge concerning the structural
basis for activity in the energy converting complexes is needed. Studies
described in this thesis were aimed at correlating aspects of structure and
function for the energy transducing enzyme, cytochrome c oxidase. In
particular, the role of one of the three largest enzyme subunits, subunit III,

and the role of associated phospholipid were investigated.

 

The Role of Qflochrome c Oxidase in Energy Conservation in

Mitochondria.

Of the major electron transfer complexes in the respiratory chain,
cytochrome c oxidase (ferrocytochrome c: oxygen oxidoreductase, EC 1.9.3.1)
has been the most extensively investigated since the initial studies by
Warburg (1924) and Keilin (1925), and yet, many aspects of its mechanism
are not understood. As the terminal member of the electron transport
pathway, cytochrome oxidase accepts electrons from cytochrome c, transfers
them through a series of intramolecular redox carriers, and ultimately
reduces oxygen to water. This enzyme is one of three sites of energy
conversion in oxidative phosphorylation. The protein is inserted through the
inner mitochondrial membrane and is in contact with both the
mitochondrial matrix and the intermembrane space. Two aspects of oxidase
function contribute to the generation of the proton electrochemical gradient.
First, the protons consumed in the reduction of oxygen to water are removed
specifically from the matrix (Sigel and Carafoli, 1980; Wikstrom, 1988).
Second, the enzyme translocates protons from the matrix to the

intermembrane space, an activity often referred to as proton pumping

(Wikstrom, 1977; Wikstrom, 1984). The net reaction can be expressed as:

4 cytochrome c2" 4- 02 + 8 H+mmx —>

4 cytochrome c3+ + 2 H20 + 4 H+imemembme,

space

assuming a proton pump stoichiometry of one proton translocated per one
electron transferred (but see Discussion in Chapter 2).

Electron Transfer Activity. The rate of oxidation of cytochrome c
(Smith and Conrad, 1956; Minnaert, 1961; Yonetani and Ray, 1965) and the
rate of oxygen consumption (Jacobs, 1960; Kimelberg and Nicholls, 1969)
are commonly used as measures of electron transfer activity in cytochrome
oxidase. These parameters can be assessed whether the enzyme is
integrated into lipid bilayers or whether it has been solubilized in a
detergent solution. When the kinetics of the reaction of oxidase with
cytochrome c are examined under steady-state conditions, a biphasic
relationship is observed (Ferguson-Miller et al., 1976; Errede et al., 1976).
The data can be fit to two hyperbolic Michaelis-Menten type reactions
(Sinjorgo et al., 1984), one with a low apparent Km and Vmax (the high
affinity reaction) and one with a high apparent Km and Vmu (the low affinity
reaction). Models to explain this phenomenon include proposals for: two
catalytically active cytochrome c binding sites (Ferguson-Miller et al., 1976;
Errede and Kamen, 1978); one catalytic and one noncatalytic site (Speck et
al., 1984; Sinjorgo et al., 1986); and a two step electron transfer at one
binding site (Antalis and Palmer, 1982).

When assayed in intact membranes, the electron transfer activity of

cytochrome oxidase is inhibited by the generated electrochemical gradient.

This phenomenon is described as respiratory control, and presumably
relates to the physiological regulation of the enzyme. However, the
mechanism and site of action of the inhibitory gradient is the subject of
controversy (see Chapter 4). The respiratory control ratio (RCR) is obtained
by dividing the activity observed in the absence of the generated gradient by
that measured in its presence. The RCR is often used as a measure of the
integrity of the membrane, and, in artificial systems, as an indication of the
completeness of the reconstitution process.

Proton Translocating Activity. Due to the nature of the proton
translocating activity, cytochrome oxidase must be incorporated into sealed
lipid bilayers that are able to maintain a transmembrane gradient in order
to assay proton pumping. Although isolated mitochondria are suitable for
this (eg. Wikstrom, 1977; Wikstrom and Casey, 1985), purified cytochrome
oxidase can be reconstituted into phospholipid vesicles to yield a simpler
experimental system (Racker, 1972; Racker et al., 1979). The enzyme thus
incorporated exhibits respiratory control (Hinkle et al., 1972; Carroll and
Racker, 1977) and has been found to be appropriate for studies of proton
translocation (Wikstrom and Saari, 1977; Krab and Wikstrom, 1978; Casey
et al., 1979a). Although the translocating activity is well documented, the
structural requirements of this function are poorly defined and many
hypothetical models have been proposed [see Chapter 2 and review by Krab
and Wikstrom (1987)].

Structural Basis of Qtochrome c Oxidase Activity.
The present knowledge concerning the structural basis of the
enzyme’s activity largely comes from examination of the purified oxidase.

Typical purification methods include solubilization of the native membranes

with various detergents followed by ammonium sulfate precipitation and/or
chromatographic steps. Investigations become complicated since
purification of oxidase by different methods (even from the same tissue)
yield enzymes that may differ in subunit composition, aggregation state,
phospholipid content and kinetic properties.

Cytochrome oxidase has been isolated from the mitochondria of many
organisms, including: yeast (Mason et al., 1973; Rubin and Tzagoloff, 1978),
slime mold (Bisson et al., 1985), plants [pea (Matsuoka et al., 1981); corn
(Dutch et al., 1987); wheat (W. Peiffer and S. Ferguson-Miller, personal
communication)], shark (Wilson et al., 1980) and mammals [rat (Hundt and
Kadenbach, 1977; Thompson and Ferguson-Miller, 1983); beef (Yonetani,
1961; Hartzell and Beinert, 1974); human (Kuhn-Nentwig and Kadenbach,
1986; Sinjorgo et al., 1987c)]. In addition, prokaryotic cytochrome c oxidases
have been isolated from bacterial cytoplasmic membranes [see reviews by
Ludwig (1987) and Anraku (1988)]. The mammalian enzyme, particularly
that isolated from beef heart, has been the most thoroughly studied.

Polypeptide Composition. While the peptide composition of
cytochrome oxidase has been shown to be as simple as two or three subunits
in the case of the bacterial enzymes (Ludwig, 1987), the purified
mammalian enzyme contains thirteen nonidentical subunits that can be
identified by high resolution SDS-polyacrylamide gel electrophoresis (Figure
1; Kadenbach et al., 1983; Takamiya et al., 1988). As shown previously for

yeast (Schatz and Mason, 1974), the three largest subunits (I, H and III) in
the mammalian enzyme are encoded on the mitochondrial DNA (Anderson
et al., 1982) and synthesized on mitochondrial ribosomes (Kolarov et al.,
1981). The smaller peptides are encoded by the nuclear DNA, synthesized
on cytoplasmic ribosomes, and imported by the mitochondrion (Koch, 1976).

FIGURE 1: Subunit composition of mammalian cytochrome c oxidase. The
subunits of cytochrome oxidase isolated from beef heart as
demonstrated by high resolution SDS-polyacrylamide gel
electrophoresis (Kadenbach et al., 1983; see Experimental
Procedures in Chapter 1). Positions of molecular weight
standards indicated at left and positions of subunits
designated at right. Several of the peptides migrate
anomalously with respect to molecular weight. The molecular
weights calculated from sequence data for twelve of the
subunits are (Buse et al., 1985): Subunit 1, 56,993; Subunit 11,
26,049; Subunit III, 29,918; Subunit IV, 17,153; Subunit Va,
10,668; Subunit Vb, 12,436; Subunit VIa, 9,419; Subunit VIb,
10,063; Subunit VIc, 8,480; Subunit VIIa, 6,243; Subunit
VIIb, 5,441; Subunit VIIc, 4,962 [using the subunit
nomenclature of Kadenbach et a1. (1983)].

Subunh

 

Except for subunit VIII, all beef heart oxidase subunits have been
sequenced from either the DNA (I, II, III; Anderson et al., 1982) or from the
protein (all except I, III and VIII; Buse et al., 1982). A molecular mass of
about 203,000 Daltons was calculated by Buse et a1. (1985), based on a
monomeric complex composed of 2 capies of subunit VIIc and one copy of all
other subunits excluding VIII which was not observed by these
investigators. [Subunits are described using the nomenclature of Kadenbach
et al. (1983)].

Three-Dimensional Structure. Due to inherent problems in
crystallization of membrane proteins (Michel 1983), there has been limited
success in preparing three-dimensional crystals of cytochrome oxidase
suitable for high resolution X-ray diffraction studies (Y onetani, 1961;
Yoshikawa et al., 1988). Thus, very little three-dimensional structural
information is available for this enzyme, making it difficult to establish
clear structure-function relationships. Most of the available knowledge
comes from the X-ray and electron microscopic analysis of the beef heart
enzyme in two-dimensional crystalline sheets (see review by Capaldi et al.,
1983). These studies have consistently predicted that the enzyme has a
lopsided Y-shape. The "tail" of the Y represents the largest protein domain
(C) and is extramembranous, extending into the intermembrane space. The
"arms" of the Y consist of two smaller domains (M1 and M2) that span the
inner mitochondrial membrane to reach the matrix space. Analyses of the
amino acid sequence have led to predictions of from 19 (Frey et al., 1985) to
23 (Azzi et al., 1985a) transmembrane a-helices. However, the three-
dimensional model may not accommodate that much protein in the
membrane domains and may not account for the total amount of protein in

the complex (Capaldi, 1989).

When the enzyme forms two dimensional crystalline arrays after
extraction of the native membranes with deoxycholate, the protein is
monomeric (Fuller et al., 1979). However, when the native membranes are
extracted with Triton X-100, the two dimensional crystalline sheets contain
oxidase dimers in which the extramembranous tails appear to be
intertwined (Deatherage et al., 1982; Frey et al., 1985). It is not known
whether dimer formation is physiologically important. Several studies have
shown that monomers in detergent solution retain electron transfer activity
and normal substrate binding and kinetic behavior (Thompson et al., 1982;
Suarez et al., 1984; Georgevich et al., 1984). However, it is proposed that
dimer formation may be necessary for proton translocation (Sone and
Kosako, 1986;Fine1 and Wikstrom, 1986; also see Chapter 2 for further
discussion).

Chemical labelling, proteolytic digestion and protein crosslinking
studies have provided some information concerning the arrangement of the
subunits within the complex and their orientation relative to the membrane
[J arausch and Kadenbach, 1985a,b; Zhang et al., 1988; also see reviews by
Capaldi et al. (1983) and Freedman and Chan (1984)]. Subunits I, II, III, V
and VIb appear to form the major portion of the cytoplasmic (C) domain.
The three large subunits are also predicted to contain transmembrane,
hydrophobic segments (Frey et al., 1985; Kuhn and Leigh, 1985) and are
believed to contribute, along with several of the nuclear-encoded peptides
(IV, V1c and VIIb; Zhang et al., 1988), to the membrane domains.

Interactions with Membrane Lipids. Cytochrome oxidase is typically
purified as a proteolipid complex (Caughey et al., 1976). For some time it
has been recognized that the enzyme’s activity may be affected by its lipid
environment (Tzagoloff and Mac Lennan, 1965) and the types of detergent

10

used to displace it (Robinson and Capaldi, 1977). It is observed that two to
three molecules of diphosphatidylglycerol (DPG) per monomer are tightly
bound to the purified enzyme (Fry and Green, 1980; Robinson et al., 1980).
In a number of studies, further delipidation by detergent exchange has been
observed to cause the loss of electron transfer activity that is recovered
when DPG is replaced (V ik et al., 1981; Robinson and Wiginton, 1986). Vik
et a1. (1981) found that the delipidated enzyme exhibited monophasic
kinetics for the reaction with cytochrome c with no apparent low affinity
phase. These results, in combination with those of Bisson et al. (1980),
suggesting that enzyme-associated DPG forms a low affinity binding site for
cytochrome c, imply that DPG is required for optimal electron transfer
activity and possibly for the low affinity binding of cytochrome c (but see
Chapter 1).

Prosthetic Groups. Cytochrome c oxidase contains at least four redox-
active metal groups - two heme irons and two copper atoms per monomer.
The two irons are part of identical heme A groups (Caughey et al., 1975).
However, the two hemes are functionally and spectroscopically distinct,
presumably because they are located in different protein environments and
coordinate with the protein using different ligands [see Wikstrom et al.
(1981a); Blair et al. (1983); Holm et al. (1987)]. They are commonly
designated heme a and heme a3, and the enzyme itself is often referred to as
cytochrome aa3. Cytochrome oxidase exhibits a characteristic redox-
dependent spectrum attributed to the heme groups (V anneste, 1966). Both
hemes contribute, although in some cases apparently unequally, to the
absorbance maxima that are observed at 420 nm (Soret band) and 598 nm
(oz-band) for the oxidized enzyme, and at 444 nm (Soret band) and 605 nm
(oz-band) for the reduced enzyme. [See Wikstrom et al. (1981) for a more

11

complete discussion] The two coppers differ in their spectroscopic
characteristics [see reviews by Wikstrom et al. (1981a) and Blair et al.
(1983)]. One copper, Cu, exhibits a near infrared spectral band at 830 nm
in the oxidized form and is EPR detectable. Along with heme a, it is believed
to be situated near the site of interaction with cytochrome c. The other
copper, CuB, is usually EPR silent and appears to function in close
association with heme as at the site where oxygen is reduced to water. The
heme a3-CuB center is the site at which oxygen and the poisonous ligands
(HCN, CO, and azide) bind (see Lemberg, 1969).

Although the specific electron transfer sequence is unclear (Wikstrom
et al., 1981a), there is evidence supporting a general scheme in which the
initial reduction of the enzyme occurs at heme a and/or CuA, followed by
intramolecular electron transfer from these centers to heme a3 and Cup,
which in turn donate electrons for the reduction of oxygen. [Several recent
reviews (Malmstrom, 1982; Hill et al., 1986; N aqui et al., 1986) examine the
present knowledge concerning the intermediate steps in this process] The
electron transfer activity is highly pH dependent and may reflect the
coupling with proton translocation (see Thornstrom et al., 1984). There is
some evidence to indicate that the electron transfer step from heme a to a3
is regulated by pH (Malmstrom and Andreasson, 1985). However, the
locations of the sites that mediate this control are unknown (see Chapter 3).

Additional metals have been reported to be consistently associated
with the enzyme in stoichiometric amounts. Einarsdottir and Caughey
(1985) observed that, in addition to heme a, heme a3, CuA, and CuB, an
oxidase monomer contains and additional 0.5 coppers, one magnesium and
one zinc atom. They proposed that the extra copper is shared by two oxidase
monomers. Bombelka et al. (1986) and Steffens et al. (1987) report that

12

there are a total of three coppers per monomer. The former group tested for
zinc and also detected one zinc atom per oxidase. As yet, there is no
evidence to indicate a function for these extra metals.

Functions of Protein Subunits. Prokaryotic aa3-type oxidases with two
(Hon-nami and Oshima, 1984; Yoshida and Fee, 1984) or three (Van
Verseveld et al., 1981; Sone and Yanagita, 1984) subunits which show
homology to the three largest subunits of the mitochondrial oxidase are
capable of electron transport and proton translocation. This is consistent
with the results of radiation inactivation studies which indicate that only a
subset of the peptides in the mammalian enzyme, comprising about one half
of the total protein in a monomer, is required for electron transfer (Suarez
et al., 1984). On the basis of these observations, in addition to evidence
suggesting that the redox centers are located in subunits I and II (Winter et
al., 1980; Holm et al., 1987), the three large peptides in the mammalian
enzyme are referred to as "catalytic" subunits while the ten smaller peptides
are considered "noncatalytic".

The functions of the "noncatalytic", nuclear-encoded subunits are
unknown, although there are indications that they may be involved in
assembly of the enzyme complex and possibly in regulation of activity.
Studies in yeast have shown that mutations in all but one of the smaller
subunits lead to an inability to produce a viable oxidase, suggesting a role in
the assembly or stability of the complex (McEwen et al., 1986). The case for
regulatory roles for some of the noncatalytic subunits is based on the
observation that several of these peptides exhibit tissue specificity
(J arausch and Kadenbach, 1982; Kuhn-Nentwig and Kadenbach, 1985;
Sinjorgo et al., 1987b; Capaldi et al., 1988), suggesting that the enzymes are
tailored to the energy requirements of certain cell types. There have been

13

reports of ATP binding to several of the smaller subunits (Bisson et al.,
1987) under conditions where ATP alters the binding of cytochrome c to the
enzyme (Ferguson-Miller et al., 1976), suggestive of a regulatory role. Upon
examination of the kinetics of the reaction of the beef liver and beef heart
isozymes with cytochrome c, Buge and Kadenbach (1986) concluded that the
catalytic activity was influenced by tissue specific subunits. However,
Sinjorgo et al. (1987b) found no major differences in the kinetics of oxidases
from various tissues.

While the catalytic activity has been attributed to subunits I, II, and
III, their individual roles are not completely defined. Subunit I (M, 56,993)
has been proposed as the location for binding of the heme a, heme a3, and
Cup metal centers, based on molecular modeling and conservation of amino
acids suitable for liganding (Holm et al., 1987). By the same criteria, as well
as chemical modification studies (Hall et al., 1988), CuA is believed to be
associated with subunit II (Holm et al., 1987; Martin et al., 1988). In
addition, chemical crosslinking and competitive binding experiments have
shown that subunit II (M, 26,049) is the major site of interaction between
the oxidase and cytochrome c (Bisson et al., 1980; Millett et al., 1984), while
subunit I contains the redox centers that form the oxygen binding site.
Thus, these two subunits, which contain the four redox centers and the
major sites for substrate binding, are fundamental to cytochrome oxidase
activity. In fact, these two peptides may be the only essential subunits for
the electron transfer activity of the enzyme, as illustrated by the fact that
prokaryotic enzymes isolated with only these two subunits are retain high
activity (see Ludwig, 1987). Thus, the role of subunit III in catalysis is
unclear and has been the subject of many studies (reviewed by Prochaska

and Fink, 1987).

14

ubunit III.

Subunit III, a product of the mitochondrial genome in eukaryotes, has
also been identified as a component of some prokaryotic cytochrome
oxidases (Sone and Yanagita, 1982; De Vrij et al., 1986). There is some
uncertainty concerning whether bacterial oxidases isolated as a complex of
only two subunits are complete enzymes or whether these enzymes have
been inadvertantly depleted of subunit HI during purification. In the case of
Paracoccus denitrificans, the purified bacterial enzyme was considered a
two subunit oxidase (Ludwig and Schatz, 1980) until recently when a gene
encoding a protein homologous to subunit III was found in the Paracoccus
genome (Saraste et al., 1986). Haltia et al. (1988) have since isolated a three
subunit oxidase from this bacteria. Therefore, until genomic testing is
complete, it is not clear whether subunit III is a fundamental constituent of
all aa3-type cytochrome c oxidases.

The amino acid sequence of subunit III (M, 29,918; Anderson et al.,
1982) indicates that the polypeptide is predominantly hydrophobic and
likely to be one of the subunits that interact with the phospholipid
membrane. Using a membrane propensity algorithm (Kuhn and Leigh,
1985), subunit III has been predicted to contain from four (Frey et al., 1985)
to seven (Azzi et al., 1985a; Wikstrom et al., 1985) transmembrane a—helical
segments. Direct evidence for the association of this subunit with the
membrane was provided by showing that this peptide was heavily labelled
in photoaflinity experiments using arylazidophospholipids (Bisson et al.,

19 79) and other phospholipid photo-spin-labels (Griffith et al., 1986). Some
information about the orientation of the peptide is available from chemical

labelling and binding studies. Derivatized forms of cytochrome c have been

15

shown to bind to cytochrome oxidase with the reducing face of the substrate
interacting with subunit II (Bisson et al., 1980; Millett et al., 1984) and the
back side of the substrate interacting with subunit III (Birchmeier et al.,
1976; Fuller et al., 1981). This suggests that a portion of subunit III
contributes to the cytoplasmic domain of the protein and may be near the
high affinity binding site for cytochrome c, or alternatively, that the oxidase
as a dimer creates a cleft with subunit III of the second oxidase molecule
approaching the back of the bound cytochrome c. While this subunit has
been consistently accessible to chemical modification (Zhang et al., 1984),
antibodies (Chan and Tracy, 1978), and proteolytic enzymes (Zhang et al.,
1984; J arausch and Kadenbach, 1985b) from the cytoplasmic side, results on
accessibility from the matrix side have been mixed (see Freedman and
Chan, 1984). It has been argued (Freedman and Chan, 1984) that only those
studies in which enzyme orientation was poorly defined (Eytan et al., 1975;
Ludwig et al., 1979) suggest that subunit III is in contact with the matrix.
Thus, although the structural predictions suggest a membrane spanning
peptide, experimental results are still equivocal.

Two basic approaches have been used to study the role of subunit III
in cytochrome oxidase activity: (1) use of reagents that react covalently with
subunit III, and (2) examination of the effects of subunit III depletion on the
enzyme’s activity. Subunit III is apparently somewhat loosely associated
with the rest of the complex and can be removed under some conditions. The
peptide has reportedly been recovered in isolated form using one type of
preparation (Bill and Azzi, 1982). However, subunit HI has yet to be
reconstituted into lipid bilayers alone or with subunit III-depleted oxidase,
presumably due to the difficulty in achieving a monodisperse preparation
that retains native-like configuration.

16

Effects of Reagents that Bind Subunit III. N, N’-
dicyclohexylcarbodiimide (DCCD), a chemical reagent specific for carboxyl
groups in hydrophobic environments, has been shown to primarily modify
subunit III in cytochrome oxidase and subunits II and IV to a lesser extent
(Casey et al., 1980; Prochaska et al., 1981). The chemistry of DCCD (see
Solioz, 1984) is such that it may interact with carboxyl groups to form a
stable adduct that can be detected by incorporation of radiolabelled reagent.
Alternatively, the reagent may also interact to form self-eliminating protein
crosslinks that cannot be detected by radiolabelling. Subsequent studies
using radioactive labelling have identified that binding to subunit III occurs
specifically at glutamic acid 90 in the beef heart enzyme (Prochaska et al.,
1981). Hydrophobicity plots predict that this residue is located in the center
of a transmembrane segment (Wikstrom et al., 1981b; Zhang et al., 1984;
Frey et al., 1985). The DCCD binding sites in subunit III from yeast and
Paracoccus denitrificans are predicted to be in homologous protein regions
(Saraste et al., 1986). Using a spin labelled analogue of DCCD [NCCD, N-
(2,2,6,6-tetramethylpiperidyl-1~oxyl)-N’~cyclohexylcarbodiimide], Casey et
al. (1981) showed that the binding site is indeed in a hydrophobic
environment.

Examination of activity in DCCD-modified cytochrome oxidase
revealed that proton translocating ability was severely inhibited (Casey et
al., 1979b; Casey et al., 1980; Prochaska et al., 1981). Electron transfer was
inhibited to varying degrees in these studies but was consistently less
afl'ected than proton pumping. Inhibition of proton translocation was found
to be proportional to the degree and the time course of incorporation of
DCCD into the whole enzyme (Casey et al., 1980; Prochaska et al., 1981),

suggesting that chemical modification was responsible for the alteration in

17

enzyme activity. Although subunit III was the most highly labelled, the
level of modification in this peptide alone was insufficient to account for the
extent of inhibition observed.

Nevertheless, the similarities between: (1) the proposed structure of
subunit III and the structure of the proton channel of ATP synthetase, and
(2) the effect of DCCD binding on proton translocation by cytochrome
oxidase and the inhibitory effect of DCCD on the proton channel of ATP
synthetase (Beechey et al., 1967; Fillingame, 1980), have led to proposals
that subunit III functions as a proton pump in cytochrome oxidase (Casey et
al.,1980; Solioz, 1984). However, it must be kept in mind that DCCD may
cause undetected crosslinks that may occur at other sites and, as pointed
out by Azzi and Nalecz (1984), DCCD modification studies cannot
distinguish between pumping, channeling, or regulatory functions for a
modified subunit. In fact, another mechanism of subunit III involvement in
proton translocation is based on results showing that DCCD causes subunit
III to detach from the enzyme (Muller and Azzi, 1985; Azzi et al., 1985b)
and that subunit III removal favors monomerization (Finel and Wikstrom,
1986). Thus, advocates of the dimer hypothesis of proton translocation,
propose that subunit III plays a secondary role in this process by stabilizing
the dimeric form (Nalecz et al., 1985; Finel and Wikstrom, 1986).

In other studies using a DCCD-modified enzyme, Moroney et al.
(1984) found that, in addition to inhibition of proton pumping, DCCD
caused the enzyme to become less sensitive to changes in the enzyme-
generated transmembrane gradient, i.e. less subject to respiratory control.
They proposed that the DCCD-binding site on subunit III was a proton
binding site that could be involved in coupling proton pumping and electron

transfer.

18

Chan and Freedman (1983) reported that anti—subunit III antibodies
had the ability to inhibit proton translocation in cytochrome oxidase vesicles
without affecting the overall rate of electron transfer. However, these
antibodies showed some cross reactivity with subunit II upon Western blot
analysis (Gai et al., 1987).

Subunit III-Depleted Cytochrome c Oxidase. Subunit III can be
removed from the cytochrome oxidase complex by a number of procedures
(Table 1). The preparation of III-deficient oxidase often includes use of high
concentrations of the detergent Triton X-100 at a pH greater than 8.
Incubation under these conditions is used in conjunction with ion exchange
(Saraste et al., 1981; Penttila 1983) or affinity (Bill and Azzi, 1982)
chromatography, native gel electrophoresis (Prochaska and Reynolds, 1986),
or proteolytic digestion (Malatesta et al., 1983; Puettner et al., 1985).
Subunit III-depleted enzymes have also been prepared under milder
conditions using the detergent lauryl maltoside [shown to maintain the
enzyme in highly active form (Ferguson-Miller et al., 1982)] in
hydroxylapatite and cytochrome c affinity chromatography (Thompson and
Ferguson-Miller, 1983) or by anion exchange fast protein liquid
chromatography in the detergent lauryldimethylamine N-oxide (LDAO;
Finel and Wikstrom, 1986).

All these methods yield preparations containing less than 15% of the
normal complement of subunit III, but also cause the removal of several of
the smaller peptides (i.e. Vb, VIa, and VIb). The heme-to-protein ratios for
the III-deficient complexes become higher in proportion to the protein lost,
indicating that the heme groups remain with the complex (Saraste et al.,
1981; Thompson and Ferguson-Miller, 1983). Some investigators have
reported slight alterations in the visible spectrum, specifically slight blue-

15)

.osanco Hocucoo madcamucoo n HHH banana» 0» o>~umaoa nofiuuaunmamo HmcoHoocsm m

 

Aowmpv soapnxa: can Honda uomno: nomp on a: cannons; acme: anmLm09meoaco ousaqa cuouoaa
uo: atom unuu omcmnxo codes an vozoanou

a.» mg on o<oq aw a" cousmnsocH

 

 

 

Ammo—v .Hm um conneonp uomuo. nomp A amp Lo>uq m an an ovanouams
Asmapv .Hm pa nogmsm umm fimasma um.o cu canon auacquum
Azom—V connaonh o osocnooumo puma: coco: ucm
Ammmpv ouaumamaaxocuan co unamcmoumeoano
LoHHHZuconszom can conaeonu
Aommpv .Hm pa munoumamz uoouom uosnoz noouow acme: m.o an on oumaozo
Ammo—V .Hm an «atom «com am an xnawcmoumaocno m<mo an
Anna—v .Hm no Locuaosa 6o30aaou m.m mg on oopnx confine
Ammmpv .Hm no munoumam: a? nu canamguosmzona and: noumoce
prmpv scan was axnmnoocm no «mm umm acme: m.» xn um
Ammo—v nuaoczom cam mxnmnooaa noon oopnx confine n~.o ea mqmocona

nocuooao How o>«umc a.o ma
um oopnx confine um cu :oHumnsocH

 

 

 

Ammm_v .Hm as nomfimz ac am» now scam: ~.s :a

Anmwmpv .Hm no :oaoze umam pm oopnx confine um.Fup sud:

Amwmpv «nn< new Hawm oopsao asmucwluacamnmoumsocno
zuucuuum o oeocnoouho ammo»

Ammm.v «Augusta “o aom mom can»: m.» mg as scaacmosmeocso

A_mmpv socunxaz cam «Hauucoa poem m<mo an oozoaaou m.m mg as
Apmmpv .Hm no mummcmm oo—ux cauuah a. c“ canumnsocH

cofiumooncmcp Hocpcoo cmumcmae moczom
noocaaouom coaoca zcoumcqammm cocuooam osmmup ceaumgaaoga uo cacao:

mmmfisuaunmqmo Hmcofisocsu

 

moomemz HZMmmmmHQ xm amm<mmma mmm<onorm mZOmIUOHwo omemnmmo I HHH thnmam mo mmnhmmmomm up m4m<H

20

shifts in the a-band of the reduced enzyme (Penttila et al., 1983; Puettner et
al., 1985).

As a whole, the III-depleted oxidases retain a high level of electron
transfer activity. In fact, using radiation inactivation to estimate the
functional size for electron transfer activity, Suarez et al. (1984) found that
a functional molecular weight of about 100,000 Daltons was obtained
regardless of whether the III-containing or III-depleted oxidase was
analyzed.

Upon incorporation of the III-depleted enzymes into artificial vesicles,
the enzymes exhibit respiratory control, indicating that enzyme-generated
electrochemical gradients are formed. However, those enzymes prepared in
Triton X—100 have poorer RCR’s than the III-containing enzymes (Penttila,
1983; Thelen et al., 1985b), while those prepared in lauryl maltoside or
LDAO have higher RCR’s than the control enzymes (Thompson and
Ferguson-Miller, 1983; Finel and Wikstrom, 1986).

In view of the implications of DCCD modification studies, evaluation
of the effect of subunit III depletion on proton translocation was of primary
importance. Indeed, several studies have indicated that pumping capability
is completely lost upon removal of subunit III (Penttila, 1983; Thelen et al.,
1985b; Prochaska and Reynolds, 1986). However, other groups, using
enzymes prepared by alternate methods, report that up to 50% of the proton
translocating activity is retained by the III-deficient enzymes (Puettner et
al., 1985; Sarti et al., 1985; Thompson et al., 1985). The latter results
indicate that subunit III is not essential to the translocating mechanism.
The fractional loss in activity could reflect that subunit III increases the
efficiency of the translocating process, or that the pumping mechanism (in

which subunit III has no role) is coincidentally disrupted by the conditions

21

used to prepare the III-depleted oxidase. The lability of this process is
evident since proton translocating activity is susceptible to heat treatment
which does not affect electron transfer activity or subunit composition (Sone
and Nicholls, 1984). An alternative interpretation to the retention of proton
translocation in some III-depleted oxidases is that the acidification detected
in the assays is an artifactual result of scalar release of protons upon
binding of substrate to the oxidase (Muller et al., 1987). However, evidence
against this interpretation is presented in Chapter 2.

In addition to variability in proton translocation, these preparations
also exhibit different states of aggregation. Some have been reported to be
dimeric (Saraste et al., 1981; Nalecz et al., 1985), while others are
monomeric (Suarez et al., 1985; Finel and Wikstrom, 1986). Monomeric
populations are consistent with the observation that subunits III (Finel and
Wikstrom, 1988) and especially VIb (Finel, 1987) are involved in the
formation of dimers.

Further characterization of these III-deficient preparations has led to
a number of interesting observations. (1) Metal analysis by Yewey et al.
(1987) revealed that the enzyme contained two irons, 2.5 coppers, one zinc,
and one magnesium in the presence or absence of subunit III. (2) Robinson
and Wiginton (1985) observed that conditions used to delipidate the enzyme
also resulted in loss of subunit III. The reverse also appears to be true since
III-depleted enzymes have been reported to have low lipid contents
(Thompson and Ferguson-Miller, 1983; Nalecz et al., 1985). (3) Changes in
ligand binding at the heme a3-Cu3 center have been attributed to subunit
III depletion based on changes in the carbon monoxide difference spectrum
(Nalecz et al., 1985) and on changes in cyanide binding kinetics (Hill and
Robinson, 1986).

22

In summary, the III-depleted preparations are similar in subunit
composition (depleted of III, Vb, VIa, and VIb) and retention of high electron
transfer activity, but differ in their ability to translocate protons and in

their state of aggregation.

Most (if not all) aa3-type cytochrome oxidases contain a peptide
showing a high degree of homology to subunit III of the mammalian
enzyme. This alone suggests that the peptide is important to the enzyme’s
function. However, so far, results of many studies have failed to present a
clear picture of its role in the transduction of energy by cytochrome oxidase.

This thesis reports on further studies of the roles of subunit III and
phospholipid in cytochrome oxidase activity. The effect of subunit III
removal on respiratory control, proton translocation, control of activity by
pH, and ligand binding at heme as was assessed for a III-deficient
cytochrome oxidase prepared from rat liver by cytochrome c affinity
chromatography in lauryl maltoside (Thompson and Ferguson-Miller, 1983),
and, in some studies, for a III-depleted beef heart oxidase. In addition, the
effect of delipidation on optimal activity was investigated. Experiments
were designed to clarify some of the inconsistencies in the existing
literature concerning the extent of subunit III involvement in energy
transduction, to evaluate the influence of the phopholipid environment on
activity, and to elucidate the complex mechanism of cytochrome oxidase

‘ function.

CHAPTER 1

PURIFICATION OF LIPID- AND SUBUNIT III-DEPLETED
CYTOCHROME c OXIDASE:
STRUCTURAL AND FUNCTIONAL CONSEQUENCES

One biochemical approach to studying the roles of particular subunits
or of phospholipid in cytochrome oxidase activity is to examine the effects of
removal of the component in question. In the case of subunit III, a variety of
treatments have been successful (eg. detergents, high salt, proteolytic
enzymes) but with variable effects on enzyme activities (see Table 1 in
Introduction). Similarly, a variety of methods for lipid removal have been
attempted including solvent extraction (Awasthi et al., 1971), detergent
exchange (Fry and Green, 1980; Robinson et al., 1980; Vik et al., 1981), as
well as, replacement with specific lipids (Yu et al., 1975; Vik and Capaldi,
1977; Powell et al., 1985) with variable consequences and degrees of su’ccess.
However, interpretation of results obtained with these III-depleted and
lipid-depleted enzymes has been difficult because of conflicting data, general
loss of activity, and the concern that altered activity may result from side
effects of the procedure used rather than the loss of a particular subunit or
lipid. This problem is difficult to overcome, except by demonstrating
retention of as many native characteristics as possible, with loss of only
specific functions, and ultimately the recovery of that function by

reconstitution with the depleted component.

23

24

Typically, cytochrome orddase depleted of subunit III has been
prepared by treatment at alkaline pH with Triton X-100 (Saraste et al.,
1981; Thelen et al., 1985b; Prochaska and Reynolds, 1986; Puettner et al.,
1985), a detergent known to inhibit oxidase electron transfer activity
(Robinson and Capaldi, 1977; Sinjorgo et al., 1987a). However, Thompson
(1984) devised a method to purify a highly III-depleted oxidase from rat
liver using the detergent lauryl B-D-maltoside which has been shown to
maintain the enzyme in a highly active form (Rosevear et al., 1980;
Ferguson-Miller et al., 1982; Sinjorgo et al., 1987a). The method involves
differential solubilization of proteins from the mitochondrial membranes
with lauryl maltoside to yield a "crude," soluble oxidase. This is followed by
chromatography in lauryl maltoside, first on hydroxylapatite and then on
an affinity resin consisting of a Sepharose 4B matrix to which horse heart
cytochrome c is covalently bound.

In addition to being greater than 95% depleted of subunit III, the
resultant enzyme is also highly delipidated, containing only 3 nmol P per
nmol of enzyme, identified as one nmol of diphosphatidylglycerol (DPG; a
phospholipid containing two phosphate groups per molecule) and one nmol
of phosphatidyl inositol (Thompson and Ferguson-Miller, 1983). Further
examination (Thompson and Ferguson-Miller, 1983; Thompson, 1984)
revealed that the subunit III-deficient orddase had an unchanged visible
spectrum, retained high electron transfer activity, and had biphasic kinetics
for the reaction with cytochrome c. Upon incorporation into phospholipid
vesicles, the III-depleted oxidase demonstrated high respiratory control
ratios, indicating successful integration with the bilayer and retention of the
ability to generate and respond to a transmembrane electrochemical

gradient. Thus, this enzyme displayed little evidence of the disruption of

25

functions that had been reported to be altered by other methods of removal
of either subunit III or lipids. Based on these findings, cytochrome oxidase
depleted of subunit III and lipid using this method appeared to provide a
promising tool for re-examination of the putative roles of subunit III and
phospholipid in cytochrome oxidase activity.

In order to carry out this investigation, it was necessary to analyze
the subunit composition of this enzyme preparation and to quantify the
extent of subunit III depletion. It was found that 2-5% remained, but an
additional step of anion exchange fast protein liquid chromatography
(FPLC) yielded a preparation with no detectable subunit III (and no loss of
activity). In addition, by extending the methodology of Thompson and
Ferguson-Miller (1983), cytochrome oxidase containing sub-stoichiometric
amounts of DPG was prepared, allowing the effect of DPG removal on
enzyme activity to be examined more critically than previously.

Studies of the effects of subunit III depletion on specific enzyme
functions are reported in the following chapters.

26

EXPERIMENTAL PROCEDURES

Mafirials, Cytochrome c (horse heart, Type VI, Sigma), repurified according
to Brautigan et al. (1978), was reduced with Na dithionite and then
chromatographed on Sephadex G-50-40 in 25 mM Hepes (pH 7.4) and 40
mM KCl to remove excess reducing agent and polymerized forms of the
protein. Lauryl B—D-maltoside was synthesized according to Rosevear et al.
(1980). The following chemicals were obtained from the sources indicated:
DEAE-Sephacel and Mono Q anion exchange column (Pharmacia, Sweden);
acrylamide and methylene-bis-acrylamide (Serva, Westbury, NY); Acrylaide
and GelBond PAG film (FMC Bioproducts, Rockland, ME); [14C]-DCCD
(Research Products International, Mount Prospect, IL); nitrocellulose filters,
goat anti-mouse alkaline phosphatase conjugate (Bio-Rad, Richmond, CA);
nitro blue tetrazolium, 5-bromo-4-chloro-3-indolyl phosphate (Sigma, St.
Louis, MO) . Sources of other materials were listed by Thompson and
Ferguson-Miller (1983).

CMChrome Oxidase Preparations, Subunit III-containing cytochrome
oxidase was prepared from beef heart according to Suarez et a1. (1984)
omitting the ethanol wash step. Subunit III-containing cytochrome oxidase
was isolated from rat liver by preparing "crude," soluble oxidase according
to Thompson (1984), followed by DEAE chromatography and anion
exchange FPLC in lauryl maltoside. The crude oxidase was bound to a
DEAE-Sephacel column that had been previously equilibrated with 50 mM
Ms-Cl (pH 8), 330 mM sucrose, 10 mM K phosphate, 15 mM KCl and 15
mm lauryl maltoside. Material that failed to bind was washed through with

27

the same buffer. Cytochrome oxidase was eluted as one fraction using 300
mM KCl in the above buffer and was then dialyzed versus 200 volumes of 25
mM Tris-Cl (pH 8) and 5% sucrose. The enzyme was chromatographed using
anion exchange FPLC on a Mono Q column in 25 mM Tris-Cl (pH 7.8), 5%
sucrose and 15 mM lauryl maltoside. The enzyme eluted in five major peaks
between 300 mM and 700mM KCl using a 0-1.0 M gradient in the same
buffer. The fourth peak, showing least contamination with other spectral or
peptide components, was concentrated by ultrafiltration.

Subunit III-depleted cytochrome oxidase was isolated from rat liver
according to Thompson and Ferguson-Miller (1983). Preparations with no
detectable subunit III were achieved by an additional step of anion
exchange FPLC as described above. The enzyme elutes as one major peak
with a trailing shoulder between 200 and 300 mM KCl when using a 01.0
M gradient. The subunit composition of fractions in the peak and the
shoulder appeared to be identical (as in Figure 5A, Lane 3).

A highly delipidated enzyme was isolated from rat liver mitochondria
starting with preparation of "crude," soluble oxidase according to Thompson
(1984). The enzyme was subjected to hydroxylapatite chromatography (as
described by Thompson and Ferguson-Miller, 1983) four times using NaCl
concentrations of 30 mM in column 1, 15 mM in column 2, 5 mM in column
3 and 0 mM in column 4. After each chromatography, the enzyme was
dialyzed overnight at 4°C versus 200 volumes of 50 mM Tris-Cl (pH 8) and
330 mM sucrose for columns 13 and 25 mM Tris-Cl (pH 8) and 5% sucrose
for column 4. The enzyme was finally submitted to anion exchange FPLC as

described above and was found to elute with the same characteristics as the

III-depleted enzyme.

28

 

SDS-Pglmgylamide Gel Electrophoresis (SDS-PAGEL Subunit composition

of the oxidase preparations was evaluated by SDS-PAGE using either of two
modifications of the procedure described by Kadenbach et al. (1983). In
Procedure 1, the gels were crosslinked with 0.38% rather that 0.56%
methylene-bis-acrylamide in a 17 cm separating gel to achieve resolution of
all 13 cytochrome oxidase subunits. In Procedure 2, the gels were
crosslinked with 0.38% Acrylaide (instead of methylene-bis-acrylamide) and
attached to GelBond PAG film. This modification enhances separation
between subunits 11 and III but fails to resolve some of the smaller peptides.
Preparation of samples for electrophoresis was performed as described by
Thompson and Ferguson-Miller (1983), but using the dissociation buffer
described by Kadenbach et al. (1983).

Densitometric scanning of Coomassie blue stained gels and
integration of the total area under the peaks reveals a ratio of subunit III to
subunit II of 1:1 for the beef heart orddase, indicating that subunit III stains
with equal intensity compared to the other peptide. The subunit III content
of other orddase preparations was calculated according to the ratio of the
area of the integrated peaks representing subunit III and subunit II.

[.129 I-DCCD Labelling of Cytochrome Oxidase. [14C]-DCCD in ethanol was
added to oxidase solubilized in 15 mM lauryl maltoside at a molar ratio of
80:1. The reaction was carried out in 10 mM Hepes-KOH (pH 7.4), 44 mM
KCl, 55 mM sucrose under the conditions described by Prochaska et a1.
(1981). The labelled oxidase was precipitated using 5% trichloroacetic acid
and then electrophoresed on a 10-22% acrylamide gradient gel using the
SDS-PAGE system of Thompson and Ferguson-Miller (1983) which resolves
the large subunits but not all of the small peptides. Quantitation of the

29

peptides and incorporated radioactivity is described in the legend to Figure
3.

Western Blot Analysis. For Western blot analysis, cytochrome oxidase was
subjected to SDS-PAGE using Procedure 2 described previously except that
the GelBond PAG film was not used. The protein bands were transferred to
nitrocellulose filters by electroblotting for 2 hr. at 150 mA in 12.5 mM Tris-
Cl (pH 8.3), 96 mM glycine, 0.005% SDS, 20% (v/v) methanol. Transferred
protein was visualized by staining in 0.025% amido black, 22.5% ethanol,
7.5% acetic acid. Strips to be treated with antibody were destained and
blocked by washing with 5% bovine serum albumin in Tris buffered saline
(TBS; 10 mM Tris-Cl, pH 7.5, 150 mM NaCl). Strips were treated with a
1:50 solution of monoclonal anti-subunit 11 antibody (02) in TBS. [Antibody
was prepared and characterized by Taha and Ferguson-Miller (1987).] This
was followed by 3 washes with 0.05% Tween-20 (v/v) in TBS. Strips were
treated with a 1:3000 dilution of goat anti-mouse alkaline phosphatase
conjugate in TBS for 1 hr. Two washes in 0.05% Tween-20 (v/v) in TBS and
one wash in TBS followed. Protein bands reacting with the primary and
secondary antibodies were visualized by staining with a solution of 0.3
mg/mL nitro blue tetrazolium (pre-dissolved in dimethyl formamide) and
0.15 mg/mL 5-bromo-4-chloro-3-indolyl phosphate in carbonate buffer (0.1 M
NaHCOa, 1 mM MgC12, pH 9.8).

Spectral Assays. Cytochrome oxidase concentrations ([aaal) were obtained
from absolute spectra (air orddized or dithionite reduced) using the

extinction coefficients of van Buuren et al. (1972): 140 mM-lcm-l at 420-490

30

nm for oxidized aa3 and 40 mM-lcm-l at 605-650 nm and 204 mM-lcm-l at
440-490 nm for reduced aa3.

Phosphorus Assays. The total phosphorus associated with purified
cytochrome oxidase was determined by ashing the enzyme and measuring
phosphorus according to the procedure of Ames (1966). Two pools of beef
heart cytochrome oxidase were prepared and analyzed several times to
evaluate the reproducibility of this assay. Pool 1 had an average value of
38.0 nmol P per nmol am; with a standard deviation of 3.0 for 14 assays
(using 0.34 nmol aa3 per assay). Pool 2 had an average value of 10.5 nmol P
per nmol am», with a standard deviation of 1.0 for 10 assays (using 0.37 nmol
aa3 per assay). In all cases, the coefficient of correlation for the standard
curve was greater than 0.99. When analyzing enzymes with much lower
phosphorus content, sufficient sample (as much as 2.5 nmol aa3) was added

to deliver more phosphorus than the lowest standard used.

Electron Transfer Activity. Steady-state electron transfer activities were
measured polarographically at 25°C using a Gilson oxygraph (Model 5\6 H;
Middleton, WI) under the conditions described in Figures 6 and 7. The
reductants, ascorbate and tetramethyl-p-phenylenediamine (TMPD) were
added to keep cytochrome c in the reduced state. Turnover numbers in
(electrons) per second were calculated as described by Thompson and

Ferguson-Miller (1983).

31

RESULTS

't Com sition of "Com lete" C chrome c Oxidase. Cytochrome
oxidase isolated from beef heart and rat liver was found to contain at least
13 different polypeptides upon analysis by high resolution SDS-PAGE
(Figure 2), as previously demonstrated for mammalian enzymes (Kadenbach
et al., 1983; Takamiya et al., 1988). Several of these peptides migrate
differently in the two samples. The differences in the migrations of subunits
VIa, VIIa and VIII have been shown to be the result of sequence differences,
apparently reflecting tissue-specific characteristics of the heart and liver
enzymes (Merle and Kadenbach, 1982; Capaldi et al., 1988). Subunit III
appears to be more loosely associated in the rat liver enzyme complex than
in the beef heart complex. A rat liver enzyme containing the same amount
of subunit III as the beef heart enzyme was not achieved. The rat liver
oxidase shown in Figure 2 contains 60% of the subunit 111 found in the beef
heart oxidase.

Subunit Commsition of Subunit III-Depleted CMChrome c Oxidase.

Electrophoresis of cytochrome oxidase using a Laemmli (1970) gel system
resolves the enzyme into only seven bands (Thompson and Ferguson-Miller,
1983; Penttila, 1983). Analysis of the rat liver enzyme purified by affinity
chromatography in lauryl maltoside using this gel system revealed that the
enzyme was highly depleted of subunit III (Thompson and Ferguson-Miller,
1983), containing only 2-5% of the amount found in other purified enzymes
(Thompson et al., 1985). The apparent depletion of subunit III in this
oxidase was confirmed by the lack of labelling in the subunit III region after

32

FIGURE 2: Subunit composition of cytochrome c oxidase from beef heart
and rat liver. Subunit III-containing beef heart and rat liver
cytochrome oxidases (100 (51g per lane) were resolved into 13
rotein bands by SDS-PA E using Procedure 1 described in
xperimental Procedures. Subunits are labelled according to
the nomenclature of Kadenbach et al. (1983).

33

Heart Liver
(BeefiiRat)

Subunfl

 

 

Figure 2.

34

reaction with [14C]-DCCD (Figure 3B). [14C]-DCCD has been shown to
covalently modify a specific glutamate residue in subunit III (Prochaska et
al., 1981) identifying the presence of this peptide following SDS-PAGE
(Figure 3A). Relative labelling of other peptides, specifically subunits II and
IV, in the III-containing and III-depleted orddase preparations was similar.

Analysis by high resolution SDS-PAGE reveals that, in addition to
subunit III depletion, this enzyme is also deficient in subunits Vb, VIa, VIb,
and VIIa (Figure 4). The loss of these smaller subunits has been reported to
accompany the removal of subunit III by other methods (Penttila, 1983;
Puettner et al., 1985; Nalecz et al., 1985; Prochaska and Reynolds, 1986),
although it is difficult to compare data due to differences in electrophoresis
systems and nomenclature of the subunits.

When using this high resolution gel system, it was surprising to find
that more subunit III appeared to be present than observed by the previous
method of analysis. Therefore, the high resolution gel system was modified
to enhance separation of subunits II and III. This procedure revealed that
the band migrating in the subunit III position can be resolved into two
separate bands (Figure 5A). Of the two, subunit III migrated farther (Lane
1) and, in the case of the III depleted-preparations (Lanes 2 and 3),
indicated subunit III contents similar to original estimates. In Lane 3, no
detectable subunit III is found in a preparation that was subjected to
additional purification by anion exchange FPLC. Western blot analysis
(Figure 5B) revealed that the extra peptide (between subunits II and III in
this gel system) reacts with a monoclonal anti-subunit II antibody,
indicating that it is a proteolytic fragment of subunit H. Thus, although
SDS-PAGE using Procedure 1 resolves the enzyme into 13 protein bands, it

is inappropriate for determinations of subunit III content due to the

35

FIGURE 3: [14C]-DCCD binding to purified cytochrome c oxidase. (A)
Subunit III-containi beef heart oxidase. (B) Subunit III-
depleted rat liver oxi ase. Purified cytochrome oxidase was
labelled with [14CJ-DCCD as described in Experimental
Procedures. Relative absorbance of Coomassie blue stained
peptides was measured at 550 nm (dotted line). Radioactivity
in 1 mm gel slices was measured by scintillation counting
(solid line). Subunits are labelled using the nomenclature of
Downer et al. (1976).

36

xmr>1<m >wm0ww>znm a. mac :3 A ....... V

 

 

 

 

 

 

 

 

 

 

 

 

 

I 1| 1 I
.. 1
9 1 9 1 V
w . ................................................ I ............................................................... ........
1- ooooooooo j
7 T N If 7 l
.l .............................. $333333” ........
m I O. m f 00000
c C 2.. ....................
.................. I I.......o.o.........o.....o..“a...” . o . o o o as. o . o. o o o
5 1. I s ....... 5 .I oooooooooo
3 r 3 r m
1' I 1' T
A B ¥
_ p p — p b _ _ .3”...
0 0 0 0 0
0 0 0 0 O
6 4 6 4 2

Alv .23 38-34;

Figure 3.

37

FIGURE 4: Subunit composition of subunit III-depleted cytochrome c
oxidase isolated fi-om rat liver. Subunit III-containing beef
heart (+HI) and subunit III-depleted rat liver (-III)
cytochrome oxidases (100 11 per lane) were resolved into their
component subunits by SD -PAGE using Procedure 1 as
described in Experimental Procedures. Subunits are labelled
according to the nomenclature of Kadenbach et al. (1983). The
suliIunits depleted in the rat liver preparation are indicated at
rig t.

38

Subunfl

Vl

 

a-
b—
c —

O
VII b;
C

VIII /

 

 

Figure 4.

39

FIGURE 5: SDS-PAGE and Western blot analysis of subunit III-depleted
cytochrome oxidase isolated from rat liver. (A) Subunit
composition as revealed by SDS-PAGE using Procedure 2 for
subunit III-containing beef heart oxidase (Lane 1) and
subunit III-depleted rat liver oxidase before (Lane 2) and
after (Lane 3) anion exchange FPLC. Each lane contained 75
pg of rotein. (B) Subunit III-depleted cytochrome oxidase
was e ectrophoresed as in (A) [but without GelBond PAG film]
and blotted onto nitrocellulose filters. Lane 1, stained with
amido black, reveals successfully transferred protein bands.
Lane 2 was incubated with monoclonal anti-subunit II
antibody, washed, and then incubated with an alkaline
phos hatase-conjugated secondary antibody. Color
deve opment as described in Experimental Procedures. Lane
3 was treated identically to Lane 2 omitting the primary
antibody. Proteolytic fragment of subunit 11 indicated with
arrows.

40

Figure 5.

IDN

_:>

955
ES

u..:>
u_>

a.o>

 

41

comigration of subunit III and a fragment of subunit II that is apparently
present in these preparations. Table 2 lists the subunit HI content for the

enzyme preparations in this study.

Phospholipid Content. Total phosphorus was used as a measure of
phospholipid content for purified cytochrome oxidase. The various enzyme
preparations exhibited a wide range of values (Table 2); the beef heart
enzyme prepared by traditional methods has a high lipid content, while the
rat liver enzymes prepared in lauryl maltoside contain very little associated
phospholipid. As noted by Thompson and Ferguson-Miller (1983),
hydroxylapatite chromatography in lauryl maltoside is especially effective
in delipidating the enzyme. When purification of rat liver oxidase included
four cycles of hydroxylapatite chromatography, the resulting enzyme

contained sub-stoichiometric amounts of associated phospholipid.

Effect of Phospholipid Content on Electron Transfer Activity. As shown in
Figure 6, the subunit III-containing beef heart enzyme exhibits the biphasic

kinetics characteristic of the reaction with cytochrome c when assayed
under these conditions (Brautigan et al., 1978). Activities were measured
polarographically in two buffer systems. In 25 mM Tris-cacodylate at pH
7.9, the high affinity phase is predominant, while in 50 mM K phosphate at
pH 6.5, the low affinity phase is predominant. A similar kinetic pattern was
observed for the III-depleted rat liver oxidase under the same conditions
(Thompson and Ferguson-Miller, 1983).

It has been reported that as the 2-3 tightly bound DPG molecules are
removed from purified cytochrome orddase, overall electron transfer activity

decreases (Awasthi et al., 1971; Fry and Green, 1980; Robinson, 1982;

42

can comaeona .ommp ..Hm no concanom ”com. .comco new ham
can» mcuasnum occupaauqonq cououquHou adcmqn.cou noumaaoamo A way oumufixo oaonzooumo Hoe: Lon can Hoe: desaxmz

.Emccumcmnmm .x up coscougoa nodumcmaoga was» sou zooms nzcosanoza a
.Amom_ .coddazuconamcma

oao as udaaaozanona canon massed» manganese on»

 

 

 

o
.noczoooocm ampcoeficoaxm ca wonficomoc no pocuagouou pcoucoo ascocnnona Hmuop u
.muoaa oopnuozuoqumm cg co>coupo nommza o
.o cannum an confluence
an Am.e may oumzanosn x :5 om ca voeaounoa «manna no» nuoaa oduocqx song voumaaoaoo Axmezpv ao>o=a=u Husuxmz a
.nocsooooaa Hmpcosficoqu c“ vengeance mm pondscouou ucoucoo HHH aficsosm a
04mm omcmnoxo coficm can»
0cm azamgmoumeogno ouuumamamxocoac m.o 0.0 onmmcqfim om: mo Lo>un
uo aoaoao gsou an coxoaaoa umm z
Aawmpv conaaozh o» unaccooow
vocmaoca onwvaxo mansaom zopzco=
04am omcmcxo coucm
up nozoaaou AmomPV Loaaqznconsmcom m.~ =.m ofimmnqfim om: no Lo>fin
new conaeoza o» unqusooom seconded pom m
onmufixo oouoaaou HHH aficsnam
04am omcmzoxo codes an carcaaou m m.m owmmcafim co: woo Lo>~n
Aaompv conaaona o» mcfiucooom pmm m
conmaoga onouqxo manage» =mczco=
can» now: Hosanna on» usonuuz pan .1 gem ohmmcafim com «00— acme:
Aamm_v .Ha no stream 0» mcuoroooa doom _
ceaumnmaona no cozuoz ommmyoao ummmaos: onofiuocfix nApunv m.o :n mHHH a oocsom coca
Hmsfixmz \aHoec mwmco>< um xmszh saunas oomufixo
acopcou cwqdaonanoca Louncmga cocuooam
pzmezou aHquomamoma 02¢
.whH>Hao< mmmmz<ma zompomnm .ezmezou HHH szamsm "mona<m<mmmm mm<ono m.m:om:uoe»o um mqm<p

43

FIGURE 6: Eadie-Hofstee plots of the kinetics of cytochrome c oxidase
with cytochrome c. Turnover numbers (TN) for subunit III-
containing cytochrome oxidase from beef heart (0.026 nmol)
were calculated from rates of oxygen consumption at 25°C in
1.75 mL of 25 mM Tris-cacodylate, pH 7.9 (0) or 50 mM K
phosphate, pH 6.5 (A) in the presence of 2.8 mM ascorbate
and 0.56 mM TMPD. The range of cytochrome c
concentrations was from 0.015 to 26.5 11M.

44

 

500

400

300*

IpM" 8 "I

 

IN/ICyI S]
3
O
I

100”

 

 

 

100 200 300 400

TN (:4)

Figure 6.

500

45

Robinson and Wiginton, 1985) and that, in particular, the low affinity phase
of the kinetics is lost (Vik et al., 1981; Nalecz et al., 1985). Furthermore, it
has been shown that both activity (V ik et al., 1981; Robinson and Wiginton,
1985) and biphasic kinetics (Nalecz et al., 1985) are restored upon addition
of DPG. Based upon these observations, it was proposed that cytochrome
oxidase had a specific requirement for 2-3 molecules of DPG per orddase (eg.
Vik et al., 1981), perhaps involved in the enzyme-substrate complex and/or
in stabilization of the active conformation. However, the kinetic data of
Thompson and Ferguson-Miller (1983) indicate that no more than one DPG
molecule per oxidase is essential to electron transfer activity.

Subunit III-depleted rat liver cytochrome oxidase was delipidated to
various extents, with average maximal DPG contents of 3, 1.2 or 0.3
molecules per oxidase (Table 2). In each case the kinetics of the reaction
with cytochrome c were evaluated under conditions in which the low affinity
phase is emphasized. As shown in Figure 7, a comparison of the results for
the preparation containing 3 DPG molecules and the preparation containing
an average of 0.3 DPG molecules per oxidase reveals no significant
differences in either maximal activity or the nature of the kinetic plots. The
kinetics of the preparation with the intermediate lipid content were nearly
identical (data not shown). The phospholipid content and the maximal
activity for each of the enzymes used in this study are listed in Table 2.

46

FIGURE 7: Eadie-Hofstee plots of the kinetics of delipidated cytochrome
oxidase with cytochrome c. Activities of rat liver cytochrome
oxidase containing 5.9 nmol P per nmol am, (A) or containing
0.6 nmol P per nmol am; (0) were assayed in 50 mM K
phosphate (pH 6.5) as described in Figure 6. Cytochrome c
concentrations ranged from 0.015 to 34 uM.

47

 

 

 

120‘
I
b
80 '5
)-
O
40- \
\
\
.- \ A
K
\
o l 4 l 1 J 1 \
O 100 200 300 400
IN (3")

Figure 7.

 

48

DISCUSSION

Cytochrome oxidase can be isolated from rat liver mitochondria using
solubilization in lauryl maltoside in conjunction with cytochrome c affinity
chromatography to yield a subunit III-depleted enzyme that retains many of
the characteristics of the III-containing enzyme (Thompson and Ferguson-
Miller, 1983; Thompson, 1984). High resolution SDS-PAGE reveals that this
orddase is also depleted of subunits Vb, VIa, VIb and VIIa. Reliable
estimates of subunit III content can be made by SDS-PAGE using a
modified gel system in which separation of subunits II and III is enhanced.
Using additional chromatographic steps, enzymes were prepared which
contained no detectable subunit III.

The conditions used to deplete cytochrome oxidase of subunit IH are
apparently also effective in delipidating the enzyme, whether Triton X-100
(Nalecz et al., 1985; Robinson and Wiginton, 1985) or lauryl maltoside is
used. Thus, these enzyme preparations are also appropriate for evaluating
the role of phospholipid in cytochrome oxidase activity.

Although most of the lipid associated with purified cytochrome
oxidase can be removed quite readily, 2-3 molecules of DPG remain tightly
bound (Awasthi et al., 1971; Robinson et al., 1980; Fry and Green, 1980).
The role of DPG has been controversial due to conflicting results and due to
technical difficulties in defining the phospholipid content of cytochrome
oxidase preparations. When DPG is removed by solvent extraction (Awasthi,
et al., 1971) or by detergent exchange in Triton X- 100 (Robinson et al., 1980;
Fry and Green, 1980; Vik et al., 1981; Nalecz et al., 1985; Robinson and

Wiginton, 1985; Zborowski and Robinson, 1985), significant inactivation is

49

observed. In some studies, the removal of phospholipid was found to
correlate with loss of activity and of interactions with cytochrome c
(Tzagoloff and MacLennan, 1965; Vik et al., 1981; Nalecz et al., 1985). These
alterations were found to be specifically reversed by addition of DPG alone
(Awasthi et al., 1971; Fry and Green, 1980; Vik et al., 1981; Robinson, 1982;
Robinson and Wiginton, 1985; Zborowski and Robinson, 1985) or in a lipid
mixture (Tzagoloff and MacLennan, 1965; Nalecz et al., 1985). These results
suggested that the specific interaction of the enzyme with DPG was
catalytically important for cytochrome oxidase activity (Vik et al., 1981;
Tzagoloff, 1982; Robinson, 1982).

In other conflicting studies, exchange of associated DPG for other
phospholipids (ie. dimyristoyl phosphatidylcholine) in the presence of
cholate did not result in changes in activity or kinetics that could be
attributed to a specific requirement for DPG (Watts et al., 1978; Powell et
al., 1985; Powell and Abramovitch, 1985). However, these results are
inconclusive because enzyme activities were not clearly reported and appear
to represent very low turnover numbers both before and after exchange of
DPG. [Reasonable conversions indicate that full activity ranges from 3 sec-1
(Powell et al., 1985) to 60 sec-1 (Watts et al., 1978) in these studies]

In contrast to the oxidase depleted of lipid by Triton X-100, the
enzyme with sub-stoichiometric amounts of associated lipid prepared in
lauryl maltoside did not exhibit loss of activity or alterations in kinetics.
This provides convincing evidence that DPG is not specifically required for
cytochrome oxidase function, since lauryl maltoside seems able to stabilize
the active form of the enzyme (exhibiting a maximal activity of
approximately 400 sec-1) with considerably less than one DPG per oxidase

available. However, it must be noted that, for an enzyme delipidated in

50

Triton X-100, high activity and biphasic kinetics could not be restored by
exchanging Triton X-100 with lauryl maltoside, while added phospholipid
could reactivate the enzyme (Nalecz et al., 1985). It is possible that only
DPG can successfully displace Triton X- 100 from a few critical (and
inhibitory) binding sites in the enzyme.

Sinjorgo et al. (1987a) demonstrated that as cytochrome oxidase
incorporates into detergent micelles, activity becomes enhanced when using
lauryl maltoside, while it becomes inhibited when using cholate or Triton X-
100. It is clear that cytochrome oxidase activity is sensitive to interactions
between the enzyme and its hydrophobic environment. However, based on
results of the delipidated enzyme prepared in lauryl maltoside, it does not
appear that DPG is a specific, essential component of the environment of
cytochrome oxidase for either interaction with the substrate cytochrome c or
for stabilization of the active enzyme form.

In summary, comparison of the subunit III-containing and III-
depleted oxidases indicates that neither removal of subunits III, Vb, VIa,
VIb and VIIa nor delipidation result in major alterations in the electron
transfer activity under the standard conditions used here. However, roles as
regulatory components that respond to hormonal stimuli or metabolic
conditions cannot be excluded by these studies. The subunit III-depleted
oxidase prepared as described here appeared to be especially suitable for
more detailed examination of the effect of subunit III removal on

cytochrome oxidase functions, as reported in succeeding chapters.

CHAPTER 2

EFFECT OF SUBUNIT III REMOVAL ON THE PROTON
TRANSLOCATING ACTIVITY OF CYTOCHROME c OXIDASE

The proton translocating activity of cytochrome c oxidase, first
demonstrated by Wikstrom (1977), was not predicted in the hypothesis of
chemiosmotic coupling of electron transport to ATP synthesis proposed by
Mitchell (1961, 1966); redox centers capable of carrying both electrons and
protons were postulated to be required for this function, but no such centers
were associated with cytochrome oxidase. Although acidification of the
extramembrane compartment was consistently observed when cytochrome
oxidase incorporated in sealed bilayers was pulsed with reductants or
oxygen, the cause of this phenomenon became a source of controversy. A
scalar release of protons as a result of binding of cytochrome c to orddase or
phospholipids was suggested to be an artifactual source of protons (Moyle
and Mitchell, 1978; Papa, 1982; Mitchell and Moyle, 1983). However, the
acidification is now generally accepted (Mitchell et al., 1985) to be the result
of vectorial translocation of protons, as concluded in many previous studies
(Wikstrom, 1977; Wikstrom and Krab, 1978; Sigel and Carafoli, 1980;
Wikstrom and Penttila, 1982; Casey and Azzi, 1983; Wikstrom, 1984;
Wikstrom and Casey, 1985).

Research is now focused on investigating the chemical mechanism of
this reaction and several models have been proposed. Consistent with

chemiosmotic principles, Mitchell and coworkers have proposed "redox loop"

51

52

mechanisms invoking the 0201202 couple (Mitchell et al., 1985) or the OH-
/HzO couple (Mitchell, 1987) as proton carriers. However, a proton pumping
mechanism involving the protein as an integral part of the proton
conducting process has been more commonly proposed. A general feature of
these models is that an oxidation-reduction event causes a localized or
overall change in protein structure resulting in alternating proton input
and output states [see review by Krab and Wikstrom (1987)]. Coupling of
the reduction and protonation events may be direct, with protonation
occuring adjacent to the redox center, or indirect, with protonation occuring
at a distance via a protein conformational change and, in either case,
involving a proton conducting pathway through the protein.

Various redox centers have been implicated in the redox events
linking proton and electron transfers. Models involving heme a have been
proposed based on the studies of Babcock and Callahan (1983; Callahan and
Babcock, 1983) showing that the strength of a hydrogen bond, between the
formyl group of heme a and a protein residue, depends upon the redox state
of heme a. A role for this center in proton pumping is supported by the fact
that the redox potential of heme a is dependent on pH (van Gelder et al.,
1977; Artzatbanov et al., 1978) and on the energy state of the mitochondrion
(Hinkle and Mitchell, 1970).

Advocates of models involving the Cut, redox center have pointed to
its unusual spectral characteristics compared to other electron-carrying
copper centers, and its proposed capacity for exchanging ligands (e.g. a
cysteine and a tyrosine) during the oxidation/reduction cycle, indicating
possibilities for involvement in proton pumping (Gelles et al., 1986; Blair et

al., 1986; Martin et al., 1988). However, it has been observed that reduction

53

of CuA is independent of pH (Wikstrom et al., 1981a) and the energy state of
the mitochondrion (Ereckinska et al., 1971; Rich et al., 1988).

Malmstrom (1985) has proposed a kinetic transition state mechanism
in which both heme a and CuA are involved such that when both centers are
reduced a conformational change occurs, representing a transition between
proton input and output states (Scholes and Malmstrom, 1986; Brzezinski
et al., 1987). Another conformational mechanism was proposed by Wikstrom
and coworkers (Wikstrom et al., 1981b; Finel and Wikstrom, 1986) in which
the input/output conversion would correspond to cycling of an interaction
between two monomers of a dimer.

Wikstrom (1989) has recently offered a model that depends upon
cooperation of two redox centers, heme a and CuB. In such a mechanism,
proton transport is facilitated by an induced dipole caused by the reduction
of CuB.

Although subunits I and II are likely to be involved in proton
pumping because of their probable role in binding the redox centers, any
involvement of the other subunits has yet to be determined. As reviewed in
the Introduction to this thesis, subunit III has been implicated in the
translocating mechanism, principally by two lines of evidence [see
Prochaska and Fink (1987)]. Firstly, some studies have suggested that
proton pumping ability is lost upon removal of the peptide from the complex
(Penttila, 1983; Thelen et al., 1985b; N alecz et al., 1985; Prochaska and
Reynolds, 1986). Secondly, chemical modification studies show that DCCD
causes loss in proton pumping under conditions in which primarily subunit
III is covalently labelled (Casey et al., 1979b; Casey et al., 1980; Prochaska
et al., 1981). Due to the indefinite nature of the models proposed thus far,

subunit III could be postulated to fill any of several roles: (1) as the location

54

of a proton binding site(s) involved in linking proton translocation and
electron transfer (i.e. Penttila and Wikstrom, 1981; Malmstrom, 1985); (2)
as a component involved in structural transitions such as association of
monomers into dimers (i.e. Wikstrom and Finel, 1986); and/or (3) as an
active or passive proton conducting channel (ie. Casey et al., 1980).

The validity of studies demonstrating the retention of proton
translocating activity upon depletion of subunit III (Thompson et al., 1985;
Puettner et al., 1985; Finel and Wikstrom, 1986) has been questioned. It has
been argued (Finel and Wikstrom, 1988) that the proton pumping III-
deficient enzyme prepared with a—chymotrypsin (Puettner et al., 1985) may
contain subunit III fragments that remain associated with the complex. In
addition, Muller et al. (1987) contend that the III-depleted oxidase can
extrude protons only during the first turnover (ie. the complete reduction of
one oxygen molecule) but is incapable of sustained pumping activity due to
the lack of an essential component.

This chapter reports further studies on the effect of subunit III
removal on proton translocation by cytochrome oxidase. Subunit HI-
depleted oxidase was prepared by hydroxylapatite and cytochrome c affinity
chromatogaphy in lauryl maltoside (Thompson and Ferguson-Miller, 1983)
with an additional step of anion exchange FPLC where indicated. After
reconstitution, cytochrome oxidase vesicles containing either the complete
or the III-depleted enzyme were evaluated and compared in terms of ability
to exhibit respiratory control and to translocate protons. The pumping
activity was also examined over a range of turnover conditions to determine
whether apparent proton translocation was maintained through multiple

turnovers of enzyme.

55

EXPERIMENTAL PROCEDURES

Materials, Asolectin (Associated Concentrates, Long Island, NY) was
repurified as described by Sone et al. (1977) using two precipitation cycles
and dithiothreitol rather than butyrated hydroxytoluene as the reducing
agent. Cytochrome c was repurified as described in Experimental
Procedures of Chapter 1. Cholic acid (U. S. Biochemical Corp., Cleveland,
OH) was recrystallized 3 times from hot 95% ethanol. Titrated 0.1N HCl
and 0.1N N aOH were obtained from J. T. Baker Chemical Co. (Phillipsburg,
NJ). Carbonyl cyanide m-chlorophenylhydrazone (CCCP) and valinomycin
were obtained from Sigma (St. Louis, MO) and stock solutions were

prepared in ethanol.

Cytochrome Oxidase Preparations Subunit III-containing cytochrome
oxidase was prepared from beef heart according to Suarez et a1. (1984)
omitting the ethanol wash step. Subunit III-depleted oxidase was prepared
from rat liver as described by Thompson and Ferguson-Miller (1983). An
enzyme containing no detectable subunit III was obtained after an
additional step of anion exchange FPLC as described in Experimental
Procedures of Chapter 1.

Baconstitution. Cytochrome oxidase was reconstituted into phospholipid
vesicles by using a modification of the cholate dialysis procedure described
by Casey et al. (1979). Asolectin (40 mg/mL) was sonicated for 1 min/mL in
2% cholate and 75 mM K-Hepes (pH 7.4). Oxidase was added at a ratio of
0.1 nmol of enzyme/mg phospholipid and the suspension was dialyzed

56

against various solutions of constant [K4] and osmolarity as follows: 4 hrs
versus 100 volumes of 75 mM K-Hepes (pH 7.4), 14 mM KCl, 0.1% (w/v) Na
cholate; 16 hrs versus 250 volumes of 75 mM K—Hepes (pH 7.4), 14 mM KCl;
8 hrs versus 400 volumes of 50 mM K-Hepes (pH 7.4), 24 mM KCl, 15 mM
sucrose; 16 hrs versus 500 volumes of 10 mM K-Hepes (pH 7.4), 41 mM KCl,
38 mM sucrose.

After reconstitution, the fraction of oxidase accessible to externally
added cytochrome c was determined. Cytochrome oxidase vesicles were
suspended in 50 mM K phosphate, pH 7.4 (approximately 0.7 11M M3) and
valinomycin (4 11M) and CCCP (20 M) were added. The amount of oxidase
reduced by 30 uM cytochrome c in the presence of 20 mM ascorbate was
measured (by absorbance at 605 nm) after anaerobiosis. Total reduction was
determined following addition of the membrane permeable reductant TMPD
(2.3 mM).

Raspiratory Control Asjsays Respiratory control of cytochrome oxidase
vesicles was determined by polarographic measurement of the rate of
oxygen consumption in the absence and presence of ionophores that
equilibrate the generated proton electrochemical gradient. Reconstituted
oxidase (10-20 nM final [aa3]) was added to 1.75 mL of 10 mM K-Hepes (pH
7.4), 41 mM KCl, 38 mM sucrose, 5.8 mM ascorbate and 0.28 mM TMPD to
determine controlled activity. Uncontrolled activity was measured after

addition of 1.1 11M valinomycin and 5.7 11M CCCP.

Proton Translocation Assays. Immediately before use, cytochrome c,
prepared as described above, was re-reduced with dithionite and

rechromatographed on a Sephadex G-15 column in 1 mM K—Hepes (pH 7 .4)

57

and 45 mM KCl to attain reduction levels of 97% or greater and to lower the
buffering capacity. Proton translocating activity at 22°C was observed by
monitoring the pH of the extravesicular solution using a combination
electrode (GK2321C) and pH meter (PHM 84) from Radiometer America.
Assay conditions are listed in the legends of Figures 9 and 10. Before assays
were performed on the oxidase vesicles, the pH of the solution of reduced
cytochrome c was adjusted with HCl or NaOH so that an injection of the
substrate into an assay mixture containing vesicles without incorporated

oxidase caused no deflection in the pH tracings.

58

RESULTS

Respiratory Control. Subunit III-containing and subunit III-depleted
cytochrome oxidases were incorporated into phospholipid vesicles using a
cholate dialysis method of reconstitution (Hinkle et al., 1972; Casey et al.,
1979a). In both cases the majority of the enzyme molecules (80-90%) were
oriented as they are in intact mitochondria, with the cytoplasmic domain
(i.e. the cytochrome c binding site) on the outside of the vesicle. Similar
results have been reported for cytochrome oxidase vesicles prepared by this
method (Casey et al., 1982; Zhang et al., 1984; Thelen et al., 1985a).

As shown by Hinkle et al. (1972), the transmembrane proton
electrochemical gradient generated by cytochrome oxidase activity inhibits
enzyme turnover in a process known as respiratory control. Ifthe gradient
is abolished by the addition of ionophores which carry ions across the
membrane, this inhibition is released. In the inhibited state, activity is
described as "controlled" or "coupled" and when the gradient is equilibrated,
is considered "uncontrolled" or "uncoupled". Figure 8 illustrates a typical
respiratory control assay in which steady-state cytochrome oxidase activity
is measured as the rate of oxygen consumption in the presence and absence
of the electrochemical gradient (i.e. before and after addition of CCCP and
valinomycin). The respiratory control ratio (RCR) is calculated by dividing
the uncoupled activity by the coupled activity. In reconstituted vesicles, the
RCR at once indicates whether the enzyme has integrated into the lipid
bilayer to form a sealed system and whether the enzyme is capable of

generating and responding to a proton gradient.

59

FIGURE 8: Demonstration of respiratory control in cytochrome oxidase
vesicles. The respirato control ratio (RCR) of vesicles
containin subunit III- epleted cytochrome oxidase was
calculate as a ratio of the uncontrolled to controlled steady-
state activit (i.e. in the presence and absence of the
ionophores CCP and valinomycin). Conditions are listed in
Experimental Procedures.

60

 

 

 

 

 

1m
1
3
7. .15
T nEOn—uqoao m.
l
A
ToxEOmo Swan—om 9 l
T253 ._
R 1
Tamnonvoqo C
R
r
_ P O
0 O
O 0 0
2 I

_E\~O .08:

MINUTES

Figure 8.

61

The subunit III-containing cytochrome oxidase (from beef heart)
reconstituted to give good RCR’s with an average value of 5.9 and a
standard deviation of 0.7 for 21 determinations (on different vesicle
preparations). The III-depleted oxidase (from rat liver) consistently
exhibited excellent respiratory control having an average RCR of 13.0 and a
standard deviation of 3.1 for 21 determinations. The degree of respiratory
control exhibited by the III-depleted oxidase was consistently, and as much
as, three-fold higher than that of the III-containing oxidase. Finel and
Wikstrom (1986) report that III-deficient enzyme prepared in LDAO also
exhibits ratios higher than the III-containing oxidase by as much as 150%
with upper values of 7.5. In contrast, III-depleted oxidase prepared using
Triton X-100 typically demonstrates lower RCR’s than the HI-containing
enzyme (Penttila, 1983; Thelen et al., 1985b; Prochaska and Reynolds,
1986), indicating that Triton X-100 treatment impairs the enzyme’s ability
to reconstitute as effectively.

Puettner et al. (1985) reported that cytochrome oxidase vesicles must
exhibit RCR’s of at least 3 to detect proton translocating activity. Vesicles
containing the complete and III-depleted oxidases used in this study exceed
this requirement to yield well coupled systems appropriate for proton

translocating assays.

Proton Translocating Activity. The many methods used to detect the proton
translocating activity of cytochrome oxidase have at least two common
features. First, in order to detect translocating activity, the assay system
must include a means of equilibrating the enzyme-generated electrical
gradient because otherwise the enzyme is inhibited too rapidly to detect
proton extrusion [see Sigel and Carafoli (1980), Proteau et al. (1983) and

62

Casey et al. (1984)]. This is usually accomplished (ie. Wikstrom, 1977;
Wikstrom and Saari, 1977; Wikstrom and Krab, 1978; Casey et al., 1979a)
by performing assays in a potassium salt solution in the presence of
valinomycin, an ionophore that transports potassium ions across lipid
bilayers. However, lipophilic cations such as tetraphenyl phosphonium ions
can also function as the charge-compensating system (Sigel and Carafoli,
1980; Casey et al., 1984).

Second, proton pumping activity has not been observed under steady-
state conditions (Papa et al., 1985; Hill et al., 1985), but is instead detected
as a transient acidification upon initiation of cytochrome oxidase activity by
a "pulse" of substrate. The substrate pulse may be: (1) a reductant such as
ferrocyanide (Wikstrom, 1977; Wikstrom and Saari, 1977; Wikstrom, 1984)
or TMPD (Wikstrom and Saari, 1977 ) added to an oxidized cytochrome
c/cytochrome oxidase test system; (2) oxygen added to a reduced, anaerobic
cytochrome c/cytochrome oxidase test system (Wikstrom and Pentilla, 1982;
West et al., 1986); or (3) reduced cytochrome c added to oxidized cytochrome
oxidase (Krab and Wikstrom, 1978; Casey et al., 1979a; Proteau et al., 1983;
Casey et al., 1984; Thelen et al., 1985a).

Proton translocating activity can be observed indirectly by measuring
the uptake of cations (ie. potassium ions in the presence of valinomycin)
that occurs in response to extrusion of protons (Sigel and Carafoli, 1980;
Penttila, 1983; Casey et al., 1984), but is most often assayed by directly
measuring pH changes. This can be done using a pH sensitive electrode
(Proteau et al., 1983; Thelen et al., 1985a; Prochaska and Reynolds, 1986), a
pH-sensitive dye such as phenol red (Casey et al., 1979a; Casey et al., 1984;
Wikstrom, 1984), or a pH-sensitive fluorescent phospholipid (Thelen et al.,
1984; Thelen et al., 1985b).

63

Proton pumping activity is expressed as the H+/e- ratio, the number of
protons translocated (as measured) per number of electrons transferred
(calculated from the amount of substrate added).

The proton translocating activity of reconstituted subunit III-
containing and subunit III-depleted oxidase was assayed by initiating
activity with a pulse of reduced cytochrome c and measuring the transient
acidification of the extravesicular solution using a pH electrode. Figure 9
shows typical assays for oxidase containing 100% subunit III (Fig. 9C),
oxidase containing 2% subunit III (Fig. 9B), and oxidase containing no
detectable subunit III (Fig. 9A). Acidification (an upward inflection),
representing proton translocating activity, was induced by injection of
reduced cytochrome c in the presence of valinomycin. Alkalinization (a
downward inflection), representing proton consumption in the reduction of
oxygen, was induced by a similar injection of reduced cytochrome c when
both valinomycin and high concentrations of CCCP were present. The H+/e-
ratios were calculated according to the acidification observed per number of
reducing equivalents injected and corrected for any quantitative deviation
observed in the alkalinization induced by cytochrome c in the uncoupled
state. Details of the assay are described in the legend to Figure 9. Upon
repeated measurement, the average H+/e- ratio for the III-containing
oxidase was 0.8 with a standard deviation of 0.1 for 12 determinations. The
III-depleted orddase exhibited an average H+/e- ratio of 0.4 with a standard
deviation of 0.1 for 8 assays. Thus, the III-depleted oxidase, including a
preparation with no detectable subunit III (Fig. 9A), exhibits 50% of the
activity of the III-containing oxidase, suggesting that it retains significant

proton translocating capability under these conditions.

FIGURE 9: Cytochrome c-induced proton translocation by cytochrome c
oxidase vesicles. (A) Rat liver oxidase containing 0% subunit
III; (B) rat liver oxidase containin 2 % subunit III; (C) beef
heart oxidase containing 100 % su unit III. Cytochrome
oxidase vesicles [0.1 mL of 2.5—4 uM aa3 in 10 mM Hepes (pH
7.4), 41 mM KCl, 38 mM sucrose] were added to 4 mL of 0.05
mM K-Hepes( H 7.4), 45 mM KC], 50 mM sucrose and 2.5
mM MgC12. Va inomycin (2 11M) and CCCP (0.25 nM) were
added to equilibrate the membrane potential. Acidification
due to proton translocation was induced by the addition of
ferrocytochrome c (c2+) in the absence of uncou ler, while a
similar addition of c2+ in the presence of 5 11M CCP resulted
in alkalinization due to roton consumption. pH changes were
calibrated by 5 nmol ad 'tions of HCl for the acidification
phase and 5 nmol additions of N aOH for the alkalinization
phase. Each standard was injected four times and the average
used for calculations; acid (solid arrows) and base (dashed
arrows). The inflections observed in the presence of 5 11M
CCCP were smaller due to the accessibility of the
intravesicular buffer. Expected levels of alkalinization
(designated by dotted lines) were used to correct the
acidification phase. If expected alkalinization was greater
than the observed alkalinization, the difference was
subtracted from the acidification phase and vice versa. H+/e-
ratios were calculated from the corrected extent of

acidification per number of reducing equivalents added via
c2+.

65

 

     
 

5“ ""10. S. 2 "’ 5.4 "MOI
\ S 2* Sumo!
\ NOOH

\ H7€=05

 

 

4.1 Turnovers

 

 

 

5.4 nmol B
.c. 2 *

Sumol

\ NaOH

\

    

5.4 nmol 5 2 *
\

      

  

 

HI/e'=0.35

 

5 nmol W
HCI

3.6 Turnovers

 

 

 

 

 

Anmol S 2+ 1
\ CCCP

 

HI/e'= 0.7

. fl 3.7 Turnovers
Snmol

HCI

 

 

 

 

 

 

 

Figure 9.

 

66

Vectorial proton translocation is expected to be more sensitive to
inhibitory transmembrane gradients and less sensitive to multiple turnover
conditions than would scalar proton release caused by the binding of
cytochrome c to the oxidase (see below). The response of the subunit III-
depleted enzyme to alterations in these parameters was characterized and
compared to the response of the HI-containing oxidase to determine the
nature of the acidification observed.

To detect proton pumping activity in cytochrome oxidase vesicles, a
charge-compensating system is required to reduce build-up of the inhibitory
transmembrane potential (i.e. Proteau et al., 1983). Thelen et al. (1985a)
observed that, in addition to the potassium/valinomycin system, a small
amount of a protonophore can further reduce the effect of the generated
gradient and optimize the acidification phase. Figure 10A shows the
relationship of the H+/e- ratio versus the concentration of the protonophore
CCCP added in the acidification phase of the proton pumping assays for III-
depleted and III-containing oxidases. In both cases, a small amount of
CCCP (0.25 nM) improved the acidification response by about 25-30%. At
higher concentrations of CCCP, the H+/e- ratios decrease since the proton
gradient is dissipated more rapidly than it is produced. The response of the
III-depleted oxidase to the inhibitory gradient and its similarity to the
behavior of the III-containing oxidase, indicates that the enzyme catalyzes
vectorial proton translocation.

It has been observed (Casey et al., 1979a; Solioz et al., 1982) that the
H+/e- ratios exhibited by dialyzed preparations of cytochrome oxidase
vesicles decrease gradually as the enzyme is subjected to more turnovers
per assay. (One turnover equals one oxygen molecule reduced per oxidase,

i.e. input of 4 electrons.) This has been attributed to high capacitance, in

67

FIGURE 10: Dependence of Ht/e- ratios on (A) charge compensation and
(B) turnovers per assay. (A) H+/e- ratios were determined as
described in Figure 9 except that the concentration of CCCP
in the acidification phase was varied as indicated. Subunit
III-containing oxidase from beef heart at 5.2 turnovers per
assay, A; subunit III-depleted oxidase from rat liver at 4.2
turnovers per assay, 0. (B) H+/e- ratios were determined as
described in Figure 9 in the absence (open symbols) or the
presence (closed symbols) of 0.25 nM CCCP in the
acidification phase. Ferrocytochrome c was injected in
amounts to catalyze a specific number of enzyme turnovers as
indicated. Subunit III-containing oxidase from beef heart (A,
A); subunit III-depleted oxidase from rat liver (0 , o).

68

 

 

 

 

 

 

 

 

 

 

0.2
o I l I L
0 1.0 2.0
[cccr] , nM
1.0
B
0.8
0.6
°" 5%
0.2
o l l I l
0 4 s s 10

Tu rnovers/ g: 3

Figure 10.

69

part due to the small vesicle size, which leads to generation of a larger
transmembrane gradient with higher turnovers (W rigglesworth, 1985).
Thus, vectorial proton translocations would be expected to decrease
gradually with increasing turnovers, while any acidification caused by
scalar proton release during the initial substrate-enzyme interaction would
be expected to decrease much more dramatically with increased turnovers.
The H+/e- ratios of the III-containing and III-depleted enzymes were assayed
over a range of turnovers in the absence and presence of an optimal amount
of CCCP (Fig 10B). Both enzymes exhibited only gradual decreases with
increasing turnovers and H+/e- ratios were improved over the entire range
in the presence of 0.25 nM CCCP. In contrast to Muller et al. (1987), who
concluded that their III-depleted oxidase was incapable of vectorial
translocation on the basis of a steep decline in H+/e- with increasing
turnovers, the III-depleted oxidase prepared in lauryl maltoside behaves in

a manner consistent with a proton pumping enzyme.

70

DISCUSSION

Vesicles containing subunit III-depleted cytochrome oxidase, that has
been prepared in lauryl maltoside according to the procedure of Thompson
and Ferguson-Miller (1983), produce substrate-induced acidification of the
extravesicular solution. The sensitivity of the acidification to the
transmembrane potential and multiple turnover conditions is consistent
with catalysis of vectorial proton translocations. Thus, in contrast to reports
in which removal of subunit III results in the loss of translocating activity
(see Table 1 in the Introduction to this thesis), this preparation retains
activity, indicating that subunit III is not essential for this function.
However, since the H+/e- ratios for the III-depleted enzyme are lower than
those exhibited by the III-containing enzyme, it is possible that subunit III
has a facilitative role in proton translocation, but equally likely that it has
no role and the proton pumping mechanism is inadvertantly disrupted by
the conditions used to remove subunit III. Since several of the smaller
peptides are removed along with subunit III, their role must also be
considered. DiBiase and Prochaska (1985) were able to remove two of these
small peptides alone with trypsin treatment and found that proton
translocation was unaffected.

Interpretation of the significance of reduced H+/e- ratios for the III-
depleted oxidase is further complicated by the uncertainty concerning the
actual stoichiometry of the translocation reaction due to technical
difficulties in measuring small changes in proton concentration. Many
investigators have reported proton-to-electron stoichiometries approaching

1 in mitochondria and reconstituted systems (Wikstrom and Krab, 1979;

71

Brand et al.,1978; Sigel and Carafoli, 1979; Sarti et al., 1985; Moody et al.,
1987). However, others obtain ratios as high as 2 in the native
mitochondrial membranes (or 4 H+/O; Azzone et al., 1979; Reynafarje et al.,
1982; Lehninger et al., 1985) and contend that lower ratios are obtained due
to inhibition of proton extrusion because of the high degree of coupling in
the test system (Azzone et al., 1985). Lehninger et al. (1985) report that
their results may suggest that cytochrome oxidase has two proton pumps,
one of which is inactivated during purification procedures. Nevertheless, by
comparison with the III-containing enzyme, under the conditions used in
this study, oxidase depleted of subunit III appears to pump protons at 50%
efficiency.

These results are in agreement with studies in which the apparent
stiochiometry of the enzyme from Paracoccus denitrificans in native
membranes, where the oxidase presumably contains subunit III, is 0.93
H+/e- (Van Verseveld et al., 1981), while the III-less purified enzyme
exhibits H+/e- ratios of 0.6 (Puettner et al., 1983). Malatesta et al. (1986)
also observe that III-deficient oxidase translocates protons at 0.5 H+/e-.
Through transient kinetic studies, they found that the initial reduction and
subsequent reoxidation of heme a were qualitatively similar with and
without III and propose that subunit III is involved in efficiently coupling
the redox reaction to proton pumping. Using a III-depleted enzyme prepared
by FPLC in LDAO, Finel and Wikstrom (1986) also observed retention of
pumping activity with H+/e- ratios of 0.4 compared to 0.9 for the III-
containing oxidase. (Indeed, oxidase that had been through the column and
retained subunit III showed H+/e- ratios of 0.4.) They correlate these
apparent stoichiometries to the percentage of dimer found following

extraction of the enzyme from the vesicles and propose that subunit III’s

72

role in proton pumping is in stabilizing the dimer. However, the importance
of the dimeric state in proton translocation is by no means clear.
Furthermore, there is no evidence to indicate that the monomerzdimer ratio
of oxidase extracted from vesicles using cholate accurately reflects the ratio
when the enzyme is in the liposomes. Thus, until the aggregation state of
the reconstituted enzyme is determined, no meaningful correlation between
dimer formation and translocating ability can be made.

Other roles for subunit III in proton translocation must also be
considered. The electron transfer activity of oxidase is regulated by pH and
the transmembrane electrochemical potential. Part of this control is
mediated by proton binding sites on the protein, one or more of which may
be critical for linking electron transfer to proton translocation. The following
chapters examine the possibility that subunit III could contain a proton
binding site(s) involved in the regulation of activity by pH (Chapter 3) or by
the proton electrochemical gradient (Chapter 4).

CHAPTER 3

THE EFFECT OF SUBUNIT III REMOVAL ON CONTROL
OF CYTOCHROME c OXIDASE ACTIVITY BY pH

Hydrogen ions participate directly in the catalytic functions of
cytochrome c oxidase. Protons are consumed in the reduction of dioxygen to
water, while in a coupled reaction the enzyme translocates protons across
the membrane in which it is incorporated (Wikstrom, 1977). In order to
clarify the mechanism of these reactions it is important to define the nature
and extent of proton modulation of the enzyme’s measurable parameters.
Steady-state electron transfer activity is highly pH-dependent, exhibiting
increasing turnover rates as pH decreases (Wilms et al.,1980). The
availability of hydrogen ions for the reduction of oxygen intermediates in
highly buffered solutions does not appear to be rate-limiting (Lindskog and
Coleman, 1973). However, analysis of steady-state kinetic data suggests
that proton binding steps may be required for intramolecular electron
transfer, and in turn, may be involved in the proton translocation
mechanism (Thornstrom et al., 1984). The location of the proton binding
sites that mediate this pH-dependent behavior has not been determined.

Subunit III has been implicated in proton translocation, as previously
discussed, and in regulation of intramolecular electron transfer by pH
(Moroney et al., 1984), activities that may be mechanistically related. The
role of this peptide in the former process appears to be facilitative rather

than essential, based on retention of up to 50% of proton translocating

73

74

activity following its removal (see Chapter 2; Thompson et al., 1985;
Puettner et al., 1985; Finel and Wikstrom, 1986). It is possible that subunit
III provides a link necessary for the efficient coupling of electron transfer to
proton translocation, such as a proton binding site or a proton conducting
channel. Indeed, studies by Moroney et al. (1984) suggest that covalent
modification of subunit III by DCCD not only inhibits proton translocation
but also interferes with a proton binding site involved in control of
intramolecular electron transfer.

The effects of subunit III depletion on pH-modulated characteristics
of cytochrome oxidase were examined to investigate the possibility that
subunit III is involved in regulation of activity by pH. In this chapter, the
pH-dependence of steady-state kinetic parameters, absorbance spectra and
DCCD inhibition of electron transfer are compared for the III-depleted and

III-containing oxidases.

75

EXPERIMENTAL PROCEDURES

Matarials. Cytochrome c and asolectin were repurified as described in
Chapters 1 and 2, respectively. DCCD was obtained from Pierce Chemical
Co. (Rockford, IL). Sources of all other materials were listed in previous
chapters.

Cytochrome Oxidase Preparations. Subunit III-containing cytochrome
oxidase was prepared from beef heart according to Suarez et al. (1984)
omitting the ethanol wash step. This preparation is considered the control
enzyme in the following studies. Subunit III-depleted cytochrome oxidase
(less than 2% subunit III) was prepared from rat liver using cytochrome c
affinity chromatography in lauryl maltoside as described by Thompson and
Ferguson-Miller (1983).

SDS-PAGE. Subunit III content was evaluated by SDS-PAGE using
Procedure 2 as described in Experimental Procedures of Chapter 1.

Reconstitution. Cytochrome oxidase was reconstituted into phospholipid
vesicles as described in Experimental Procedures of Chapter 2 except that
the final dialysis buffer was 25 mM potassium Hepes, pH 7.4, 40 mM KCl,
20 mM sucrose for the kinetic studies. After reconstitution, the fraction of
oxidase accessible to cytochrome c was determined as described in
Experimental Procedures of Chapter 2 and enzyme activities were corrected

accordingly.

76

pH Dependence of Cflochrome Oxidase Astivity. The uncoupled electron
transfer activities of the reconstituted enzymes (containing 1.1 uM

valinomycin and 5.7 uM CCCP) were measured using a Gilson oxygraph
(Model 5\6H; Middleton, WI) at 25°C. The high ionic strength (I=0.55M)
conditions used by Thornstrom et al. (1984) were modified for polarographic
analysis by adding ascorbate (2.9 mM) and TMPD (0.6 mM) to maintain
complete reduction. All buffers were prepared so that ionic strength was
constant over the entire pH range. Before use, repurified cytochrome c was
reduced with sodium dithionite and chromatographed on a Sephadex G-50-
40 (Pharmacia, Uppsala, Sweden) column in 50 mM K2804, 10 mM MES,
pH 6.0, for the pH 5.0-6.5 assays or in 50 mM K2804, 10 mM Hepes, pH 7.6,
for the pH 7.0-8.5 assays.

Spectrophotometric assays of electron transfer activity were
performed using the conditions of Thornstrom and coworkers using a 0.5 cm
pathlength cell with a Perkin-Elmer Lambda 4B spectrophotometer. After
addition of cytochrome oxidase vesicles, enzyme concentration was 5.5 nM
in the pH 7.5 assays and 9.3 nM in the pH 8.5 assays. Initial rates were
calculated using the integral method described by Yonetani and Ray (1965)
and then divided by enzyme concentration to give turnover rates.

Kinetic parameters and their standard deviations were estimated by
linear regression analysis with the computer program WILMAN4 (Brooks
and Suelter, 1986) using the intermediate weighting option (Cornish-
Bowden, 1976).

77

RESULTS

pH Dependence of Kinetic Parameters of Reconstituted Cytochrome Oxidase
With and Without Subunit III. The effect of pH on the kinetics of steady-
state electron transfer activity in reconstituted cytochrome oxidase with and
without subunit III was studied at high ionic strength (I =0.5M) in the
presence of uncouplers, conditions similar to those described by Thomstrom
et al. (1984). Reconstitution into lipid vesicles stabilizes the enzyme at lower
pH values (data not shown) and eliminates any complicating effects of
differing lipid contents of the isolated enzymes. However, vesicle
aggregation at low pH appeared to be a problem which necessitated the use
of high ionic strength in order to obtain reproducible data.

To determine the assay method most advantageous for this study,
kinetic parameters of subunit III-containing cytochrome oxidase were
obtained at several pH’s using spectrophotometric and polarographic assays
and then compared (Table 3). The maximal activity (TNm) and apparent
Km values at pH 7.5 and 8.5 calculated from polarographic data were found
to be very similar to those computed from spectral data at corresponding
pH’s. At low pH, however, very high concentrations of cytochrome c (at least
400 11M) are needed to obtain good estimates of kinetic parameters due to
the high Km values for cytochrome c under acidic conditions. The resulting
high absorbances interfere with spectrophotometric activity determinations
making the data relatively inaccurate, whereas the polarographic assay is
not limited by high substrate concentrations and thus offers a significant
advantage for accurately assessing the kinetic constants over a wide pH

range. Addressing the reservations expressed concerning the use of the

78

.Gmma .8285m 98 3389 wig Batman 83988
2: mafia poem—:33 083 8:32:86 @8988 ~85 9.8 8889889 38:08

89588.5 ~388v8nxm E @2588“. 8.8 $5488 .3833

8838 8895038 mamfificcoéfl £555 “6333288 8m 3:83?

 

 

N a 8 a a S a. a 8 a. a 8“ 0388828
I sass
m a mu m a we a u... 8 ...2 + own -8888st
93 as 2-3 SE. 33 an 2-3 329
Emma.
3 ma 3 ma

 

om><mm< Ommgwogom Q72 UHEHEOBOEQOEUmmm Mrm
DHZH<BmO mmmfigmafl Drag E<Emuwfl<mfim «HO ZOmHmAEEOO ”m mafia.

79

polarographic assay in this type of study (Smith et al., 197 9; Wilms et al.,
1980; Thornstrom et al., 1984), we found that cytochrome c remained
greater than 97% reduced under the high salt conditions used. Therefore,
based on its superior accuracy at low pH and good agreement with the
spectral method, the polarographic assay was deemed more desirable for
investigation of pH dependence. (See Discussion for further consideration of
apparent discrepancies between these assay systems.)

To determine the pH range over which the enzyme was stable, the
reconstituted enzymes were incubated in buffers ranging from pH 4.5 to 8.5
for 10 minutes and then assayed for activity at pH 7.4 (Figure 11). Each
enzyme tested retained greater than 85% of its activity after incubation at
pH’s 5.0 through 8.5, establishing the pH range used in subsequent studies.
At pH 4.5, significant inactivation occurred.

When the kinetic data at various pH’s are visualized by double
reciprocal plots (Figure 12), all forms of the oxidase (beef heart, rat liver, i
subunit III) display qualitatively similar patterns. Under the high ionic
strength conditions of these assays, the plots at all pH’s were monophasic,
probably representing the interaction of cytochrome c at only the high
affinity site of the enzyme (Sinjorgo et al., 1986). Data obtained at pH 5.0
through pH 6.0 form a series of plots of decreasing slopes while data from
pH 6.5 to 8.5 describe a series of nearly parallel plots. This phenomenon
results in nonlinear log (TNnm/Km) versus pH plots (See Discussion; Figure
16B).

For all reconstituted oxidase preparations tested, TNm (Figure 13A)
and K1m values (Figure 13B) for uncoupled electron transfer activity increase
continuously with decreasing pH, as observed by Thornstrom et al.(1984)

using the reconstituted oxidase and previously by Wilms et al.(1980) using

80

FIGURE 11: Stability of reconstituted cytochrome c oxidase versus
incubation pH. Reconstituted oxidase (0.03-0.06 nmol) was
incubated in buffer (0.3 mL) of designated pH for 10 min at
25°C. The incubation suspension (0.25 mL) was then added to
1.5 mL of 100 mM K-Hepes (pH 7.4), 22 mM K2804 and
activity (TN) was measured polarographically in the presence
of 2.85 mM ascorbate, 0.6 mM TMPD, 37 M cytochrome c,
1.1 11M valinomycin, and 5.7 11M CCCP. Buffers composed of
167 mM K2804, 50 mM MES were used for incubations at pH
4.5-6.5 and 167 mM K2804, 50 mM Hepes at pH 7.0-8.5.
Subunit III-containi beef heart orddase (0); subunit III-
depleted rat liver oxi se (A).

300

81

 

 

 

l l l I

l J l l I

 

4.5 5.0 5.5 6.0

6.5 7.0 7.5 8.0 8.5
pH

Figure 11.

 

FIGURE 12:

82

Double reciprocal plots showing steady-state uncoupled
activity of cytochrome c oxidase as a function of cytochrome c
concentrations at various pH values. (A) Subunit III-
containing cytochrome oxidase from beef heart; (B) subunit
III- depleted oxidase from rat liver. Rates of oxygen
consumption were measured as described 1n Experimental
Procedures using cytochrome oxidase vesicles (0. 015- 0. 03
nmol mg) in 1.75 ml 167 mM K280, and 50 mM MES for pH
5. 0-6. 5 or 50 mM Hepes for pH '7. 0- 8. 5. The reaction mixture
contained 2. 9 mM ascorbate, 0. 6 mM TMPD, 1.1 uM
valinom cin, and 5. 7 11M CC.CP Cytochrome c concentration
ranged om 18 to 410 11M.

83

 

 

 

0.03 'r
0.025 —

0.02 -
/

 

-0.03 _-0.02 -0.01

 

 

0.01 0.02 0.03 0.04 0.05

Mg] (HM-1)

0

 

 

 

/ , .
B ./ ~/, / ,1, n~555n~5
/ , x 4 556788
/ , . HHHHHH 1
/ , , « PPPPPP
/ /’, « ooaaoo
/ a, I x.
x /, x
/ 5 / ’A a. I
x / fix 5
/ / I I
/ I a
/ ’ II ..o
/6 ./ 0A ..’
x .o. {a r.
// / .
0 (. w
x / ... ..
/ ./ f w
,.
o/ A 3.;
x . say...
/ a I.»
) O I.» I.
I I
z» o. A
I I
_ . _ . _ _ _ . .0 1..
9 7 5 3 / /1/’..
o. o. o. o. x .o. 2
O 0 0 O l0 J
G.
E E: J.

 

 

0.02 0.03 0.04 0.05

1/[91 (uM'l)

0.01

-0.01

Egmmm.

84

FIGURE 13: The pH dependence of (A) maximal activit (TN ) and (B)
Km for reconstituted cytochrome oxidase. gubunit‘III-
containing beef heart enzyme, (0); subunit III-depleted rat
liver enzyme, (0). Kinetic parameters and standard
deviations were estimated by linear regression analysis.

85

 

 

7.0 8.0
pH

6.0

 

 

 

5.0

 

 

8.0

7.0
pH

6.0

 

5.0

 

 

Figure 13.

86

detergent-solubilized enzyme. The degree of stimulation of activity at the
lowest pH (5.0) was independent of the presence of subunit III, but the
activity of the subunit IH-depleted enzyme decreased more dramatically as
pH increased from 5.0 to 5.5 and was consistently lower than the activity of
the III-containing enzyme from 5.5 to 8.5. In contrast, Km values are higher
for the subunit IH-depleted oxidase when compared to the complete
enzymes, an effect that is accentuated at pH 5.0. The plot of log (TN nun/Km)
(Figure 163) is similar for the two forms of the oxidase Q subunit III) but
shifted down by 0.5 log units for the HI-depleted enzyme as a result of its
lower TNmax and higher Km values.

Kinetic studies at lower ionic strength with the different forms of the
enzyme in reconstituted or soluble states show biphasic kinetics typical of
cytochrome oxidase (Ferguson-Miller et al.,1976). Interestingly, while the
Km values are dramatically different (30 to 40-fold greater at I :05 M
compared to I =0.07 M), the maximal activities are very similar (within 15%)
for a particular pH. This observation is in agreement with the findings of
Wilms et al. (1980) showing that at a single pH, TNmax is unaffected by

increases in ionic strength.

Effe t of H n the e tral Pro erties f C chr me Oxidase With nd
Without Subunit HI. The possibility was considered that the increase in
TNmax at low pH might be due to an alteration in the redox potential of one
of the metal centers. A candidate was suggested by the observations of
Callahan et al. (manuscript in preparation) which indicate that low pH
disrupts a hydrogen bond between the protein and the formyl group of the
heme a2+ porphyrin ring, resulting in a blue shift of the Soret and a-peaks

and a predicted lower redox potential of heme a. When the absorbance

87

maximum for the a-peak of reduced, reconstituted oxidase at pH 7.4 was
compared to that of the same enzyme exposed to pH 5.0 for one minute, a
small blue shift (from 605 to 604 nm) was observed that increased to 3 nm
after five minutes. These shifts were minimal compared to those seen with
the soluble enzyme (immediate shifts of 4 nm at pH 5.5 and 10 nm at pH
4.5) and the presence or absence of subunit HI had no effect. In addition, the
results were independent of the presence or absense of uncouplers which
equilibrate pH across the vesicle membrane. It is evident that the lipid
environment of the vesicle affords the enzyme a large degree of protection
against the pH effect observed by Callahan and coworkers. This suggests
that the increase in maximal activity at low pH, measured during the first
minute, is not the result of loss of this protein-to-heme a hydrogen bond. It
is interesting to note that the a-peak of the reconstituted subunit HI-
depleted oxidase (603 nm) is already blue shifted 1 to 2 nm when compared
to subunit III-containing oxidase, in agreement with findings of Penttila
(1983) and Puettner et al. (1985) for the Triton X-100 solubilized oxidase
depleted of subunit III. However, the III-depleted enzyme in lauryl
maltoside does not show a blue shift and resonance Raman analysis reveals
that the formyl hydrogen bond is still intact (Babcock et al., 1981). Thus it
appears that the lipid environment not only protects against extremes of pH
but also stabilizes a different conformation of the enzyme than that seen in
lauryl maltoside.

Cpmparison of Beef Heart and Rat Liver Oxidasas Depleted of Subunit III.
In the studies discussed above, subunit III-depleted oxidase originated from

rat liver while subunit III-containing oxidase was prepared from beef heart.

Buge and Kadenbach (1986) have proposed that tissue isozymes of

88

cytochrome orddase demonstrate kinetic differences (but see Sinjorgo et al.,
1987b). To examine the possibility that the observed differences were
related to species or tissue variations rather than subunit HI content, the
beef heart enzyme was depleted of subunit III using lauryl maltoside and
affinity chromatography as described by Hill and Robinson (1986). Different
extents of removal were achieved by varying detergent and salt
concentrations during treatment. As subunit IH content decreased,
respiratory control ratios increased accompanied by some reduction in
proton translocating ability (figure 14), characteristics similar to those
exhibited by subunit III-depleted oxidase from rat liver. Under all
conditions tested, the kinetic parameters of the highly III-depleted beef
heart enzyme closely resembled the III-depleted rat liver oxidase. Kinetic
studies of reconstituted beef heart preparations of varying subunit IH
content were carried out at pH 5.0 and pH 7.4 (Figure 15). At pH 5.0,
decreasing levels of subunit III resulted in increases in Km values without
changes in maximal activity, while at pH 7 .4, the Km was only slightly
affected by removal of subunit IH whereas maximal activity decreased
significantly. The effect of subunit III removal on Vm at pH 7.4 is
consistent with the observed difference in activity between the beef heart
control oxidase and the subunit III-depleted rat liver enzyme. Our results
indicate that the lower activity of the rat liver oxidase at higher pH, as well
as its higher Km at low pH, are related to subunit III removal rather than
species or tissue specific differences. The values of the kinetic constants for
the different forms of beef heart oxidase are summarized and compared to

the HI-depleted rat liver enzyme in Table 4.

89

FIGURE 14: Subunit III-depleted beef heart cytochrome c oxidase.
Densitometric scans of Coomassie blue-stained SDS-
polyacrylamide gels for beef heart cytochrome oxidase before

11 after treatment to remove subunit III. (A) Untreated
subunit III-containing oxidase. (B) Oxidase after a 24 hour
incubation 1n 1% lauryl maltoside (9m per mg protein) and
0. 5 M NaCl followed by cytochrome c a nity
chromatograph .(C) Oxidase after a 24 hour incubation 1n
1. 9% lauryl ma toside (15 mg per mg Ergtein) and 1 M NaCl
followed by affinity chromatography. nges of values
obtained for the functional properties of these enzymes are:
(A) RCR: 4. 2-7. 0, H+/ -= 0. 7- 0. 9 (using 2 to 4 turnovers); (B)
RCR: 9. 2- 9. 4, H+/ -= 0. 5- 0. 6 (using 5 turnovers), (C) RCR:
9. 4-14. 0, H+/e-= 0. 5 (using 4 turnovers).

90

 

A 96m RCR H‘7e'

J 100 5.9 0.8
I U L

n 39 9.3 0.6
I I w
«IL - UIJUL

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Absorbance 535
r i

23 11.8 0.5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I II III IV V VII
VI VIII

Figure 14.

91

FIGURE 15: Double reciprocal plots for reconstituted beef heart
cytochrome oxidase of varying subunit III content at (A) pH
5.0 and (B) pH 7.4. Activities were assayed as described in
Figure 12. Untreated enzyme (100% subunit III), (0); oxidase
with 39% subunit III, (0); oxidase with 23 % subunit III, ([2]).
grgelatic parameters and standard deviations are listed in
a e 4.

92

 

 

 

 

 

 

 

 

 

 

 

A I'
.05-
Q
2.03-
'—
> )-
.01-
o 0.01 l 0.02 '
1/IQI (pM‘l)
B .05
8.03-
z
'—
> ..
.01
o 0.02 0.04
1/IQI (nM'l)

Figure 15.

.8me .8285m 9.8 goose VZSZAHB :8an 88988 2:
mafia @83380 8.83 83838". p8p88 .83» 38 8.88893 £85K s

 

93

 

8H y 82 d a. a 82. a s5
ofi 83 O2 82 at am
a. c 88 y a c 88 a
a: 88 o2 8% s8
3 y 88 fl 3. t 83. c
o: 88 8a 8% s8
8 t 88 c a. y 82 3 8mm
2: 88. on” 82: 82 tom
3 ma 3 ma 3 mg on ma

928.200 momaom
:23 am as 529 E .3288 momma.

mm<QUAO mzommoogo m0 gain CHEMZUH
ME 20 EEOC SH BEDmDm UEM<> .mO 9099mm a. mafia.

94

Effect of DCCD on the Kinetic Parameters of Cyt_ochrome Oxidase With and
Without Subunit H1. The effects of DCCD treatment on the kinetics of
cytochrome oxidase were examined to determine whether glutamate 90 of
subunit III, which is highly modified by DCCD and postulated to have a role
in proton translocation (Casey et al.,1980; Prochaska et al.,1981), could be
implicated in the observed pH-dependent kinetic difi‘erences between the
III-containing and III-depleted oxidases. The reconstituted enzymes were
treated with a 300 molar excess of DCCD using the method of Prochaska et
al. (1981). Upon DCCD modification, the two enzymes Q subunit III)
showed similar changes in kinetics at pH 5.0 and pH 7.4 (Table 5). Each
exhibited significant losses in activity at pH 5.0 (approximately 50%) with
lesser inhibition at pH 7.4 (25%), suggesting a loss of sensitivity to pH
stimulation. The results show that the kinetic alterations and attenuation of
pH effects resulting from DCCD treatment cannot be attributed to

modification of subunit III.

95

.353 .uBHmsm van axooumv $722135

EEMPE HBsano 23 mama: £33938 0.83 mcofimgmc canvafim .393 98 3308359 335M 9

.28: .3 so «flanged 3 3283 saves £8. ma moon 338 8m 5...; 58.5. s

5:8 22853 .2828 23 mama: ecofiewob QUUQ no.“ 35:00 a

 

8H 3 82. HV
0: comm

8m 3 Sena 5
com econ

G. 3 8cm ..3

Am 5 83 ..J
0: com m6 moon

2: y 83. fl
o2 82 I: 3.580 22.13

Q 3 83 fl

 

2: 82 SN 82 3 Soon
8 fl 8% fl 2: ..3 83. fl
2: 8%.. 8m 98% 3 29530 c:¢§
3 mg 3 ma 3 ma o.» ma mom @2245? 32m
93 am :3 SE.

 

H: :ZDmDm. S053? DZ< $.53 mm<DHNO HEOMEOOEWO

QEmZOUHm m0 gamgm 03.3 20 GOOD ho BUMEH ”m mAmQB

96

DISCUSSION

Many studies have demonstrated that the electron transfer activity of
cytochrome oxidase is pH dependent. Although pH optima have been
reported by a number of investigators (Y onetani, 1961; Davies et al., 1964;
Ferguson-Miller et al., 1976; Maurel et al., 1978), more recent studies
indicate that maximal activity increases continuously with decreasing pH,
exhibiting no pH optimum (Wilms et al., 1980; Thornstrom et al., 1984; and
Sinjorgo et al., 1986). The observation of an optimal pH can be attributed to
one or more of the following: 1) the use of only one substrate concentration
over the pH range without compensation for increases in Km with
decreasing pH; 2) instability of the enzyme preparation at pH extremes; 3)
inability to achieve the high substrate concentrations needed to obtain the
most accurate kinetic parameters; or 4) insufficient ionic strength to
prevent enzyme or particle aggregation. Studies of the pH dependence of
cytochrome oxidase activity in Keilin-Hartree particles at low ionic strength
(I=0.01-0.05M) demonstrated that activity was Optimized in the alkaline
region (Ferguson-Miller et al.,1976). Although Thornstrom et a1. (1984)
attributed this to the use of the polarographic assay, Maurel et al. (1980)
observed the same relationship using the spectrophotometric assay with
submitochondrial particles at low ionic strength (I =0.0 1M). In both cases it
is likely that the apparent pH optimum is strongly influenced by
aggregation of the membrane particles at low pH and low ionic strength. In
the polarographic assay, the high pH optimum is further accentuated by a
major contribution to the activity by the rapid turnover of bound cytochrome

c (through efficient re-reduction by TMPD) at high pH and low cytochrome c

97

concentrations (Brautigan et al., 1978). At the high ionic strengths and high
substrate concentrations used in the present study, neither of these
phenomena influence the measured activity. Indeed, in agreement with the
recent spectrophotometric analysis of the pH dependence of cytochrome
oxidase activity, we find that TNmu increases continuously with decreasing
pH when rates of oxygen consumption are measured in the presence of
added reductants (ascorbate and TMPD) under conditions otherwise similar
to those described by Thornstrom et al. (1984). At lower ionic strength
(I=0.075M; data not shown) we observed some discrepancies in TNmu
between the two assay methods, but these differences disappear at high
ionic strength and the kinetic parameters obtained by the two methods
become nearly identical.

A plot of log TNlam versus pH (Figure 16A) reveals a complex
relationship from which pK. values are not readily determined. However,
when the model of Dixon and Webb (1964), as adapted by Wilms et al.
(1980), is applied to this data, the relationship of activity to pH is found to
be described best by at least three pK. values: <5.0, 6.2, and 7.8 for the
subunit III-containing enzyme, and <5.0, 6.0, and 7.9 for the subunit III-
depleted enzyme. Our values agree with those obtained by Wilms et a1.
(1980) for the detergent-solubilized enzyme (4.8, 6.5, and 8.0) and by
Thornstrom et al. (1984) for the reconstituted enzyme (4.5, 6.8, and 7.8).
The significance of the pK. values has been discussed by these
investigators, but there is no consensus or experimental basis for assigning
them to specific groups or relating them, as yet, to specific functions.
However, our studies at least exclude subunit III as the location of the
protonation sites since there is no significant change in the pK. values or

the magnitude of the response in its absence.

98

FIGURE 16: pH dependence of (A) log (TN ) and (B) log (TNmn/Km)
using data from Figure 13. Su unit III-containing oxidase
from beef heart (control), (C); subunit III-depleted oxidase
from rat liver, (0 ).

 

 

99

 

O
0

JL-- ------
O O
I l
7.0 8.0
pH

+lll
- .o- -

4
I
I
IL
I

 

.6 .

I
B . a.
5

 

 

 

 

 

x 2. . 6 w 5
«pie mo. .Ex \xaEzt mo.

 

Figure 16.

5.0

100

Analysis of the kinetic data reveals a marked pH dependence of the
Km as well as TNm, These studies, which were carried out at ionic
strengths of 0.5M, show monophasic kinetics in agreement with the
observation that the low affinity phase of the kinetics disappears with
increasing ionic strength (120.1M; Veerman et al., 1980; Brooks and
Nicholls, 1982; Sinjorgo et al., 1986). Therefore, only one Km, that of the
high affinity or primary site, is obtained at each pH (from monophasic plots
shown in Figure 12). Parallel to the change in TNm, the Km is also found to
increase continuously with decreasing pH (see Figure 13B). In this case,
however, a comparison of III-containing and III-deficient enzymes reveals
that the latter has significantly higher Km values at low pH, regardless of
tissue source. Although the major site of interaction of cytochrome oxidase
with cyotochrome c appears to be subunit II (Bisson et al., 1982; Millett et
al., 1983), subunit III has also been implicated in cytochrome 0 binding
(Birchmeier, et al., 1976; Fuller et al., 1981). Previous studies have shown
that, at pH 7.4, there is no change in Km for cytochrome c upon removal of
subunit III (Prochaska and Reynolds, 1986). However, our studies suggest
that in the highly protonated form of the enzyme, subunit III (or subunit
Vb, VIa, VIb, or VIIa which are also removed when III is depleted) helps to
maintain a higher affinity binding domain for cytochrome c. Indeed, in
contrast to Thornstrom et al. (1984) who report that log lam/Km is
independent of pH (an unusual observation given the three pK. values
found in the log km versus pH plot), we find that this parameter becomes
pH dependent in the more acidic region with a slope close to +1 (Figure
16B), indicating that a group with a pK value of approximately 5.7 to 5.8
affects the rate of association of the substrate and enzyme. Thornstrom et

al. (1984) suggest that the lack of pH dependence of the lam/Km in their

101

studies indicates that pH does not affect the rate of combination of enzyme
and substrate. In the present studies, removal of subunit III increases the
effect of low pH on the Km, but it does not alter the apparent pK. observed
in the log TNmlem versus pH plot, indicating that the group or groups
involved in determining this change in association are not on this subunit.
Analysis of pH dependence data using a minimal kinetic scheme, as
carried out by Malmstrom and Andreasson (1985), indicates that the rate of
dissociation of cytochrome c is faster than the km and, therefore, not
limiting in the lower pH range when using high ionic strength. When
treated similarly, our data also suggest that the rate of dissociation, while
very similar to maximal turnover rates at high pH, becomes increasingly
faster than the TNm as pH is decreased. In contrast to Sinjorgo et al.
(1984) who state that substrate dissociation is rate-limiting under steady-
state conditions and suggest that decreasing pH allows increases in Tme
as a result of increases in this dissociation (Sinjorgo et al., 1986), our
results as well as those of Malmstrom and coworkers (Thornstrom et al.,
1984; Malmstrom and Andreasson, 1985) suggest that increasing activity
with increasing [H+] reflects control of one or more intramolecular electron
transfer events. Since the subunit III-depleted enzyme, which is only 50%
efficient at proton translocation, shows no major changes in its response to
pH, our data do not provide direct support for the further hypothesis that
regulation by pH occurs at an electron transfer step that is involved in
coupling that process to proton translocation. However, since removal of
subunit III may not, in fact, alter the basic mechanism that couples proton
translocation to electron transfer, this lack of change in pH dependence may

not be a definitve test of the relationship.

102

In summary, the pH dependence of the kinetic parameters of
cytochrome oxidase support the conclusion that the protonation state of the
enzyme regulates the internal electron transfer step. However, since
subunit III-depleted oxidase retains a sensitivity to pH similar to that of the
III-containing oxidase with respect to maximal velocity and inhibition by
DCCD, it is clear that subunit III does not contain the protonatable group(s)
responsible for this regulation. Moreover, the inhibitory effect of DCCD on
the pH sensitivity suggests that a site of modification on a peptide other
than subunit III may mediate this control. Thus, subunit III can be ruled
out as the primary structure responsible for pH regulation of electron

transfer by bulk pH.

CHAPTER 4

CONTROL OF RESPIRATION IN CYTOCHROME c OXIDASE VESICLES:
INDEPENDENT REGULATION BY pH AND ELECTRICAL GRADIENTS
AND EFFECT OF SUBUNIT III DEPLETION

Hinkle et a1. (1972) demonstrated that cytochrome c oxidase (EC
1.9.3.1) incorporated into phospholipid vesicles exhibits respiratory control
behavior apparently similar to that observed in the mitochondrial
membrane. It can be shown that vesicular cytochrome oxidase generates a
transmembrane electrochemical gradient which appears to inhibit turnover
rates, resulting in controlled activity. Equilibration of this gradient by
addition of ionophores increases electron transfer activity giving rise to
uncontrolled or ’uncoupled’ rates. The gradient in right-side-out vesicles,
reconstituted such that the enzyme accepts electrons from externally added
cytochrome c, is positive and acidic on the outside. Various factors have
been proposed to be the dominant rate-controlling element of the
transmembrane gradient: 1) the lack of substrate protons in the
intravesicular space; 2) the electrical component (Aw) of the gradient; or 3)
the H+-concentration component (ApH) of the proton gradient.

If the concentration of intravesicular protons needed for continued
reduction of oxygen to water and for proton translocation were the sole rate-
limiting factor (Nicholls, 1982), coupled activity would be a function of the
rate at which protons "leak" back into the intravesicular space and control

would depend on the degree of membrane impermeability. However, studies

103

104

indicate that enhancing the rate of the return of protons to the vesicle
interior by adding nigericin does not relieve the control (Moroney et al.,
1984; Brunori et al., 1985). In addition, the supply of protons in highly
buffered solutions does not appear to limit the reduction of oxygen
intermediates in the even higher turnover carbonic anhydrase system
(Lindskog and Coleman, 1973). This suggests that some aspect of the
intrinsic activity of the protein is altered in response to an electrochemical
gradient (Brunori et al., 1985). [But see Konstantinov et al. (1986) for
discussion of possible electrogenic control of proton access to the heme a3-
Cu]; site.]

Using whole mitochondria, Hinkle and Mitchell (1970) observed that
the transfer of electrons from cytochrome c to heme a was inhibited in the
presence of an electrical gradient (positive on the outside) due to an
apparent change in the redox potential of heme a. In later studies, Brunori
et al. (1985) examined the rapid kinetics of cytochrome c oxidation by
cytochrome oxidase vesicles and found that equilibration of the electrical
gradient alone, but not the pH gradient, increased the rate of orn'dation of
cytochrome c. This group concluded that the major regulator of oxidase
activity is the electrical component of the gradient and proposed that it
functions by mediating a change in protein conformation which alters the
kinetics of electron transfer. Similar conclusions have been reached by Azzi
and coworkers (Muller et al., 1987).

While the studies discussed above suggest a dominant effect of the
electrical gradient in respiratory control, other investigations indicate an
important role of the pH gradient in control of respiration in cytochrome
oxidase vesicles. Moroney et a1. (1984) observed that elimination of the pH

gradient decreased the level of steady-state reduction of heme 0, indicating

105

a release of electron transfer from heme a to heme as. They concluded that
pH was controlling this step, in addition to the electrical-potential control of
electron transfer between cytochrome c and heme a demonstrated by Hinkle
and Mitchell (1970). In agreement with Artzatbanov et al. (1978) they
observed that control was imposed specifically by intravesicular pH.
Nicholls and coworkers (Shaughnessy and Nicholls, 1985; Nicholls et al.,
1988) studied the effects of ionophores on enzyme activity and on the
formation of pH and electrical gradients in cytochrome oxidase vesicles and
concluded that the pH gradient imposed a greater inhibitory effect on
steady-state oxidase activity than did the electrical gradient. Control of
oxidase activity by pH is also consistent with kinetic studies which show
that the maximal velocity of cytochrome oxidase decreases with increasing
pH (see Chapter 3; Wilms et al., 1980; Thornstrom et al., 1984).

Since subunit III may facilitate proton translocation, thus
contributing to generation of the membrane potential, it is important to
investigate the possible role of subunit III in mediating the response to the
electrochemical gradient. However, this study is complicated by the
conflicting results and controversy surrounding the relative contributions of
components of the membrane potential to respiratory control in cytochrome
oxidase vesicles. This chapter reports the results of experiments designed to
clarify the regulatory roles of the electrical and pH components of the
potential and thus allow a more meaningful evaluation of the role of subunit
III. Changes in the reduction levels of heme a and cytochrome c in addition
to the rates of oxygen consumption and cytochrome c oxidation under
steady-state conditions were examined following elimination of the pH
gradient or the electrical gradient. The results demonstrate that control by
the pH and electrical gradients is exerted predominantly at different steps

106

in the electron transport sequence of cytochrome oxidase under steady-state
conditions. The same qualitative relationships between the components of
the membrane potential and oxidase activity were observed with and
without subunit III, and any quantitative differences correlated with the
degree of respiratory control exhibited by the vesicle preparation rather
than subunit III content of the enzyme.

107

EXPERIMENTAL PROCEDURES

Materials. Cytochrome c was repurified as described in Experimental
Procedures of Chapter 1 except that the final gel filtration chromatography
was carried out in 10 mM K-Hepes (pH 7.4), 40 mM KCl. Sources of other

materials are listed in previous chapters.

Cytochrome Oxidase Preparations. Subunit III-containing cytochrome
oxidase was purified from beef heart according to Suarez et al. (1984) but
without the described ethanol wash step. Subunit III-depleted oxidase was
isolated from rat liver according to Thompson and Ferguson-Miller (1983).
These preparations typically differ from each other in respiratory control
ratios (RCR) following reconstitution: RCR=5-8 and RCR=8-15, respectively.
Several other preparations were used where indicated in addition to the two
already described: 1) a preparation of beef heart oxidase from which subunit
III was partially removed (75% depleted) according to the method of Hill
and Robinson (1986) (RCR=9-14); 2) a preparation of rat liver oxidase
containing about 30% subunit III, obtained by the procedure of Thompson
and Ferguson-Miller (1983) omitting the hydroxylapatite chromatography
(RCR=8-9); and 3) another preparation of beef heart oxidase (subunit III-
containing) purified from "green fraction" [a product of the fractionation of
beef heart mitochondria (Hatefi, 1978), a generous gift from Dr. Y. Hatefi]
using the method of Suarez et al. (1984) starting at the ethanol wash step
(RCR=2-4).

108

Reconstitution. Cytochrome oxidase was incorporated into asolectin vesicles

as described in Experimental Procedures of Chapter 2.

Ass—ays of Electron Tra_nsfer Activity. Enzyme activities and respiratory
control ratios were measured polarographically (Thompson and Ferguson-
Miller, 1983) or spectrophotometrically in buffer identical to the last dialysis
solution. Specific conditions are described in the figure legends. Turnover
rates fi'om spectroscopic assays were determined by calculating the initial
rate using the integral method described by Yonetani and Ray (1965) and
then dividing by the enzyme concentration.

Steady-state kinetic analysis of the dependence of oxidase activity on
cytochrome 0 concentration was carried out under the conditions described
in the legend to Figure 21. The data was analyzed according to the formula
of Sinjorgo et al. (1984), assuming a two phase mechanism. A computer
program (by Dr. R. J. Miller, Physics Department, Michigan State
University) was employed that used an iterative procedure to correct each
phase for the contribution of the other until the best fit to the data was

achieved.

Steady-State Reduction Studie_s_. Steady-state reduction spectra of heme a
and cytochrome c were recorded with an Aminco DW2a spectrophotometer
using 1 cm pathlength cells. Assays were performed in buffer identical to
the last dialysis solution to avoid gradients generated by non-enzymatic
mechanisms. Cytochrome c was added at levels which support sufficient
oxidase activity to create inhibitory transmembrane gradients but not so
much as to obscure the spectral signal of heme a and drive high levels of

activity that would too quickly consume available oxygen. Ascorbate and

109

TMPD were added to keep the substrate reduced. The ratios of enzyme to
substrate used (3-4 mol cyt c/mol heme (103)111 the measurements of steady-
state reduction levels, although not saturating for uncoupled respiration,
allowed 3 to 8-fold increases in activity upon uncoupling. Other conditions
are described in the figure legends. Heme a was assumed to contribute
100% of the absorbance at 605 nm minus 630 nm, under steady-state
conditions where (13 is oxidized (Gibson et al., 1965), and 80% of the
absorbance after complete reduction [see Wikstrom et al. (1981a)]. The
reduction of cytochrome c was measured by the absorbance at 550 nm minus
542 nm. Heme (103 and cytochrome c were considered completely reduced
five minutes after the cell became anaerobic.

In the steady-state difference spectra (steady-state minus oxidized) of
the cytochrome oxidase vesicles when coupled or when nigericin was added,
red shifts were observed in the 605 nm peak of heme M3 and the 550 nm
peak of cytochrome c. In addition, anomalous absorption was found in the
630 nm region which largely disappeared upon addition of valinomycin or
upon anaerobiosis. It appears that a small amount of TMPD+ (W urster’s
blue, resulting from oxidation of TMPD) electrophoretically migrates to the
vesicle interior where it cannot be re-reduced by external ascorbate and
thus contributes to the absorbance spectrum (Nicholls et al., 1988). Ifthe
absorbance of TMPD+ in the 500 to 650 nm region is subtracted in
proportion to the absorbance observed at 630 nm, the cytochrome oxidase
peaks appear in their normal position and reduction levels of heme a and

cytochrome c can be determined.

110

RESULTS

Independent Elimination ef An; ad ApH, Valinomycin, a potassium

ionophore, has been shown to equilibrate transmembrane electrical
gradients with some increase in ApH, while nigericin, a proton/potassium
exchanger, equilibrates the pH gradient with a concomitant increase in the
electrical gradient (Singh and Nicholls, 1985; Nicholls et al., 1988). Thus,
independent elimination of the Aw and ApH can be achieved by addition of
the appropriate ionophore. However, each of these ion carriers becomes less
specific in its behavior at higher concentration (Ferguson et al., 1971;
Pressman, 1976). In addition, there are some conflicting reports regarding
whether one or other of these compounds more effectively stimulates
oxidase activity when added to proteoliposomes (Brunori et al., 1985;
Shaughnessy and Nicholls, 1985), including results suggesting that
valinomycin alone can inhibit oxidase activity (Nicholls et al., 1988).
Therefore, preliminary to other studies, the minimum concentrations of
nigericin and valinomycin needed in combination to obtain maidmal enzyme
stimulation (uncoupling) were determined for a particular set of conditions.
For example, Figure 17 shows the stimulation of activity upon addition of
increasing amounts of nigericin in the absence and presence of a fixed
amount of valinomycin and compares it to the maximum stimulation
achieved by adding CCCP (carbonyl cyanide m-chlorophenylhydrazone).
When nigericin is added alone at concentrations insufficient to fully release
respiratory control even in the presence of valinomycin, activity is increased
by 20%, and as concentrations begin to exceed the amount needed to

uncouple vesicles when valinomycin is present, activity is stimulated to a

111

FIGURE 17: Stimulation of activity in cytochrome oxidase vesicles versus
nigericin concentration. Turnover numbers (TN) for
reconstituted subunit III-containing beef heart oxidase (0.03
nmols aa3) were calculated from rates of oxygen consumption
at 25°C in 1.75 mL of 40 mM KCl/IO mM Hepes-KOH (pH
7.4)/38 mM sucrose in the presence of 35 M cytochrome c, 5.6
mM ascorbate, and 0.28 mM TMPD. Ionophore-induced
stimulation is represented as the ratio of stimulated activity
divided by unstimulated activity (TNionophmJ'I‘Ncon led): with
increasing concentrations of nigericin (o ); upon addition of
valinomycin (1.4 M; O ), and after the addition of CCCP (6.9
M; A). In these assays 1 uM nigericin corresponds to 4.4
nmols nigericin/mg phospholipi .

 

112

.21 450.3ng
o.u 3 0.. md

 

 

_ — u A q a

 

 

 

m

‘-
paudnoo N1 /e:oqdouo! N 1

Ln

 

 

Figure 17.

113

greater degree, presumably representing equilibration of both the ApH and
AV by an electrogenic mechanism (Ferguson et al., 1971; Pressman et al.,
1976). Nigericin at 0.8 M (or 3.5 nmol/mg phospholipid) was used to
preferentially equilibrate ApH in subsequent experiments using this vesicle
preparation. The minimal concentrations of valinomycin required were
determined in a similar manner. The optimal ionophore concentrations
varied somewhat between vesicle preparations and, while the levels
required increased with the total lipid content of the assay (as might be
expected for these hydrophobic reagents), thisrelationship was not linear
(lower ionophore-to-lipid ratios were needed when lipid concentration was
high).

Using the appropriate minimal concentrations, the addition of
valinomycin alone or nigericin alone stimulated electron transfer activity in
cytochrome oxidase vesicles only slightly (1.2 to 1.8-fold), and both were
needed to completely release respiratory control, thus allowing 5 to 17-fold
increases in rates of oxygen consumption. This was observed to be true
whether steady-state activities were measured polarographically as rates of
oxygen consumption (with ascorbate and TMPD added as reductants) or
spectrophotometrically as rates of ferrocytochrome c oxidation (Figure 18).
In the latter case, no artificial reducing agents were present that might

afl‘ect the rate-limiting steps in the steady-state activity.

Qhanges in Steady-State Reduction of Heme a and QMchrome c in
Response to Alterations in the Membrane Potential. To investigate the

nature of the regulatory effects of the individual membrane potential
components on the activity of cytochrome oxidase, reduction of heme a and

cytochrome c was monitored for responses to specific elimination of Au; or

114

FIGURE 18: Effect of independent elimination of Aw and ApH on oxidation
of ferrocytochrome c by cytochrome oxidase vesicles:
semilogarithmic plots of ferrocytochrome c concentration
versus time of reaction. The absorbance of 26.9 M
cytochrome c (reduced minus oxidized) in 10 mM Hepes-KOH
(pH 7.4)/40 mM KCl/ 38 mM sucrose at room temperature
was monitored at 550 nm (pathlength 1 cm) after addition of
cytochrome oxidase vesicles (16 nM subunit III-containing
beef heart aa3). Assays contained: no ionophores (O ); 1 uM
valinomycin (5 nmol/mg phospholipid, A); 0.2 uM nigericin (1
nmol/mg phospholipid, D); or 1 uM valinomycin and 0.2 uM
nigericin (0). An RCR of 6.3 was calculated from this assay.

115

 

 

0.60

  

60

20

 

 

 

 

0.50

.
O
4.
0

6255.683»: .5. omm<

 

p

O
3
0

d 0.20”

0.15

Time (s)

Figure 18.

116

ApH by ionophores. Previous studies have examined changes between
coupled and uncoupled vesicles (Wrigglesworth and Nicholls, 1978) or the
changes following specific equilibration of Aw or ApH but using
hexaamineruthenium(II) as an artificial substrate (Moroney et al., 1984). In
the studies reported here, low levels of the natural substrate cytochrome c
were used in conjunction with ascorbate and TMPD.

Starting with coupled vesicles, the transitions in steady-state
reduction levels of heme a and cytochrome c upon equilibration of the
electrical gradient (using valinomycin) and the total gradient (using
valinomycin and CCCP) are shown in Figure 19A. The spectrum of the
completely reduced system (anaerobic) is included for comparison. Heme a
was found to be 30-55% reduced in the coupled state while cytochrome c was
70-90% reduced. After equilibration of Aw, the fraction of total heme a
reduced increased to 60-80%, while there was very little change (<5%) in the
level of reduction of cytochrome c. These changes occurred within 20
seconds of the addition of valinomycin. Only slight increases in enzyme
turnover were observed at this stage (1.3 to 1.7-fold). This result suggests
that the electrical gradient exerted some regulation on the transfer of
electrons from cytochrome c to heme a that was released upon addition of
valinomycin, but that electron flow remained inhibited beyond heme a.
When CCCP (a protonophore) or nigericin plus valinomycin (data not
shown) were present to uncouple the vesicles, both heme a and cytochrome c
became more oxidized and activity increased 3 to 8-fold, indicating that
elimination of the pH gradient released control on electron transfer within
the enzyme.

The alterations in redox levels when the pH gradient is equilibrated

first are shown in Figure 19B. When nigericin was added to coupled vesicles

117

FIGURE 19: Effect of independent elimination of Aw and ApH on the
reduction of heme a and cytochrome c: difference spectra
(steady-state minus oxidized) of cytochrome oxidase vesicles.
Cytochrome oxidase vesicles (III-containing beef heart
oxidase) were diluted to approximately 0.6 uM aa3 in 10 mM
Hepes-KOH (pH 7.4)/4O mM KCl/38 mM sucrose and then,
cytochrome c (2 pM final) was added. The suspension was
maintained at 18°C during assay. Baseline spectra (oxidized
versus oxidized) were subtracted from steady-state spectra.
The steady-state spectra of coupled vesicles
(_ _ _ _) were obtained in the presence of 12 mM ascorbate
and 0.3 mM TMPD. In (A) the electrical gradient was
abolished by the addition of 4 “M valinomycin (0.7 nmol/mg

phospholipid, ___), and in (B) the pH gradient was
abolished by the addition of 5.5 uM nigericin (0.9 nmol/mg
phospholipid, ___). The remaining pH gradient in (A)

was equilibrated by adding 20 uM CCCP (- ° - - - - -), and the
remaining electrical gradient in (B) was equilibrated by
addingr 4 uM valinomycin (- - - ~ - - -). Spectra were also
recorded 5 minutes after anaerobiosis in each system ( ----- ).

 

118

 

 

 

 

 

 

590 620 650
n m

560

530

 

 

 

500

 

1—

0.01 A

db

 
 

‘0.--.----.1

 

 

 

0.02A

650

620

590

nm

560

530

 

 

 

500

Figure 19.

119

to eliminate the H+-concentration gradient, heme a became more oxidized
while the reduction level of cytochrome c was unchanged. Oxygen
consumption increased only slightly. This suggests that removal of the pH
gradient releases electron transfer between heme a and a3, but that the
electrical gradient is still inhibiting the input of electrons from cytochrome
c. The levels of heme a reduction increased when valinomycin was added to
these pH-equilibrated vesicles and cytochrome c became more oxidized. This
was accompanied by full stimulation of activity.

Figure 20 is a summary of the changes in levels of reduction of heme
a and cytochrome c in the same vesicle preparation with sequential
equilibration of gradient components. The independent elimination of Aw or
ApH causes significant changes in the reduction level of heme a, but they
are of opposite direction apparently due to the differential release of control
of electron flow to or away from this redox center (i.e. from cytochrome c to
heme a and from heme a to a3, respectively). Reduction of cytochrome c,
however, is largely unaffected by the elimination of only one gradient
component, since enzyme turnover increases very little. The level of
reduction of heme a in the uncoupled state returns to a level close to, but
usually more reduced than, that of the coupled state, while cytochrome c is
significantly more oxidized in the uncoupled state reflecting full stimulation
of activity under these substrate-limited conditions. This is compatible with
the results of Wrigglesworth and Nicholls (1978) who observed a slight
increase in reduction of heme (1 along with a substantial oxidation of
cytochrome c when the steady-state spectra of controlled and uncontrolled

vesicles were compared.

120

FIGURE 20: Changes in levels of reduction of heme a (hatched bars) and
cytochrome c (o n bars) upon independent equilibration of
Axfi and A H. ( ) Equilibration of t e electrical potential
fo owed y the fpH gradient (as in Figure 19A). (B)
Equilibration o the pH gradient followed by the electrical
gradient (as in Figure 19B). Percent reduction of cytochrome c
and heme a in cytochrome oxidase vesicles is shown for: (C)
the coupled state; (V) after elimination of only the Aw by
addition of valinomycin; (N) after elimination of only the ApH
by the addition of nigericin; and, (U) after the vesicles were
uncoupled by equilibration of both components of the proton
electrochemical gradient. Levels of reduction were calculated
as percent of the totally reduced system upon anaerobiosis, as
described in Experimental Procedures. The levels of reduction
and the magnitudes of the changes agree within 10% upon
replication or a given vesicle preparation, but the the sizes of
the changes observed in different vesicle preparations are
dependent on the RCR. (See Figure 24.)

121

 

 

 

 

 

 

Zw

/‘
/ C

 

 

 

c.5253. 2.3.0.—

 

 

 

 

 

 

 

 

 

Figure 20.

122

Effect of Am and ApH on Steady-State Kinetics of the Reaction of
Cuchrome c with CMQhrome Oxidase. To determine whether the
inhibition of oxidase activity by either of the membrane potential
components could be accounted for by a change in affinity for substrate,
studies were undertaken to examine the separate and combined effects of
nigericin and valinomycin additions on rates of oxygen consumption as a
function of cytochrome c concentration.

Analysis of the data by Eadie-Hofstee plots reveals that the kinetics
of the vesicular enzyme with cytochrome c are biphasic in coupled, partially
equilibrated, and uncoupled states. Figure 21 shows the data for subunit
III-containing beef heart oxidase vesicles. Assuming that the kinetic
behavior is the result of at least two separate interactions of cytochrome c
with the oxidase, it can be analyzed in a manner described in Experimental
Procedures to yield kinetic constants for the two phases (coupled: Tme 1
=27 s-l, Km 1=7x10'3 M, TN“m 2:44 s-l, Km 2=1><1O'6 M; uncoupled: TNmax 1
=110 s-l, Km 1=15x10'8, TNmu 2:210 s-l, Km 2=1x10'6 M). Although these
values give a good fit to the kinetic curves, they must be considered
’apparent’ constants because of the uncertainty concerning the number and
nature of the interactions of cytochrome c with cytochrome oxidase
(Ferguson-Miller et al., 1976; Errede and Kamen, 1978; Antalis and Palmer,
1982; Speck et al., 1984; Sinjorgo et al., 1984). Nevertheless, a strong
correlation between the apparent Km values and the binding constants for
cytochrome c has been shown (Ferguson-Miller et al., 1976; Ferguson-Miller
et al., 1979). In view of this, comparison of the Km values for the coupled and
uncoupled states may give some information about the effect of the
membrane potential on the binding affinity of cytochrome c. In fact, the

apparent Km values for the low affinity phase were relatively unaffected by

123

FIGURE 21: Eadie-Hofstee plots of activity of reconstituted cytochrome
oxidase (subunit III-containing enzyme fiom beef heart)
under coupled and uncoupled conditions. Coupled activity (0)
in absence of ionophores; valinomycin-stimulated activity ([3);
nigericin-stimulated activity (A); uncoupled activit (O) with
nigericin plus valinomycin. Conditions as describe in the
legend to Figure 17 with 0.04 nmols aaa, 1.1 uM valinomycin
(4.8 nmol/mg phospholipid), and 0.8 ni ericin (3.5
nmol/mg phospholipid). The range 0 cytoc rome c
concentrations was from 0.1 to 33 1.1M. The dotted lines
indicate the two components of the coupled and uncoupled
kinetic curves which were derived by computer analysis as
described in Experimental Procedures.

1 000

TN/[c] (“W1 s")

800

600

400

zoo '-

124

 

 

 

 

 

 

125

the partial or complete removal of the transmembrane gradient, while those
for the high affinity reaction were somewhat lower (approximately 2-fold)
for the coupled compared to the uncoupled state. While these changes in Km
values cannot be interpreted unequivocally, the results suggest that, if
anything, there is an increase rather than decrease in affinity for
cytochrome c in the presence of a membrane potential, which would not be
expected to contribute to control of electron transfer from cytochrome c to

heme 0.

Effect of Subunit III Content on the Control of Stead - te uc ' n
Eels of Heme a. In studies of the effects of ApH and Aw on the electron
transfer activity of reconstituted cytochrome oxidase, Moroney et al.(1984)
found that DCCD treatment lessened the sensitivity of the redox state of
heme a to additions of proton equilibrating agents. From these observations
the authors proposed a role for subunit III in mediating the regulation of a
proton-limited step between heme a and (13.

To test this possibility, the relative changes in levels of reduction of
heme a and cytochrome c that occured upon sequential elimination of Aw
and ApH were compared for vesicles containing subunit III-depleted or
subunit III-containing cytochrome oxidase. The two forms of the enzyme
gave similar responses to perturbations of the electrical and pH components
of the gradient (Figure 22). Thus, subunit III removal did not prevent the
enzyme’s response to these regulatory effectors. In fact, the magnitudes of
the changes in heme a reduction were often greater for the III-depleted
enzymes (see below for further discussion).

To further investigate the possibility of subunit HI involvement in

control of oxidase activity, the reconstituted enzyme with and without

FIGURE 22:

126

Changes in levels of reduction of heme a and cytochrome c
upon independent elimination of Aw and ApH in the presence
and absence of subunit III. Reduction of heme a (open bars)
and cytochrome c (shaded bars) were determined from
difference spectra (steady-state minus oxidized) as described
in Experimental Procedures. (A) Subunit III-containing
oxidase from beef heart. (B) Subunit III-depleted oxidase from
rat liver (<2% III). Ascorbate (12 mM) and TMPD (0.3 mM)
were added to maintain steady-state activity of 0.6 uM
vesicular cytochrome oxidase in the presence of 2 CM
cytochrome c. Reduction levels are presented for: , coupled
activity; V, activity after electrical equilibration b the
addition of valinom cin (4 nM); and, U, uncouple activity
after subsequent a dition of CCCP (20 nM).

127

 

 

A

 

 

fl w

V

 

_ c

 

 

 

1C”)

.
0
8

. .
O O
6 4

O
3: 2‘)”
u

0
v

 

 

V/flfl
////////////Afl
/////////////4Mc

_ U
_ v

C

 

 

 

 

 

 

1C")

P P p
o o o O
8 4 2

a. 6K)-

.cou.o

ty

t
y
C

Figure 22.

128

subunit III was treated with DCCD over a range of 0-1900 moles DCCD per
mole oxidase. Both preparations showed loss of activity with increasing
DCCD, but the subunit III-depleted on’dase was somewhat more affected.
Examination of the steady-state levels of reduction of heme a, measured by
the absorbance at 605 minus 630 nm (Figure 23), revealed that DCCD
treatment caused increased reduction of heme a in all phases (coupled,
electrically equilibrated, and pH equilibrated). As observed for the loss in
activity (also noted by Prochaska et al., 1981), the increase in reduction was
dependent on the concentration of DCCD used for the modification. This
may indicate that both phenomena are due to increasing modification of a
residue essential for intramolecular electron transfer. It is also apparent
that the redox state of heme a becomes less responsive to changes in the Aw
and ApH following DCCD treatment. These effects are observed in the
presence or absence of subunit III and indicate that these DCCD-induced

changes are not caused by modification of subunit III.

Cerrelation of Heme e Reduction Levele and the Respiratory Control R_atio.

The comparison of the changes in heme a reduction levels with alterations
in gradient components between oxidases with and without subunit III,
while showing no qualitative difference, suggested that the magnitude of
the changes may be related to the respiratory control ratio (RCR) of the
vesicle preparation. The RCR is used as an indicator of the responsiveness
of cytochrome oxidase in vesicles to a membrane potential. It presumably
reflects the degree of correct insertion and the level of the membrane
potential developed in steady-state. To further characterize the relationship
suggested by the previous observation, various types of enzyme

preparations exhibiting differing respiratory control ratios were tested for

129

FIGURE 23: Effect of DCCD treatment on the steady-state level of reduced
heme a (monitored at 605-630 nm) in the presence and
absence of ionophores. Cytochrome oxidase vesicles (0.5 - 0.6
nmol (103) were added to 0.85 ml buffer containing 3.5-5.5 11M
cytochrome c. Ascorbate (12.5 mM) and TMPD (0.2 mM) were
added to maintain steady-state activity (coupled).
Valinomycin (Val; 4 nM) was added to abolish Aw followed by
CCCP (20 M) to equilibrate theA H. Complete reduction
was measured five minutes after e system became
anaerobic. A. Reconstituted subunit III- depleted oxidase (rat
liver). B. Reconstituted subunit III-containi oxidase (beef
heart). Solid traces from untreated vesicles; shed traces
fromaDCCD-treated vesicles (1400 mol DCCD per mole
0x1 se .

130

 

 

F
[ ........ P---“

i 3'“ ‘ Anaerobic
‘ CCCP
Val

 

 

 

 

 

~ Asc
TMPD

 

 

   

x \ Anaerobic

‘
Val CCCP

 

0.004 A

 

 

‘ Asc 20 sec
TMPD

 

Figure 23.

 

131

responses to alterations in the membrane potential. Figure 24A shows the
correlation between RCR and the percent of heme a reduced in the coupled
state. The high levels of reduction with lower respiratory control may reflect
the partial release of (or inability to maintain) the electrical gradient. Upon
elimination of gradient components, we found that all vesicle preparations
showed the same qualitative changes in heme a reduction (an increase upon
addition of valinomycin followed by a decrease upon addition of CCCP), and
the magnitude of the change was dependent on the magnitude of the
respiratory control ratio (Figure 24B), indicating the relationship between

these redox changes and the control of activity.

132

FIGURE 24: The relationship of RCR and the level of steady-state
reduction of heme a. (A) Plot of respiratory control ratios
(RCR) versus the percent of heme a reduced in the coupled
state (no ionophores). (B) Plots of respiratory control ratios
(RCR) versus the magnitudes of the changes in percent
reduction of heme a upon addition of ionophores. In (B) the
solid symbols represent the increase in percent reduction
upon elimination of the electrical potential by adding
valinomycin to coupled vesicles and the open 3 mbols
represent the decrease in percent reduction (p otted as a
positive value) when electricall equilibrated vesicles become
completely uncoupled by the a dition of CCCP. The reduction
of heme a was monitored at 605 minus 630 nm under the
conditions described in the legend of Figure 19. A range of
RCR values was achieved by reconstituting enzyme
preparations that typically vary in this characteristic: subunit
III-containing beef heart oxidase (O , o); rat liver oxidase, 95-
98% depleted of subunit III (I,D); beef heart oxidase, 75%
depleted of subunit III (A, A); rat liver oxidase, 70% depleted
of subunit III (0,0); and subunit III-containing beef heart
oxidase prepared from " een fraction" by a method involving
an ethanol wash (*utr). e Experimental Procedures for a
more detailed description of oxidase preparations and for the
method of RCR determination.

133

 

 

 

_ _ _

_

15

13

11

 

 

o o o
5 4 3
v w. 050... ..o 20:0:qu—

0
2

RCR

 

 

 

 

I I EU
A
I2 1 _u 1
O 6
0 .00
*1 .w
B
_ _ _ _ _
O O O O O O
5 4 3 2 1

a 080: he cozozuom g E omcmco

 

15

13

11

1

RCR

Figure 24.

134

DISCUSSION

A Deg Mechanism for Respiratog Centrol. The lack of any significant
stimulation in oxygen consumption rates upon addition of nigericin to
cytochrome oxidase vesicles demonstrates that: 1) enzyme activity in
coupled vesicles is not simply limited by a lack of substrate protons in the
vesicle interior bulk phase, and 2) elimination of the pH gradient alone
cannot relieve the inhibitory effect on the enzyme. Similarly, release of the
electrical gradient by valinomycin alone has little power to increase enzyme
turnover rates. When these two ionophores are used in combination,
however, the enzyme becomes fully stimulated. These results show that
each of the two gradients has the ability to regulate activity. In addition,
kinetic analysis of activities in the absence and presence of ionophores
indicates that control is not mediated by decreasing the aflinity of the
enzyme for substrate. The studies of ionophore-induced alterations in
steady-state heme a reduction suggest that the electrical gradient
predominantly regulates electron transfer from cytochrome c to heme a,
while the pH gradient controls the heme a to a3 transfer. These conclusions
are consistent with those discussed by Papa (1988) and Maison-Peteri and
Malmstrom (manuscript in preparation). [It is interesting to note that each
of the two gradient components are also believed to inhibit activity of
reconstituted ubiquinone-cytochrome c oxidoreductase (the enzyme
immediately preceding cytochrome oxidase in the mitochondrial respiratory
chain). However, both gradients appear to control an intramolecular

reaction (Rich and Clarke, 1982).]

135

Figure 25 depicts a model that attempts to explain the results of the
studies presented here in terms of a thermodynamic mechanism of
regulation in which the controlling forces are represented entirely in terms
of changes in reduction potentials of the heme groups. In reality, the control
may reflect one of several response mechanisms: 1) a real change in the
redox potential resulting from some change in protein structure in the
immediate environment of the heme; 2) an apparent change in the redox
potential resulting from an altered contribution of the membrane potential
to the driving force perceived by the electron; or 3) a real change in the
kinetics of electron transfer due to an altered pathway (length, chemical
nature) through which the electron must travel. Both 1) and 3) would
require some conformational change in the protein, while 2) would simply
reflect the degree of electrogenicity of the electron transfer step.

Since there is considerable variation in the redox potential values
reported in the literature, the redox potentials of the heme centers are
represented by values from references that used conditions most relevant to
the reconstituted system at steady-state. Relative changes that may
correspond to regulatory responses are emphasized rather than specific
values.

Considering first the effects of Aw, the reduction potential of
cytochrome c is apparently insensitive to the electrochemical gradient
(Wikstrom et al., 1976) but does shift from 280 mV to 230 mV upon binding
(Dutton et al., 1970). Evidence for electrical gradient-dependent shifts in the
reduction potential of heme a was provided by Hinkle and Mitchell (1970) in
studies using whole mitochondria. They concluded that the midpoint
potential of heme a was modified by the electrical gradient in a continuous

manner according to the equation:

136

FIGURE 25: Hypothetical scheme for the dual control of respiration in
atochrome oxidase vesicles by the electrochemical gradient.
e DISCUSSION for detailed description. Redox centers
[cytochrome c (c); heme a (a); and heme a3 (a3)] are arranged

according to midpoint potentials (Em, scale at left) in the
presence and absence of ApH and Aw. Controlled electron
transfer steps, light arrows; uncontrolled electron transfer
steps, bold arrows; changes in transmembrane gradients,
dashed arrows. Path of e ectron transport: in controlled
conditions (C); in presence of nigericin (N); and, in the
presence of valinomycin (V).

137

 

 

300 -

320 5

mV

figure 25.

138

Em" = Em’ - 0.5 Aw (I)

where a positive Aw reflects an electric potential that is positive on the
outside of the membrane and Em’ and En," are the midpoint potentials at
electrical gradients of zero and some finite value, respectively. Thus, if the
redox potential of heme a in the uncontrolled state is 285 mV [as calculated
by Wikstrom et al. (1976) for uncoupled mitochondria in aerobic steady-
state] and the Aw across the membrane in oxidase vesicles is approximately
150 mV in the coupled state and 175 mV when nigericin is present [as
reported by Singh and Nicholls (1985)], the Em" values would be shifted to
about 210 mV and 198 mV, respectively. This Em value for heme a in the
coupled state corresponds well to the 212 mV potential obtained from
mitochondria in State 4 respiration (Wikstrom et al., 1976). The reduction
potential of CuA has been shown to be about 20 mV more positive than that
of cytochrome c and to be insensitive to the presence of a membrane
potential (Rich et al., 1988). Although there is some question whether
cytochrome c transfers electrons initially to Cu); (Rich et al., 1988) or heme
a [see Wikstrom et al. (1981a)]. this model can accommodate either
sequence. (Hence, Cu); is not depicted in this model.) The redox potentials
shown for heme as in the coupled and electrically equilibrated states are
represented as somewhat less positive than for heme a (Dutton and Wilson,
1974; Wikstrom et al., 1976), although more positive values for heme as
have been reported for other conditions (eg. Wilson and Dutton, 1970;
Dutton et al., 1970; Wilson et al., 1972). A shift in Em similar to that
proposed for heme a is suggested for heme a3 upon equilibration of the

electrical potential; however, the relatively slight increase in oxidase

139

activity observed upon addition of valinomycin may indicate a smaller
alteration in this center. The observation that the steady-state level of
reduction of heme a is higher in the absence of an electrical gradient is
consistent with an enhancement of the cytochrome c to heme a electron
transfer resulting from a change in reduction potential. The electrical-
gradient dependence of the redox potential of heme a has been
demonstrated in mitochondria (Hinkle and Mitchell, 1970; Rich et al., 1988),
and the observations on oxidase vesicles reported here provide the
additional evidence that the redox state of heme a is correlated with the
degree of respiratory control.

Another mechanism of Aw control has been proposed by Wikstrom
(1989), involving negative cooperativity in the reduction of heme a and Cu};
and stabilization by the electrical gradient of a hypothetical dipole created
by reduction of CuB. This model would also predict an increase in the redox
potential of heme a upon removal of the electrical gradient. Although there
has been recent speculation about a role of Cu]; in regulation of oxidase
activity (Wikstrom, 1989; Nicholls, 1989), direct evidence regarding its
redox state and kinetics of reduction has been difficult to obtain.

A different mechanism of regulation by the electrical gradient has
been suggested (Konstantinov et al., 1986) whereby the presence of Aw
impedes access of protons to heme a and/or heme a3. prrotons are required
for heme a reduction, removal of Aw would increase its apparent redox
potential.

Considering the effects of ApH, the control of electron transfer from
heme a to a3 is also illustrated as a shift in reduction potential, but
specifically at heme as. The pH dependence of the midpoint potential of this
center has been documented (Wilson et al., 1972; Blair et al., 1986).

140

However, it is likely that the pH gradient exerts an additional kinetic
control of this step by maintaining an alkaline intravesicular pH (see
Chapter 3; Thornstrom et al., 1984; Malmstrom and Andreasson, 1985).
Thus, elimination of the pH gradient by nigericin is represented as causing
a slight decrease in the reduction potential of heme a due to an increase in
Aw and release of the regulation on transfer from heme a to as. The
oxidation of heme a reflects the increase in transfer from a to as, while the
minimal increase in activity reflects the continued block in electron transfer
from cytochrome c to heme a.

The final state represented in the model is that where the addition of
valinomycin plus nigericin has removed all regulatory barriers,
thermodynamic and/or kinetic.

In studies of changes in oxygen consumption by sonicated
proteoliposomes, Nicholls and coworkers (Shaughnessy and Nicholls, 1985;
Nicholls et al., 1988) observed that nigericin released respiratory control
more effectively than valinomycin. Indeed, these authors found that
valinomycin caused an increase in ApH and a decrease in enzyme activity,
and concluded that the ApH was the dominant regulatory component in
their system. In contrast, in the present studies valinomycin addition
resulted in a small increase in activity and an immediate large increase in
steady-state reduction of heme a. This combination of effects would argue
against an increase in ApH being the predominant cause of the observed
valinomycin-induced changes.

Moroney et al. (1984) studied the effects of ionophores on the
reduction of heme a, in addition to the stimulation of activity of
reconstituted oxidase, but in a system substituting hexaammineruthenium

for cytochrome c. The levels of heme a reduction they observed in the

141

coupled state were significantly higher than those observed in this study
(80% versus 35-55%) and, although they also found that valinomycin did not
stimulate activity, they found that the levels of cytochrome a reduction were
insensitive to elimination of Aw by this ionophore. The lack of response to
alteration in Aw in this system can be explained by the low redox potential
of hexaammineruthenium (78 mV; Meyer and Taube, 1968) which allows
this artificial substrate to drive the reduction of heme a in the presence or
absence of an electrical gradient and thus by-pass the control by the
electrical gradient. 9

In agreement with the findings reported here, Moroney and
coworkers observed the oxidation of heme a upon elimination of ApH.
However, their results show that nigericin caused a 3-fold increase in
activity when used at a level causing only a 30% increase when using
cytochrome c as substrate. Again, due to the low potential of the
hexaammineruthenium, this can be predicted by the model proposed in
Figure 25 since the remaining Aw in their system would not inhibit the
donation of electrons to heme a as it would if cytochrome c were the donor.
Thus, both sets of data demonstrate the regulation of the electron transfer
from heme a to a3 by ApH (or the intravesicular pH).

By studying the differential effects of ionophores on the rapid kinetics
of cytochrome c oxidation by reconstituted cytochrome oxidase, Brunori et
al. (1985) found that elimination of the pH gradient by nigericin failed to
release oxidase activity, while equilibration of the electrical gradient by
valinomycin alone resulted in nearly complete stimulation. They concluded
that the electrical component of the gradient constituted the major control
on activity in cytochrome oxidase vesicles. However, the results shown in

Figures 18 and 21 show no significant release of activity by valinomycin

142

alone, whether assayed as oxygen consumption rates or by spectral
measurement of ferrocytochrome c oxidation rates. A possible explanation
for this discrepancy would be the lack of maintenance of a pH gradient in
the vesicles used for the experiments reported by Brunori and coworkers,
due to dialysis or assay conditions, or the nature of the lipid vesicles.

It is apparent that the vesicle membrane is not simply a barrier that
limits the re-entry of substrate protons (for the reduction of oxygen) but a
capacitor that sustains transmembrane pH and electrical gradients which
directly and differentially affect specific steps in the electron transfer
process of cytochrome oxidase. The reconstituted enzyme provides a
valuable system in which to investigate the structural requirements and
regulation of the electron and proton transport mechanisms of cytochrome

oxidase, as well as, the involvement of subunit III in these processes.

Effeet of eebunit III removel on the eentrol of Steedy-Sgte Reductien ef
Heme a. Subunit III-depleted oxidase responds to alterations in the

electrochemical gradient in a manner similar to the III-containing oxidase,
consistent with a dual control mechanism. Although subunit III was
implicated in the control response by the enzyme’s loss in sensitivity to the
gradient upon DCCD modification (Moroney et al., 1984), sensitivity to
DCCD is not diminished by removal of subunit III. These results conflict
with the hypothesis (Moroney et al., 1984) that subunit III contains the
proton binding site(s) involved in gradient control of electron transfer. In
addition, the retention of DCCD inhibition indicates that other subunits
must be involved in mediating the DCCD effect on pH and electrical
gradient sensitivity. This conclusion is strongly supported by our kinetic
analysis which shows that DCCD attenuates H+-stimulation of activity in a

143

subunit III-independent manner (see Chapter 3), indicating a DCCD-
modifiable residue elsewhere on the enzyme involved in mediating this
effect. Indeed, in addition to its major modification site in subunit HI,
DCCD has also been shown to covalently label cytochrome oxidase in
subunits II and IV (Chapter 1; Prochaska et al., 1981) and may form protein
crosslinks in self-eliminating reactions [see Nalecz et al. (1986)].

In summary, it is clear from these studies that subunit III is not
responsible for mediating the inhibition imposed by the components of the
electrochemical gradient. Moreover, the inhibitory effect of DCCD on the
control response suggests that a site of modification on a peptide other than

subunit III may be important in this control.

CHAPTER 5

THE EFFECT OF SUBUNIT III REMOVAL ON THE BINDING OF
CARBON MONOXIDE, OXYGEN AND CYANIDE TO
CYTOCHROME c OXIDASE

Heme a3 has been identified as the site of binding for the ligands, CO
and HCN (Keilin and Hartree, 1939), as well as the substrate, oxygen
(Lemberg, 1969). This iron center is believed to be closely associated with
the copper redox site, C113, and the two are often referred to as the binuclear
center. Indeed, these two prosthetic groups, both predicted to be bound to
subunit I (Holm et al., 1987), display strong magnetic interactions (Van
Gelder and Beinert, 1969; Tweedle et al., 1978) and appear to be separated
by a distance of less than 5 A (Powers et al., 1981).

Many procedures for purification of cytochrome oxidase yield an
oxidized enzyme that is characterized by slow reduction of heme a3
(Antonini et al., 1977) and slow, multi-phasic binding with HCN (Van
Buuren et al., 1972). Oxidase exhibiting these traits is typically designated
as the "slow" or "resting" form of the enzyme. Spectroscopic studies indicate
that this enzyme form has a ligand that bridges the oxidized centers of the
binuclear site. The bridging ligand, which possibly occupies the substrate
binding site (Powers, et al., 1981), has not been identified conclusively, but
several candidates have been proposed: a sulfur atom (Powers et al., 1981);
an oxygen atom (Blumberg and Peisach, 1979); the imidazole nitrogen from
a histidine residue (Palmer et al., 1976); and a chloride ion (Scott et al.,

144

145

1989). In other physical studies, Baker et al. (1987 ) found that the presence
of a specific EPR signal (g’=12) correlated with slow cyanide binding.

When the purified enzyme has been subjected to redox cycling (eg.
has been completely reduced and reoxidized several times), the oxidase
converts to a "fast" form1 showing rapid electron transfer from heme a to
heme a3 (Antonini et al., 1977 ) and rapid, monophasic binding with HCN
(Naqui et al., 1984). The comparison of the "slow" and "fast" forms by
physical methods have indicated that significant conformational changes,
especially in the heme a3 environment, coincide with this conversion
(Palmer, 1988). The "fast" form notably lacks the ligand bridging heme a3
and Cu}; (Powers and Chance, 1985) and the g’=12 EPR signal (Baker et al.,
1987). It has been shown that the "fast" conformation can be maintained in
the oxidized state for some time before it reverts to the "slow" form (Baker
et al., 1987).

Oxygen and CO apparently will not bind to oxidase at heme as when
the enzyme is in either the "slow" or "fast" form, unless the heme a3 and Cu};
centers are reduced (see review by Hill et al., 1986). In the "slow" enzyme,
reduction of the binuclear center displaces the bridging ligand and provides
the correct redox state of the metal centers for 02 or CO binding. In the
"fast" enzyme conformation, the ligand-bridged state is not present and
reduction is only necessary to provide the correct redox state of the metals.

Cyanide, on the other hand, will only bind to oxidized heme a3
(Yonetani, 1960). However, the binding rate is apparently greatly influenced
31223132323$:3335,2203?ifieixifiléti’iifii‘i‘iliiifi‘fmi‘iiilfih7x
"oxygenated" (Sekuzu et al., 1959); also see Lemberg, 1969]. However, due
to the difficulty in correlating all of this spectroscopic and kinetic data,

these) forms will be referred to in a group as the "fast" form (see Palmer,
1988 .

146

by the conformation of the enzyme. Cyanide binds slowly to the "slow",
oxidized form of the enzyme (Van Buuren et al., 1972), and more quickly to
the "fast", oxidized enzyme (Baker et al., 1987), but exhibits much higher
binding rates with the enzyme during redox cycling (Jones et al., 1984). It
has been proposed that during turnover, a transient conformational state
exists that has high affinity for cyanide (Jones et al., 1984; Thornstrom et
al., 1988).

Several research groups have studied various aspects of ligand
binding at heme a3-Cun for subunit III-depleted cytochrome oxidase. Nalecz
et al. (1985) report that the carbon monoxide difference spectra for the III-
depleted enzyme show alterations that might suggest modification in the
redox properties of the enzyme upon removal of this subunit. Malatesta et
al. (1986) observed no changes in CO-difference spectra for the III-depleted
enzyme when they examined absorbance in the Soret region, and found no
difference in the rate of CO binding to heme as between the complete and
III-depleted oxidase. Hill and Robinson (1986) studied the kinetics of
cyanide binding to III-depleted oxidase and found that after treatment to
remove subunit III, the purified enzyme reacted with HCN in a manner
similar to that observed for the "fast" enzyme form. These authors suggest
that subunit III may supply the bridging ligand which normally prohibits
HCN binding to the "slow" enzyme.

To clarify the effect of subunit III removal on the characteristics of
CO and 02 binding, the kinetics of binding were examined at low
temperature after photolysis of the CO-inhibited cytochrome oxidase with
and without subunit III. In addition, the cyanide binding kinetics of the
oxidized, III-depleted enzyme were compared to those of the III-containing

oxidase.

147

EXPERIMENTAL PROCEDURES

Cyt_o_chrome Oxidase Preparat_i ens. Subunit III-containing beef heart
oxidase was purified according to Suarez et al. (1984), omitting the ethanol
wash step. Subunit III-depleted rat liver oxidase was prepared by the
method of Thompson and Ferguson-Miller (1983). Rat liver oxidase
retaining 60% of its subunit III was obtained by preparing crude oxidase as
described by Thompson (1984), followed by DEAE chromatography and
FPLC in lauryl maltoside, as described in Chapter 1.

Carbon Monexigle-Difference Speetre. Cytochrome oxidase in Tris-Cl (pH
7.8) was reduced with cytochrome c in the presence of ascorbate. The sample
solution was added to a mixture of buffer and ethylene glycol that had been
pre-saturated with CO at 20°C. CO was bubbled through the mixture in the
cuvet for an additional 5-10 minutes to ensure CO saturation and
anaerobiosis, and then the sample was frozen in ethanol and dry ice at
-78°C. Final concentrations of sample constituents were as follows: 3 uM
cytochrome oxidase , 7 mM ascorbate, 5 11M cytochrome c, 20% ethylene
glycol, and 38 mM Tris-Cl (pH 7.8).

The sample cuvet (path length 2mm) was transferred to the
thermostated chamber of the Johnson Foundation dual wavelength
spectrophotometer (Chance et al., 1975) and equilibrated to -130°C. The
baseline spectrum of a blank solution was scanned from 400-800 nm and
stored in computer memory. The absolute spectrum (minus baseline) was
recorded for the CO-complexed enzyme. The sample was then irradiated
with a xenon lamp for 1 minute (to photolyze the CO-enzyme complex) and

148

then rescanned from 400-800 nm. Since oxygen was not present in these
samples and since recombination of CO and enzyme is negligible at -130°C,
the stable photoproduct is reduced, uncomplexed oxidase (Chance et al.,
197 5). The spectrum of CO-oxidase was subtracted from that of reduced
oxidase by computer to yield CO-difference spectra.

Carbon Monoxide-Binding Assefi. Samples were prepared as described for
CO-difference spectra and then equilibrated to -80°C in the
spectrophotometer sample chamber. Baseline absOrbance at 440 minus 460
nm was determined by a short exposure to incident light. Samples were
irradiated for 1 minute and the rate of CO recombination was monitored at
this wavelength pair until absorbance reached baseline levels. Pseudo-first
order rate constants were obtained from the slopes of semilogarithmic plots
of absorbance versus time, and bimolecular rate constants were calculated

for a 980 1.1M CO content.

Oxygen-Binding Assays. Sample solutions were prepared in cuvets in a 0.5
mL volume containing 6 uM aa,,, 14 mM ascorbate, 10-30 1.1M cytochrome c,
and 25% ethylene glycol in Tris-Cl (pH 7.8). After reaching anaerobiosis, the
mixture was saturated with CO and then equilibrated at -30°C. From this
point until the assays were completed, samples were handled in the dark or
under red light. An oxygen-saturated solution of ethylene glycol and buffer
(0.5 mL) was added to the cuvet at 20°C (650 pM oxygen, final), mixed
thoroughly, and the solution was frozen to -78°C. Cuvets were placed in the
spectrophotometer sample chamber and equilibrated at -100°C, where CO-
recombination rates were negligible but oxygen combination was still rapid

(Chance et al., 1975). Assays were conducted as described for CO-binding

149

assays but were irradiated 10 seconds to achieve photolysis. Rate constants
were calculated as previously described for the appropriate oxygen content.

Cyanide-Binding Assays. Cytochrome oxidase was diluted to

5 M in 50 mM Tris-Cl (pH 7.8) and 2 mM lauryl maltoside. Cyanide
binding was determined spectrophotometrically by measuring the
absorbance at 430 nm over time after addition of freshly prepared KCN (5
mM, final). The increase in absorbance at 430 nm occurs upon the shift in
the Soret absorbance maximum from 420 nm for oxidized cytochrome

oxidase to 428 nm for cyanide-bound cytochrome oxidase (Van Buuren,

1972).

150

RESULTS
Carbon Monexide-Differenee Speetre. Carbon monoxide and reduced

cytochrome oxidase combine to form a photodissociable complex. CO-
difference spectra can be obtained by subtracting the absorbance of the CO-
oxidase complex before photolysis, from that of the reduced enzyme after
photolysis. Figure 26A and 26B show the spectra for subunit III-containing
beef heart and rat liver oxidase, respectively. The beef heart enzyme
exhibits increases in absorbance at 447 and 611 nm and losses in
absorbance at 430 and 590 nm that can be attributed to the dissociation of
CO (see Vanneste, 1966 and Wikstrom, 1981a). The III-containing rat liver
oxidase shows similar changes with peaks at 448 and 613 nm and troughs
at 430 and 592 nm. The removal of subunit HI from rat liver oxidase by
cytochrome c affinity chromatography in lauryl maltoside (Figure 26C)
caused no significant alterations in CO-difference spectra. Absorbance
maxima were detected at 448 and 613 nm and minima at 430 and 592 nm.
The Soret maximum and minimum are in agreement with the results of
Malatesta et al. (1986) for a III-depleted enzyme prepared by chymotrypsin
treatment in 1% Triton X-100 at pH 8.5. In contrast, altered spectra were
obtained by Nalecz et al. (1985) when subunit III was removed by yeast
cytochrome c aflinity chromatography in 1.5% Triton X-100 at pH 7.2. These
authors observed a 4 nm red shift in the a-peak maximum and a 4 nm blue
shift plus a prominant shoulder on the lower wavelength side of the Soret
peak. The retention of the appr0priate absorbance maxima and minima, in

the case of our own studies and those of Malatesta and coworkers, indicates

151

FIGURE 26: Carbon monoxide difference spectra of cytochrome c oxidase
with and without subunit III. Absorbance of reduced oxidase
minus CO-oxidase for: (A) subunit III-containing beef heart
oxidase; (B) subunit III-containing rat liver oxidase; and (C)
subunit III- depleted rat liver orddase. Absorbance maxima
and minima are as indicated. Conditions are described 1n
Exfierimental Procedures. [The absorbance scales are as

n 'cated, except that the entire (B) spectrum was recorded at
the more sensitive scale.]

152

 

 

447 IAA= 0.] IA=A 0.025

449

430

(A

448

431

430

400 440 480 570 610 650

nm n m
Figure 26.

 

 

153

that the removal of subunit III causes minimal disruption in the heme

environments.

Rates of Carbon Monoxide Binding. Using stopped flow and flow-flash
kinetics, Gibson and Greenwood (1963) obtained rate constants of 5.6x104 to
8.0x 104 M-ls-l for the combination of reduced cytochrome oxidase with CO at
room temperature. This reaction can be observed on a much slower time
scale at low temperature (Chance et al.,1975). The following comparison of
CO-binding rates in III-containing and III-depleted oxidase was conducted
at -80°C in the presence of ethylene glycol as a cryogenic agent. The CO-
oxidase complex was dissociated by flash photolysis causing an increase in
absorbance at 440 minus 460 nm. Recombination of the enzyme and ligand
was monitored by the decay in absorbance following the flash. Table 6 lists
the half times for the return of absorbance to the pre-photolysis level and
the calculated bimolecular rate constants for the reaction. No marked
difference was observed between the III-containing and III-depleted
enzymes. Malatesta et al. (1986) reached the same conclusion for their III-
depleted enzyme using room temperature stopped-flow kinetics.

Rates of Oxygen Binding. Low temperature studies also allow examination
of the combination reaction of reduced cytochrome orddase with oxygen by
using the "triple trapping" method of Chance et al. (1975). The enzyme is
initially saturated with CO and then equilibrated at -30°C in the dark.
Since dissociation of CO is minimal under these conditions, oxygen may be
added to the suspension without reacting. The temperature of this mixture
is then lowered to -100°C. After flash photolysis, the absorbance at 440

minus 460 nm again increases. At this temperature, CO recombination is

154

Sum—Emozvoamp mo cosmom 5 088 oz.» 3 33:2: v.8 .38 3953 mg .8.“ 3.35:8“:

mzmomumflfim .8: A 2.5? anew—arrow cumcgm .3335 E 33333 cumvcfim
65 98 885:6qu E 333525ch no .8355: 23 8:3 :03» 0.8 839, owns}?

.3: .3338 3mm $383259

 

 

 

3o fl :d j .834
8V m6 3. 23 m4 mm 5cm I
5o fl :6 d .834
A8 m6 mv av ma mdm 9cm +
:d 3 team
A: oA mm £8 ma 9mm Comm +
2.2.5 393 2.925 663
e S a i a: a momDOm E
gmmfl. .EZDmDm
ZOHE<ZHB>HOD NO ZOHB<ZHmEOUmm OD
Zmowxo Q24 HEXOZOE

ZOmm<U ES mm<QUmO EOmmOOEO m0 ZOEZEmEZOD
m0 mWBé ME 20 ZOHBMAQHQA: 9:43me m0 SCHEME ”w mafia.

155

very slow and the decay in absorbance reflects the combination of oxygen
with the enzyme (Chance et al., 1975). Since rates of electron transfer are
minimal at this temperature, the oxygen-oxidase species detected is
proposed to be a complex of reduced orddase and dioxygen (Compound A)
before the reduction of oxygen begins (Chance et al., 1975). The bimolecular
rate constants and half times for decay of absorbance to the prephotolysis
baseline are given in Table 6. The rates of combination for the rat liver
enzymes were only slightly lower than that of the beef heart oxidase and the
average of the results for the III-containing versus HI-depleted rat liver
orddases were very similar. However, the III-depleted oxidase exhibited a
larger variation in rates in repeated assays which may reflect this enzyme’s
increased susceptibility to the conditions used to prepare the samples (ie.

high ethylene glycol).

Cyanide Binding Figure 27 shows the time course for the binding of cyanide
to cytochrome oxidase with and without subunit III. Each preparation
exhibited at least two kinetic phases. In the initial, fast phase,
approximately 75% of the oxidase combined with cyanide in the first 2 to 4
minutes. The subunit III-depleted rat liver oxidase showed a binding rate
intermediate to the faster rate of the III-containing rat liver enzyme and the
slower rate of the III-containing beef heart enzyme. The reactions continued
in a second, slower phase and after 3 hours 90% or more of each enzyme was
in the cyanide-bound form (data not shown).

These results show a high degree of similarity in the kinetics of the
subunit III-containing and subunit III-depleted oxidases. This is in contrast
to the results of Hill and Robinson (1986) in which only 30% of the III-

containing oxidase combined with cyanide (1 mM final concentration) at 15

156

FIGURE 27: The kinetics of cyanide binding to cytochrome c oxidase with
and without subunit III. Semilog lot of the change in
absorbance at 430 nm at time (t) ivided by the maximal
change in absorbance versus time. Subunit III-containing beef
heart oxidase (A); subunit III-containing rat liver oxidase (O );
and subunit III-depleted rat liver oxidase (O). Assays
performed as described in Experimental Procedures.

 

157

 

 

 

 

 

 

('v-°V)/7“v-‘v)

Figure 27.

0.1

TIME (min)

158

minutes, while 85% of the III-depleted oxidase became cyanide-bound over

the same time period.

159

DISCUSSION

Hill and Robinson (1986) observed that treatment of purified
cytochrome oxidase with lauryl maltoside and affinity chromatography to
remove subunit III resulted in a subunit III-depleted oxidase that bound
HCN in a rapid, monophasic reaction, while the untreated enzyme bound
HCN in the slow, multi-phasic reaction typical of the "slow" enzyme. They
suggested that subunit III provided-the bridging ligand detected in the
"slow" enzyme. However, since displacement of the bridging ligand requires
only reduction of the binuclear center while the conversion of "slow" to "fast"
form requires several cycles of reduction and oxidation, the two events may
be separate and there may be other explanations for these results. In fact,
there is some direct evidence suggesting that removal of subunit III does not
remove the bridging ligand. Preliminary analysis of the copper edge
structure of subunit III-depleted oxidase (D. A. Thompson, L. Powers and B.
Chance, unpublished results) by extended X-ray absorption fine structure
(Powers, 1982) shows the same results as a "slow", subunit III-containing
enzyme (Naqui et al., 1984).

The cyanide-binding kinetics reported in this chapter reveal no major
differences in the behavior of the subunit IH-containing and III-depleted
oxidases. The majority of the enzyme in both types of preparations
combined with cyanide in a fast reaction. These findings are consistent with
recent observations that modification of purification procedures by
minimizing or using alternatives to ammonium precipitation will yield a
"fast" (but subunit-complete) form of the oxidase (Baker et al., 1987; Brandt,
et al., 1988). It has been noted that the enzyme isolated as a "fast" form does

160

not convert to a "slow" form with time and also that the "slow" form of the
enzyme has yet to be detected in intact mitochondria (Baker et al., 1987).
The implication is that the "slow" conformation may be an artifact of
traditional purification procedures (Brandt et al., 1988). In the case of the
studies by Hill and Robinson (1986), lauryl maltoside treatment, rather
than subunit HI depletion, may cause the change in cyanide-binding
kinetics by providing the necessary environment to promote or maintain the
"native" (fast) conformation of the enzyme.

The studies of carbon monoxide binding reported in this chapter
confirm the findings of Malatesta et al. (1986) that subunit III removal had
no significant effect on the characteristics of CO-difference spectra and on
the rates of CO binding. Based on this result, these authors predicted that
the enzyme’s reactivity with oxygen would also be unaffected by the loss of
this subunit. Indeed, the rate constants for the combination of oxygen with
oxidase (ie. the formation of Compound A) presented here substantiate that
prediction. These studies indicate that: (1) subunit HI removal causes no
changes in the accessibility of oxygen and CO to the binuclear center, and
(2) the reduced oxidase is equally reactive with oxygen (or CO) in the
presence or absence of this subunit. Thus, it appears unlikely that subunit
III is involved in regulating the binding of substrate oxygen at the binuclear

complex active site.

SUMMARY

Subunit HI and enzyme-associated diphosphatidylglycerol (DPG)
have been implicated in various aspects of cytochrome c orn'dase function
based on loss of or alterations in activities upon removal of these
components and, in the case of subunit III, upon chemical modification of
the enzyme with N N-dicyclohexylcarbodiimide (DCCD). However, studies
in this thesis show that the enzyme can be depleted of subunit III and DPG
by alternative methods and still retain characteristics previously attributed
to these components. Furthermore, in some cases, DCCD has been shown to
cause the same changes in the behavior of the enzyme independent of the
presence or absence of subunit HI, illustrating a lack of specificity in the
reaction of this inhibitor that is not always considered.

Studies in this thesis reveal that, although difficult to remove, DPG
can be replaced with detergent (lauryl maltoside) and cytochrome oxidase
still retains high activity and normal substrate interactions.

While subunit III was found not to be essential for proton
translocating activity, these studies do not distinguish between a facilitative
role in this process or total lack of involvement. Investigation of the
aggregation state of oxidase in lipid bilayers will be critical in evaluating
proposals of dimeric models for proton translocation and possible
involvement of subunit III in promoting dimer formation.

Subunit III was excluded as a major regulatory site involved in
control of electron transfer activity by pH or by the proton and electrical

gradients. However, no correlation could be made between the magnitude of

161

162

these regulatory responses and apparent efficiency in proton translocation,
although these two functions might be expected to be linked. From other
studies, it also appears unlikely that subunit III contributes to binding of 02
or inhibitory ligands at heme a3.

Thus, it remains to be determined how this large, apparently well
conserved peptide contributes to cytochrome oxidase function in general and
the efficiency of proton translocating activity in particular. Possible roles yet
to be investigated include involvement in assembly of the enzyme complex
and involvement in association of the protein with other enzymes that
participate in oxidative phosphorylation. More definitive answers may be
forthcoming as it becomes possible to manipulate the genes for the
mitochondrially encoded subunits. Continued investigation of the role of
this and other cytochrome oxidase subunits will help to advance our

understanding of energy transduction mechanisms.

APPENDIX
Publications and Abstracts

163

PUBLICATIONS:

Thompson, D.A., Gregory, L. and Ferguson-Miller, S. "Cytochrome c

Oxidase Depleted of ubunit III: Proton-Pum ing, Respiratory Control, and
H Dependence of the Midpoint Potential of ytochrome a," (1985) J. Inorg.
iochem. 2_3, 357-364.

Gregory, L. and Ferguson-Miller, S. " H-dependence of the Maximal

Velocity of Cytochrome c Oxidase: Re ation to Proton Translocation and

Respiratory Control," (1988) in Advances in Membrane Biochemistry and

Bioenergetic , (Kim, C. et al., eds.) pp. 301-309, Plenum Press, New York.

Grego , LC. and Ferguson-Miller, S. "The Effect of Subunit III Removal on
gggygs of Cytochrome c Oxidase Activity by pH," (1988) Biochemistry 21,
- 14.

Gregory, L. and Fer son-Miller, S. "Control of Cytochrome c Oxidase
Activity by pH and t e Electrical Potential Gradient Occurs at Separate
Electron Transfer Steps and Does Not Require Subunit III," (1989) Annals
of New Yorit Acad. Sci.. in press.

Gregory, L. and Ferguson-Miller, S. "Independent Control of Respiration in
Cytochrome c Oxidase Vesicles by pH and Electrical Gradients," (1989)
Biochemistry, in press.

ABSTRACTS:

Gregory, L. and Ferguson-Miller, S. "Is the pH-dependence of the Maximal
Turnover of Cytochrome c Oxidase Related to its Proton Translocating
Function?", Symposium: The Mitochondrion 1986, The Johns Hopkins
I1J9niversity School of Medicine, Baltimore, MD, June, 1986, Abstract No. G-

Gregory, L. and Ferguson-Miller, 8., "pH Dependence of Kinetic Parameters
of Proton Translocation-com etent and -incompetent Cytochrome c
Oxidase," Abstract for ASB /DBC, ACS Meeting, Federation Proceedings
(Fed. Amer. Soc. Exp. Biol.) 45, 1555 (1986).

Gregory, L., Rajarathnam, K., and Ferguson-Miller, 8., "Complete Removal
of Subunit III from Cytochrome Oxidase with Retention of Hi h Electron
Transfer and Proton Translocatin Activities," Abstract for A BC Meeting
1987, Federation Proceedings (Fe . Amer. Soc. Exp. Biol.) 46, 1931 (1987).

Gregory, L. and Ferguson-Miller, S. "Altered pH-dependent Characterisitcs
of Cytochrome Oxidase Depleted of Subunit III." Gordon Conference:
Energy Coupling Mechanisms, June, 1987.

164

Gre ory, L. and Ferguson-Miller, S. "Separate Control of Cytochrome c
Oxi ase Activity by Transmembrane pH and Electrical Gradients." Abstract
for ASBMB Meeting 1989, submitted.

BIBLIOGRAPHY

BIBLIOGRAPHY

Ames, B. N. (1966) Methods Enzymol. 8, 115-118

Anderson, S., de Bruijn, M. H. L., Coulson, A. R., Eperon, I. C., Sanger, F.
and Young, 1. G. (1982) J. Mol. Biol. 1_56, 683-717

Anraku, Y. (1988) Ann. Rev Biochem. 57, 101-132
Antalis, T. M. and Palmer, G. (1982) J, Biol. Chem. 257, 6194-6206

Antonini, E., Brunori, M., Colosimo, A., Greenwood, C. and Wilson, M. T.
(1977) Proc. Natl. Acad. Sci. USA 74, 3128-3132

Artzatbanov, V. Y., Konstantinov, A. A. and Skulachev, V. P. (1978) FEBS
Lett. _8_'Z, 180-185

Awasthi, Y. 0., Chuang, T. F., Keenan, T. W. and Crane, F. L. (1971)
Biochim. Biophys. Acta 226, 42-52

Azzi, A. and Nalecz, M. J. (1984) Trends Biochem. Sci. _9_, 513-514

Azzi, A., Bill, K., Bolli, R., Casey, R. P., Nalecz, K. A. and O’Shea, P. (1985a)
in Structure and Properties of Cell Membranes (Bengha, G., ed.) pp.
105-138, CRC Press

Azzi, A., Muller, M., O’Shea, P. and Thelen, M. (1985b) J. Inorg. Biochem.
23, 341-347

Azzone, G. F., Pozzan, T. and Di Virgilio, F. (1979) J. Biol. Chem. 254,
10206-10212

 

 

Azzone, G. F., Zoratti, M., Petronilli, V. and Pietroban, D. (1985) J. Inorg.
Biochem. 23, 349-356

Babcock, G. T. and Callahan, P. M. (1983) Biochemistry fl, 2314-2319

Babcock, G. T., Callahan, P. M., Ondrias, M. R. and Salmeen, I. (1981)
Biochemistry 20, 959-966

Baker, G. M., Noguchi, M. and Palmer, G. (1987) J Biol Chem. 262, 595-
604

Beechey, R. B., Robertson, A. M., Holloway, C. T. and Knight, F. C. (1967)
Biochemistry 6, 3867-3879

 

Bill, K. and Azzi, A. (1982) Biochem. Biophys. Res. Commun. 106, 1203-
1209

165

166

Birchmeier, W., Kohler, C. E. and Schatz, G. (1976) Proc. Natl. Acad. Sci.
USA 72, 4334-4338

Bisson, R., Montecucco, C., Gutweniger, H. and Azzi, A. (1979) J. Biol.
Chem. 2_54, 9962-9965

Bisson, R., Jacobs, B. and Capaldi, R. A. (1980) Bioehemistfl _1_9, 4173-4178

Bisson, R., Steffens, G. C. M., Capaldi, R. and Buse, G. (1982) FEBS Lett.
M., 359-363

Bisson, R. Schiavo, G. and Papini, E. (1985) Biochemistry 2_4, 7845-7852

Bissogégté, Schiavo, G. and Montecucco, C. (1987) J. Biol. Chem. 2652. 5992-

Blair, D. F., Martin, C. T., Gelles, J ., Wang, H., Brudvig, G. W., Stevens, T.
H. and Chan, S. I. (1983) Chem. Scr. 21, 43-53

Blair, D. R, Ellis, W. R. Jr., Wang, H., Gray, H. B. and Chan, S. I. (1986a) L
Biol. Chem. 261, 11524-11537

Blair, D. F., Gelles, J. and Chan, S. I. (1986b) Biophys. J. 55, 713-733

Blumbezg, W. E. and Peisach, J. (1979) in Qfiochreme Qxidese (King, T. E.
., eds.), pp. 153-159, Elsevier/North Holland, Amsterdam

Bombelka, E., Richter, F.-W., Stroh, A. and Kadenbach, B. (1986) Biochem.
Biophys. Res. Commun. m, 1007-1014

Brand, M. D., Harper, W. G., Nicholls, D. G. and Ingledew, W. J. (1978)
FEBS Lett. 9_5, 125-129

Brandt, U., Schagger, H. and Von J agow, G. (1988) EBEC Reports 5, p. 89

Brautigan, D. L. Ferguson-Miller, S. and Margoliash, E. (1978) Methods
Enz o_ym_l_. 53, 128- 164

Brooks, S. P. J. and Nicholls, P. (1982) Biochim. Biophys. Aeta fig, 33-43
Brooks, S. P. J. and Suelter, C. H. (1986) I.nt J. Bio-Med. Cement. t1_9, 89- 99

Brunori, M., Sarti, P., Colosimo, A., Antonini, G., Malatesta, R, Jones, M.
G. and Wilson, M. T. (1985) EMBO J, _4, 2365-2368

Brzezinski, P., Thornstrom, RE. and Malmstrom, B. G. (1987) FEBS Lett.
l.9_49 1'5

Buge, U. and Kadenbach, B. (1986) Eur. J. Biochem 1_61, 383-390
Buse, G. Steffens, G. C. M, Steffens, G. J., Meinecke, L., Biewald, R. and

Erdweg, M. (1982) 2nd Eur. Bioenerg. Conf. Short Reports, L. B. T. M. -
C. N. R. 8., pp. 163-164, Villeurbanne, France

 

167

Buse, G., Meinecke, L. and Bruch, B. (1985) J, Inorg. Bioc5em. a, 149-153
Callahan, P. M. and Babcock, G. T. (1983) Biochemistg 22, 452-461
Capaldi, R. A. (1989) Annals N. Y. Acad. Sci., in press

Capaldi, R. A., Malatesta, F. and Darley-Usmar, V. M. (1983) Biochim.
Biophys. Acta 7_28, 135-148

Capaldi, R. A., Halphen, D. G., Zhang, Y.-Z. and Yanamura, W. (1988) L
Bioenerg. Biomembr. 20, 291-311

Carroll, R. C. and Racker, E. (1977) J. Biol. Chem. 2Q, 6981-6990
Casey, R. P. and Azzi, A. (1983) FEBS Lett 15_4, 237-242
Casey, R. P., Chappell, J. B. and Azzi, A. (1979a) Biochem. J. 182, 149-156

 

Casey, R. P., Thelen, M. and Azzi, A. (1979b) Biochem. Biophys. Res.
Commun. 81, 1044-1051

Casey, R. P., Thelen, M. and Azzi, A. (1980) J. Biol. Chem. 2_55, 3994-4000
Casey$3R P., Broger, C. and Azzi, A. (1981) Biochim. Biophys, Acta 8%, 86-

Casey, R. P., Ariano, B. H. and Azzi, A. (1982) Eur. J. Biochem. m, 313-318

Casey, R. P., O’Shea, P. S., Chappell, J. B. and Azzi, A. (1984) Biochim.
Biephys. Acta 765, 30-37

Caughey, W. S., Sm e, G. A., O’Keefe, D. H., Maskasky, J. E. and Smith,
M. L. (1975) J. iol. Chem. _25_0, 7602-7622

Caughey, W. 8., Wallace, W. J ., Volpe, J. A. and Yoshikawa, S. (1976) in _T_h_e
gnzkmes (Boyer, P. D., ed.) Vol. 13, pp. 299-344, Academic Press, New
or
Chan, S. H. P. and Freedman, J. A. (1983) FEBS Lett. _lfl, 344-348
Chan, S. H. P. and Tracy, R. D. (1978) Eur. J. Biochem 8_9, 595-605

Chancgh 173., Saronio, C. and Leigh, J. 8., Jr. (1975) J. Biol. Chem. 250, 9226-
9

 

Cornish-Bowden, A. (1976) in Principles pf Eprme Kinetic , pp. 168-193,
Butterworths Inc., Boston

Davies, H. C., Smith, L. and Wasserman, A. R. (1964) Biochim. Biophys.
Acta 85, 238-246

Deatherage, J. F., Henderson, R. and Capaldi, RA. (1982) J. Mol. Biol. 158,
501-514

168

De Vrij, W., Poolman, B., Konings, W. N. and Azzi, A. (1986) Methods
Enmol. _128, 159-173

Di B1233, V. and Prochaska, L. J. (1985) Arch. Biochem, Biephys. 24_3, 668-

Dixon, M. and Webb, E. C. (1964) Enryrpes, pp. 116-145, Longman, London

Downer, N. W., Robinson, N. C. and Capaldi, R. A. (1976) Biochemistry 15,
2930-2936

Dutch, R., 1e, R., Centeno, J ., Peiffer, W., Babcock, G. and Ferguson-
Miller, . (1987) in Plant Mitoch ndria (Moore, A. L. and Beechey, R.
B., eds.) pp. 301-304, Plenum, New York

Dutton, P. L. and Wilson, D. F. (1974) Biochim. Biophys. Aeta 848, 165-212

Dutton, P. L., Wilson, D. F. and Lee, C. P. (1970) Biochemistry 8, 5077-5082

 

Einarsdottir, O. and Caughey, W. S. (1985) Biochem. Biophys. Res.
Commun. 1_2_9, 840-847

Erecinska, M., Chance, B. and Wilson, D. F. (1971) FEB§ Lett. _1__6_, 284-286
Errede, B. and Kamen, M. D. (1978) Bioehemistry ll, 1015-1027

Errede, B., Haight, G. P., Jr. and Kamen, M. D. (1976) Proc. Natl. Acad. Sci.
[18A _7_8, 113-117

Eytan, G. D., Carroll, R. C., Schatz, G. and Racker, E. (1975) J. Biel. Chem.
250, 8598-8603

Ferguson, S. M. F., Estrada-0., S. and Lardy, H. A. (1971) J. Biol. Chem.
M. 5645-5652

Ferguson-Miller, S., Brautigan, D. L. and Margoliash, E. (1976) J. Biol.
Chem. 25_1, 1104-1115

Ferguson-Miller, S., Weiss, H., Speck, S. H., Brauti an, D. L., Osherhoff, N.
and Margoliash, E. (1979) in Cytocih. Jrne Oxi se (King, T. E. et al.,
eds.) PP. 281-292, Elsevier/North _ 'o‘ and Biomedical Press,
Amsterdam

 

 

 

 

 

 

 

 

Ferguson-Miller, S., Van Aken, T, and Rosevear, P. (1982) in Electron
Trgrsport and nygen UtilizeQ’on (Ho, C., ed.) Pp. 297-303, Elsevier
North Holland, Amsterdam

Fillingame, R. H. (1980) Annu. Rev. Bioehem, 49, 1079-1114

Finel, M. (1987) Biochim. Biophys. Aeta M, 174-179

Finel, M. and Wikstrom, M. (1986) Biochim. Biophys. Acta &, 99-108

Finel, M. and Wikstrom, M. (1988) Eur. J. Biochem. 176. 125-129

169

Freedlrggn, J. A. and Chan, S. H. P. (1984) J. Bioenerg. Biomembr. E. 75-

Frey, T. G., Kuhn, L. A., Leigh, J. S. Jr., Costello, M. J. and Chan S. H. P.
(1985) J. Inorg. Biochem. 23, 155-162

Fry, Méférd Green, D. E. (1980) Biechem. Biophys. Res. Commun. fl, 1238-

Fulleré,2§. D., Capaldi, R. A. and Henderson, R. (1979) J. Mol. Biol. 134, 305-

Fuller, S. D., Darley-Usmar, V. M. and Capaldi, R. A. (1981) Biochemistry
A), 7046-7053

Gai, W. Z., Vanaja, B. and Chan, S. H. P. (1987 ) Fed. Proc. fl, 2205

Gelleg,0 .; .,21316air, D. F. and Chan, S. I. (1986) Biochim. Bi0phys. Acta fl,

Georgevich, G., Darley-Usmar, V. M., Malatesta, F. and Capaldi, R. A.
(1984) Biochemistry 22, 1317-1322

Gibson, Q. H. and Greenwood, C. (1963) Biochem. J. 88, 541-554

Gibson, Q. H., Greenwood, C., Wharton, D. C. and Palmer, G. (1965) J. Biol.
Chem. 2_48, 888-894

Griffith, O. H., McMillen, D. A., Keana, J. F. W. and Jost, P. C. (1986)
Biochemistry 25, 574-584

Hall, J ., Moubarak, A., O’Brien, P., Pan, L. P., Cho, I. and Millett, F. (1988)
L Biol. Chem. 268, 8142-8149

Haltia, T., Puustinen, A. and Finel, M. (1988) Eur J. Biochem 112.. 543-546

 

 

Hartzell, C. R. and Beinert, H. (1974) Biochim. Biophys. Acta fl, 318-338
Hatefi, Y. (1978) Methods Enzymol. 53, 3-4
Hill, B. C. and Robinson, N. C. (1986) J. Biol. Chem. 26_1, 15356-15359

Hill, B. C., Greenwood, C. and Nicholls, P. (1986) Biochim. Biophys. Acta
85_3, 91-113

Hill, H. A. 0., Walton, N. J ., Whitford, D. and Coleman, J. O. D. (1985) Jr
Inom. Biochem, 23, 303-309

Hinkle, P. and Mitchell, P. (1970) J. Bioenerg. _1_, 45-60
Hinkle, P. C., Kim, J. J. and Racker, E. (1972) J. Biol. Chem. eg, 1338-1339
Holm, L., Saraste, M. and Wikstrom, M. (1987) EMBO J. 8, 2819-2823

170

Hon-nami, K. and Oshima, T. (1984) Biochemistg 23, 454-460

Hundt, E. and Kadenbach, B. (1977) Hopm-Seyler’s Z. Physio]. Chem. 35_8,
1309-1314

Jacobs, E. E. (1960) Biochem. Biophys. Res. Commun. _3, 536-539

J arausch, J. and Kadenbach, B. (1982)Hoppe-Sey1er’s Z. Physio]. Chem.
fl, 1133-1140

J arausch, J. and Kadenbach, B. (1985a) Eur. J. Biochem. $8, 211-217
J arausch, J. and Kadenbach, B. (1985b) Eur. J. Bioehem. L48, 219-225

Jones, M. G., Bickar, D., Wilson, M. T., Brunori, M., Colosimo, A. and Sarti,
P. (1984) Biochem. J. 228, 57-66

Kadenbach, B., J arausch, J ., Hartman, R. and Merle, P. (1983) Anal.
Biochem. £8, 517-521

Keilin, D. (1925) Proc. R. Soc. London Ser. 888, 312
Keilin, D. and Hartree, E. F. (1939) Proc. Roy. Soc. B127, 167-191

Kimelaberg, H. K. and Nicholls, P. (1969) Arch. Biochem. Biophys. fl, 327-
35

Koch, G. (1976) J. Biol. Chem. 251. 6097-6107
Kolarov, J ., Kuzela, S., Wielburski, A. and Nelson, B. D. (1981) FEBS Lett.

L38, 61-65
Konstantinov, A., Vygodina, T. and Andreev, I. M. (1986) FEBS Lett. 20_2,
229-234

Krab, K. and Wikstrom, M. (1978) Biochim. Biephys, Aeta 594, 200-214
Krab, K. and Wikstrom, M. (1987) Biochim. Biephys. Age 885, 25-39
Kuhn, L. A. and Leigh, J. S. Jr. (1985) Biochim. Biophys. Acta M, 351-361
Kuhn-Nentwig, L. and Kadenbach, B. (1985) Eur. J. Bioehem. E8, 147-158

Kuhn-Nentwig, L. and Kadenbach, B. (1986) J. Bioenerg. Biomembr. 18,
307-314

Laemmli, U. K. (1970) Nature (London) 221, 680-685

Lehninger, A. L., Reynafarje, B. and Costa, L. (1985) J. InorgJiochem. 28,
335-340

Lemberg, M R. (1969) Physiol. Reysr 48, 48-121

171

Lindszkgicgé S. and Coleman, J. E. (1973) Prec. Natl. Acad, $1, USA 18, 2505-

Ludwig, B. (1987) FEMS Microbiel. Qv. 48, 41-56
Ludwig, B. and Schatz, G. (1980) Proc. Natl. Acad. Sci. USA fl, 196-200

Ludwig, B., Downer, N. W. and Capaldi, R. A. (1979) Biochemistry 18, 1401-
1407

Maison-Peteri, B. and Malmstrom, B. G., manuscript in preparation
Malatesta, F., Georgevich, G. and Capaldi, R. A. (1983) in Structure and
Function of Membrane Protefis (Quagliariello, E. and Palmieri, F.,
ecls.) Vol. 6, pp. 223-235, Elsevier, Amsterdam

Malatesta, F., Antonini, G., Sarti, P. and Brunori, M. (1986) Biochem. J.
M: 569'572

Malmstrom, B. G. (1982) Ann. Rev. Biochem. 5_1, 21-59

 

 

 

Malmstrom, B. G. (1985) Biochim. Biophys. Acta 811. 1-12
Malmstrom, B. G. and Andreasson, L. (1985) J. Inorg Biochem. 23, 233-242

Martin,2C. T., Scholes, C. P. and Chan, S. I. (1988) J. Biol. Chem. _26_3, 8420-
84 9

Mason, T. L., Poyton, R. O., Wharton, D. C. and Schatz, G. (1973) J. Biol.
Chem. 248, 1346-1354

 

Matsuoka, M., Maeshima, M. and Asahi, T. (1981) J. Biochem. @, 649-655

Maurel, P., Douzou, P., Waldmann, J. and Yonetani, T. (1978) Biochim.
Biophys, Aeta 525, 314-324

McEwen, J. E., Ko, C., Kloeckner-Gruiessem, B. and Poyton, R. O. (1986) J_.
Biol. Chem. 2_6_1, 11872-11879

Merle, P. and Kadenbach, B. (1980) Eur. J. Biochem. 1Q5, 499-507
Merle, P. and Kadenbach, B. (1982) Eur. J. Biochem, Q5, 239-244
Meyer, T. J. and Taube, H. (1968) Iperg, Chem, 1, 2369-2379
Michel, H. (1983) Trends Biechem. 8g. 8, 56-59

Millett, F., de J ong, C., Paulson, L. and Capaldi, R. A. (1983) Bioehemistry
22, 546-552

Minnaert, K. (1961) Biochim. Biophys. Age 50, 23-34
Mitchell, P. (1961) Nature 181, 144-148

 

172

Mitchell, P. (1966) Chemiesmotic CeuEling in Oxidative and Photosypthetic
Phosphogletion, Glynn Researc Ltd., Bodmin, U. K.
Mitchell, P. (1987 ) FEBS Lett. 222, 235-245

Mitchell, P. and Moyle, J. (1983) FEBS Lett. _1_5_1_, 167-178

Mitchell, R, Mitchell, R., Moody, A. J ., West, I. C., Baum, H. and
Wrigglesworth, J. M. (1985) FEBS Lett. _188, 1-7

Mood , A. J ., Mitchell, R., West, I. C. and Mitchell, P. (1987) Biochim.
iophys. Acta &, 209-227

Moroney, P. M., Scholes, T. A. and Hinkle, RC. (1984) Biochemistry 28,
4991-4997

Moyle, J. and Mitchell, P. (1978) FEBS Lett. 88, 268-272

Muller, M. and Azzi, A. (1985) FEBS Lett. 184. 110-114

Muller, M., Labonia, N., Sclapfer, B. and Azzi, A. (1987) in Cytochrome
Systems (Papa, S., Chance, B. and Ernster, L., eds.) pp. 239-246,
Plenum, New York

Nalecz, K. A., Bolli, R., Ludwig, B. and Azzi, A. (1985) Biochim. Biophys.
Acta 808, 259-272

Naleczégl. J ., Casey, R. P. and Azzi, A. (1986) Methods Enrymol. gs, 86-
1

Naqui, A., Kumar, C., Ching, Y., Powers, L. and Chance, B. (1984)
Biochemistg 28, 6222-6227

N aqui, 121., Chance, B. and Cadenas, E. (1986) Ann. Rev. Bieehem, 55, 137-
16

 

 

Nicholls, D. G. (1982) Bioenergeties: An Ingreduction 59 the Chemiesmotic
Theog, Academic Press, New York

Nicholls, P. (1989) Annals N. Y. Acad. Sci., in press

Nicholls, P., Coo er, C. E. and Kjarsgaard, J. (1988) in Advances in
Membrane iochemistg end Bioenergetics (Kim, C. H., Tedeschi, H.,
IY)iwl:m, J. J. and Salerno, J. C., eds.) pp. 311-321, Plenum Press, New

or

Palmer, G. (1988) EBEC Reports é, p. 35

Palmer, G. Babcock, G. T. and Vickery, L. E. (1976) Proc. Natl. Acad, Sci.
USA 7_3_, 2206-2210

 

Papa, S. (1982) in Membranes ad Trensmrt (Martonosi, A. N., ed.) Vol. 1,
pp. 363-368, Plenum, New York

173

Papa, S. (1988) 1n @dases and Related Redox Sr stems (King, T. E., Mason,
HS and Morrison, M. eds.) pp. 707- 730, Alan R. Liss, Inc., New York

 

Papa, S., Capitanio, N., De Nitto, E. and Izzo, G. (1985) J. Inorg. Biochem.
5, 317-325

Penttila, T. (1983) Eur.J .Biochem. 133. 355- 361

Penttila, T. and Wikstrom, M. (1981)1n Vectorial Reactions 1n Electron and
Ion Transport 1n Mitochondria and Bacteri_a (Palmieri, F. et al., eds.)
pp. 71- 80, Elsevier/North Ho ‘and Biomedical Press, Amsterdam

Powell, G. L. and Abramovitch, D. A. (1985) Fed. Prop. .44, 478

 

 

 

 

 

Powell, G. L., Knowles, P. F. and Marsh. D. (1985) Biechim, BiOphys. Aeta
m, 191-194

Powers, L. (1982) Biochim. Biophys. Acta 683, 1-38

 

Powers, L. and Chance, B. (1985) J. Inorg. Biochem. fl, 207-217

Powe£%,5L4§8hance, B., Ching, Y. and Angiolillo, P. (1981) Biophys. J. _34,

Pressman, B. C. (1976) Annu. Rev. Biochem. 45, 501-530
Prochaska, L. J. and Fink, RS. (1987) J. BioenegBiomembr. 18, 143-166
Prochaska, L. J. and Reynolds, K. A. (1986) Biochemistry g, 781-787

Prochaska, L. J ., Bisson, R., Ca aldi, R. A., Steffens, G. C. M. and Buse, G.
(1981) Biochim. Biophys. cta fl, 360-373

Proteasi;é (365Wrigglesworth, J. M. and Nicholls, P. (1983) Biochem. J, 218,
1 -

Puettner, I., Solioz, M., Carafoli, E. and Ludwig, B. (1983) Eur. J. Biochem.
134, 33-37

Puettéi'zegélq Carafoli, E. and Malatesta, F. (1985) .1 Biol. Chem. 260. 3719-

Racker, E. (1972) J. Memb. Biol. 18, 221-235

Racket, E., Violand, B., O’Neal, 8., Alfonzo, M. and Telford, J. (1979) Arch.
Biochem. Biophys. L98, 470-477

Reynafa 'e, B., Alexander, A., Davies, P. and Lehninger, A. L. (1982) Proc.
Nat . Acad. Sci. USA 18, 7218-7 222

Rich, P. R. and Clarke, S. D. (1982) FEBS Lett. 148, 54-58
Rich, P. R., West, I. C. and Mitchell, P. (1988) FEBS Lett. 2_33, 25-30

174

Robinson, N. C. (1982) Biochemistry 21, 184-188

Robinson, N. C. and Capaldi, R. A. (1977) Biochemistry 1_6, 375-381
Robinson, N. C. and Wiginton, D. (1985) J. Inorg. Bioehem. fl, 171-176
Robinson, N. C., Strey, F. and Talbert, L. (1980) Biochemistry 18, 3656-3661

Rosevear, P., Van Aken, T., Baxter, J. and Ferguson-Miller, S. (1980)
Biochemistry fl, 4108-4115

Rubin, M. S. and Tzagoloff, A. (1978) Methods Enrym_51. 53, 73-79

Saraste, M., Penttila, T. and Wikstrom, M. (1981) Epr, J, Biechem. E,
261-268

Saraste, M., Raitio, M., J alli, T. and Peramaa, A. (1986) FEBS Lett. 208,
154-156

Sarti, P., Jones, M. G., Antonini, G., Malatesta, F., Colosimo, A., Wilson, M.
T. and Brunori, M. (1985) Proc. Nat_l. Acad. Sci. USA 82, 4876-4880

Schatz, G. and Mason, T. L. (1974) Annu. Rev. Biochem. 43, 51-87
Scholes, C. P. and Malmstrom, B. G. (1986) FEB8 Lett. 198, 125-129
Scott, R. A., Li, RM. and Chan, S. I. (1989) Annuls NY. Acad. Sci., in press

Sekuzu, I., Takemori, S., Yonetani, T. and Okunuki, K. (1959) J. Biochem.
(Tokyo) 48, 43-49

Shaughnessy, S. and Nicholls, P. (1985) Biochem. Biophys. Res. Commun.
1_28, 1025-1030

Sigel, E. and Carafoli, E. (1979) J. Biol. Chem. 254, 10572-10574
Sigel, E. and Carafoli, E. (1980) Eur. J. Biochem. 111, 299-306

Singh, A. P. and Nicholls, P. (1985) J. Biochem. Biophys. MethLds Q, 95-
108

Sinjorgo, K. M. C., Meijling, J. H. and Muijsers, A. O. (1984) Biochim.
Biophys. Age 782, 48-56

Sinjorgo, K. M. C., Steinebach, O. M., Dekker, H. L. and Muijsers, A. O.
(1986) Biochim. Biophys. Agra fl), 108-115

Sinjorgo, K. M. C., Durak, I., Dekker, H. L., Edel, C. M., Bieleman, A. H. L.,
Back, N. B. T., Hakvoort, T. B. M. and Muijsers, A. O. (1987a) mam
Biophys. Acta 883, 241-250

Sinjorgo, K. M. C., Durak, I., Dekker, H. L., Edel, C. M., Hakvoort, T. B. M.,
Van Gelder, B. F. and Muijsers, A. O. (1987b) Biochim. Biophys. Acta
Q3, 251-258

175

Sinjorgo, K. M. C., Hakvoort, T. B. M., Durak, I., Draijer, J. W., Post, J. K.
P. and Muijsers, A. O. (1987c) Biochim. Biophys. Acta 858, 144- 150

Slater, E. C. (1987) Eur. J. Biochem. 1%, 489-504

Smith, L. and Conrad, H. (1956) Arch. Bioehem. Biophys. 63, 403-413
Smith, L., Davies, H. C. and Nava, M. E. (1979) Bighemistry 18, 3140-3146
Solioz, M. (1984) Trends Biochem, Sci. 8, 309-312

Solioz, M., Carafoli, E. and Ludwig, B. (1982) J. Biol. Chem. 2&, 1579-1582
Sone, N. and Kosako, T. (1986) EMBO J. 5, 1515-1519

Sone, N. and Nicholls, P. (1984) Biochemistg 23, 6550-6554

Sone, N. and Yanagita, Y. (1982) Biochim. Biophys. Acta 882, 216-226
Sone, N. and Yanagita, Y. (1984) J. Biol. Chem. 2_5_9, 1405-1408

Sone,51;Ié, Egghida, M., Hirata, H. and Kagawa, Y. (1977) J, Biochem. 81,

Speck, S. H., Dye, D. and Margoliash, E. (1984) Proc. Natl. Acad. Sci. USA
fl, 347-351

Steffengs, (3.003. M., Biewald, R. and Buse, G. (1987) Eur. J. Biochem. & ,
2 5- .

Suarez, M. D., Revzin, A., Narlock, R., Kempner, E. S., Thompson, D. A. and
Ferguson-Miller, S. (1984) J. Biol. Chem. 258, 13791-13799

Taha, T. S. M. and Ferguson-Miller, S. (1987) Fed. Proc. 48, 1931

Taka12r17i'37'aé88é, Lindorfer, M. A. and Capaldi, R. A. (1987) FEBS Lett. m,

Thelen, M., Petrone, G., O’Shea, P. S. and Azzi, A. (1984) Biechim. Biophys.
AQLQ lfifi. 161-168

Thelen, M., O’Shea, P. S. and Azzi, A. (1985a) Biochem. J. 2_27, 163-167

Thelen, M., O’Shea, P. S., Petrone, G. and Azzi, A. (1985b) J. Biol. Chem.
&, 3626-3631

Thompson, D. A. (1984) Doctoral Thesis, Michigan State University, East
Lansing, Michigan

Thompson, D. A. and Ferguson-Miller, S. (1983) Biochemistry 22, 3178-3187

Thompson, D. A., Suarez-Villafane, M. and Ferguson-Miller, S. (1982)
Biophys. J. 31, 285-292

176

Thompson, D. A., Gregory, L. and Ferguson-Miller, S. (1985) J. Inorg.
Biochem. fl, 357-364

Thornstrom, P., Soussi, B., Arvidsson, L. and Malmstrom, B. G. (1984)
Chem. Scr. 2_4, 230-235

Thornstrom, P., Nilsson, T. and Malmstrom B. G. (1988) Biochim. Biophys.
Acta $5. 103-108

Tweedle, M. R, Wilson, L. J ., Garcia-Iniguez, L., Babcock, G. T. and Palmer,
G. (1978) J. Biol. Chem. 253, 8065-8071

Tzagoloff, A. (1982) Mitochondria, pp.124-127, Plenum Press, New York

Tzagotlgg', A. and Mac Lennan, D. H. (1965) Biochim. Biophys. Acta 88, 476-

van Buuren, K. J. H., Nicholls, P. and Van Gelder, B. F. (1972) Biochim.
Biophys. Acta 256, 258-276

van Gelder, B. F. and Beinert, H. (1969) Biochim. Biophys. Acta 188, 1-24
van Gelder, B. F., van Rijn, J. L. M. L., Schilder, G. J. A. and Wilms, J.
(1977) in Structure and Function of Energ-Transducing Membranes

(van Dam, K. and van Gelder, B. F., eds.) PP. 61-68, Elsevier/North-
Holland, Amsterdam

Vanneste, W. H. (1966) Biochemistry 5, 838-848

Van Verseveld, H. W., Krab, K. and Stouthamer, A. H. (1981) Biochim.
Biophys. Acta 635, 525-534

Veerman, E. C. I., Wilms, J. Cateleijn, G., and van Gelder, B. F. (1980)
Biochim. Biophys. Acta 5%, 117-127

Vik, S. B. and Capaldi, R. A. (1977) Biochemistry 1_6, 5755-5759

Vik, S. B., Georgevich, G. and Capaldi, R. A. (1981) Pres. Natl. Agd. Sci.
11S_A 18, 1456-1460

Warburg, O. (1924) Biochem. Z. 152, 479

 

Watts, A., Marsh, D. and Knowles, P. F. (1978) Biochem. Biophys. Res.
Cemmun. fl, 403-409

Wikstrom, M. (1977) NM £5, 271-273

Wikstrom, M. (1984) Nature 308. 558-560

Wikstrom, M. (1988) FEBS Lett. 231, 247-252
Wikstrom, M. (1989) Annals N. Y. Acad. Sci., in press

177

Wikstrom, M. and Casey, R. (1985) FEBS Lett. 1_83, 293-298
Wikstrom, M. and Krab, K. (1978) FEBS Lett. 81 , 8-14

Wikstrom, M. and Krab, K. (1979) Biochim. Biophys. Acta £8, 177-222
Wikstrom, M. and Penttila, T. (1982) FEBS Lett. M. 183- 189

Wiksgrgin, M. K. F. and Saari, H. T. (1977) Biochim. Biophys. Acta 4Q, 347-

Wikstrom, M. K. F., Harmon, H. J ., Ingledew, W. J. and Chance, B. (1976)
FEBS Lett. 85, 259-277

Wikstrom, M., Krab, K. and Saraste, M. (1981a) C chrome Oxidase: A
Sypthesi , Academic Press, London-New Yor

Wikstérggégg" Krab, K. and Saraste, M. (1981b) Annu. Rev. Biochem. 58,

Wikstrom, M., Saraste, M. and Penttila, T. (1985) in The Enzmes of
Biological Membranes (Martonosi, A. N., ed.) Vol. 4, pp. 111-148,
Plenum, New York

Wilms, J ., Van Rijn, J. L. M. L. and van Gelder, B. F. (1980) Biochim.
Biophys. Acta £3, 17-23

Wilson, D. F. and Dutton, P. L. (1970) Arch. Biochem. Biophys. 136. 583-584

Wilson, D. F., Lindsay, J. G. and Brocklehurst, E. S. (1972) Biochim.
Biophys. Acta 258, 277-286

Wilson, M. T., Lalla-Maharajh, W., Darley-Usmar, V., Bonaventura, J .,
Bonaventura, C. and Brunori, M. (1980) J. Biol. Chem. 255, 2722-2728

Winter, D. B., Bruyninckx, W. J ., Foulke, F. G., Grinich, N. P. and Mason,
H. S. (1980) J. Biol. Chem. 255, 11408-11414

Wrigglesworth, J. M. (1985) J. Inorg. Biochem. 2_3, 311-316
Wrigglesworth, J. M. and Nicholls, P. (1978) FEBS Lett, 81, 190-193

Yewey, G. L. and Caughey, W. S. (1987) Biochem. Biophys, Res. Commpn.
148, 1520-1526

Yonetani, T. (1960) J. Biol. Chem. &, 845-852

Yonetani, T. (1961) J. Biol. Chem. 238, 1680-1687

Yonetani, T. and Ray, G. S. (1965) J. Biol. Chem. 248, 3392-3398
Yoshida, T. and Fee, J. A. (1984) J. Biol. Chem. 258, 1031-1036

178

Yoshikawa, S., Tera, T., Takahashi, Y., Tsukihara, T. and Caughey, W. S.
(1988) Proc. Natl. Acad. Sci. USA 85, 1354-1358

Yu, C.-A., Yu, L. and King, T. E. (1975) J . Biol. Chem. 258, 1383-1392
Zborowski, J. and Robinson, N. C. (1985) Fed. Proc. 44, 478

Zhanngé-gézGeorgevich, G. and Capaldi, R. A. (1984) Bigchemistry £1,
- 1

Zhangbgg-Zlégaindorfer, M. A. and Capaldi, R. A. (1988) Biochemistry fl,