1
Functional and Time-Resolved Structural Studies of
Bacteriorhodopsin
Mutants Following Expression in
Halobacterium
halobium
by
Ramin Mollaaghababa
B.S., Chemistry
State University of New York at Stony Brook
Submitted to the Department of Chemistry in Partial Fulfillment of
the Requirements for the Degree of
DOCTOR OF PHILOSOPHY
at the
Massachusetts Institute of Technology
June
1995
© 1995 Massachusetts Institute of Technology
Signature of Author .....
Department of Chemistry
February 15, 1995
Certified
by.
...............
....
-.-.....
H. Gobind Khorana, Alfred P. Sloan Professor of Biology
and Chemistry Emeritus
Departments of Biology and Chemistry
Thesis Advisor
Accepted by .............................
,
Dietmar
Dietmar Seyferth
Seyferth
Chairman, Departmental Committee on Graduate Studies
MASSACHJJS.S
,
ttl
JUN 12 1995
LIRARIES
2
This doctoral thesis has been examined by a Committee of the
Department of Chemistry as follows:
Lawrence J. Sterr.
Professor
.............................
Chairman
H. Gobind Khorana
Professor.............................
Thesis Supervisor
Phillip A. Sharp
Professor........................
,/
I-
3
Acknowledgments
As I began to write this section, so many people came to my
mind without whom the completion of this thesis would not have
been possible. I would like to start by mentioning my deep gratitude
to my advisor, Professor H. Gobind Khorana, for accepting me into his
research group to complete my graduate studies. My exposure to
such highly stimulating research environment combined with
Gobind's invaluable guidance culminated in an important maturation
I would also like to
experience, both scientifically and personally.
thank Professor Uttam L. RajBhandary for his help and his interest in
my research throughout my graduate work. The discussions with
Tom during the Monday seminars and other occasional scientific
exchanges
with him were always illuminating.
The insightful editorial comments of my wonderful wife,
Florence Davidson, and her help with preparing the figures improved
the quality of the thesis tremendously. Her unparalleled support
provided me with the time and focus which was needed to finish this
thesis.
Moreover, throughout the past several years, my exposure to
Florence's great scientific talent and skills and also her unique
personal qualities have been among the main reasons for my
scientific and personal growth. The ability to perform many of the
experiments described in this thesis was a direct result of
collaborative efforts with the laboratories of Wayne L. Hubbell at
UCLA and Maarten P. Heyn at Freie University in Berlin. I am very
grateful to both Wayne and Maarten for allowing such flourishing
The names of the very able
collaborations to be established.
researchers from these laboratories who were involved with the
presented work are listed in the respective chapters throughout the
thesis.
During my stay at MIT, I have had the privilege of interacting
with, and learning from, many experienced scientists.
In particular, I
would like to thank Mark (Captain Kirk) Krebs for teaching me,
among other things, the art of molecular biology. His quick wit was
always refreshing, and our long collaboration was extremely fruitful.
My benchmates Sriram Subramaniam, Duncan Greenhalgh, and Tom
Marti were always available to teach, guide, criticize, and share good
times.
I hope Sriram's "hydroxylRamin" experiments will always
come out beautifully. Many thanks go to Kevin Ridge for sharing
Judy Carlin
with me his technical expertise in protein biochemistry.
to
occasions
duty
on
numerous
call
of
the
went above and beyond
provide me with her excellent administrative assistance. I am also
very much indebted to Professor Jonathan King and Professor Chris
4
Kaiser of the MIT biology department for always finding time to
teach me science and to provide invaluable advice throughout my
graduate career.
Twelve years ago, I left Iran during the war time in the
pursuit of a dream, one which has come largely true. I am deeply
grateful for the help and encouragement of my parents, teachers, and
friends back home which made it possible for me to embark on my
journey. Along the way, I have had the amazing fortune of crossing
paths with so many kind, interesting, and inspirational people. The
Ascher family, which I consider my own, provided me with a home
away from home during my stay in Paris when I finished my last
year of high school studies. During my college years at Stony Brook,
it would have been impossible to overcome the financial difficulties
and the pain of the abrupt loss of my mother without relying on
friends like Bagher Bahavar, Hamid Sodeifi, Hadi Sadrosadat, Richard
Olness, Hamed Eslami, and Elaine Scanlon, to name a few. My
brother, Reza, was the main reason I was able to overcome these
I have always been
difficulties and finish my undergraduate studies.
guided in my life not only by taking note of his stellar academic
accomplishments but also of his compassion and kindness. He
greatly sacrificed of himself in order to ensure that I can complete
my undergraduate program and for that I will always be grateful to
him. I could have never wished for a more wonderful brother. I
also benefited tremendously from the guidance and caring of the
faculty members in the Chemistry Department at Stony Brook. In
particular, I would like to acknowledge Robert Kerber, Charles
Springer, Richard Porter, Steven Rokita, Sei Sujishi, Gerald Whitten,
and Glenn Prestwich.
I dedicate this thesis to the memory of my mother, Afagh
Hakami, who taught me the true value of education.
Functional and Time-Resolved Structural Studies of
Bacteriorhodopsin Mutants Following Expression in Halobacterium
halobium
by
Ramin Mollaaghababa
Submitted to the Department of Chemistry on March 13, 1995 in
partial fulfillment of the requirements for the degree of Doctor of
Philosophy.
Abstract
Bacteriorhodopsin (bR) is a light-driven proton pump which is
found in the plasma membrane of the archaeon Halobacterium
halobium. The extrusion of protons into the extracellular space by
bR is coupled to ATP synthesis and several other cellular processes,
Previously, biochemical and
such as amino acid transport.
biophysical analyses of a large number of site-directed mutants of bR
have been performed following the reconstitution of purified mutant
proteins in mixed lipid/detergent micelles. These studies identified
the residues R82, D85, D96, and D212 as the most important for the
function of bacteriorhodopsin.
For comparison, mutant proteins containing site-directed
substitutions at these sites were produced in the purple membrane
form following the development of a homologous expression system.
The substitutions included different neutral replacements and the
exchange of each of the aspartic acid residues with glutamic acid.
Proton translocation assays, photovoltage studies, and photocycle
measurements of the purified mutant proteins in the purple
membrane form confirmed the functional importance of the residues
R82, D85, D96, and D212.
Time-resolved structural studies of bR were also performed
using a series of cysteine substitution mutants purified in the purple
membrane form and subsequently modified with specific structural
probes. Time-resolved electron paramagnetic resonance (EPR)
studies of spin-labeled mutants allowed an analysis of the localized
conformational changes during the bR photocycle in real time. These
measurements indicated that sequential molecular motion occurs in
6
separate localized regions of bR during the formation and decay of
the M intermediate. These light-dependent structural changes are
reversible. The largest movement was observed near site 101
during M decay.
Based on a structural model of spin-labeled bR, the
observed change was interpreted as light-dependent motion of the CD and/or E-F interhelical loops at the cytoplasmic side.
Consistent
with this conclusion, preliminary time-dependent X-ray analysis of
at site 101 indicates a
bR modified by p-chloromercuribenzoate
during
the formation of the M
the
mercury
label
displacement of
state. The extent of this motion is roughly 5-7 angstroms.
Thesis Supervisor: Professor H. Gobind Khorana
Title: Alfred P. Sloan Professor of Biology and Chemistry Emeritus
7
Table
of
Contents
Acknowledgments
3
Abstract
Table of Contents
List of Figures and Tables
5
7
11
Chapter One: Introduction
15
1.1 General Properties of Halobacterium
halobium and
the Purple Membrane
1.2 Structural Features of Bacteriorhodopsin
1.2.1 The Primary Structure of Bacteriorhodopsin
1.2.2 The Secondary Structure of Bacteriorhodopsin
1.2.3 The Tertiary Structure of Bacteriorhodopsin
1.2.4 Chromophore Properties
1.3 Photochemical Properties of Bacteriorhodopsin
1.3.1 Bacteriorhodopsin Photocycle
1.3.2 Photocycle Intermediates
1.3.3 Chromophore Changes Associated with the
Photochemistry of Bacteriorhodopsin
1.3.4 Protein Changes Associated with the
Photochemistry of Bacteriorhodopsin
1.4 The Mechanism of Proton Translocation
1.4.1 Deprotonation of the Schiff Base
1.4.2 Reprotonation of the Schiff Base
Chapter Two: Homologous Expression of
Bacteriorhodopsin
Mutants in Halobacterium
16
18
18
18
19
24
24
24
25
27
30
31
31
34
36
halobium
2.1 Background
36
2.2 Development of a Homologous Expression System for
Bacteriorhodopsin Mutants
39
Chapter Three: Characterization of Bacteriorhodopsin
Mutants which Affect Proton Translocation Activity
43
3.1 Introduction
43
3.2 Experimental Procedures
3.2.1 Bacterial Strains and Culture Conditions
3.2.2 Plasmid Construction
3.2.3 Purification and Analysis of Constructed Plasmids
45
45
45
49
8
3.2.4 H. halobium Transformation and Isolation of
Recombinants
3.2.5 Purification of Wild-Type and Mutant
Bacteriorhodopsin in Purple Membrane Form
3.2.6 Proton Translocation Assays
3.2.7 Time-Resolved Absorption Spectroscopy
Studies
3.2.8 Photovoltage Measurements
3.3 Results
3.3.1 Construction and Analysis of Expression
Vectors
3.3.2 Gene Replacement at the Chromosomal bop
50
51
52
53
56
56
56
58
Locus; H. halobium Transformation and
Isolation of Recombinants
3.3.3 Isolation of MPK40 Recombinants Carrying
Mutations in the bop gene
3.3.4 Purification of Wild-Type and Mutant Proteins
and their Characterization by UV/Visible
Spectroscopy and Proton Translocation Assays
3.3.5 Measurements of Photocycle Kinetics
3.4 Discussion
3.4.1 Gene Replacement in H. halobium
3.4.2 Phenotypes of Proton Translocation Mutants
61
65
68
78
78
79
Chapter Four: Structural Studies of Bacteriorhodopsin:
I. Preparation and Characterization of Cysteine
Substitution
81
Mutants
4.1 Introduction
81
4.2 Experimental Procedures
4.2.1 Bacterial Strains and Culture Conditions
4.2.2 Plasmid Construction
4.2.3 Purification and Analysis of Constructed Plasmids
4.2.4 H. halobium Transformation and Isolation of
Recombinants
4.2.5 Purification of Mutant Proteins in the Purple
Membrane Form
4.2.6 Derivatization of Cysteine Mutants with
Fluorescein and Flash Spectroscopy Analysis
4.2.7 Proton Release Measurements
4.3 Results
4.3.1 Isolation of Mutant Recombinants
84
84
84
85
85
86
86
86
87
87
9
4.3.2 Purification of Mutant Proteins in the purple
Membrane Form
4.3.3 Derivatization of Cysteine Mutants with
Fluorescein and Flash Spectroscopy Analysis
4.3.4 Proton Release Kinetics
4.4 Discussion
Chapter
Five:
Structural
Studies
92
92
97
101
of Bacteriorhodopsin:
II. Time-Resolved Electron Paramagnetic
Studies of Spin-Labeled Mutants
106
Resonance
5.1 Introduction
106
5.2 Experimental Procedures
5.2.1 Derivatization of Mutant Proteins
5.2.2 Instrumental Design for Time-Resolved Electron
Paramagnetic Resonance Studies
5.2.3 Flash Spectroscopy: Monitoring Absorbance
Changes at 410 nm and 570 nm
107
107
108
5.3 Results
115
5.3.1 Time-Resolved EPR Measurements
5.3.2 Flash Spectroscopy: Monitoring of Absorbance
Changes at 410 nm and 570 nm
5.4 Discussion
115
123
115
127
Chapter Six: Structural Studies of Bacteriorhodopsin:
III. Time-Dependent
X-ray Measurements of a
Cysteine-Containing
Mutant and Its
Mercury
130
Derivative
6.1 Introduction
130
6.2 Experimental Procedures
6.2.1 Derivatization of the Purified Mutant Proteins
131
131
with p-Chloromercuribenzoate
(PCMB)
6.2.2 Derivatization of A160C and G231C Mutants with 133
' 4C-Labeled Methylmercury iodide ( 4C-CH 3HgI)
6.2.3 X-ray Diffraction Measurements
6.2.4 Data Analysis
134
136
6.3 Results
137
6.3.1 Derivatization of the Purified Mutant Proteins
with PCMB and '4C-CH 3HgI
6.3.2 X-ray Diffraction of D96A/V10O1Cand
D96A/V101C-MB
6.4 Discussion
137
143
149
10
References
Curriculum
152
Vitae
165
11
List of Figures and Tables
Figure 1.1: Secondary structure model of bacteriorhodopsin
Figure 1.2: A helical wheel projection model of
bacteriorhodopsin viewed from the cytoplasmic
side of the membrane
Figure 1.3: A three dimensional model of a bacteriorhodopsin
monomer
20
Figure 1.4: The structure of the all-trans retinal chromophore
23
of bacteriorhodopsin
Figure 1.5: A model of the bacteriorhodopsin photocycle
Figure 2.1: Structure of mevinolin
Figure 2.2: Alignment of the native and synthetic bop gene
sequences
Figure 3.1: Locations of the residues Arg82, Asp85, Asp96 and
Asp212
Figure 3.2: Construction of mevinolin-resistance plasmids and
bop expression vectors
Figure 3.3: DNA sequence of the synthetic bacterio-opsin gene
Figure 3.4: A diagram of the photocycle apparatus
Figure 3.5: Sequence of the region flanking the 3' end of the
bop gene in the 1.6 kbp fragment from pBRB01
21
22
28
40
41
44
46
47
54
57
Figure 3.6: Southern blot analysis of MevR transformants
Figure 3.7: Southern blot analysis of MPK5 transformants and
59
60
recombinants
Figure 3.8: Products of homologous recombination in MPK5
transformed with pMPK62 derivatives
Figure 3.9: Structure of pMPK66
62
Figure 3.10: Southern blot analysis of the bop deletion strain
64
63
and Pum+ recombinants
Figure 3.11: SDS-PAGE analysis of wild-type and mutant
Figure 3.12:
Figure 3.13:
Figure 3.14:
Figure 3.15:
Figure 3.16:
bacteriorhodopsin proteins
Spectra of PM purified from wild-type and
mutant recombinants
Photocycle kinetics of the bacteriorhodopsin
mutant D85N at pH 7.3
Photocycle kinetics of the bacteriorhodopsin
mutant D85N at pH 11.0
Photovoltage traces of the bacteriorhodopsin
mutant D85N at pH 2.4, pH 6.7, and pH 10.8
The rise and decay of the M intermediate in the
bacteriorhodopsin
pH 7.3, and pH 8.5
mutant R82A at pH 5.0,
66
67
69
71
72
73
12
Figure 3.17: Photocycle kinetics of the
mutant D96N at pH 7.3
Figure 3.18: Photocycle kinetics of the
mutant D212N at pH 7.3
Figure 3.19: Photovoltage traces of the
mutant D212N at pH 2.8,
bacteriorhodopsin
75
bacteriorhodopsin
76
bacteriorhodopsin
pH 5.1. pH 8.9, and
77
pH 11.0
Figure 4.1: Secondary structure model of bacteriorhodopsin
Figure 4.2:
Figure 4.3:
Figure 4.4:
Figure 4.5:
Figure 4.6:
showing the cysteine substitution sites
Southern blot analysis of recombinants of G72C,
V1O1C, Q105C, and G72C/D85N
Southern blot analysis of recombinants of A103C
and G231C
Southern blot analysis of recombinants of S35C,
A160C, R82A/G72C, R82A/A160C, and
D96A/V101C
Southern blot analysis of recombinants of S35C
and V130C
SDS-PAGE analysis of PM purified from the
wild-type and cysteine mutant strains
Figure 4.7: UV/visible absorption spectra of dark-adapted
83
88
89
90
91
93
94
and light-adapted purified PM for S35C, G72C,
V1OIC, and D96A/V10O1Cmutants
Figure 4.8: UV/visible absorption spectra of dark-adapted
95
and light-adapted purified PM for A103C, Q105C,
V130C, and A160C mutants
Figure 4.9: UV/visible absorption spectra of dark-adapted
Figure 4.10:
Figure 4.11:
Figure 4.12:
Figure 4.13:
and light-adapted purified PM for G231C mutant
Flash-induced absorption changes at 410 nm for
wild-type bR and G72C-MF, A160C-MF, and
G231C-MF
Comparison of the rise and decay of the M
intermediate with the kinetics of proton release
for G72C-MF, A160C-MF, and G231C-MF
Arrhenius plot of the time constants for the
formation of the M intermediate and for the
proton release for G72C-MF
Cartoon of proton flow along the surface and
around the edge of the purple membrane
Figure 5.1: Structures of methanethiosulfonate derivatives
used for spin labeling of bR mutants
96
98
100
102
104
108
13
Figure 5.2: A block diagram of the apparatus used for the
time-resolved EPR studies of spin-labeled bR
mutants
Figure 5.3: A diagram showing the relative phases of the
pulse generator signal and the reference signal
in the time-resolved EPR apparatus
Figure 5.4: The photocycle of bacteriorhodopsin
Figure 5.5: Ground state and difference EPR spectra for
109
113
114
116
C1O1R1, C101R2, G72R1, and C105R1
Figure
5.6: Ground state and difference EPR spectra for
C35R1, C103R1, and C130R1
Figure 5.7: Ground state EPR spectra and light-dependent
EPR transient changes for D96A/C1O1R1, C160R1,
and C231R1
Figure 5.8: Normalized optical densities and EPR spectral
amplitudes as a function of time for ClOR1 and
C101R2
Figure 5.9: The structure of bR backbone with the side chain
R1 shown as a stick model at positions 72, 101, and
117
118
119
121
105
Figure 5.10: Normalized optical densities and EPR spectral
amplitudes as a function of time for C35R1, and
C130R1
Figure 5.11: Normalized optical densities and EPR spectral
amplitudes as a function of time for D96A/C10O1R1,
and C160R1
Figure 5.12: Normalized optical densities and EPR spectral
amplitudes as a function of time for C103R1, and
124
125
126
C231R1
Figure 6.1: A block diagram of the X-ray apparatus
Figure 6.2: UV absorbance spectra of PCMB and cysteine-MB
Figure 6.3: Drivatization of D96A/V10O1C mutant bR with
excess PCMB
Figure 6.4: Derivatization of D96A/V101C mutant bR with
substoichiometric amounts of PCMB
Figure 6.5: Time courses of derivatization of A160C and
D96A/V101 C mutant proteins with PCMB
135
138
139
Figure 6.6: X-ray diffraction patterns of D96A/V101C lattice
144
140
142
in the ground state, with and without mercury
labeling
Figure 6.7: X-ray diffraction patterns of D96A/V101C lattice
in the M state, with and without mercury labeling
145
14
Figure 6.8: Location of the mercuribenzoate label in
D96A/V 10C-MB
Figure 6.9: Location of the mercuribenzoate label in
A103C-MB
Figure 6.10: Detection of a conformational change in bR upon
M formation by X-ray diffraction analysis of the
mercuribenzoate-labeled D96A/V101C mutant
protein
Table 3.1: Spectral characteristics and proton pumping of
wild-type and mutant bR in the membrane form
Table 4.1: Time constants and amplitudes for the kinetics of
the M intermediate and proton release, with
fluorescein attached to various positions on bR and
with pyranine in the bulk phase
147
148
150
163
164
15
Chapter
One
Introduction
Overview
The ability to harvest light energy and transform it into
chemical energy is essential to the survival of many living organisms.
Blue-green algae and plants carry out this task by a well-known
process of photosynthesis using their related photosynthetic reaction
centers. By contrast, certain species of halophilic bacteria achieve
the same goal by an alternative photosynthetic
process.
They rely on
the light-dependent protein bacteriorhodopsin for production of a pH
gradient across the cell membrane, which is used by the bacteria for
ATP synthesis (1).
Bacteriorhodopsin has been the subject of intense biochemical
and biophysical research over the past two decades (2, 3,4, 5) . By
far one of the best characterized membrane proteins to date, it was
first discovered by Oesterhelt and Stoeckenius in 1971 (6). These
investigators demonstrated that bacteriorhodopsin undergoes a
photocycle upon light absorption, resulting in net transport of
protons from the cytoplasmic side of the cell to the extracellular
space. Over the years, the main question spawning much of the
investigative effort on bacteriorhodopsin has been the following: how
does it succeed in generating an electrochemical gradient across the
cell membrane following absorption of light?
Although great
progress has been made in identifying the key functional residues
involved in the proton transport mechanism (5) and also in
understanding the structure and photochemical properties of
bacteriorhodopsin
(7, 8), the above question remains to be answered
in specific details. This is especially the case in regard to the nature
of structural transitions that bacteriorhodopsin undergoes in order to
achieve its function.
In this chapter, a description of bacteriorhodopsin is presented,
with focus on its known structural and functional features. Several
comprehensive reviews on bacteriorhodopsin are available (e.g., 3, 4,
9, 10). A number of reviews emphasizing the use of specific
biochemical and biophysical techniques in studying
bacteriorhodopsin
15).
have also been published (e.g., 5, 8, 11, 12, 13, 14,
16
1.1 General Properties
Purple
Membrane
of Halobacterium
halobium
and the
Bacteriorhodopsin is a light-driven proton pump which exists
within the plasma membrane of the archaeon Halobacterium
halobium' (16).
It is organized into a two-dimensional crystalline
lattice known as the purple membrane (17). The name
bacteriorhodopsin derives from its many similarities with the visual
pigments, or opsins, of animal photoreceptor cells. H. halobium
belongs to the phylum archaeobacteria.
As the name implies, H.
halobium are halophilic and depend on the presence of a high
concentration of salt (>3 M NaCl) for survival and growth.
Consequently, they may be found in salt marshes, salterns, and salt
lakes (bacteriorhodopsin molecules present in the halophilic bacteria
produce the deep purple color of many salt lakes, familiar to air
travellers to the San Francisco bay area). H. halobium cells lyse
easily when they are placed in a milieu with only moderate or low
ionic strength, partly due to a change in osmotic pressure (18). The
major cell surface component of H. halobium is a glycoprotein which
forms a regular hexagonal lattice (19). This rigid matrix is believed
to be responsible for the maintenance of the characteristic rod shape
of halobacteria in their natural environment (20). In addition to the
purple membrane, these bacteria also contain what are referred to as
the red membrane
and the brown membrane.
red membrane is lycopene (21).
cytochrome
The pigment of the
The brown membrane contains
b and some bacteriorhodopsin
(22).
If the cells are
grown in the presence of nicotine in order to inhibit retinal synthesis,
all the bacterio-opsin produced by the cell is found in the brown
membrane (22). The brown membrane isolated from "nicotine cells"
does not show any crystalline order (23, 24). However, upon the
addition of retinal to the brown membrane and subsequent
reconstitution of bacteriorhodopsin, the formation of a crystalline
lattice similar to that of the purple membrane takes place (22, 23).
This demonstrates the importance of retinal in aiding the formation
of the purple membrane lattice. However, because the crystalline
structure formed from the reconstituted brown membrane does not
show the same level of order as in the purple membrane, factors
'This organism is also refered to as Halobacterium
salinarium
17
other than the presence of retinal must also play a part in
determining the organization of the purple membrane.
When the amount of oxygen available to H. halobium falls
below the level required for oxidative phosphorylation, an induction
of purple membrane biosynthesis takes place, in order to allow
production of energy by harvesting light (3). In nature, such a
phenomenon often takes place in hot salt water. The synthesis of
purple membrane is also induced by light.
As much as 75% of the
cell surface can be covered by this specialized membrane under
optimum conditions of synthesis (25). The purple membrane patches
have an average diameter of 0.5 [im and are usually round or slightly
oval in shape (17). The thickness of the purple membrane is about
50 angstroms. It has been shown that, depending on their age, the
purple membrane patches can vary in size and shape (26). In the
early stages of cell growth, these patches are small and separate, and
as the culture grows toward saturation they fuse to form larger
sheets (26). The two faces of purple membrane are asymmetric, so
that they can be distinguished from each other unambiguously by
examination with an electron microscope after metal decoration (27).
A diether analog of phosphatidylglycerophosphate
(PGP) and
glycolipid sulfate are the major components of the lipids in the
purple membrane (21).
Fortunately, purple membrane, which contains
bacteriorhodopsin as its only protein component, can be prepared in
highly pure form subsequent to cell lysis (18). It shows remarkable
stability in that it can be stored at room temperature indefinitely in
the absence of bacterial growth. Because the protein to lipid ratio is
unusually high in the purple membrane (3:1 protein:lipid), aqueous
suspensions of it show a low level of light scattering and are
therefore amenable to spectroscopic studies. In addition, suitable
detergents capable of complete solubilization of the purple
membrane are available (28). Bacteriorhodopsin solubilized in
detergent can be removed from the lipids of purple membrane by
gel filtration and be studied in monomeric form (28, 29).
Existence
in the lattice form of the purple membrane is not a functional
requirement for bacteriorhodopsin (29, 30). However, it is important
to point out that differences in structural stability and photocycle
kinetics do exist between solubilized bacteriorhodopsin and the
purple membrane (e.g., see Chapter 3, section 3.3.5). For instance,
the solubilized protein is much more susceptible to an irreversible
loss of the chromophore, a process referred to as bleaching.
18
1.2
1.2.1
Structural
The
Features
of
Bacteriorhodopsin
Primary Structure
of Bacteriorhodopsin
The polypeptide chain of fully processed bacteriorhodopsin
contains 248 amino acid residues and has a molecular weight of
26,000.
The sequence of bacteriorhodopsin has been obtained both
at the protein (31) and at the DNA level (32). The protein is
modified post-translationally by the removal of an N-terminal presequence which is 13 residues long (33). Additionally, an aspartic
acid residue is removed from the C-terminus, and the N- and C-
termini are covalently modified to yield pyroglutamic acid and
acetyl-serine respectively.
1.2.2
The
Secondary
Structure
of
Bacteriorhodopsin
Hydropathy plots of the obtained sequence for
bacteriorhodopsin indicate the presence of seven putative
transmembrane regions (34). The most compelling evidence in
support of this deduction has come from extensive electron
cryomicroscopic work of Henderson and colleagues (35). Their
diffraction patterns, obtained from the purple membrane at various
tilt angles, indicate the presence of seven rod-shaped high density
regions which are approximately 35-40 angstroms long and span
across the membrane.
Currently, projection maps with an in-plane
resolution of 2.8 angstroms and a resolution of about 6 angstroms
normal to the plane of the membrane have been obtained by the
technique of electron microscopy.
The improvement in resolution
has been obtained by refinements in image processing, sample
preparation, and data analysis (36, 37, 38). Several lines of work
have identified the seven transmembrane regions of
bacteriorhodopsin as a-helices.
Circular dichroism measurements
show the presence of about 70% a-helix
(39, 40, 41), displaying high
correlation with the total percentage of the protein which is confined
to the membrane region (70-80%), as observed by electron
microscopy
(42).
In addition, FTIR and Raman spectroscopic studies
of bacteriorhodopsin show that it is predominantly a-helical,
based on its vibrational spectrum in the amide region (43, 40, 44).
More recently, direct evidence for the presence of a-helical
within helices D and E of bacteriorhodopsin
motifs
has been obtained by
19
electron paramagnetic resonance studies of a series of spin-labeled
cysteine substitution mutants (45, 46). Furthermore, these studies
have provided more precise information on the boundaries of
membrane-embedded helix E on the extracellular side, and helices C
and D on both sides of the membrane (45, 47). A current secondary
structural model of bR derived from a synthesis of the available
strucrtural data is shown in Figure 1.1.
1.2.3 The Tertiary
Structure
of Bacteriorhodopsin
The helices observed in the electron diffraction pattern of bR
correspond to the helices in the polypeptide sequence as proposed by
Engelman and colleagues (48). Thus, helices A through G correspond
to the density positions numbered 1, 7, 6, 5, 4, and 2, respectively
(see Figure 1.2).
Neutron diffraction studies which reveal that
helices A and B correspond to densities 1 and 7 (49) are also in
agreement with this helix assignment. Furthermore, proteolysis (50),
and antibody (51) experiments along with energetics calculations
(52) provide a sequence for the linkage of helices compatible with
that deduced from structural studies.
As can be observed from Figure 1.3, the seven membraneembedded helical segments of bR come together in a bundle such
that a juxtaposition of helices A and G occurs, with the helices tilted
0-20 ° relative to the plane of the membrane and to each other. In
Henderson's model of bR, the orientation of helix D is the most
uncertain because it does not contain any aromatic residues such as
those that were used to discern the orientations of the other helices.
The orientation of helix D was therefore based on the results of site-
directed mutagenesis at site 115. Recently, the techniques of sitedirected mutagenesis and electron paramagnetic resonance were
combined to provide much greater information on the orientation of
this helix (46). The results of the spin-labeling experiments are
consistent with Henderson's proposal. In addition, the proposed
orientation of helix E is in aggreement with spin-labeling
experiments performed on bR mutants introduced in this helix (45).
Site-specific mutagenesis studies have suggested a bR structure
consistent with Henderson's model, in terms of whether certain
residues in the helices face the interior of the protein, or whether
they point outwards towards the lipid chains (5).
20
k) (V
-l
(Ve
ci
or)
y0
C
m
O
_
r
-'
0~C DII
>
H'
r-C
__
Hr
0<
i
ms
(O
:;
>_M
H
_
>
m
(V .
>
D
__
m
,, .
o
,
^
r0
I
--i
_I
G
< <
U)
C
i
;5
"-
''
mn
r
-- -
--
<
<
5'
<.<
D
n~
_
"1
0
FN
o7 r
r
o0
ct
-
_
-
>
· ~
Hr
Q
%r
Cfj
I
<
0
<
-<
-
0-4 >
_l
.
>:
F
2.
>
-
>
I,
j
.'
3
r
,-
r
. D_'
-~
-'
)I
T
r
!
)C
m
_
_
Z <
U) <
-C
<
OI
-i C C
U) rr '
<
p-
QC)
>= i
t
_-
I
r
C K
K
G
1
o -I
O 0
NG
Go C
7l
0
c-
t
M
-J
r
<K
U)
7I
(n
<
_-
N
(V~
_
N
r
c DK C
c
C/)
0
-l
rl
> N
>U) G)
P
r
D
>
N
Ct
MU)
Ir>i
-m
2 1
~-
z
c
° t~ D
OtV
o
_.o
_F
c5 OJ'
' _
<.oc r Cq
_.
C)
cr 05
C)C
C)r(
C) ~
C)
h
e_~.
0
C)
-,
C0P
"'(:
_
,
C!
)
O
_e
TC
C,
cn
CD
ct:
C)
_. ,.,...5-
crQC
o
yD
uz
C
D
_
C)
r ,C3 ,--C
LG
n
OQ
cr
p:
acz
o
C;
0 E3
O
n. _
=j
OCt
~
.
_
CD
O:
C_
-~
,_
22
G
Figure 1.3.
A three dimensional model of a bacteriorhodopsin
monomer. showing the packing of the transmembrane ca-helical
segments of the protein (taken from reference 42i.
23
The helical bundle of bR forms a pocket for the retinal
In the light-adapted state (bR56 8), almost everv
chromophore.
contains one bound molecule of allbacteriorhodopsin
molecule of
trans retinal. covalentlv attached via a Schiff's base moeit- at the Eamino group of lysine 216 (53). The almost exclusive preference
CH 3
s 1o
CH3
6
5
CH3 H
'2
3l ,- ._
4
H
LYSINE-216
N'
I
H
The structure of the all-trans
Figure 1.4.
retinal chromophore of
bacteriorhodopsin.
for one retinal isomer in the light-adapted form shows the spatial
rigidity imposed on the chromophore in the retinal binding pocket.
At least twenty amino acid residues line the retinal binding pocket of
Of these residues,
bacteriorhodopsin.
thirteen
are conserved
between six retinal-binding proteins found in halophilic bacteria
(54).
these conserved residues
Therefore,
similar functions in these proteins.
are thought to serve
Various techniques have been
Heyn and
used to find the position of the bR568 chromophore.
coworkers have shown the position of the Schiff base nitrogen to be
approximately halfway between helices C and G by neutron
diffraction studies (55). These experiments also showed that the
cyclohexene ring of the retinal chromophore is positioned adjacent to
helix E (55).
The ionone-ring end is tilted in the direction of the
extracellular surface (56, 57, 58). More recent neutron diffraction
studies using partially deuterated retinals (59) have provided
further information
chromophore.
on the transmembrane
location
of the
One interesting result of these measurements is that
the N-- H bond of the Schiff base points toward the extracellular
surface of bacteriorhodopsin, an observation consistent with previous
linear dichroism
studies (60).
The membrane plane and the plane of
24
the chromophore are nearly perpendicular to each other (55, 61, 62).
Based on linear dichroism experiments (60, 63, 64), the polyene
chain of the chromophore is tilted by approximately 20° with respect
to the membrane plane.
1.2.4
Chromophore
Properties
The main function of bacteriorhodopsin
chromophore
is to
harvest light energy for transformation into chemical energy by the
retinal-protein complex. The binding of the retinal chromophore to
the apoprotein bacterio-opsin
results in a red shift of its absorption
maximum value. Whereas model protonated Schiff bases have a
of about 440nm, the bR chromophore absorbs maximally at
approximately 570 nm. This phenomenon is referred to as the "opsin
%max
shift" (65) and is caused by specific retinal-protein interactions.
of significance
that a very similar phenomenon
binding visual pigments of photoreceptor cells.
It is
occurs in the retinal-
Several factors which
together can account for the bR opsin shift have been identified.
The
most prominent contribution comes from a weak hydrogen-bonding
interaction between the Schiff base and its protein counterion (66,
67, 68, 69, 70). A complex counterion model has been invoked based
on NMR studies (70, 71). This proposal is consistent with the results
of two-photon spectroscopy studies of bR5 68 (72), and FTIR studies on
model Schiff bases (73). The complex counterion is believed to be
comprised of the charged residues R82, D85, and D212, in addition to
at least one bound water molecule. The principle counterion of the
Schiff base, however, appears to be Asp-85, based on extensive site-
directed mutagenesis studies, a point which will receive further
elaboration in section 1.4.1. Another factor which makes a
considerable contribution to the opsin shift is the cis to trans
isomerization of the C6 -C 7 bond of the chromophore after it binds
A third contributing element may be the presence
the protein (74).
of a dipolar residue in the vicinity of the f-ionone ring of retinal (72,
74).
1.3
1.3.1
Photochemical
Properties
Bacteriorhodopsin
of
Bacteriorhodopsin
Photocycle
A sample of bacteriorhodopsin equilibrated under illumination
at room temperature, referred to as "light-adapted", absorbs
maximally at approximately 570 nm.
The incubation
of bR in the
25
dark produces the "dark-adapted" species, which has lower
extinction coefficient and a
max
of about 560 nm (6).
The light-
adapted form of bR contains >98% all-trans-retinal isomer whereas
the dark-adapted form has a mixture of 67% all-trans
isomers (75).
and 33% 13-cis
Following absorption of a green photon, isomerization
of the all-trans-retinal
to the 13-cis isomer occurs (76), initiating a
cycle of transformations within bacteriorhodopsin which allows the
net transport of one proton across the cellular membrane and into
the extracellular milieu (4, 77). Cyclic photoreactions exist for both
the 13-cis and the all-trans components.
However, the proton
pumping activity of bR is exclusively due to the all-trans cycle. A
branch pathway from the 13-cis cycle to that of all-trans, with a slow
backreaction, leads to the rapid disappearance of the 13-cis
component in the light. The bR photocycle is discussed in more detail
in the next two sections.
1.3.2
Photocycle
Intermediates
One of the most intensely studied aspects of bacteriorhodopsin,
the question of which model satisfactorily explains the bR photocycle,
is still debated vigorously (12). The first photocycle model was
proposed by Lozier et al. based on light-induced absorbance changes
at low temperature.
According to that model, the photointermediates
K, L, M, N, and O exist sequentially in an unbranched unidirectional
cycle (78), although the existence of the N intermediate was in doubt
at the time:
bR -- K -
Independently,
L -- M --> (N) -- O -
bR
evidence for the existence of the intermediates
K, L
and O were found in other reasearch laboratories (79, 80, 81).
Experiments performed at high salt concentrations and pH have since
shown the existence of the N intermediate (77, 82). Together, these
and numerous other studies of the photocycle kinetics have
demonstrated
12).
the presence
of the above five photointermediates
(9,
The differences in the absorption maxima of the intermediates
arise from changes in either the chromophore, or the protein, or a
combination thereof; a detailed discussion of what is currently
known with regard to these changes is provided in sections 1.3.3 and
1.3.4. The sole light-dependent step of the phtocycle is the
isomerization of the all-trans chromophore to the 13-cis species, with
all the subsequent steps occurring thermally. The intermediates are
26
produced on time scales ranging from 10-12 seconds for K to 10-5
seconds for L and M to 10-2 seconds for the rest of the photocycle
species. Recent studies have indicated that a photoproduct named J
forms prior to K on a femtosecond time scale and converts to K in
about 3 picoseconds (83, 84).
The simple photocycle model that emerged from the pioneering
work of Lozier et al. was extremely informative. Nevertheless, it
could not provide satisfactory explanations for all of the
experimental data available at the time (78). Considerable difficulty
exists in finding a satisfactory kinetic model for the photocycle based
on absorption spectroscopy. This arises in part because the
absorption spectra of most of the intermediates overlap strongly
with one another and the bR spectrum. Therefore, their amplitudes,
which must be known for a kinetic analysis of the photocycle, cannot
be determined precisely.
Differences in experimental conditions
have also added to the complexity of analyzing the photocycle data.
For example, variations in salt concentration (85), humidity (86),
illumination conditions, pH, or temperature (4) have considerable
effects on bacteriorhodopsin photochemistry. Accordingly, numerous
modifications of the photocycle have been proposed since the
discovery of the main set of bR photointermediates.
For instance, in
order to account for the biphasic kinetics of L and M, several
investigators have suggested kinetic schemes involving parallel
photocycles of multiple bacteriorhodopsin species (87, 88, 89, 90).
Others have proposed branched photocycles containing parallel
reaction sequences (91, 92). The presence of a multiphoton
photocycle has been suggested by Kouyama et al. (77).
Recently,
kinetic analyses of the photocycle have been performed using timeresolved resonance Raman spectroscopy (93). The advantage of this
technique is that probe beams of appropriate wavelengths can be
used to separately probe the intermediates which have distinct
absorption maxima. The kinetic resonance Raman studies have
shown that the photocycle scheme must include back reactions
(93),
consistent with the findings of other research groups (94, 95) which
indicate that a unidirectional photocycle model should be excluded.
Studies of the photocycle have also demonstrated that most of the
intermediates
can convert to the bR 568 species directly,
thus
introducing short-circuits into the photocycle (96, 97, 98).
Because the different kinetic analyses of the photocycle all
have had to be based on certain underlying assumptions, a totally
unbiased and unique solution to the photocycle is not available at
27
present.
For example, the relaxation time-constants derived from
the global fit analysis (94, 99) are functions
constants of photocycle reactions.
decomposition
of model-dependent
rate
Or, in the method of singular value
(SVD) (100), which allows a calculation of basis
difference spectra for the intermediates, each basis spectrum is
derived from model-dependent combination of the spectra of the
intermediates. Despite the existence of such difficulties in the kinetic
analysis, time-resolved FTIR (99), Raman (93) and absorption
spectroscopy studies (12, 101) have all yielded basically the same
photocycle model. This model is shown in Figure 1.5 and it shows
the simplest way to conceptualize the photocycle without a need to
invoke elaborate schemes. It is important to note, however, that this
model does contain certain oversimplifications. For instance, studies
performed in several different laboratories have shown the
heterogeneity of the M population (81, 92, 96, 102, 103, 104). Thus,
several different M species are present, some of which are kinetically
distinguishable.
The inclusion of one M species in the scheme shown
in Figure 1.5 does not take this finding into account. Varo and Lanyi
have discussed the heterogeneity in the M state in terms of two
distinct M species, called Ml and M 2, based on a kinetic analysis of
the photocycle (105, 106, 107) and also a considerable shift in the
absorption spectrum of the bR mutant DllS15N (108). Based on their
analysis, they have argued that an irreversible M,- M 2 transition
takes place during the photocycle, and that this transition represents
the photocycle switch responsible for re-orienting the Schiff base
from the extracellular
side to the cytoplasmic
half of the photocycle.
side during the second
Such re-orientation would raise the pKa of the
Schiff base and thus allow its reprotonation.
Double pulse excitation
experiments have been performed on wild type bR at room
temperature to demonstrate the existence of two M forms (109). It
is also clear from time-resolved structural studies (see Chapters 5
and 6) that a conformational change in bR occurs during the life-time
of the M intermediate.
If the observed conformational changes
represent the photocycle switch, a kinetic scheme
of the photocycle should clearly be able to account for such structural
transitions.
1.3.3
Chromophore
Photochemistry
of
Changes
Associated
with
the
Bacteriorhodopsin
Given that the bR chromophore plays a critical role in the
proton translocation pathway, characterization of the structural
changes it undergoes during the photocycle is crucial for a better
28
/7
0640
Cytoplasmlc
surface
N=6o
K5 9 0
*~CI
M412
Figure 1.5.
L550
Extracellular
H+
surface
A model of the bacteriorhodopsin photocycle.
The
transitions associated with proton release and uptake processes are
designated. Subscripts denote the absorbance maxima of the
photocvcle intermediates and bR568 species.
29
understanding of the pumping mechanism.
The structure of the
chromophore in each of the intermediates has been studied using
FTIR, NMR, and resonance Raman techniques. The first Raman study
of bR, performed by Lewis et al. (110), demonstrated the presence of
a protonated Schiff base in bR 56 8 and an unprotonated Schiff base in
the M intermediate.
Furthermore, the change in the isomeric state of
the retinal from all-trans to 13-cis in the bR--M transition has
been established by several studies (111, 112, 113). It has been
shown that this isomerization is complete within the life-time of the
K intermediate (76). The chromophore then remains in the 13-cis
conformation until the formation of the O intermediate, when
reisomerization of retinal back to all-trans takes place in the N--O
transition (114).
The spatial rigidity of the retinal binding pocket
causes the chromophore to conform to a twisted conformation in K as
shown by FTIR (115) and resonance Raman studies (76). The K to L
transition brings about a relaxation of the steric constraints imposed
upon the chromophore, producing a more planar 13-cis
conformation (116). An increased deuterium shift of the Schiff base
C=N frequency during this step shows stronger hydrogen bonding
interaction between the Schiff base and its counterion (117, 118),
implying closer proximity between the two moieties.
From a
functional point of view, this provides one reason why the Schiff
base pKa is sufficiently lowered in L to allow Schiff base
deprotonation in the L to M transition. Even though several different
spectral M species exist in the photocycle, the only change in the
chromophore during the L to M step is the Schiff base deprotonation
(110, 119), and an upward displacement of one end of the
chromophore by 0.5 angstrom (120). The absence of any structural
changes in the chromophore in the L-- M step persists through the M
to N transition during which the reprotonation of the Schiff base
takes place (9). Significant differences in the intensities and
frequencies of the C-C and C=C stretch modes of the chromophore are
observed in comparing the L and N states (93). Because the
chromophore isomer and the protonation state are identical in L and
N, the observed differences in Raman spectra indicate that the
protein environments of the chromophore in these intermediates are
not the same.
The return to the bR5 68 species is a prerequisite for
completion of the cycle and therefore reisomerization of the retinal
chromophore back to the all-trans isomer must take place. This
thermal reisomerization takes place in the N -, O transition (114).
30
1.3.4 Protein Changes
of Bacteriorhodopsin
Associated
with
the
Photochemistry
Although mechanistic studies of bR have been performed in
abundance in regard to many features of its photocycle, little is
known about the localized structural transitions bR undergoes in
order to achieve its function as an active transport system.
Given
that new approaches for structural studies of bR at specific locations
of the protein have recently become available (e.g., see Chapters 5
and 6), it is safe to predict that this area of bR research will blossom
in near future. It is known that molecular motion in
The existence of
bacteriorhodopsin occurs during the photocycle.
these conformational changes point to the importance of detailed
structural studies of each of the photointermediates if a complete
picture of the proton pumping mechanism is to emerge. One
frequently used approach for this purpose is to trap a specific
The intermediates K, L, and M can
intermediate at low temperature.
be trapped in this manner. The prolonged isolation of the M
intermediate is the easiest to achieve, by comparison.
Mutations
which increase the lifetime of M have been discovered (95,121, 122,
123), making this intermediate more amenable to structural analysis.
Similarly, site-directed mutagenesis studies have opened the door for
structural analyses of the O intermediate, which cannot be trapped
Specifically, it has been found that substitution of the
thermally.
residue Leu93 by alanine or threonine results in the formation of a
long-lived O intermediate (124). The use of site-directed
mutagenesis studies has also been essential for the new approaches
described in Chapters 5 and 6, making it possible to introduce
specific sites for derivatization with structural probes.
Several studies have shown that a conformational change in bR
takes place during the bR to M transition at specific locations within
the protein. These include time-resolved cryoelectron microscopy
(125), X-ray diffraction (126, 127; see also chapter 6), neutron
diffraction (128), and electron paramagnetic resonance (see Chapter
As mentioned previously, several different M species
5) studies.
exist in the photocycle and it is not entirely clear at this time which
M state has been trapped for analysis in each of the diffraction
methods listed here. However, the results of these separate
diffraction analyses are similar to a large extent, indicating overall
that structural changes associated with the bR --- M transition occur
predominantly in the vicinity of helices G, and E. The occurrence of a
conformational change concomitant with M formation has also been
31
suggested based on other types of studies.
For eaxmple, a large
enthalpy decrease associated with M formation was interpreted to
signify the occurrence of a conformational transition (129). Changes
in the amide I and amide II modes in the FTIR analysis of
the bR --> M step also indicate that a protein structural change takes
place (130). In addition, FTIR analysis of site-directed mutants of
several membrane-embedded proline residues has shown that
proline 186 undergoes structural changes in the bR to M transition
(13 1). It will be interesting to see whether any of these changes
reflect the irreversible conformational "switch" postulated by Varo
and Lanyi (105, 106, 107).
Time-resolved FTIR spectroscopy has identified structural
changes in the protein backbone during the decay of M to N (103,
132). The first time-resolved EPR analysis of bR has also shown that
a conformational change takes place during this step, specifically in
the vicinity of the E-F and C-D interhelical loops (133; see also
Chapter 5). The EPR transient changes were interpreted as possibly
reflecting an opening and closing of a pathway from the aqueous
solution to Asp 96 residue which is very near the top of helix C and
is involved in reprotonation of the Schiff base in the M --> N step.
1.4
1.4.1
The Mechanism
of Proton
Deprotonation
of the Schiff
Translocation
Base
The bR photocycle can be conceptually described as consisting
of two halves. In the first part of the cycle, the deprotonation of the
Schiff base takes place in the L to M transition. During the second
half of the cycle, the Schiff base reprotonates and the pump is set for
initiation of another round of light-induced proton translocation.
The
existence of mutations which only affect the rate of M decay but do
not perturb the kinetics of the M rise justifies proposing such a
division in the photocycle (122, 134). The identification of the
specific residues involved in the proton translocation pathway has
been made possible by biochemical and biophysical
directed mutants of bR (135, 136).
studies of site-
The availability of purified,
folded mutants was made possible by several achievements.
An
earlier important discovery was that denatured bacterio-opsin can
be refolded properly to fully restore pumping activity in vitro (137,
138, 139). Subsequently, methods were developed for the
expression of bacterio-opsin in the heterologous host Escherichia coli
32
(140, 141) and for extraction and purification of the apoprotein from
the E. coli membranes in a denatured state (142). In addition to the
discovery of methods which allow expression, purification, and
refolding of bR mutants, the design and creation of a synthetic
bacterio-opsin gene containing numerous restriction sites (143) also
facilitated the production of site-specific mutants.
As it was mentioned previously, the Schiff base moiety of bR
interacts
with a complex counterion.
However, the mutagenesis
studies have shown the principle counterion of the Schiff base to be
residue Asp-85. Of a very large number of mutants tested, only
substitution of this residue with nonprotonable groups leads to
complete abolishment of the proton translocation activity with a
significant slowing down of the M rise (144, 145). FTIR resonance
peaks have been assigned to this residue by the use of mutants
(146).
The FTIR studies have demonstrated that the kinetics of the
Schiff base deprotonation match the kinetics of protonation of Asp85 (147), consistent with the effect of D85 mutations on the M rise
and the central role of this residue in Schiff base deprotonation.
Another piece of evidence which supports the role of Asp-85 as the
principle counterion and the acceptor of the Schiff's base proton is
the failure of the Schiff base to deprotonate at very low pH values at
which D85 is protonated from the bulk (148).
Furthermore,
mutations at this site affect the absorption spectrum and the pKa of
the protonated Schiff base, suggesting a close proximity between
these two moieties (145, 149). This spatial proximity is also
suggested in the structural model of bR where the distance between
Asp-85 and the protonated Schiff base is about 0.4 nm (35). An
elegant demonstration of the requirement of Asp-85 for the Schiff
base deprotonation is that the mutant D85C, which is inactive in
proton transport, can be reactivated by derivatization of C85 with
iodoacetic acid, so as to re-introduce a carboxylate group at this site
(150).
Postulating a direct transfer of the proton from the Schiff
base to the D85 residue would be consistent with the results
discussed above. However, Raman (151), FTIR (152) and NMR
studies (70) suggest that this proton transfer may take place
indirectly through a bound water molecule.
FTIR evidence shows that after acquiring the Schiff's base
proton, D85 does not relinquish it until the recovery of the bRs68
species (99, 132, 146, 147). Measurements of transient absorbence
changes of pH-indicator dyes either in the bulk (153, 154), or
covalently attached to different specific sites at the protein surface of
33
cysteine mutants (154, 155), show that the proton release to the
external medium is associated with M formation. This is also
observed when analogous studies are carried out with cysteine
mutants in the purple membrane form (134, 156). Accordingly, the
proton released to the extracellular side of bR and detected by pHsensitive dyes thereafter cannot be the same one that is transferred
to D85 from the Schiff base. This proton originates from a separate
group refered to as XH (157). A very good candidate for XH is R82,
which is one of the components of the Schiff's base complex
counterion.
Khorana and colleagues have postulated the existence of
a salt bridge between R82 and D85 residues based on similar
absorption maxima values for D85N, neutral substitutions of R82 at
low pH, and the blue membrane of bR which forms upon acidification
or deionization (145, 149, 158). This is in accordance with one of the
two likely positions of R82, in close proximity to D85 and the Schiff's
base (35, 47). The existence of ionic interactions between R82 and
D85 would lower the pKa of R82 residue, a requirement for the
function of XH. The strongest evidence that R82 plays a crucial role
in proton transfer to the extracellular side, whether or not it is XH, is
that neutral substitutions of this residue lead to delayed proton
release (145), so that the proton uptake actually precedes it. These
substitutions are only partly defective in proton translocation
activity. Therefore, either a second, less efficient pathway for proton
release exists in the absence of R82, or XH is part of a hydrogenbonded complex that includes R82, as suggested by Lanyi (10).
Besides R82 and D85, residue D212 is also a component of the
complex counterion as mentioned previously.
replacement
Although the
of D212 with a neutral residue causes a red shift of the
chromophore Xmax (159, 160), similar to the effect observed with
neutral substitutions of D85, these two residues do not play
equivalent roles.
The red shift caused by replacement of D85 with a
neutral residue such as asparagine is much larger than that produced
by a similar substitution of D212. Furthermore, such substitutions
for D85 cause a lowering of the pKa of the Schiff base by several units
(145) whereas similar replacements of D212 do not affect this pKa
value (149, 159). Although neutral substitutions of D212 greatly
reduce proton translocation activity (144, 159, 160), they do not
abolish it totally as in the case of similar D85 mutants. The exact
function of D212 has been difficult to discern. It may be that D212
helps the formation and stability of the protonated Schiff base by
performing an important structural role. Indeed, neutral
substitutions of this residue are unstable and bleach during
illumination,
or do not form any chromophore at all (144, 149).
34
1.4.2
Reprotonation
of the Schiff
Base
During the second half of the photocycle, the retinal Schiff base
is reprotonated and retinal reisomerization takes place in order to
complete the cycle and return to the bR568 species. The reprotonation
step takes place in the M---N transition, and mutagenesis studies
have shown that residue Asp96, located near the top of helix C, is
critical for this process.
Neutral substitutions of D96 severely affect
the proton pumping activity, lowering it to a level of 2-3% of that for
the wild type bR (95, 144). These mutations decrease the rate of M
decay by >1 order of magnitude in mixed lipid/detergent micelles
(95, 121, 122), and by >2 orders of magnitude in purple membrane
patches (see Chapter 3). While the kinetics of M decay is at least
biphasic in the wild type protein, it is monophasic in D96A or D96N
mutants. The rate of M decay in these mutants is also pH-dependent
in a linear fashion so that at low enough pH values the effect of the
mutations on photocycle kinetics can be overcome (95, 123). These
observations indicate that in the absence of the internal proton donor
D96 the Schiff base is directly reprotonated from the cytoplasm.
The
mutation effects can also be reversed by the addition of diffusable
proton donors such as azide (95, 123). D96 mutants have been used
in the assignment of FTIR resonance peaks for this residue (146, 147,
161, 162, 163).
The FTIR measurements indicate that D96 does
indeed deprotonate during the M->N transition (99, 162, 164, 165),
in agreement with its postulated role in Schiff base reprotonation.
It
is not clear whether a direct transfer of proton from D96 to the
Schiff's base takes place. This would require a shortening of the 12angstrom distance between the two sites during M decay. It is more
likely that this transfer is indirect and involves a hydrogen-bonded
network which includes side-chains of specific residues and water
molecules.
The presence of water molecules in the channel between
D96 and the Schiff base has been suggested based on electron
microscopy studies (35).
In addition, neutron diffraction studies
have shown the presence of a strongly bound water molecule in the
vicinity of the Schiff base (166). Additional evidence that water may
play an important role in the reprotonation process exists. For
instance, kinetic studies of the photocycle show that the removal of
water by the use of osmotically active solutes and perturbants
specifically inhibits the M to N step (167). Removal of water by
dehydration also inhibits this transition (86). As for the involvement
of other specific amino acids in the M to N transition, residues such
35
as T46, T89, Y185, and D212 have been postulated to be involved in
a hydrogen-bonded network containing D96 and the Schiff base, with
the D212 residue responsible for the direct reprotonation of the
Schiff base (168).
However, the question of what the specific
mechanism of Schiff base reprotonation is remains to be answered.
After proton transfer to the Schiff base in the M to N step, the
reprotonation of D96 takes place during the N to O transition, based
on FTIR studies (11). Mutagenesis studies (158) and structural data
(35) suggest that an interaction between D96 and the positively
charged R227 may exist. It would be expected that such an
interaction would modulate the pKa of D96 during the photocycle. An
interesting possibility to consider is that reprotonation of D96 in the
N-->O step is mediated by R227 residue.
As discussed above, several studies indicate the presence of
conformational
changes during the bR-->M and M-->N steps.
These
studies also indicate that these structural transitions reverse during
the return to bR 568 as part of the completion of the cycle. The initial
state (bR 568 ) is recovered as D85 deprotonates during the O to bR
transition (132).
36
Chapter
Two
Homologous Expression of Bacteriorhodopsin
Mutants in Halobacterium
halobium
2.1
Background
The mutagenesis studies of bacteriorhodopsin using the
E. coli
heterologous
expression system have provided a wealth of
biochemical and biophysical information in regard to the function of
bR, discussed in detail in the previous chapter. Briefly, in these
studies, bR mutants were expressed in E. coli, purified, and then
regenerated with all-trans retinal in mixed lipid/detergent micelles.
The mutants were used in a variety of biochemical and biophysical
analyses, including reconstitution in lipid vesicles to study proton
translocation.
Although these studies led to important achievements in our
understanding of bR mechanism, it was of great interest to develop a
system for studying the mutant phenotypes in vivo. It had been
recognized that the ability to express bR mutants homologously in H.
The
halobium would offer several important advantages.
development of such a system allows purification of bR mutants in
the purple membrane form. Therefore, it would be possible to study
the functional properties of the mutant proteins in the native lattice
environment
for comparison
with the observations
mixed lipid/detergent micelles.
underscored
made in the
The importance of this comparison is
by the fact that rate constants of proton release and
uptake steps are significantly altered in detergent solubilized bR
monomers as compared to when bR exists in the purple membrane
lattice. The crystalline structure of the lattice also provides a
wonderful opportunity for structural analyses of the mutant
proteins.
techniques
structural
which the
The presence of a lattice form is essential for the use of
Some
such as X-ray, neutron, or electron diffraction.
studies of bR can be performed by using techniques for
However, in
presence of a lattice is not a requirement.
general it is desired to have the proteins in purple membrane
patches, especially if a study of protein conformational changes
associated with the photocycle is the objective. The difference in the
photocycle rate constants between monomeric bR and bR patches
suggests that differences in some of the photocycle-linked stuctural
transitions of bR may exist between the two cases.
It is worth
37
mentioning here that recent experimental evidence has shown such
differences in the case of the mammalian photoreceptor rhodopsin.
Depending on whether the sample is solubilized (169) or present in
the ROS membrane (170), localized differences in the observed
conformational changes are seen in comparative studies of the Meta
II state and the so called ground state of rhodopsin by EPR. In the
case of some bR mutants, experimental evidence shows that some
properties of their bR568 species differ depending on whether the
protein is solubilized, or is present in the lattice form (e.g., see
discussions in Chapters 3 and 5).
Another important advantage of a homologous expression
system for bR is that screening of bR mutants is facilitated by such a
method and studies of bop gene expression in vivo are made
possible.
A very interesting and little-studied area of research which
would also benefit greatly from such an expression system is the
investigation of how the purple membrane lattice assembles.
For
instance, it is not known what types of interactions contribute
to the
formation of lattice trimers, or whether trimer and lattice assembly
are separate step-wise processes. Mutant patches which show lower
stability compared to wild type purple membrane, or mutant
proteins that are partially, or completely, defective in lattice
formation
would be invaluable tools in such studies.
The development of an expression system for bR in H.
halobium
had been hampered until very recently by the
unavailability of the necessary genetic tools. Studies of
Halobacterium
genetics were initiated in the 1970's. It was
discovered that these organisms harbor multiple copies of large
endogenous plasmids (171, 172), pHH1 in the case of H. halobium.
Another observation was the existence of genetic instability with
respect to phenotypic markers in some of the species. For instance,
mutations which affect gas vacuole (Vac) synthesis occur with a
rather high frequency of 1-2% (172).
Spontaneous mutations which
affect bacterioruberin (Rub) or purple membrane (Pum) synthesis
also occur, albeit at a lower frequency of about 10'4 (173).
The
existence of these mutations was subsequently traced to the
presence of frequently transposing insertion elements (IS) within the
cells (174). The spontaneous occurrence of bacteriorhodpsindeficient mutant strains has been very beneficial in developing an
efficient homologous expression system of bR, where a bop- strain
can serve as a convenient negative backgorund. The loss of purple
38
membrane in a bop- mutant, however, cannot be easily detected in
as the change in colony
cells which are bacterioruberin-positive,
Accordingly, the
color is masked by the main carotenoid pigment.
availability of bacterioruberin-negative mutants was also very
helpful since it allowed much easier screening for Pum- mutants.
Studies by our laboratory and others of different mutants which
contain insertion elements within the bop gene, or in its upstream
region, have shown that the defect in purple membrane synthesis in
these strains occurs at the level of transcription (175, 176, 177).
Such analyses of those mutants in which the insertion is outside the
open reading frame of the gene provided insight regarding the
location of the promotor region of bop (175, 178). This knowledge,
together with the availability of strains deficient in
bacteriorhodopsin
and bacterioruberins
and the isolation
and
characterization of potential cloning vectors, facilitated the
development of a homologous expression system for bR described
below.
However, the development of such a system was still not
possible without an efficient transformation method for halobacteria.
The absence of such a method was the main reason that bR mutants
were originally produced in Halobacterium by in vivo random
mutagenesis, followed by selection for radically reduced or
completely absent bR function (179, 180).
The development of an efficient transformation procedure by
Cline and Doolittle (181) was therefore a very important step in
allowing Dr. Mark P. Krebs in our laboratory to develop a system for
the expression of wild type bR in H. halobium (182). The method
involved transformation of halobium cells with a plasmid carrying
the bop gene.
Halobacterium
This plasmid contained pGRB1, a plasmid isolated from
sp. (183), the insertion element ISH11 (184), and a
1.6-kilobase pair fragment from H. halobium S9 (185) carrying the
bop gene. The plasmid pGRB1 had been chosen as the parent vector
for construction of the bop expression plasmid because it has a small
size (1.8 kbp) and a high copy number. In addition, its sequence had
Insertion events in pGRB1 were screened,
been determined (186).
using H. halobium R cells transformed with this vector, in order to
identify a cloning site which does not interfere with its maintenance.
A suitable derivative of pGRB1, called pMPK29, which contained the
characterized insertion sequence ISHl1, was obtained. This vector
was further manipulated through several steps to yield a plasmid
containing the bop gene by virtue of having a 1.6-kbp Bam HI
fragment from pBR-B01 (185) which carries bop . Transformation of
39
a bop- insertion mutant of H. halobium (MPK5) with this plasmid
yielded purple (Pum-) colonies, and it was shown that the plasmid
The purple
was mainly responsible for the Pum+ phenotype.
membrane purified from these transformants contained bR in a
hexagonal lattice with the same unit-cell dimension as observed in
the wild-type patches. The absorption and visible CD properties of
this purple membrane were also similar to those expected of a wildtype bR lattice. The purple membranes from wild-type and
transformed strains showed similar proton translocation activity
upon reconstitution into vesicles.
The next section in this chapter introduces and outlines the
work aimed at the development of a system for the expression of
mutants of bR in H. halobium. The logic of the methodologies is
described along with a summary of the obtained results. The
experimental methods and results, however, are given in Chapter 3,
together with the findings on the phenotypes of several functionally
important mutant proteins which were thus prepared in the purple
membrane form.
2.2 Development
of a Homologous
Bacteriorhodopsin
Expression
System
for
Mutants
After joining our laboratory, I started collaborating with Dr.
Mark Krebs on the development of an homologous expression system
for bR mutants. We were interested in finding a selectable marker to
allow the efficient identification of transformants.
It had been
discovered near that time that the mevinolin resistance gene (MevR)
of Haloferax volcanii can be used as a selectable marker in H.
halobium (187, 188). Mevinolin (Figure 2.1) is a very potent
competitive inhibitor of the enzyme hydroxymethylglutaryl-CoA
reductase
(189).
This enzyme catalyzes the formation of the
precursor for the isoprene intermediate isopentenyl pyrophosphate
during cholesterol biosynthesis. Mevinoline has been successfully
used as a cholesterol-lowering agent for the treatment of
In halobacteria, the inhibitory action of
hypercholesterolemia.
mevinolin interferes with biosynthesis of many lipids, an effect
which can be detrimental to the cells. We confirmed the previous
observations regarding the use of the Haloferax volcanii MevR gene in
H. halobium, and subsequently developed a method for direct
selection of cells transformed with a MevR plasmid. This was done by
determining the optimal outgrowth period in the absence of the drug
to allow for expression of the resistance gene, and the optimal drug
40
H 3C
HO
0O
CO
0-"
Figure 2.1.
Structure of mevinolin, a competitive inhibitor of the
enzyme HMG-CoA reductase.
concentration for selection. It was found that the highest
transformation efficiency was obtained after a 12 hour period of
outgrowth, and the optimal drug concentration was 4 ig of mevinolin
per ml of media. The next step was to develop an expression vector
which could be propagated both in E. coli and H. halobium, so that
one could easily and quickly obtain large amounts of purified
plasmid.
In previous work aimed at the development of a suitable
bop expression vector, Mark had observed that E. coli sequences
affected plasmid stability in H. halobium. Therefore, these sequences
had been deleted before transforming the organism. Fortunately, we
found that with mevinolin selection, the deletion of E. coli sequences
is not a necessity, although such a treatment does increase resistance
to the drug.
As mentioned in Chapter 1, the mutagenesis studies of bR using
the E. coli expression system benefited tremendously from the design
and construction of a synthetic gene containing numerous restriction
sites.
A large number of mutant synthetic opsin genes were thus
available because of these studies.
Since we were interested in
expression of some of the same mutants in H. halobium, it was
advantageous to use the existing mutations in the synthetic gene.
For this purpose, it proved possible to use several restriction sites
which are conserved between the native and synthetic bop
sequences (Figure 2.2), and swap a fragment of the native gene with
the corresponding synthetic fragment carrying the desired mutation.
It was hoped that by using the common restriction sites, which cut
the gene into fragments of 212 basepairs (bp) or smaller in size, if
the perturbation of the native gene sequence would be minimal. At
41
Top strand - native bop
Bottom scran - synthetic Dbop
·
Nhel .
AatII
.
532
433 CAGGCCCAGATCACCGGACGTCCGGAGTGGATCTGGCTAGCGCTCGGTACGGcGCTAAGGGACTCGGGACGTCTATTTCCTCGTGAAAGGGATGGGCG
IIII
11 1 1111111
1H
11
I
11 II
11
I I11
ll
1 11 11 11
11111
I11111
i
11111
100
1 CAAGCTCAAATTACTGGACGTCCGGAATGGATCTGGCTAGCTCTGGGCACCGCTCTGATGGGTCGGGCACCCTGTACTTCCTGGTTAAAGGTATGGG
632
533 TCTCGGACCCAGATGCAAAGAAATTCTACGcCATCAcGACGCTCGTCcAGCCATCGCGTTCAcGATGTACcTCTCGATGCTGCTGGGGTATGGCCTCAC
11 II11111111111 IIII
Illllill
II
II 1111111
11111
I
I
II
1 11111111111
I
II 11
200
GCGGCTATCGCATTCACCATGTACCTGTCTATGCTGCTGGGTTACGGTCTGAC
101 TTTCGGATCCGGATGCGAAAAAATTCTACGCTATCACCACCCTGGT
Asp718.
.
.
.
732
633 AATGGTACCGTTCGGTGGGGAGCAGAACCCCATCTACTGGGCGCGGTACGCTGACTGGCTGTTCACCACGCCGCTGTTGTTGTTAGACCTCGCGTTGCTC
I 1 111
I I 11 11 111
H
iII IIilIIIIIIIIIIhll
HI
I 111111!
IllillhIll
II
I[IIhIIhIHMll
300
201 CATGGTAcCGTTCGGTGGTGAACAGAACCcGATCTACTGGGCCCGTTACGCTGACTGGCTGTTCACCACCccGCTGCTGCTGCTAGATCTGGCTCTGCTG
BHI
SP
832
CGT GGTGCGACGGATCATGATCGGGACCGGCCTGGTCGGCGCACTGACGAAGGTCTACTCGTACC
733 GTTGACGCGGATCAGGGAACGATCCTTGCGCT
1111 11111
11111 1111
11
11111111111
II
II I111111
11111111 11111111 11 IIIII 1
400
301 GTTGACGCTGATcAGGGCACCATCCTGGCTCTGGTTGGCGccGACGGTATCATGATCGGCACcGGCCTGGTTGGCGCGCTGACCAAGGTTTACTCTTACC
932
833 GCTTCGTGTGGTGGGCGATCAGCACCGCAGCATGCTGTACATCCTGTACGTGCTGTTCTTCGGGTTCACCTCGAAGGCCGAAAGCATGCGCCCGAGGT
11
-11
1
11 1111 1 1111
1 1111
11111111
!II11111111!!11
!1 !!1
11111II1111111 I i:i
500
401 GTTTCGTTTGGTGGGCTATCTCTACTGCAGCTATGCTGTACATCCTGTACGTACTGTTCTTCGTTTCACCTCTAAAGCTGAAAGCATGCGTCCGGAAGT
Scal .
1032
933 CGCATCCACGTTCAAAGTACTGCGTAACGTTACCGTTGTGTGTGGTCCOCGTATCCCGTCGTGTGGCTGATCGGCAGCGAAGGTGCGGGAATCGTGCG
111
III
111111I
11 I11
11 11
I II11 11III
1111!1111!!
IIIIIIl
11 11 II 1111111111 11
600
501 TGCCGTCGACCTTCAAAGTACTGCGTAACGTTACCGTTGTTCTGTGGTCGCTTACCCAGTTGTTTGGCTGATCGOTTCTGAAGGTGCCGGCATTGTTCC
CGGCTTCGGCTCATCCTCCTGCAGTCGTGCGATCTTCGGCGAAGCCG1132
1033 CTGAACATCGAGACGCTGCTGTTCATGGTGCTTGACGTGAGCGAAGT
111II11111
iI
II
II
11 II I lll1
I I 11II ( il llll11111 :! III
1I
GGTGAAGA CG
GTTTCGGTCTGATCC TGCTGTTCTT
601 CTGAATATTGAAACCCTGCTGTTCATGGTTCTAGACGTTTCTGCTAAAGTT
.NOtl
1133AAGCCGCCGAGCCGTCCaCCGGCGACGSCCGCCGCGAC 1173
il;l ill 1 ! li
IIII Ilil ! 11111111 ii11
71
7'1 AAGC'CCGGAACCGTCCGCGGGTGACGGTGCGGCGCSTACC
Alinment of the native and synthetic bop gene
Fi2ure 2.2.
sequences. Several of the common restriction sites for the two
species are shown.
T
1I
G 700
42
this juncture, having assembled the necessary components and
determined a strategy, the construction of a series of vectors aimed
at testing bR expression in H. halobium was begun. These plasmids
contained the MevR gene as a selectable marker, wild-type or mutant
bR genes, and sequences to allow their selective maintanence in both
H. halobium and E. coli.
Initially, the strain MPK5 was transformed with the
constructed vectors. MPK5 (Rub- Pum-) is a derivative of MPK1
(Rub-), itself a carotenoidless derivative of the R1 strain, and contains
an insertion within the bop gene. The presence of the insertion
results in the inability of the strain to express bR, which causes its
Pum- phenotype. Pum* revertants of MPK5 occur at a low frequency;
'
<5 x 105.
Thus, the loss of the chromosomal insertion within the bop
gene is a rare occurrence. However, we were concerned with the
possibility that the plasmids carrying the bop gene might recombine
at the chromosomal locus. A possible outcome of such recombination
events would be the generation of an intact bop gene, resulting in
contamination of the mutant proteins with the wild-type product.
Southern blot analysis of genomic DNA obtained from the
transformants indicated that indeed rearrangements at the bop locus
occurred.
The fragment sizes were consistent
with the occurrence of
homologous recombination between the plasmid and the
chromosomal bop genes. As importantly, it was discovered that a
second recombination event took place in the absence of selection,
which led to the loss of the integrated plasmid. These recombination
phenomena suggested the means of producing stable H. halobium
recombinants expressing bR mutants.
rearrangements
In order to avoid
at the bop locus that could regenerate
a
chromosomal wild-type gene, a bop deletion strain was constructed
which could recombine with our construction vectors through the
flanking sequences of bop. This new strain, named MPK40, was
transformed with our constructed expression vectors and produced a
series of mutants subsequently used for analysis of their effects on
proton translocation activity. A more in-depth description of the
methodology described above and an analysis of the mutant
phenotypes analyzed by various biochemical and biophysical
techniques
are presented
in the next chapter.
43
Chapter
of
Bacteriorhodopsin
Proton
Translocation
Characterization
3.1
Three
Mutants
which
Affect
Activity
Introduction
Among the best characterized
bR mutants are those which
contain substitutions at the key functional residues described in
Chapter 1. Site-directed mutagenesis studies using the E. coli
expression system identified these residues to be R82, D85, D96, and
D212. Their locations are shown in the secondary structure model of
bR in Figure 3.1. The residues R82, D85, and D212 are the critical
components of the Schiff's base counterion. R82 and D85 are also
involved in the deprotonation pathway of the Schiff base. The
residue D96 participates in the reprotonation of the Schiff base
D212 may also be involved
during the second half of the photocycle.
in this process, although the phenotypic features produced by
neutral substitutions at this site are not easy to interpret, and
therefore the exact role of this residue remains to be determined.
For the reasons discussed in the previous chapter, we were
interested in expression of these well-characterized mutations in H.
halobium . Specifically, we wanted to compare the phenotypes of
these mutations in vivo with the observations made previously using
the in vitro reconstituted system. Our main focus was to see whether
the observed alteration of proton translocation activity by these
mutations had been affected by the reconstitution
procedure.
Except
for the photocycle and photovoltage measurements, the experimental
data presented in this chapter including the results of the proton
pumping assays which I performed have been published (reference
160).
44
CD
0 -
0
o
-.
.--
r
-.--
It
0
z7
r ,,,
.
-_
I ,
,.+
"0-
-t -,
UQ O
00 Kc,=
° C) C) Cr
G,~r~
L- r <C
r --
PI
T
0
Cd)~=f
Wf
rr
(
..
i
>
>r.
Co> pn
%~otO
cn
U
>
-_
C)
9
C)
'o_
Z
-r>
r
>
<U)
<,,
m
<_
< r
<>,-> >m
C)
Rl>m
In
45
3.2
3.2.1
Experimental
Procedures
Bacterial Strains and Culture Conditions.
H. halobium strains, MPK1 (Rub-), MPK5 (MPK1 Pum-), and
MPK40 (Pum-) and their derivatives were grown in a rich medium
which contained the following ingredients per liter: NaCl, 250.0 g;
MgSO 4 . 7H1120,20.0 g; trisodium citrate . 2H20, 3.0 g; KCI, 2.0 g; and
Oxoid Bacteriological Peptone, 10 g. This rich medium will be
referred to by the abbreviation "RM". For the purpose of indefinite
storage of each strain, a small volume of a saturated culture (1.2 ml)
which contained 25% glycerol (v/v) was quickly frozen in liquid
nitrogen and stored at -80°C. The strains were also grown on slants
of the rich medium containing 1.8% agar, and subsequently
In order to obtain satisfactory growth, the agar
maintained at 4C.
plates were incubated at 37°C for a period of at least 5 days. They
were placed inside a closed glass chamber during this incubation
time in order to prevent crystallization
of the medium salts as a
result of dryness.
For isolation of purple membrane, a small volume of saturated
cell culture was prepared first, by transferring a single colony of the
desired strain to the rich medium and allowing growth at 37°C for a
This was used as inoculum for a larger shake
period of 5 days.
culture (a 5 ml inoculum was used for 2 liters of the rich medium).
Subsequently, the inoculated cell culture was grown to saturation at
37°C under conditions of low oxygen tension and constant agitation at
a speed of 150 rpm.
In order to further induce the biosynthesis of
purple membrane during growth, the culture was kept under
constant illumination
by a bank of Daylight fluorescent
tubes
(Sylvania, Danvers, MA).
3.2.2
Plasmid
Construction
Mutations in the bop gene were introduced
in pMPK62 by
casette mutagenesis. Restriction fragments from the native gene
were replaced with the corresponding fragments from the synthetic
gene of the mutant derivatives of pSBO2 (143) by directional
pMPK62 is a derivative of pMPK54 which is also the parent
for pMPK56 (Figure 3.2). pMPK54 is an E. coli H. halobium
vector and contains pGRB1 and the insertion element ISH1.
cloning.
vector
shuttle
In
46
pGRBI
AR
ISH II
bop
Amp
ACeBp
S
J06
0
,
N
*
. I
"*
BI
;H//
MevR
.
.
S,
S
Figure 3.2.
Construction of mevinolin-resistance plasmids and bop
expression vectors. pMPK54 contains pGRB1 and insertion element
ISH11 from pMPK35, and the mevinolin-resistance gene and E. coli
vector pAT153 from pWL102.
The pAT153 portion was removed in
pMPK56. A 1.6-kbp fragment bearing the bop gene was combined
with pMPK54 to yield pMPK62. pMPK62 derivatives were made by
replacing fragments in the native bop
gene with corresponding
fragments from the synthetic gene. A, Asp 718; B, Bam HI; Bc, Bcl I;
Bp, Bsp HI; H, Hind III; Sp, Sph I; N, Not I; AmpR, ampicillin
resistance.
4?7
CD
D
,.~~~~~~~~~~~~~~D
CD
0
0 0~
0
O0
X~~~~~
Z
..~~
~
~
QO
3G)O
EGG)
O
Y
W
0C
r~~~~~~~~~
-00
O
0
t >
i
B·
· 002f·a~tn
db )YCn~o,
--OO ~~~~~~~~~~~~OQ
a*-~~~~~~rna
~ O <C
C)
oC
~
I.
o~
,o~~~~~~~~~~~~~~~C
-
CL~~~~~~~~~~~~~~~~~~~~~
m,. ~
00
0
ai
- 4. ...
I"~
0
O~
O
.~~~~~~~
·
D.
CI
c
~a ,
,, 3 Cp
C
b
0O
Q't~
Io~·
~,
00~nr
on
-.
0. 0' .
c0
u
-I,,,.q0
0 0 a)00,
"S"~~~~~~~~~~~~~~C
n0
o
~n ~~~~~
Y
0
oo~~~~~~~~~~~~~~~~~~~~~~0I
0
8 ~ ~~~~~~~
z
-0
n-'T'
,
~n~.
.='
..C~,=~o.
'80
0QC) 0
Ic o
O
0
0
r~~~~~~~~~~0
sr-r·
C
Ii , °
0-
~
: .
C
rn
O~~~~~~~~~~
~~~~
1))
~~-,4-
.l~
·
-~
:....~
='O~~~~~~~
...0
~~OOR Oa
L9~~~~~~~ICan
)C~~~~~~~~~
J,~~~~~
rna
n 63
C) n
'"
0
~ ~
C~~~~~~~~
-"
O
.,F'()
-0
0 IDa ilr~a
--
ct O
'
C)C)
CD
cn ~
o
_
,
v~~~~~~~~~-
0
.,~
,~~
>n
V)O---
w
'
,
O-
o
Y)-(
_-n
'4
~~~~~~~~~~.
00
C
-v
n
0). a)
·- C)
G) 00~~~~~~~~~~~~~~~~~~~~~a
0
~n l<
·
) O 0 .
. _ ~~
--
~~fi
P
-1~~~<
,
sgn
0L
8
,~~~~~~~~~~~~~~~M0n
p
eo
O
O:'"
-OC]~~~~~~~~~~~~~
48
addition, it carries ampicillin-resistance
and mevinolin-resistance
genes and E. coli vector pAT153 from pWL102 (190). Removal of the
pAT153 portion yields pMPK56. In order to construct pMPK62, a
1.6-kilobase-pair (kbp) fragment bearing the bop gene was
combined with pMPK54 (Figure 3.2).
For construction of pMPK62-WT (SN), -D212E, and -D212N, the
242-bp Sph I-Not I synthetic fragment from pSBO2 derivatives was
combined with the following fragments in a four part in-gel ligation:
the 920-bp Barn HI-Sph I and the 2.4-kbp Not I-Hind III fragments
from pMPK39 and the 6.7-kbp Hind III-Bam HI fragment from
pMPK54. Four part in-gel ligation was also performed to construct
pMPK62-WT (AB), -R82A, -R82Q, -D85A, -D85E, -D85N, -D96A, D96E, and -D96N. The 145-bp Asp 718-Bsp HI synthetic fragment
from pSBO2 derivatives was combined with the 640-bp Barn HI-Asp
718 fragment and the 2.7-kbp Bsp HI-Hind III fragment from
pMPK39 and the 6.7-kbp Hind III-Bam HI fragment from pMPK54
for these constructions (Figure 3.2). In order to perform digests with
enzyme Bsp HI, the plasmids were
the dam methylase-sensitive
purified
from
the damrn-E. coli strain CSH26.
propagated in and
The 145-bp Asp 718-Bsp HI and the 242-bp Sph I-Not I
fragments of pSBO2 used for the ligations were isolated using 4%
Seaplaque agarose in TAE buffer. The rest of the fragments which
are larger in size were isolated using 1% low melting point agarose.
For each ligation, the gel pieces containing the separate purified
fragments were melted separately at 680C and were then combined
in equimolar amounts in a microfuge tube, kept at 37°C, to a total
volume of 10 il. Ice-cold ligation buffer (10 ul) containing 100 mM
Tris-HC1, pH 7.6, 20 mM MgC1,, 2 mM DTT, 2 mM ATP, and 0.5 units
of T4 ligase was subsequently added to each tube and the contents
were mixed well. The tubes were allowed to sit at room temperature
for at least 4 hours to allow the ligations to proceed. For each
ligation reaction, negative control tubes missing either the synthetic
fragment or the enzyme ligase, and otherwise identical to the other
reaction tubes, were included. The ligation mixes were melted at
68°C and diluted with 180 l of ice-cold TCM buffer (10 mM Tris-HC1,
pH 7.5, 10 mM MgCl2, 10 mM CaC12). Competent E. coli cells DH5a,
were transformed by addition
prepared by the CaCl, method (191),
of 50 tl of the diluted ligation mix to 200 tl of the cells. The tubes
were then kept on ice for 90 minutes with occasional mixing and the
cells were subsequently heat-shocked for 2 minutes at 370C. Luria
49
Broth (1 ml) which had been warmed up to 37°C was then added to
the tubes and the cells were incubated at this temperature for one
hour to allow expression of the ampR gene carried by the constructed
plasmid.
Each outgrowth (200 l) was spread over the surface of an
LB plate containing 50 g of ampicillin per ml of LB and the plates
were incubated at 370 C overnight to allow growth of the transformed
cells. In addition to inclusion of negative control tubes for
assessment of the success of ligation, both negative and positive
control tubes for the transformation process were also included.
The
negative control tube was treated identically to the other tubes and
contained the same amount of competent cells and buffer except that
no plasmid was present in the mix. For the positive control tube, 1 l
of a purified plasmid solution in TE buffer (1 jLg/gl) was added to the
cells. In a separate experiment,
the optimum amount of purified
plasmid needed to obtain the highest transformation efficiency was
measured.
3.2.3
Purification
and
Analysis
of
Constructed
Plasmids
From each LB plate containing transformed E. coli cells, three or
more colonies were selected depending on the overall observed
efficiencies for ligation and transformation.
Each colony was used as
inoculum for a 5 ml shake culture of LB medium containing 50 tg
ampicillin per ml. After growth to saturation at 370 C (>18 hours), the
cells were pelleted and the plasmids purified using alkaline lysis
(191).
Purified plasmids were subjected to restriction enzyme
analysis and the results of the digests were visualized on 1% agarose
gels after RNase treatment. The presence of the synthetic DNA piece
carrying the mutation in each plasmid was verified by using
enzymes which digested unique sites on the synthetic fragment as
compared to the native gene sequence. The analysis also verified
whether each isolated plasmid contained ligation fragments of the
expected size.
Based on the results of restriction enzyme analysis,
correct transformants were selected for plasmid purification on a
larger scale by CsCl gradient (191), or Qiagen columns (Qiagen Inc.,
Chatsworth, CA).
The entire synthetic DNA duplex within each purified palsmid
was sequenced by the dideoxy chain termination method, using
Sequenase ® version 2.0 reagents (United States Biochemical,
Cleveland, Ohio) and 6% polyacrylamide-urea gels, in order to verify
the presence of the mutation. In one case, the entire bop gene was
50
sequenced
in order to be sure that no spontaneous mutations
occurred during plasmid work-up.
The 443-bp region flanking the
3'-end of the bop gene in the 1.6-kbp Bam HI fragment from pBRBO1 (140) was also sequenced, for the first time, in both directions.
Four primers, RMOO1, RM002, BoplO and Pucl were designed,
synthesized, and purified for use in this sequencing, along with
primer PGRB2, which had been previously constructed and purified
(Krebs, M. P., unpublished data). Automated oligonucleotide
synthesis was performed using an Applied Biosystems DNA
synthesizer model 380B. The oligonucleotides were deprotected with
concentrated amonium hydroxide, dried in vacuo by Speedvac
overnight, and the DNA pellets dissolved in TE (10 mM Tris.HCl, pH
8.0, 1 mM EDTA). Product yields were calculated by UV absorption
spectroscopic analysis of the oligonucleotides.
The oligonucleotide
products were then purified by gel electrophoresis on denaturing
polyacrylamide gels followed by Sep-Pak' column purification. The
purity of the final products was assessed by UV absorption
spectroscopy.
3.2.4
H.
halobium
Transformation and Isolation of
Recombinants
The procedure followed for transformation contains several
modifications
to that described by Cline and Doolittle (181).
Cultures
were used for transformation after reaching a density of - 1 x 109
colony-forming units per ml. The cells were harvested by
centrifugation at 1100 x g for 10 minutes at room temperature and
in a rotor warmed up to 37°C. They were subsequently resuspended
gently in 0.1 volume of a solution used for formation of spheroplasts
referred to herein as solution A (2 M NaCl, 27 mM KCl, 50 mM
Tris.HCl, pH 8.75, and 15% sucrose [wt/vol]). An 0.2 ml amount of
the concentrated cells was added to 10 gtl of 0.5 M EDTA in solution A
with brief and very gentle agitation. After an incubation time of 1
minute, 1 ug of purified plasmid in a volume of 15 gl of solution A
was added to the spheroplasts and the mixture was incubated for 5
minutes at room temperature (tubes included for negative control
were treated in an identical fashion except that solution A which did
not contain any plasmid was added at this step). An equal volume of
60% PEG 600-40% solution A (wt/vol) was subsequently added to
each tube with gentle shaking and the mixture was incubated for 30
minutes (PEG-600 was purified as described (192) prior to use in this
Spheroplasts were then diluted by addition of 5 ml of
procedure).
51
medium salts (RM) containing 15% sucrose (wt/vol). The diluted
spheroplasts were centrifuged at 3000 x g for 15 minutes at room
temperature, and the cell pellets were resuspended in 10 ml of
medium containing sucrose.
After an incubation period of 12 hours,
the cells were plated on medium which contained mevinolinic acid
(Figure 3.3) at a concentration of 4 Lg/ml. Mevinolin was converted
to the -hydroxy acid following the described procedure (193).
In order to allow for the growth of mevinolin-resistant
(MevR)
colonies, the plates were incubated inside a desiccator at 37°C for up
to 7 days. For isolation of recombinants, MevR colonies were
resuspended and plated on medium lacking mevinolin. Southern blot
analysis was then performed to verify that the desired recombinants
had been isolated. For Southern analysis of colored (Pum+)
recombinants,
genomic DNA was digested with Pst I and a
combination of Pst I and Bcl I which cleaves a unique restriction site
within the Asp 718-Bsp HI synthetic fragment carrying the mutation.
The preparation of genomic DNA was performed following previously
described procedures (194). Digested genomic DNA was visualized
by ethidium bromide staining, and was subsequently denatured and
transferred to GeneScreen Plus membranes (NEN) for hybridization,
following published procedures (191). For hybridization, a 32p_
labeled probe was prepared by random-hexanucleotide priming
(195), using as template the 1.6-kbp Bam HI fragment described in
section 3.2.3. This fragment carries the entire bop gene plus its
flanking sequences.
3.2.5
Purification
Bacteriorhodopsin
of Wild-Type
in
and Mutant
Purple Membrane
Form.
Cell cultures of H. halobium producing wild type and mutant
bacteriorhodopsins
were prepared as described
in section 3.2.1.
A
time-course of cell growth was obtained by monitoring cell density at
different stages of growth via A660 nm measurements. The procedure
used for purple membrane (PM) purification was basically that of
Oesterhelt and Stoeckenius (18). All solutions were made using
milliQ water. Harvested cells from a two-liter culture were
resuspended in 20 ml of media salts (no peptone). DNase I (500
units) was added to the suspension followed by transfer to a dialysis
tube (45 cm by 2.5 cm width; Spectrum Laboratory Products). The
suspension was dialyzed against 7 liters of 0.1 M NaCl for 6 hours.
The collected dialysate was centrifuged in a Ti45 rotor (Beckman) at
52
23,000 rpm for 40 minutes at 4C.
The pellet was resuspended in 0.1
M NaCl by filling up the centrifuge bottle, and the wash was repeated
3 times or until the supernatant was colorless. The final pellet was
resuspended in 5 mM NaCl and centrifuged at 27,000 rpm for 40
minutes at 4C in a Ti45 rotor. This step was repeated twice and the
final pellet was resuspended in 2-3 ml of milliQ water. This
suspension was applied to a 38 ml linear 30 to 50% sucrose density
gradient in 5 mM NaC1. The gradients were then centrifuged at
28,000 rpm using the swinging bucket rotor SW28 (Beckman) for 20
hours at 15°c. The purified purple membrane band was then
collected, diluted with 50 ml of 5 mM NaC1, and washed by
centrifugation at 25,000 rpm in a Ti45 rotor for 30 minutes at 4C.
This step was repeated twice to achieve complete removal of the
sucrose. The final pellet was typically resuspended in 1-2 ml of
milliQ water containing 0.025% sodium azide (w/vol) and stored at
4 0 C.
The purified protein was subjected to SDS-PAGE analysis and
was also analyzed by UV/visible spectroscopy using a Hitachi U-3110
spectrophotometer modified with a head-on photomultiplier.
3.2.6
Proton Translocation
The procedure
Assays
used for assays of proton translocation
was a
modified version of that described originally by Racker for formation
of lipid vesicles (196).
Initially, soybean phospholipids
(Sigma) were
purified according to the procedure of Kagawa and Racker (197).
Briefly, 25 grams of soybean lipids were added to 500 ml of dry
acetone containing 1 mM DTT, and the mixture was stirred for three
days under argon at room temperature, in a flask covered with
aluminum foil in order minimize light-induced oxidation.
The
insoluble material was collected by filtration, dissolved in 100 ml of
dry diethylether, and the ether-insoluble material was removed by
centrifugation
at low speed (2000 x g for 5 minutes) in glass
stoppered centrifuge tubes.
Subsequently, the soluble lipids were
slowly added to 900 ml dry acetone containing 1 mM DTT by stirring.
The mixture was then placed on ice under argon and the precipitate
was collected by filtration.
The lipids were resuspended in
diethylether and the precipitation process was repeated.
The final
precipitate
as
was taken up in diethylether
described above.
evaporator,
and centrifuged
The solvent was subsequently using a rotary
and the dried lipids were stored under argon at -20°C for
further use.
53
The purified soybean lipids (20 mg) were dissolved in 0.4 ml of
chloroform.
The bulk of the added chloroform was evaporated away
by placing the suspension under a steady flow of argon while
rotating the tube rapidly, in order to form a thin film of lipid at the
bottom of the tube. The lipids were then lyophilized overnight to
remove any residual solvent and were either used immediately, or
stored at -20°C under argon in the dark. A 1 ml-volume of 0.15 M
KC1 was added to the lipids and the suspension was vortexed
The suspension
vigorously for 2 minutes and stored at 4C overnight.
of lipids in KC1 was then sonicated to clarity (8 minutes) using a bath
sonicator (Laboratory Supplies Co.).
It was subsequently cosonicated
with bR sheets, free of any azide, for another 8 minutes, in order to
incorporate the purified purple membrane sample into lipid vesicles.
A lipid to protein ratio of 40:1 [(w/w); 198] was used (100 l of
sonicated lipids and 50 tg of purified protein in the purple
For both sonications,
membrane form were mixed for cosonication).
a plastic frit was placed on top of the suspension to reduce splashing
of the liquid onto the inside wall of the tube during sonication. The
vesicles were diluted 1:10 in 0.15 M KCl and adjusted to pH 6.5 with
NaOH or HC1 for proton translocation measurements. Lightdependent pH changes were measured with a type MI-410 pH
electrode (Microelectrodes, Inc.).
The initial rate of pumping was
determined at room temperature by a linear fit of pH changes
occuring between 1 and 5 seconds after illumination (>495 nm).
electrode was calibrated by addition of 1 nmol of HCl for each
The
measurement.
3.2.7
Time-resolved
Absorption
Spectroscopy
Studies
The photocycle measurements by flash spectroscopy were
performed by Dr. Harald Otto and Dr. Ulrike Alexiev in Professor
Maarten Heyn's laboratory (Freie Universitat Berlin). A diagram of
the home-made flash spectrophotometer used for kinetic analysis of
the photocycles is shown in Figure 3.4. The light-adapted samples in
10 mM MOPS buffer and 150 mM KCl were excited by a dye laser
(rhodamine 6G) which has an output wavelength of 581 nm, with a
20 nm width in the absence of wavelength selection. The dye laser
was pumped by the excimer laser (3-6 mJ) with 5-10 nanosecond
flashes.
The intensity of the laser was measured using the
photodiode shown in the diagram. The measuring light beam was
provided by a 100 W halogen lamp (Osram Co.). This beam was
54
L
P
L
IRF
Figure 3.4.
A diagram of the photocycle apparatus. IF. interface:
IRF, infrared filter: L. lens: P. polarizer: P.. Dhotomultiplier.
55
passed through a monochromator and then focused by a lens on the
sample cuvette. The monochromator was used to allow monitoring of
different wavelengths during the photocycle at chosen 10 nm
intervals.
The range of scanned wavelengths was between 370 nm
and 700 nm. Any measuring light transmitted from the sample
(probe) was detected by a photomultiplier which has an output
voltage directly proportional to the intensity of the transmitted light.
The intensity changes measured by the photomultiplier after the
flash were recorded with a computer equipped with a 40-Mhz two-
channel A/D converter with 8-bit resolution and 128-kbyte memory.
In the two channels, the data points were analyzed simultaneously,
with 0.1 sec per point for the first channel and 100 ptsec per point
for the second channel. A global fit at all wavelengths was obtained
with one set of exponentials.
Alternatively, the data at a single
wavelength were fitted with a sum of exponentials.
Time-resolved absorption spectroscopy allows the
measurement of light absorption by the sample at different specific
time-points.
The difference in light absorption by bR before and
after excitation by a laser pulse can be described as follows:
AA = log (V 0 /V(t))
where V0 is the voltage measured by the photomultiplier before the
flash and V(t) is the measured voltage at a specific time-point
following the flash. An offset voltage, VG, is provided to allow for
detection of small voltage changes during the photocycle.
Accordingly,
the observed values of V0 and V(t) are expressed by the
following terms:
V 0 '= VO + VG
and
V'(t) = V(t) + VG
Therefore, the amount of light absorption
flash can be measured as follows:
at a specific time after the
AA= -log [1 + (V'(t) - V,')/V 0 ]
56
3.2.8
Photovoltage
Measurements
Photovoltage measurements were performed by Dr. Stephan
Moltke in Professor Heyn's laboratory. Purple membrane sheets
were oriented via adsorption to lipid-impregnated polyethylene
sheets in a saline buffer solution containing 3 mM HEPES, 3 mM
acetate buffer, 3 mM Trizma, and 150 mM KCl at pH 7.0, and at 25°C.
Transient charging of the membrane and the polyethylene sheet can
be observed due to charge movements within the adsorbed purple
membrane after excitation of bacteriorhodopsin molecules with a 10
nanosecond laser flash (580 nm). The accompanying voltage change
was recorded using a high shunt resistance of 5-30 G.
Depending
on the shunt resistance, the system discharges bi-exponentially after
1-10 seconds. The analog bandwidth of the system is about 5 Mhz.
Voltage data were collected in two channels simultaneously with an
8-bit amplitude resolution, a total memory of 128 kb, and sampling
times of 50 nanoseconds and 1-2 milliseconds covering a total time
scale from 50 ns to almost 200 s. The data were fitted by sums of
exponentials with gaussian distributed rate constants (199).
Signal
to noise ratios were 100-200 to 1. The change in the pH of the
sample to low, or high values was achieved by using H 2SO 4 and KOH,
respectively.
3.3
Results
3.3.1 Construction and
3.2.
Analysis of Expression Vectors
The strategy for construction of the plasmids is shown in Figure
Sequencing of the purified plasmids showed that each carried
the expected mutation, and the restriction enzyme analyses of miniprepped plasmid samples confirmed that each had the expected
fragment sizes and that each contained the expected synthetic piece.
Primers RM001, RM002,
Bop 10, and PUC 1 were utilized for
bidirectional sequencing of the region downstream of the stop codon
for the bop gene within the 1.6 kbp fragment used in construction of
our vectors (Figure 3.5). This sequencing effort led to the correction
of the sequence at the transcription termination site of the bop gene
where two nonexistent G residues had been previously included at
positions 37 and 39 following the stop codon for bop (175).
The
inclusion of these G residues had erroneously placed them as part of
the hairpin structure of the transcription termination signal.
Accordingly,
in light of this correction,
the hairpin structure of the
+
c'
.3
CZ
C,,
Ci
U
u
C
D
U
4:
U
C
U
U'
C,
U
u
U'
U
a
C
+
C
CL
Cl
L
r,U,
U
.
C,
4:I
4:
-.4:
3
C,
t3
I,
a
II
C
4:
a
C,
I
U
-
4.
C
L) +
C
i
-
C3
U
I
C,
C
C
I
i
I
4:
C
U
i
U3
U
I
U
U
U
C
a
C,
U
+
a-
a
C
C
C,
C,
U
C
I
L,
C..
C,
I
.3
4:
c:
C.3
-
I
I
C,
C3(,
\z
LO
U.
0
U
U
C,
U
-
_,
Ul
C
C
I
C,
+
(U
C,
La
-C
U
C
C,
U
c
c3
C
U
U
U
,
C
C.)
u
:
U
4:
3
U
Cc
C,
C,
u4:
C
C,
+
o
- -W
U
a
U
U
004
._
:
4:
:
C,
C,
U
I
1
1
u
_3
u
I
=bn
4C,
Cr
C
-
I
+I
U
.
C,
cC
C,,+
U
U
-U
C,
CL
C,
C
C:
S
U3
C.
C,
Ci
U
C,3
C,
U +
-
tz
L)
U
i
.
CL
a
C,
C
L3)
U
L)
CL
+
U
C
u
u
U
I
+
C,
C
C3
L3D
S
C,
C
U
--Ecl
C
C
cs
U
u0-
a
C
u
C
r..
L,
C
(3
ua
,U
L:
,
L
U
C:
C,
tz
C,
C
4i
r,
CL
C
CO,
I
U ,0
C
C
C
iU
C;3
C
U'
U
U'
i
C:
U
c,
U
U
U
U
CD
C
i
I
a-
C3
aL4
C,
a-
U
u
L)
C
-a
I
U
a
U
Ci
I
C,
+
I
4:'
U.
C3
U
C3
C,
4:
4:'
Uj
U
C
iC
U
C
a
U~
4:
4:
4:
4:
C
U
4
0
t
U
c
_
C,
I
U
I
L
I
U
U..
I
U
C
U
U
C
C)
i
U
C
C-
u
uC)
.
a-
c.3
U.
II
_
I
I
-
_
c3
G
._
_
I+
V-
,.
iU
C,
4-
C,-
.I
.
C
4:
_,
i
O
C:
I
I
I
C
4,.
U
C,
u
'0
,
C
4:
p
a.
a-
Ud +
I
+
4:
U
4:
cl
C.
U
4:
C:
.
U
r:,
iI
O
4+
_
C
_
58
termination signal only contains the underlined sequence which is
G+C-rich followed by a run of thymines (Figure 3.5).
This is very
similar to p-independent transcription termination signals in E. coli
(200).
3.3.2 Gene Replacement at the Chromosomal bop Locus; H.
halobium
Transformation and Isolation of Recombinants
H. halobium MPK5 was transformed efficiently with both
pMPK54 and pMPK56 (up to 106 MevR colonies per g of DNA) and
Southern blot analysis indicated that the transformants carried
plasmids of the expected size (Figure 3.6).
pMPK54 occurs in lower
abundance compared to pMPK56, suggesting that pAT153 reduces
the stability of pGRB1 derivatives in H. halobium. A similar effect
has been observed with another E. coli vector, pUC18 (160).
However, it was decided to use pMPK54 as the parent plasmid for
construction of the expression vectors in order to allow maintenance
and propagation in E. coli.
MPK5 was transformed with pMPK62 and pMPK62-WT(AB).
Transformants were obtained at high efficiency (105 MevR colonies
per g of DNA) and expressed bR at about 75% of the level obtained
for MPK1. Figure 3.7 shows a Southern blot analysis of Pst IThe results
digested genomic DNA obtained from the transformants.
shown in lanes 3 and 4 indicate that rearrangements at the bop locus
In addition to the expected
have occurred in these transformants.
6.2-kbp chromosomal bop fragment, observed also in untransformed
MPK5 (Figure 3.7, lane 2) and a 7.6-kbp plasmid bop fragment, other
bands are also present. The sizes of these additional Pst I fragments
are 3.8-, 4.9-, and 8.9-kbp, consistent with the sizes predicted from
integration of the plasmid by homologous recombination between the
chromosomal and plasmid bop genes (Figure 3.8).
In the absence of mevinolin selection, it was discovered that a
second recombination event took place, resulting in the loss of the
integrated plasmid. Transformants showed three different
phenotypes on medium lacking mevinolin, purple color (Pum+),
orange color (Pum-), and sectored (mixed Pum+/Pum-), at respective
frequencies of 4, 16, and 80%.
The diagram in Figure 3.8 shows how
these phenotypes can arise from homologous recombination between
copies of the bop gene flanking the integrated plasmid.
If
accompanied by plasmid loss, recombination would generate Pum+
59
1234
.23.1
kbp
-s -9.4
,I
-W
- 6.6
--
,A
- 2.3
Figure 3.6.
Southern blot analysis of MevR transformants.
Total
genomic DNA from MPK5/pMPK54 and MPK5/pMPK56 grown in the
presence of mevinolin (1.3 Lg/ml) was digested with XhoI, which
cuts once within the mevinolin-resistance
gene. After
electrophoresis on a 0.7% agarose gel and transfer to nitrocellulose,
the DNA was probed with 3 2 P-labeled pGRB1. Molecular size
markers were obtained from a Hind III digest of bacteriophage X
DNA.
Lanes: 1 and 2, MPK5/pMPK54; 3 and 4, MPK5/pMPK56.
60
1 2 34
W *
F
(I
5 6 7 8 9 10
o
j 8.5 kbp
j 7.2
-- 6.4
-- 5.7
-- 4.8
4.3
" 3.7
4, - 2.3
- 1.9
Figure 3.7.
Southern blot analysis of MPK5 transformants and
recombinants.
MPK5 was transformed with pMPK62 or pMPK62WT(AB) and recombinants were isolated. Genomic DNA was prepared
from cultures grown in the presence (lanes 3 and 4) or absence
(lanes 1, 2 and 5-10) of mevinolin (1.0 gtg/ml). After digestion with
Pst I (lanes 1-6), or Pst I and Bcl I (lanes 7-10), the samples were
electrophoresed on a 0.7%c agarose gel, transferred to nitrocellulose,
and probed with a 3 2 P-labeled 1.6 kbp BarnamHI fragment containing
hop.
Lanes: I and 7, MPK1: 2 and 8. MPK5; 3. MPK5/pMPK62
MPK5/pMPK62-W'T(AB);
5 and 9, MPK5/pMPK62
10, MPK5/pMPK62-WT(AB).
recombinant;
4.
6 and
Molecular size markers were obtained
from a Bst ElI digest of bacteriophage 2. DNA. The fragments in lane
site in insertion element ISH1.
8 (MPK5) arise because of a Bcl
61
colonies carrying an intact bop gene or Pum- colonies retaining the
original insertion. In support of this, Pum colonies contained a single
5.1-kbp bop fragment (Figure 3.7, lanes 5 and 6) that co-migrated
with the bop fragment in wild-type (Figure 3.7, lane 1). To exclude
the possibility that Pum+ colonies were due to MPK1 contamination,
the presence of a Bcl I site within the synthetic bop fragment was
tested for recombinants obtained with pMPK62-WT (AB). The site
was present in these recombinants (Figure 3.7, lane 10) but not in
MPK1 (Figure 3.7, lane 7), or in recombinants obtained with pMPK62
(Figure 3.7, lane 9), as expected. The Bcl I digestion pattern was also
distinct from that of MPK5 (Figure 3.7, lane 8).
The above results suggested a strategy for obtaining stable H.
halobium
recombinants expressing bR mutants.
In order to avoid a
regeneration of the wild-type gene from sequential recombination
events at sites flanking the insertion element (Figure 3.8, sites a and
b), we constructed a bop deletion strain, MPK40. MPK1, a Pum+
strain, was transformed with pMPK66, in which the bop coding
sequence of a 5.1-kbp chromosomal bop fragment was replaced by a
synthetic oligonucleotide duplex (Figure 3.9).
The transformation
efficiency was at a high level of 105 MevR colonies per }zg of DNA. To
isolate bop- recombinants, we screened transformants for Pumcolonies in the absence of mevinolin. Of six colonies tested, four
yielded Pum- colonies at a frequency of 5-95% whereas the
remaining two yielded only Pum+ colonies. Southern blot analysis of
Pst I-digested genomic DNA indicated that the 5.1-kbp bop fragment
in MPK1 (Figure 3.10, lane 1) was replaced with a 4.5-kbp bop
fragment in Pum- colonies (Figure 3.10, lane 3; only the clone MPK40
is shown). Bgl II cleaved the 4.5-kbp Pst I fragment in MPK40
(Figure 3.10, lane 4), as expected from the site present in the
synthetic linker of pMPK66. Bgl II did not cleave the 5.1-kbp bop
fragment of MPK1 (Figure 3.10, lane 2). These results indicated that
homologous recombination involving pMPK66 led to the deletion of
the chromosomal
3.3.3
bop gene in MPK40.
Isolation of MPK40 Recombinants Carrying Mutations
in the bop
Gene
Mutations in the bR coding sequence were introduced into the
bop deletion strain MPK40 using the gene replacement method
described above. This strain was transformed with the wild-type
and mutant derivatives of pMPK62 which encode WT(AB), WT(SN),
62
ISHI
ae c p
M
R
Mev
X; mutation
7.6
,
P
*I
P
P
P
a
3.7
P
b
10.1
\
P
P
P
m
4.8
MevR
9.0
~-----fIIIIV
p
P
P
4.8
9.0
MevR
U
P
P
Pum + , wild type
a,b
5.1
P
P
O,c
Pum+ , mutant
5.1
b .c
Figure 3.8.
Pum-
6.2
6.2
Products of homologous recombination in MPK5
transformed with pMPK62 derivatives.
The chromosomal and
plasmid bop genes are indicated by the solid and open boxes,
respectively. Wavy lines indicate flanking chromosomal DNA, the
inverted triangle corresponds to insertion element ISH1, and the X
(Arrow I) Reciprocal strand
indicates the site of a mutation.
exchange (dotted lines) at sites a, b, or c results in the corresponding
single crossover products.
(Arrow II) A second crossover event and
loss of the plasmid yields a single copy of bop. P, Pst I; numbers, size
of fragments homologous to the bop probe.
63
Aatll
W'IC'AGCCCAGATCAt
k u
ClaI
Bgll Notl
'GI
G
ICTCA1CGA1TGAGATCTGC
GGCCGCGACCAG'ACIC.A
AG(GTCCGGGTCTAGTIG,,
C i GCAGAGTAGCTACTCTAGACGCCGG CGCTGGTCGCTGACT
L
J
-
n
n
I
P
F
E
A
H
AmpR
MevR
Aa N
Figure 3.9.
Structure of pMPK66. In pMPK66, a chromosomal Pst I
fragment from MPK1 containing the bop gene was combined with the
mevinolin-resistance gene and E. coli vector pAT153 from pWL102.
The sequence shown corresponds to the 5' and 3' regions of the bop
gene. The boxed sequence corresponds to the synthetic
oligonucleotide duplex used to replace the bop open reading frame.
The arrow and asterisk correspond to the N- and C-terminal codons
of the mature bR respectively. Aa, Aat II; Ap, Apa I; Bx, Bst XI; E, Eco
RI; H, Hind III; AmpR, ampicillin resistance.
64
2 3 4 5 6 7
9 0 Ii 213 i4151617
8 5 kbp
j72
,,4
m
-5 7
-48
'4.3
-3.7
-2.3
-- I.9
--1.4
Figure 3.10.
Southern blot analysis of the bop deletion strain and
Pum+ recombinants.
Genomic DNA was prepared from MPK1, the
bop deletion strain and pum+ recombinants derived from
transformation of MPK40 with pMPK62 derivatives. After digestion
with Pst I (lanes 1,3 and 5-17), or Pst I and Bgl II (lanes 2 and 4),
the DNA was analyzed as in figure 3.7. Lanes: I and 2, MPKl; 3 and
4. MPK40; 5, WT(AB); 6, WT(SN); 7, R82A; 8, R82Q; 9, D85A; 10, D85E;
11, D85N; 12, D96A; 13, D96E; 14, D96N; 15, D212E; 16, D212N; 17, an
aberrant rearrangement obtained during screening for D212N.
Molecular
bacteriophage
size markers were obtained from a Bst ElI digest of
?XDNA.
65
R82A, R82Q, D85A, D85E. D85N, D96A, D96E, D96N, D212E, and
MevR transformants were grown in the absence of mevinolin,
D212N.
and recombinants were identified by colony color. The presence of a
single 5.1-kbp Pst I fragment was verified for each Pum*
recombinant (Figure 3.10, lanes 5-16), and in addition the presence
of the synthetic DNA piece within this Pst I fragment was verified in
almost all cases by performing restriction analysis at a site unique to
the synthetic sequence within the gene. The recombinants were
sensitive to mevinolin, confirming that the integrated plasmid had
been lost. About 10% of the Pum+ colonies contained additional
fragments homologous to the bop probe (Figure 3.10, lane 17) which
may arise from integration of the plasmid at other sites in the
genome. Such recombinants were not used for further study because
of the possibility of rearrangements between the different copies of
the bop gene.
Purification of Wild-Type and Mutant Proteins and
Spectroscopy and
by UV/Visible
Characterization
3.3.4
their
Proton
Translocation
Assays
The protein samples were purified as described above.
In the
last step of the purification, the protein samples were subjected to
It was found that in all cases a
sucrose density centrifugation.
membrane fraction at the same buoyant density as the wild-type
purple membrane was obtained.
shifted
max
For the mutants which showed
values at neutral pH, a large fraction of the protein was
also located in a membrane fraction of lower buoyent density than
purple membrane. This fraction corresponds either to a
noncrystalline membrane form, analogous to a noncrystalline
precursor of purple membrane observed in the wild-type cells (21),
or to a crystalline form less extended than the purple membrane, or
both. Because of the location of this fraction in the gradients, it was
not included in our preparations for further study, as it was likely to
contain contaminating proteins. Accordingly, the yield of the bR
mutants with shifted absorption maxima was lower (2-10 mg/liter)
as compared to the yields obtained for the D96 mutations, WT(AB),
The wild-type and mutant
and WT(SN) samples (10-15 mg/liter).
proteins showed the same mobility on denaturing polyacrylamide gel
electrophoresis (Figure 3.11).
The purified protein samples were examined by UV/visible
absorbance spectroscopy (Figure 3.12 and Table 1). The wild-type
samples and D96 mutants had similar spectra in the dark-adapted
66
WT
kDa
-44
-29
-18
Figure 3.11.
SDS-PAGE analysis of wild type and mutant
bacteriorhodopsin proteins.
the protein bands.
The same mobility is observed for all
67
o
_
0
0
CD
0~ao
a
0 .
t3
o
:3
0wB
CD
P)
C-
C
Lc
0
4
0
cO 0
Absorbance
0
4
0
0
0
o
0
m
68
state, and they showed upon illumination
kmax
a 10-nm increase in the
and an increase in extinction characteristic of purple membrane.
In contrast, the Xmax of R82, D85, and D212 mutants was red-shifted
by 10-50 nm in the dark-adapted state. The shift in the mx,, of these
mutants upon light-adaptation was not to the same extent as in wildtype bR and was accompanied by aberrant changes in their
extinction coefficients.
The chromophore peaks in these mutants
were significantly broader than in wild type, most probably due to
the presence of both purple and blue forms of the proteins, as will be
evident from discussion of the photocycle and photovoltage
measurements below. All of the light-induced changes were fully
reversible except in the D212N mutant which lost about 3% of its
chromophore after 4 min of illumination.
As shown in Table 3.1, all of the mutants showed reduced
proton translocation ativity compared to wild-type bR.
No light-
induced proton pumping was observed for D85A, D85N, and D212E
mutants.
3.3.5
Measurements
of
Photocycle
Kinetics
For all the purified mutants, detailed photocycle kinetic and
and photovoltage measurements were performed. However, only the
photocycle behavior of a neutral substitution at each of the selected
sites is shown herein, along with the results of the electrical
measurements for D85N and D212N mutations, since these data
adequately illustrate the effect of the removal of the functional
groups at these sites on proton translocation
activity.
The remainder
of the photocycle and photovoltage data will be published elsewhere.
The data on neutral substitutions also illustrates the many
similarities and a few of the differences that are observed between
the results in purple membrane and the results of the same
measurements performed in mixed detergent/lipid
micelles.
Prior to the presentation of the results, it may be helpful to
reiterate that the "blue" form of bR can be obtained by lowering the
pH of the membrane suspension (201).
The addition of salts can then
cause the return of the purple color at low pH, conditions under
which bR is referred to as being in the acid-purple form. Based on
these observations and mutagenesis studies of the R82, D85, and
D212 residues in the lipid/detergent micelles, it is believed that
lowering the pH leads to neutralization of D85 and D212 in native bR,
69
-r
I
..
uCD
:
_"t W
.
Ur
0
- _
nrn
?
I
A,~~~~~~~~~~
I-
t-.
c
Of
I
_,
-\
I
__~~~~~~~~
.
r
C'n
O
- _
i
6
cc
=r
CC
·
i
v
pv
r
3"
·.
G
-e
_
J
i
o
cc
(O
c!
r.
O
c
=i
O
i
i
':
i
i
-.-
I
P,
,
ao
-i
I
1
v,
3
3
cc
P,
3
A
i
c
V
OG
·^
\
·
ii
-
I,
V
~~~
a
sc~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
c
Lh)
3,
coo
L
C_=
I
;
~
~
,
/7~~~~~~~~~~~~~~~~~~
r~~~~~~~~~-L
-
I
C
~~~miiiiG
~~~~~~~~~~~
__
70
and this is what causes the change in color to the blue (148, 202).
The salt effect then appears to be due to the ability of exogenous
anions to substitute as counterions to the protonated Schiff base
while D85 and D212 remain neutralized at low pH (202). There is no
net electrogenic transport associated with the blue form, because the
proton release and uptake both take place from the cytoplasmic
side.
In the electrical measurements, a downward signal signifies proton
transport in the forward (pumping) direction.
Therefore, the sign of
the electrical signal is reversed in the blue form. As the pH is
increased and the purple form is restored, the signal reverses and a
net forward proton translocation occurs.
A discharging of the system
returns the signal to the baseline, with a positive overshoot, in about
1 second. The photocycle features of the blue form are the
occurrence of an intermediate at 530 nm (K/L) and a depletion at
around 650 nm.
D85 Mutants:
Figures 3.13 and 3.14 show the photocycle kinetics
of D85N following an excitation flash at pH 7.3, or pH 11 respectively.
The residue D85 is the most critical component of the pathway for
Schiff's base deprotonation.
In contrast to wild-type bR, and
consistent with the micelle studies (145), the D85N mutant produces
very little M (410 nm) at neutral pH (Figure 3.13). This is in
accordance with the postulated direct role of this residue in Schiff
base deprotonation. The positive absorbance change observed for
this mutant at around 530 nm at neutral pH is due to the formation
of the L/N intermediate.
In addition, a depletion signal is observed
at around 630 nm. These features of the photocycle of D85N at
neutral pH are very typical of those for the blue form of bR. In
agreement with this observation, the electrical signal for this mutant
is that of a blue form at low pH, which changes to that of a purple
state when the pH is raised (Figure 3.15). These observations are
consistent with the models discussed above regarding the nature of
the blue and acid-purple forms of bR.
The most prominent feature of the D85N photocycle at pH 11 is
the presence of a long-lived 650-nm species (Figure 3.14). A similar
absorbance change is observed for the 13-cis cycle of bR. The
photocycle behavior for the mutant D85A is very similar to that of
D85N mutation.
R82
Mutants: Arginine-82 is an important residue whose role in
proton release to the extracellular region was discussed in detail in
Chapter 1. The main photocycle characteristics for the mutant R82A
71
-
-C)
c
DD,
CO
OF
N o
0
C)
Po
C)
~
ov:C)~C
O
L-
0zo D
(D
r
O
OF
a:
cn
I
,
i,
r,
2,\
=
'
I
L
ci^
I
tA
I__
;_
Pul
C:
:c!cmiiliO -
72
0.6
0.5
0.4
0.3
E 0.2
0.1
0
-0.1
-n 9- 2
10
10-
100
102
10 4
10 6
time after excitation / ps
108
Figure 3.15.
Photovoltage traces of the bacteriorhodopsin mutant
D85N at three different pH values in standard buffer with 150 mM
KCL at 25C
: (1) pH 2.4. (2) pH 6.7. and (3) pH 10.8.
adjusted using H2SO4 and KOH.
The pH was
,3
'
D
CD
-'o
"0
-
de La-s-Rbs/
0
w
j
D- 0
u
_"
040
3
ct
,
CD
'D
0
=.eD
r~D
0,
.,.
0
*o
3
0
0
.
C _+
5'
~
_ECD
0
t~- 5
N
aT
oD
o
D
Co
-3 5
o
O.
O _i
oc
'=tA
5
V r.,
_g
74
in the membrane patches are shown in the rise and decay of the M
intermediate at pH 5.0, pH 7.3, and pH 8.5 in Figure 3.16. Similar to
D85A and D85N mutations, a pH-dependent purple to blue transition
exists for R82A and R82Q mutants, with a higher pKa of transition as
compared to the wild-type bR . This phenomenon, also observed in
the micelles, is postulated to be due to an increase in pKa of D85 in
the absence of its salt bridge with R82. As Figure 3.16 shows, the
amplitude of the M signal, fitted with a sum of 5 exponentials,
increases dramatically at higher pH values, consistent with a
transition from a blue to a purple cycle. The rise of the M
intermediate is accelerated in these R82 mutants regardless of the
pH. Two rise times of 0.5 isec and 9 usec are found for M in R82A,
as compared to three rise times of 1.1, 45, and 152 sec for the wild-
type protein under the same conditions.
These features of the R82A
photocycle, also observed for R82Q mutant, are consistent with the
observations made using mixed micelles (145), and support the
importance of R82 in the proton release pathway. A very interesting
feature of these R82 mutations, which has also been observed in the
mixed micelles (145), is that the proton uptake and release processes
appear to be reversed based on measurements with the pH indicator
dye pyranine (data not shown). The significance of this observation
in terms of the possible role of R82 in proton release was discussed
in Chapter 1.
D96 Mutants:
The photocycle kinetics of the D96N mutant are
shown in Figure 3.17. The main features of the photocycle are the
same for the D96A mutant.
Consistent with the postulated role of
this residue in Schiff base reprotonation, D96N has a very long-lived
M intermediate.
The M decay is slower in the purple membrane
form of this mutant than in micelles and lasts for about three
seconds. The effect of this residue on M decay is consistent with the
previous observations made in mixed micelles (95, 122). In addition,
the effect of the mutation on M decay is reversed with a decrease in
pH, as had also been observed previously in micelles (95), although
the fast wild-type times are not reached (data not shown). This is in
accordance with the previous observation that the Schiff base can be
reprotonated directly from the cytoplasmic side in the absence of the
proton donor residue D96 (95).
D212 Mutants: The exact role of D212 in the photocycle is yet to be
discovered. However, the photocycle data show that it does play a
role in Schiff base deprotonation.
Figure 3.18 shows the photocycle
__
I -
0
3
I
z
.
I
o
-
o
t
:....-~~-.~£;.--
Ej
7
-,
0t
0
I
.
r;IoON
C
I
7
C
I
vi
.' .
~
-
O
-c
L
D
\_I·-----·
o
o
e--
.&
,
,,
f'
-
~ ~ ~ ~ ~
-
w~~~~~~~~.
..f.d.
,.> :
..
.
_
._ .~ ~,
.;
3
&D
.
,.
,,.,
. :
:.. :
.~' ? % k
.
I
E3
_;I
C
...~~ .....
_
mn fA
'\
°·--
-
·-
;
\_
tP
C
CS
l
~ ~ ~ ~
.. ~~~~~~~~~~~~~~"'-¥
G~~~~~~~~~~~L~
O
·
.,
. .-,¢ / -'>-
v
o v
m
/.
.%
Y
,`··,
o
6
^
di':a miiiiOD%
.:_
76
,
_.
,.w
(D e ,
D
CD
:~
I-
C
I
· ~~~~~~~~
-
_,,
.
\Ic
i
-
I
I-
I
I
a
a
C
"'"'
z O
_x.)
c~~~~~ '
_
Z,
~~~~~~
_F
5
,
.Le~~;..
,,
;-'V
Lr6
;
O
C
I
E
X`
f
tz, G
X
,.y
u" O \ /,
0
0
CL
\
,
\
C
C
a"
cn
E
-
i
-
11.
z
t\,.
I
eel
I I
-
I
I
I
.
I
-
1
1
7 1:::,
cre
r rI
-,,~~~
w
v
;-
_,.,
~,..~~~
'1~:
mriliOn
;
~~~~~~~~~~~~~~~
77
0.6
0.4
0.2
E
0
D) -0.2
-0.4
-0.6
-n8R
10
10°
10 2
106
10 8
time after excitation / .ts
Figure
D212N
KCl at
arrows
3.19.
Photovoltage traces of the bacteriorhodopsin mutant
at four different pH values in standard buffer with 150 mM
25°C: (1) pH 2.8, (2) pH 5.1, (3) pH 8.9. (4) pH 11.0. The two
labeled T1lsYs and sys mark the time constants of the passive
svstem discharge.
78
kinetics of the D212N mutant at pH 7.3. The photocycle has the
features of the blue form of bR and no M intermediate is present.
Lowering the pH to 5.1 produces a photocycle similar to that of the
purple form of bR although the amount of M present is unusually
small (data not shown). This behavior of D212N has also been
observed by Needleman et al. (159), who argue that the existence of
a branched cycle directly from L to bR for a portion of the molecules
explains this phenotype.
The pH-dependent blue to purple transition
is also shown to occur by photovoltage measurements of this mutant
(Figure 3.19). As discussed above, the existence of a blue form of bR
at low pH is attributed to a blocking of the Schiff base deprotonation
due to protonation of D85 and D212 residues. In agreement with this
hypothesis, the appearance of a blue form for the D212N mutant at
near neutral pH shows that the presence of the carboxylate group of
D212 is important for transfer of the Schiff base proton to D85.
This
observation provides further support for the existence of a complex
counterion environment for the Schiff base, with residue D85 acting
as the principle counterion.
3.4
Discussion
3.4.1 Gene Replacement
in H.
halobium
Certain features of H. halobium and the expression plasmids we
constructed made the use of the gene replacement method feasible.
The instability of the pMPK54 derivatives or the lack of a
halobacterial origin of replication in pMPK66 causes the integration
of these vectors into the chromosome so that the cells can maintain
resistance to mevinolin. Once excised, these plasmids are quickly
lost. Because the only extensive homologous
or near the bop gene, the integration events
exclusively in these regions. Thus, for about
recombinants, the crossover events take place
sequences are located at
take place almost
90% of the
at the bop locus. The
apparent existence of an efficient homologous recombination
mechanism in H. halobium is another important advantage which
allows isloation of desired recombinants with relative ease. Stable
recombinants are produced from transformants containing integrated
plsmids at a frequency of 5-95%.
An efficient mechanism for
homologous recombination may exist for halobacteria in general, as
suggested by the results of marker-rescue experiments in Haloferax
volcanii (203, 204). The presence of high crossover frequency in H.
halobium emphasizes the importance of removing the wild type
alleles in studies of mutantions of cloned halobium
genes.
Otherwise,
79
as discussed previously, reversions causing the reconstiution of the
wild type gene may occur. We achieved this goal by the use of a bop
deletion strain. This concern, however, has not been addressed in
several published studies of bop mutants in H. halobium (108, 159,
167).
The gene replacement method described in this chapter may be
applicable to study of genes for which phenotypic markers are not
available in H. halobium.
In such cases, the recombinant candidates
may be selected based on sensetivity
to mevinolin and be subjected
to Southern blot analysis to confirm their identity, a feasible
approach given that the recombination frequencies are likely to
exceed 5%.
3.4.2
Phenotypes of Proton
Translocation
Mutants
Our studies underscore the importance of comparing the
phenotypes of bR mutants in vivo with the previous observations
made in vitro. For several mutants, certain differences have been
observed. For instance, the light-adapted form of D85N has a Xmax
which is red-shifted by 45 nm in the purple membrane (Figure 3.12),
compared to only 28-31 nm in the reconstituted form (149, 205).
The light-adapted form of D115N mutant does not show any
considerable shift in its absoption maximum in the purple membrane
(108) compared to a 17 nm blue shift it displays in the reconstituted
form (144). The difference observed for the D212N mutant was
already discussed above. It is especially important to consider such
differences in the studies which aim to gain a better understanding
of the retinal-protein interactions.
Despite the existence of such differences, the studies of bR
mutants in the purple membrane form have confirmed the majority
of the previous findings from the E. coli expression system.
Most
importantly, it is clear that the previous studies have correctly
identified the residues most critical for pumping activity of bR.
Thus,
consistent with the previous analyses, we find that mutations at
residues R82, D85, D96, and D212 all affect proton translocation
activity considerably. These results demonstrate that the presence
of the purple membrane lattice is not a requirement for bR function.
Perhaps, it is present simply to provide higher stability for bR, and
therefore has solely a structural role.
We also find that the main
photocycle features for the mutations of the above residues are very
similar to those observed previously using the reconstituted form of
80
the same mutants, confirming the postulated roles of the above
amino acids in the bR photocycle.
81
Chapter
Four
Structural Studies of Bacteriorhodopsin: I. Preparation and
Characterization of Cysteine Substitution Mutants
4.1
Introduction
A somewhat unexplored area of bR research and yet of intense
interest deals with the following question: What conformational
changes in the protein are associated with the formation of the
distinct photointermediates of the bR photocycle? Information that
conformational changes occur has come from time-resolved neutron,
X-ray, and electron diffraction studies of mutant and/or wild type bR
obtained in the lattice form, as discussed in Chapter 1. In addition,
as it was also pointed out, time-resolved FTIR studies have identified
While
structural changes in the bR backbone during the photocycle.
all of the above approaches have been invaluable in gathering
information regarding light-dependent conformational changes of bR,
none of them have provided localization of the structural changes. I
was interested in the use of cysteine mutants of bR to help localize
some of its light-dependent conformational transitions. The presence
of introduced sulfhydryl groups at specific sites permits the
introduction of structural probes at known locations within the
protein and thus allows monitoring of movement near selected
regions. Our H. halobium expression system for bR mutants is
especially
suitable for use in such experiments.
As mentioned
earlier, this expression system allows the purification of the mutants
in the native lattice structure. Accordingly, any effects that in vitro
reconstitution of the mutants may have on the light-dependent
conformational changes of bR are avoided. The first use of cysteine
mutagenesis to study the ground state structure of bR in the
reconstituted form was made by Dr. Sabine Flitsch in the Khorana
laboratory (206, 207).
By use of the gene replacement method described in the
previous chapter, I isolated and characterized H. halobium
recombinants for a series of cysteine mutants of bR.
This series
contains single cysteine substitution mutants and also double
mutants in which replacement with cysteine at a site is coupled to
mutation of a functional residue. These mutants are S35C, G72C,
V1O1C, A103C, Q105C, V130C, A160C, G231C, R82A/A160C,
D85N/G72C, and D96A/V101lC. All the cysteine residues were
introduced in the putative loop regions of bR to allow accessibility of
82
the sulfhydryl groups to derivatization by structural probes.
I chose
to focus on a subset of this series of mutations for subsequent
studies. This subset consists of the mutants S35C, G72C, V1O1C,
A103C, Q105C, V130C, A160C, G231C, and the double mutant
D96A/V101C (Figure 4.1). All of the mutations in this subset are
single cysteine substitutions except for the double mutant
D96A/V101C. The mutant D96A/V1O1C was constructed because the
neutral substitution at position 96 causes a long delay in the decay of
the M intermediate.
structural
perform.
contains
mutants
Because
This makes measurements of localized
changes in M at site 101 (i.e., at the top of helix C) easier to
It is worth mentioning in this connection that helix C
the key functional residues R82, D85, and D96. The set of
shown in Figure 4.1 were chosen for several reasons.
they are located in different loop regions of bR, they allow
structural studies of the protein at different localized sites.
In
addition, with the exception of D96A/V101C, they are all single
substitutions of uncharged residues, therefore introducing minimal
perturbation of the photocycle kinetics. This is particularly
convenient for the analysis of proton release an uptake
measurements described below. The previous characterization of
these mutants in the reconstituted form is also advantageous as it
allows a comparison with their study in purple membrane.
In this chapter, a description
of the preparation
and
characterization of these mutants is given. Additionally, the results
of proton release and uptake measurements during the photocycle of
these mutants are presented and discussed. These studies were
performed by both the attachment of pH-indicator dyes to the
introduced sulfhydryl groups of the mutated sites, and the use of the
pH-indicator dye pyranine in the bulk phase.
Such measurements
have been previously performed in mixed micelles using the same
mutant proteins and experimental conditions (155). Having available
the gene replacement method for homologous expression of bR
mutants, we were greatly interested in performing these studies
with the cysteine mutants in the purple membrane form.
The
advantages that purple membrane patches offer in regard to such
measurements will be discussed. The findings of the proton release
and uptake measurements have been published (reference 134).
83
o~oI
C
_.
I
0>
-
r'- -I
--
<
{-r > >
r " ~(.<
~r- r> - > 7
~ _00
r
In D
-< - 1
<
E > <--i
-o
,
-n
-
-
C,
'
C)
0
r-
Cb
O
*i
_,7
(A
CD
D
C)
(
m
C
C.CD
OZa
rr
0
-
:C:
S
E
'll
(
es
0
ab
I
r-<
7m
)
la
z
-n T
i
I---- E<"
ir Z:U) <
-.
c,
r
<
I _ c > r
m
C
C)
,-- r
-,
,'
>e
is:
(n
<=o D
-
(b
I
_
0
-
IZ
I-I-
Ve MM 5 -
,
_
r-
r -" n
<I
L'
a
)
X
<
(JI<
_.
r)
cn
>
cr
>
c)
CD(b
MnF
E
I rn-a
nM r-c
CI
r-
m
0
O
;
I
>
>
:>
-n >
-n
Dci,
11
M
co
-LMO1]F
m>
84
4.2
Experimental
Procedures
4.2.1 Bacterial Strains and Culture Conditions
The H. halobium bop deletion strain MPK40 was used for
isolation of the recombinants.
The culture conditions and the
conditions for growth and storage of the cells were as described in
Chapter 3, section 3.2.1.
4.2.2
Plasmid
Construction
The cloning strategy and ligation procedures were the same as
described in Chapter 3, section 3.2.2.
The DNA fragments used for
construction of the expression vector for each mutant are listed
below:
S35C:
This construct was obtained by combining the 6.7-kbp Barnm
HI-Hind III fragment of pMPK54, 476-bp Bam HI-Nhe I and 2.9-kbp
Asp 718-Hind
III fragments of pMPK39, and 171-bp Nhe I-Asp 718
fragment of pSBO2.S35C.
G72C, V1O1C, A103C, Q105C, D85N/G72C, and D96A/V101C:
The restriction fragments used to construct these vectors were the
same as described for the mutations in Chapter 3 with the Asp 718-
Bsp HI synthetic piece carrying the desired mutation.
R82A/A16OC: This double mutant had not been previously
prepared using the E. coli expression system. Therefore, pSBO2
carrying these mutations was first constructed by a two-part in-gel
ligation of 333-bp Nhe I-Bss HI fragment of pSBO2.R82A and 2.7-kbp
Nhe I-Bss HI fragment of pSBO2.A160C. The constructed
pSBO2.R82A/A160C was cloned in E. coli and then purified. A 320bp Asp 718-Psp 14061 synthetic fragment of this plasmid was then
combined with 6.7-kbp Bam HI-Hind III fragment of pMPK54, 2.5-
kbp Psp 1406I-Hind III fragment of pMPK39 and 639-bp Bam HIAsp 718 fragment of pMPK39 by four-part in-gel ligation.
V130C,
and A160C: To obtain these constructs, the 6.7-kbp Bam
HI-Hind III fragment of pMPK54, and 2486-bp Hind III-Psp 14061
and 781-bp Barn HI-Bsp HI fragments of pMPK39 were combined
with the177-bp Bsp HI-Psp 14061 synthetic fragments of
A 2452-bp fragment is
pSBO2.V130C and pSBO2.A160C respectively.
generated by Hind III-Psp 14061 double digest of pMPK39. It was
85
necessary to resolve this fragment on preparative agarose gel away
from the desired 2486-bp fragment which is also produced by the
double digest. This was accomplished by performing an additional
restriction
digest using Sma I.
This enzyme cuts the 2452-bp
fragment into two smaller pieces of similar size, rendering the 2486bp fragment resolvable on preparative agarose gel.
G231C: The four-part in-gel ligation performed to obtain this
construct involved the following fragments; 6.7-kbp Bam HI-Hind III
fragment of pMPK54, 2.4-kbp Hind III-Not I and 920-bp Sph I-Bam
HI fragments of pMPK39, and 244-bp Sph I-Not I synthetic fragment
of pSBO2.G231C.
4.2.3 Purification
and Analysis of Constructed
Plasmids
The procedures followed for identification of correct E. coli
transformants and subsequent purification and analysis of the
expression vectors were the same as described in Chapter 3, section
3.2.3.
4.2.4 H.
halobium Transformation
Recombinants
and
Isolation
of
The procedures described in Chapter 3, section 3.2.4 were
followed for transformation of MPK40 strain with expression vectors
and isolation of recombinants. However, the nonradioactive methods
of Genius DNA labeling and detection system from Boehringer
Mannheim (Figures 4.2, 4.3 , and 4.4) and the enhanced
chemiluminescence detection method developed by Amersham
incorporation (Figure 4.5) were used for Southern blot analyses.
In
the former method, the DNA probe is labeled by random-primed
incorporation of digoxigenin-labeled dUTP instead of 32 P-dCTP, and
the DNA fragments are visualized by enzyme-catalyzed color reaction
subsequent
to binding of an antibody-conjugate
to the hybrids.
In
the recently developed chemiluminescence method, the probe is
directly labeled with the enzyme horseradish peroxidase. During the
detection step, reduction of the substrate hydrogen peroxide by the
enzyme is coupled to a light producing reaction involving luminol,
which on oxidation produces blue light.
on a blue-light sensitive film.
The emitted light is detected
86
4.2.5
Purification
Membrane
of Mutant
Proteins
in the
Purple
Form
Purification of the mutant proteins was performed as described
in section 3.2.5 of Chapter 3.
The mutant proteins were analyzed for
purity by SDS-PAGE and UV/visible spectroscopy as described.
4.2.6
and
Derivatization
Flash
Mutants
of Cysteine
Spectroscopy
with
Fluorescein
Analysis
These experiments were performed as part of a collaborative
effort on proton release and uptake measurements with Dr. Ulrike
Alexiev in Professor Maarten Heyn's laboratory. Mutant protein (0.1
mM) in 150 mM KC1, 4 gtM EDTA, 0.1 mM dithiothreitol, and 50 mM
Tris buffer, pH 7.0, was allowed to react with the pH-indicator dye 5(BMF) from Molecular Probes (1 mM)
(bromomethyl)fluorescein
under argon at room temperature. The excess reagents were
removed by chromatography on Sephadex G-25 (Pharmacia).
Column preequilibration and elution were performed using highly
purified Milli-Q water (Millipore). The labeling stiochiometry was
determined spectroscopically by using an extinction coefficient of =
68,000 M-'.cm - ' (Molecular Probes) for the alkaline form of
fluorescein at 495 nm and of e = 63,000 M' 1 .cm' for bR at 570 nm
The apparent pKa of methylfluorescein bound to position 72
(208).
(G72C-MF) was 7.0 in 150 mM KC1.
Flash spectroscopy
and data analysis with a sum of
exponentials were performed as described in Chapter 3, section 3.2.7.
The dye kinetics and the kinetics of M rise and decay were typically
measured by averaging 70-100 signals and 30-50 signals
respectively.
4.2.7
Proton
Release
Measurements
These studies were performed
using both the pH-indicator
pyranine (8-hydroxy-1,3,6-pyrenetrisulfonic
acid trisodium salt;
Serva) in the aqueous bulk phase, or fluorescein
protein surface.
dye
attached to the
For pyranine experiments, flash-induced absorbance
changes at 450 nm (209) were monitored in purple membrane
suspensions containing 4-15 CiMbR mutant and 45 p.M pyranine in
150 mM KC1, pH 7.3, at 22°C. The light-induced absorbance changes
at 450 nm were also measured in the absence of any pyranine in
87
otherwise identical samples and under identical conditions. The
difference between the light-induced absorbance changes at 450 nm
with and without pyranine (AAA450) was used to measure proton
release and uptake processes in the aqueous bulk medium.
The
light-induced proton concentration changes detected with fluorescein
attached covalently to the surface of the cysteine mutants were
determined by measuring the flash-induced absorbance difference at
495 nm (210) between samples with and without 10 mM Tris buffer
at pH 7.3, at 220C in 150 mM KC1.
4.3
4.3.1
Results
Isolation
of Mutant Recombinants
The results of Southern blot analyses presented in Figures 4.2
through 4.5 indicate that recombinants showing the correct digest
patterns were found for each of the mutants. The purified genomic
DNA samples from all the isolated recombinants were subjected to
digestion by Pst I enzyme. Except for V130C, A160C, and
R82A/A160C recombinants, this digestion should produce a single
5.1-kbp fragment containing the bop gene. This was indeed the case
(Figure 4.2, even-numbered lanes; Figure 4.3A, lane 3; Figure 4.3B,
lanes 1,2; Figure 4.4, lanes 1-3; and Figure 4.5, lanes 1-3). Following
PstI digestion, the purified genomic DNA samples from the
recombinants were subjected to a second restriction analysis.
This
second digestion was performed at sites unique to the introduced
synthetic sequences within the bop genes of isolated strains,
verifying in each case the presence of the synthetic fragment
carrying the desired mutation. Thus, Bcl I cuts within the synthetic
fragments of the purified genomic DNAs from mutant recombinants
shown in Figure 4.2, with the exception of Q105C, producing two
fragments with the expected sizes of about of 2.7 and 2.2 kbp. The
substitution at position 105 modifies the recognition sequence for Bcl
I such that it can no longer cut at this site.
Similarly, Xho I makes an
additional cut within the synthetic fragment which carries the G231C
mutation, as expected (Figure 4.3B, lanes 3 and 4). In the absence of
the synthetic fragment used to introduce the G231C mutation, Xho I
makes a single cut within the Pst I fragment. This cut produces an
approximately 300-bp piece which is too small to be visualized on
the blot, and therefore only the remaining fragment of about 4.8 kbp
can be observed.
The synthetic fragments which carry V130C and
A160C mutations contain within them an additional Pst I site. For
these mutants and also the double mutant R82A/A160C which also
88
1 2 3 4 5 6 7 8 9 10111213
Kbp
_8.5
J7.2
_
am
l
_M
_
_=
_
_
SW
-
_-
4.8
4.3
_
SW_
6.4
5.7
3.7
- 2.3
-1.9
-
-1.4
1.3
Figure 4.2.
Southern blot analysis of purified genomic DNA samples
prepared from MPK40 recombinants of G72C. VO11, Q105C, and
G72C/D85N mutants.
The analyses of two separate recombinants are
shown for both G72C and G72C/D85N. Lane 1 and the evennumbered lanes showNthe results of P i destion.
The remaining,
odd-numbered lanes show the results of double di2estion with Pst I
and Bcl 1. Lanes:l MPK40 enomic DNA:
and 3, G72C(a): 4 and 5.
G72C(b): 6 and 7. G72C/D85N(a;:
V101C: 12 and 13. Q105C.
8 and 9 G72C/D85N(b):
10 and 11.
89
3, w
7,r
' 3
=
C
CL "
dq
r,
,)
=re
_.
,
mL Cg
,
3 n
&5
3,
_
C
L "
cm,
C)
CC
,C
cr
P-P
(D
" -'
.L
* 'I
-A
-L
"
rv
(D i4
P -:-- C Cn 4 C X
.b vOC
,
V
11
.K,,
11 I
sj . io
)))W
O a
=
V0
f
6
V)
I
I-!O
wi"
V
Or
e
Cct
L
no .
c
WclCD
YOa
C3
"
0-3
OQ
CD
CD39
500
c
=.
o
P
r,
e
S
B,
=f!
o
CD
Y·
o
tlt3
tZZ
Zgg
rr
ct
P
rd
hQ
C3
=33
Ca
nCD
Cli
ci
i I
5=0
=jl
Cd
,
p
o
c,
C
u·
e
t3
r
Cr
cl
3
e3
3
c
H"
rsr
f3
,,
C;
-r-3
r-f3
-J
Y·
v
C
3
sre
3
a
II
rV
w
C,
J:
iI
M)
cCs
dI ((
I.
P
=·
U
Pz n n
g
"a
E
,
w
I
a
V
90
1 2 3 4 5 6 7 8 9 1011 12
Kbp
7.1
6.1
- 5.0
- 4.0
" .... 3.0
-
2.0
1.6
1.0
Figure 4.4.
Southern blot analysis of purified genomic DNA from
MPK40 recombinants for R2A/G72C, D96A/VO11C S35C. A160C,
and R82A/A160C.
The analysis of three separate recombinants is
shown for the R82A/A160C double mutant. Lanes 1-7 show the
results o P.sr i digestion. Lanes 8-12 show the results of double
digestion with Pst I and BcI I. Lanes: I and 8, R82A/G72C; 2 and 9.
D96A/V101C: 3. S35C (see also Fiour
.5): 4. A160C: 5 and 10.
R82A/A160C(a):
6 and 11. R82A/A160C(b):
7 and 12,
R82A/A160C9c0..
91
1 23
4 5 6 7 8
9
Kbp
\--
8.5
_ 7.2
_ 6.4
"
·_.
_
..
.
\ 4.3
3.7
- 2.3
_
_WL-
Figure 4.5.
Southern
MPK40 recombinants of
G72C were included as
recombinants are shown
Pst I digestion. Lanes
5.7
4.8
- 1.9
1.4
1.3
blot analysis of purified genomic DNA from
S35C and V130C. Genomic DNA digests for
a control. The analvses of two separate
in each case. Lanes 1-5 shown the results of
6-8 showNthe results of double digestions with
Pst I and Bsm . Lanes: I and 6. S35C(a): 2 and 7, S35C(b), 3 and S
G72C: 4. 13OC(a': 5, Vi3OC(b}) 9. molecula- weight markers made
from a Bsr I digest of . Dhnae DN.
92
carries this additional site, Pst I digestion should produce two
fragments which are about 3 kbp and 2 kbp in size. This was
observed (Figure 4.4, lanes 4-7; Figure 4.5, lanes 4 and 5). For
R82A/G72C and D96A/V101C mutations, an additional digest with
Bcl I within their synthetic fragments produces the two expected
pieces of about 2.7 and 2.2 kbp (Figure 4.4, lanes 8 and 9). Bcl I cuts
only within the approximately 3-kbp PstI fragment of R82A/A160C
which contains the synthetic sequence carrying the BclI site. This
digestion produces a fragment of about 120 bp in size which is too
small for visualization on the blot shown. Finally, a digestion of the
Pst I fragment for the S35C mutant with Bsm I produces two
fragments with the expected sizes of approximately 2.8 and 2.3 kbp
(Figure 4.5, lanes 6 and 7). As expected, Bsm I does not produce a
cut within the Pst I fragment for the G72C mutant, since the
synthetic fragment used for this mutation does not carry the
recognition site for this enzyme (Figure 4.5, lane 8).
4.3.2
Purification
Membrane
of
Mutant
Proteins
in the
Purple
Form
The dark-adapted (DA) and light-adapted (LA) forms of the
purified purple membranes for all the mutants showed max values
similar to the values for wild-type bR (DA, =561 nm; LA, =568 nm).
The spectra for the dark-adapted and light-adapted forms of the
mutants are shown in Figures 4.7, 4.8, and 4.9. Similar to the wild-
type purple membrane, the absorbance ratio
1.8 in value for the light-adapted
membranes.
A 2 8 0 nm/A568 nm
was 1.7-
forms of the mutant purple
This indicates that the mutant proteins fold properly
and their level of purity is high. In addition, the purity of the
proteins were checked by SDS-PAGE. A typical gel is shown in Figure
4.6 and indicates that the mutant and wild-type proteins had the
same mobility on denaturing polyacrylamide gel electrophoresis.
of Cysteine Mutants
4.3.3 Derivatization
and Flash Spectroscopy Analysis
with
Fluorescein
The labeling stiochiometry obtained was 0.65-0.95 mol of BMF
per mol of bR mutant. By using the same lableing conditions, wildtype bR, which lackes cysteine, bound < 0.05 mol of BMF per mol of
bR. The absorbance maxima of the mutants in the visible region
were the same before and after labeling, identical to the max for the
wild-type bR in the visible in both cases.
93
1
2 3 4 5
6 7
Kd
-.
- 46
·
- 30
- 21.5
1- 14.3
- 6.5
Figure 4.6.
SDS-PAGE analysis of purple membrane purified from
wild type and mutant strains. PM proteins were electrophoresed
through 12% polyacrylamide and visualized with Coomassie Brilliant
Blue. The estimated molecular mass of prestained molecular weight
markers is shown to the right. Lanes: 1, wild type bR; 2,
D96A/V101C; 3, A160C; 4, G231C; 5. G72C; 6, VOIC; 7, molecular
weight standards.
94
:= a,
Absorbance
> ( C
'I e
.
O
v
b
00~ ee
\-
~
~
tc-)t
pJ
'V
=t
IIco
CCC~
3"
~t
F-i^ r-
?7
.
r=
Of
Il
O
1D
0
CD
=
©
c*
,)
?>
X
·
N-P1~
V
v
0 3> e.
CL 0
i.
O
o.
.d
=t
95
*"
Absorbance
w~
0o
- ell,
i
C_
00
0o
O,
o
C
C)
rD
8 o
,
_
a.
CD
6
C
acn
r!, I
'"
;
0p
(OD
_-f
r
0
o""
0D
-O
00
o
P
tfD =p-p
'-.
e.
0D
P9
eF
CT
Cm
0
X
° °- r
r3
cn
P
a
0-
_', a
CD
L
96
a:
-Q
0
.2
-Q
0
o0
o400
Wavelength (nm)
UiV/Visible absorption spectra of dark adapted (DA) and
light-adapted (LA, Durified purpl membrane the G231C mutant.
The mar,- values fo: DA (561 nm) and LA (568 nm) forms of the
Figure 4.9.
mutant
ar-. the same
a-, for- w ild type
bacteriorhodopsin.
97
The photocycle kinetics were measured before and after
derivatization, and it was observed that they were not altered by
fluorescein labeling. The kinetics of the rise and decay of M are
shown for derivatized G72C, A160C, and G231C in Figure 4.10 (the
labeled mutants are designated as G72C-MF, A160C-MF, and G231CMF). The M kinetics for G72C are identical to those of the wild-type
bR, similar to the V130C mutation, which is also on the extracellular
side. By contrast, A160C and G231C mutations, which are on the
The
cytoplasmic surface of the protein, perturb these kinetics.
mutants V101C-MF and Q105C-MF, present also within the
cytoplasmic domain, alter the photocycle kinetics as well, although
their effects are quite mild. Only the cysteine substitution at position
35 within the cytoplasmic region does not show any such effects.
Table 4.1 provides a comparison of the time constants and relative
amplitudes which describe the kinetics of M between the wild-type
bR and the mutants on the cytoplasmic surface which alter these
kinetics. It is important to note that with the exception of G231C all
the mutants only affect the decay of M. For G231C the rise of M is
accelerated and the decay contains an additional slow component
(Figure 4.10).
4.3.4
Proton
Release
Kinetics
Models of localized energy coupling between proton sources
and sinks require a transient confinement of released protons near
the membrane surface, along which lateral transport of protons
would take place (21 1). In order for significant lateral diffusion to
occur, the rate of proton release into the bulk medium must be
slower than the rate of transfer along the surface so that the protons
can be retained at the membrane surface for a sufficient period of
To determine these rates, dynamic methods of
time (212, 213).
The purple membrane
sufficient time resolution are required (214).
patches offer an ideal model system for such studies. The purple
membrane sheets have of a diameter of approximately 0.5-km and
contain about 104 light-activated bacteriorhodopsin proton pumps.
Because a simultaneous activation of bR molecules within the patches
can be achieved by a nanosecond light pulse, a transient proton
concentration jump at the extracellular membrane surface takes
place due to simultaneous proton release by the bR molecules. In H.
halobium cells, H+-ATP-synthase acts as the proton sink for bR. The
bR-induced transient increase in proton concentration occurs within
the 50-100
sec time range.
This process of proton release and the
98
0.05
0.04
C,
Co
0.03
-o
0.02
cn
0.01
0
-0.01
10
10
10
1 05
1 07
log (time/us)
rtgure4.10.
Fiasih-inaucec aosorbance cnanges at 410 nm for vwildtype bR and the cvsteine mutants G72C. A160C and G231C after
derivatizatior. with fluorescein.
G72C-MF.
mnI
KC'
.
A1 60C,-MF. G23 i C-MF.
Tim
T ioarithmic
Th; drivatives
arc designated
Conditions: pH 7.3. 22C. and 150
tims scale i fron: 10- 1 to 10'7 us.
99
subsequent proton uptake can be monitored spectroscopically, either
by the use of water soluble pH-indicator dyes in the aqueous phase
(95, 145, 153, 215,216)
or by attachment of pH-sensitive dyes to the
protein surface (154, 155, 217,218).
In the case of bR,
measurements using pH-indicator dyes also allow studies of kinetic
coupling between the proton release process and the formation of the
M intermediate.
We used the single substitution
mutants described
in this
chapter to attach the pH-indicator dye fluorescein at desired
positions within the loop regions of bR at either side of the
membrane. The proton detection times of the surface-bound
fluorescein dyes were compared with detection times in the aqueous
bulk phase by using the water-soluble dye pyranine.
Additionally, a
comparison of the kinetics of the M rise and the kinetics of dye
response to protons released at the extracellular side was made in
each case. In Figure 4.11 the kinetics of M rise and decay are
compared with the kinetics of proton release as detected by
fluorescein (DIFMF) and pyranine (DIFPY). The time traces for M
were fit by six exponentials, the rise times being marked by the
vertical arrows. The proton release time constants obtained by use
of the two dyes are listed in Table 4.1. Figure 4.11A shows that
fluorescein attached at position 72 on the extracellular side detects
the proton release, which occurs on the same side, 71 ,gsec after the
flash. On the other hand, it takes 850 ,usec before pyranine in the
bulk medium registers the same proton release.
The proton release
time of 71 tsec falls between the two major rise times of M, 2 and
(see Table 4.1). As it was pointed out previously, the photocycle
kinetics remain unperturbed in this mutant.
3
Surprisingly, the results
for detection of proton release with the dye attached on the
cytoplasmic side (i.e., the opposite side) were quite similar (Figure
4.11B and C; A160C, G231C, and Table 4.1; V1O1C and Q105C). In
these cases, the detection time also lies within the two main rise
times of M.
These experiments demonstrate that regardless of the
location of the bound-dye on the surface, the detection time for
proton released on the extracellular side is 10 times faster than in
the bulk.
The bulk detection times were about the same for all the
mutants and close to the wild-type value (Table 4.1). Of the current
set of mutants on the cytoplasmic side, only G231C alters the kinetics
of M rise, causing an acceleration of M formation (Figure 4.10, and
Table 4.1). It is nothworthy in this connection that the proton
release detection time is also slightly faster when fluorescein is
100
152us
G72C-MF
~
Extracellular Side i
,S-~sX
4 5us/
\
0 03
1 ii..S
M-intermediate
0 01
-0 01
i
D-lnicatCBO
7I
(Proton-Signal)
850u
-O003
At60C-ME
17us
0 05
-
Cvooilaamc: Side ]
475
0 03
0 01
'
&
74u5
~oH-lnoIcalor
p
(Proton Signal)
850us
-0 03
C
0 05
G231C-MF
lO
101s
CytootasmicSide
0 03
14
24
M-lntermediate
-i
0 01
-0 01
(Hotnigartar
|$
'l
60us
820us
.
1
(Proton Slgnet)
-0 03
10
10
C
s
1C
10
Time (log s)
Figure 4.11.
Comparison of the rise and decay of the M
intermediate with the kinetics of proton release. as detected by
fluorescein bound on the extracellular side [G72C(A)], or cytoplasmic
side [AI60C(B) and G231C(C)] of the purple membrane. Conditions:
pH 7.3. 22°C. and 150 mI KCl. The upper and lower parts of each
panel contain the positive absorbance change AA at 410 nm due to M
and the negative absorbance changes AAA due to fluorescein at 495
nm, (DIFMF! and nvranine ar 450 nm (DIFPY,. respectivxel. The
vertical arrows ndicate the time constants obtained from a
multiexponential fit. The fluorescein sianal was scaled b a factor
between
s: nal.
2 and
to,
i
e
it th
sanm- ampnitude
a
the pranine
101
attached to G231C (60 ptsec), although we cannot unequivocally
state
a direct correlation considering the magnitude of the experimental
error. For the rest of the mutants, the proton release times are
virtually identical, with an average value of 76 + 5 tsec on the
cytoplasmic side and 71 + 4 isec on the extracellular side.
The existence of a kinetic correlation between the M rise and
proton release was tested by investigating the temperature
dependence of these processes between 5C and 40°C for G72C.
Figure 4.12 shows the Arrhenius plots for the two major rise times of
M and the H*-release time. The plots for 2, and 3 give nearly
1 kJ/mol and are linear. The
equivalent activation energies of 68
plot for proton release, however, contains a break between 20°C and
25°C and appears to approach the straight line for 2 tangentially at
Therefore, the proton release time equals the
low temperatures.
second rise time of M at 5°C whereas it is slower than the slowest
component of the M rise at 40°C.
4.4
Discussion
The most intriguing finding of the described proton release
measurements is that a proton released on the extracellular side of
the membrane is detected on the opposite side about as quickly as it
is detected on the release side and 10 times faster than in the bulk
phase.
The difference in proton detection times between fluorescein
on the cytoplasmic
side and pyranine in the bulk had been
previously observed from analogous studies performed in the mixed
micelles (154, 155, 218). We were very interested in performing
these studies with bR present in its native extended purple
membrane sheets, which are much bigger than micelles and contain
The above observations made in the
different lipid composition.
patches provide evidence for the existence of lateral proton flow at
the membrane surface and give support to hypotheses of local proton
coupling. Thus, a model can be provided for how an efficient
coupling between proton translocation by bR and ATP synthesis by
the H*-ATP-synthase may be achieved in H. halobium cells. The fact
that proton mobility from one side to the other of bR in micelles
varies with lipid head group compositions that affect pKa values and
proton dwell times (154, 155) corresponds with the notion of proton
migration along the surface. Yet, interestingly, the much bigger
diameter of the purple membrane sheets does not cause sufficient
delay in lateral proton diffusion at the membrane surface so that
102
10"
102
10
4
It
3.1
3.2
3.3
1/T
3.4
3.5
3.6
3.7
(10- 3 1/K)
Figure 4.12.
Arrhenius plot of the time constants 2: (0) and 3 (x)
for the formation of the M intermediate and for the proton release
(,) measured with fluorescein bound to position 72 (G72C-MF). The
sample was at pH 7.3 in 150 mMl KCI. From the slopes of the two
linear plots for the rise of M. activation energies of 66.9 k/mol ()
and 68.1 kJ/mol (x) were obtained. When the curved plot for the
proton release was analyzed by assuming linear low- and hiahtemperature branches with a break around 220 C. the activation
energies were 46.5 and 30.4 kJ/mol. respectively.
103
differences in detection times could be observed between the two
sides of the membrane.
The response time of fluorescein to protons
released to the extracellular side is virtually the same regardless of
its position on either side of the protein. This implies that for
protons the magnitude of the surface diffusion constant D (212, 219)
is large. A rough estimate of this constant was obtained based on
assuming an average value of 0.25
m for purple membrane radius
and a difference of <5 jisec between proton detection times on both
sides of the membrane. The obtained value is >3 x 10-5 cm 2 /sec,
indicating that surface diffusion of proton may occur even faster
than diffusion into the bulk water phase for which the value of D is 9
x 10 5 cm2 /sec. These observations are consistent with the
hypothesis that released protons are trapped initially at the
interfacial surface layer on the extracelluar side and they
subsequently equilibrate along the surface much more rapidly (75
gtsec) than they diffuse into the bulk medium (=880 tsec). In such a
scenario, the high buffer capacity of the purple membrane and its
high concentration of negatively charged lipid head groups (220,
221) would allow initial confinement of the released protons and
their subsequent equilibration at the membrane surface. Figure 4.13
shows a schematic representation of the above hypothesis.
In comparing the studies performed with the micellar system
and those reported here, the observed difference between the proton
detection times at the cytoplasmic surface and in the aqueous phase
is much more pronounced in purple membrane sheets. The large
difference in the proton detection times of pyranine between the two
systems (125 tsec in micelles versus 800 .isec in purple membrane)
may be attributable to differences in surface charge, pKa of the lipid
headgroups and buffer capacity.
For example, whereas the molecules
of the mixed micelles used in previous studies are zwitterionic, the
major lipids of the purple membrane are negatively charged. On the
other hand, the fluorescein response times on the cytoplasmic side of
the membrane match very closely for micelles and purple membrane
patches.
For instance, the response times of fluorescein attached at
position 160 in micelles and in purple membrane sheets are 60 and
74 sec respectively.
Considering that bR micelles are much smaller
than the extended purple membrane patches, this observation
suggest that the rate-limiting factor for proton detection on the
cytoplasmic
side may be the step around the purple membrane edge.
104
f
-
I
Cartoon of proton flow along the surface and around
Figure 4.13.
the edge of a purple membrane patch and into the bulk. The figure
is not to scale. The lateral dimension of the membrane is about 100
The interfacial layer between
times larger than the thickness.
membrane surface and bulk is marked in gray. The pH indicator
dyes, fluorescein and pyranine are indicated by FLU and PYR,
Protons released on the extracellular side (bottom) are
respectively.
detected b
covalently bound fluorescein on the cytoplasmic surface
at position 160 well before they are detected by pyranine in the bulk
(74 s after the flash versus 880 us). The proton release time on the
extracellular surface at osition 72 i.s 71 us. Arrows indicate
schematically the lateral proton migration along the surface and the
delayed transfer from the surface laver into the bulk. In the actual
experiments. each bR molecul-- has one bound surface dye on one
side onlyv.
105
Regarding the existence of a coupling of the kinetics of proton
release and M rise, it is important to consider that the Schiff base
deprotonation is a two-step process. As mentioned in Chapter 1, the
proton transferred to D85 from the Schiff base during the L to M
transition stays bound to this residue until the very last step of the
photocycle, and the actual proton release is accomplished by the XH
group. Accordingly, there is no a priori requirement for the rates of
Schiff base deprotonation (i.e., M rise) and fluorescein protonation to
be the same. The Arrhenius plots shown in Figure 4.12 show that
these rates in fact differ.
However, at lower temperatures the
Arrhenius plot for proton release approaches that of
rise time for M, and at 5C they have the same value.
2,
the second
These findings
suggest that whereas Schiff base deprotonation is rate-limiting at
lower temperatures, a subsequent reaction becomes rate limiting
when the temperature is raised.
This suggestion fits well with the
idea of a unidirectional two-step proton release process discussed
above with 2 assigned to the first step, the deprotonation of the
Schiff base. It is interesting to note that with the exception of G231C,
all the mutants on the cytoplasmic side with altered photocycle
kinetics only affect the kinetics of M decay specifically. Because M
decay is associated with proton uptake from the cytoplasmic side of
the protein, these observations suggest that the cytoplasmic region of
bR is dynamically involved during the second half of the photocycle.
This indeed seems to be the case based on the experimental results
discussed in the next two chapters.
106
Chapter
Five
Structural Studies of Bacteriorhodopsin: II. Time-Resolved
Electron Paramagnetic Resonance Studies
of Spin-Labeled Mutants
5.1
Introduction
Several independent studies have demonstrated the occurrence
of conformational
changes of bR during the photocycle.
Both neutron
and electron diffraction studies on bR cold-trapped during the
photocycle have revealed changes in the projected structure (125,
128). Time-resolved X-ray diffraction studies on mutants with a
slow photocycle have detected similar changes (126, 127).
Furthermore, structural changes in the protein backbone during the
M to N transition
have been identified
by time-resolved
Fourier
transform infrared (FTIR) spectroscopy (130). However, none of the
above approaches have provided both real-time resolution and
localization of the structural changes. I was interested in exploring
the possibility of obtaining such information by performing timeresolved site-directed spin labeling (SDSL) studies. The characterized
cysteine mutants described in Chapter 4 are ideal for this purpose.
The substituted cysteine residues provide accessible and specific
attachment sites for the spin-labels. In addition, the mutant proteins
are purified in a native two-dimensional lattice form similar to that
of the wild-type bR, eliminating possible occurrence of artifactual
transitions due to in vitro reconstitution in micelles. The availability
of the lattice form of the mutants also allows the use of SDSL to
obtain information on the ground state structure of bR trimers. The
results can be compared with previous EPR studies of the ground
state structure of some of the same mutants in the monomeric form
(45, 47, 207). Time-resolved SDSL allows monitoring of the localized
conformational changes of bR in real time using aqueous suspensions
of purple membrane under physiologically relevant conditions.
conditions cannot be completely maintained with diffraction
Such
techniques.
In those studies, the bR sample is partially dehydrated.
In addition, the bR sample has to be frozen so that the M
intermediate can be trapped long enough to perform the
measurements.
The importance of complementing low-temperature
structural studies with room temperature measurements has been
shown by FTIR studies of bR. Although the room-temperature rapidsweep difference spectrum of the bR568- M transition (130) is similar
107
to a static FTIR spectrum obtained at -20°C (222), the rapid-sweep
spectrum has significant new features in the amide I (1660 cm-1)
and amide II (1555 cm-1) regions. Accordingly, structural changes
of the protein backbone associated with M formation are observed at
room temperature but not under frozen conditions. Recent
experiments
have shown how dehydration
of bR crystals or their
cooling results in selective entrapment of different M species (223,
224).
This point will be elaborated upon in this chapter with regard
to interpretation of observed structural changes based on the
extistence
of different M species in the photocycle.
This chapter describes the results of the first time-resolved
EPR studies of spin-labeled bR mutants. The measurements were
performed in collaboration with Professor Wayne Hubbell at the
Jules Stein Eye Institute, UCLA.
instrumental set-up.
measurements
Dr. Juergen Steinhoff designed the
Under his tutelage, I performed the
for several of the mutants described
Professor Hubbell's laboratory at UCLA.
presented for spin-labeled
in this chapter in
The experimental results
G72C, VIOIC, and Q105C have been
published (reference 133).
5.2
5.2.1
Experimental
Derivatization
Procedures
of Mutant Proteins
Spin-labeling of purified purple membrane sheets was
performed in the presence of a five-fold molar excess of the label for
12 hours in 200 mM sodium phosphate buffer, pH 7.0 at 25°C and
under dim light conditions. The stock label solution was prepared
using highly pure acetonitrile as solvent (American Burdick and
Jackson) and was stored at 4C under dim light conditions.
Excess
free label was removed by successive centrifugal washing of the
membranes
(6 washes).
For EPR measurements,
the purple
membrane was resuspended in 0.1 M sodium phosphate buffer, pH
7.0, 0.1 M NaCl. The stiochiometry of labeling was 0.7 to 0.9 mol of
spin-label per mol of bR mutant except for G231C for which a
stiochiometry of 0.14 was found. Under the same conditions, no
derivatization was observed for the wild-type bR which lacks
cysteine residues. The concentration of each mutant sample was
calculated based on its
max
value in the light-adapted
form, using the
extinction coefficient =63,000 M.cm'
(208). For measurement of
labeling stiochiometry, EPR spectral integration was performed for
108
each mutant after calibrating the EPR machine with a known amount
of the water soluble spin-label TEMPO. The mutants were modified
at their single reactive sulfhydryl group with (-oxyl-2,2,5,5to give the
tetramethylpyrroline-3-methyl)methanethiosulfonate
nitroxide side chain R1.
The mutant V1O1C was also labeled with (1-
'Il
wich gives the
oxyl-2,2,5,5-tetramethylpyrroline-3-trans-propene)
nitroxide side chain R2.
r - nrnpoin
'"'
S-S
S-S.",
II
g- nrotein
%_1 P
r
I
O
0
R2
R1
Figure 5.1.
Structures of methanethiosulfonate derivatives used for
spin labeling of bacteriorhodopsin mutants. The nitroxide side chains
introduced by derivatization of bacteriorhodopsin with these spin
labels are designated as R
5.2.2
Instrumental
Paramagnetic
and R2.
Design
Resonance
for
Time-Resolved
Electron
Studies
A block diagram of the instrument used for the time-resolved
studies is shown in Figure 5.2.
Before describing the details of the
set-up it is useful to briefly review the theory of electron spin
resonance. Fundamental physics provides the principles upon which
E.S.R. is based. The unpaired electron of the spin-label, which orbits
around a nucleus, also revolves around what can be thought of as an
This spinning
internal imaginary axis and has spin momentum.
charge can act as a very small magnet and give rise to a rotating
magnetic field. The external magnetic field of an EPR instrument can
therefore orient the unpaired electron in discrete directions. If the
electron orients itself with its magnetic field in the opposite direction
of the externally applied field, an spin quantum number (M,) value
of +1/2 is assigned to the electron. The electron is in a highest
energy position in this unstable state. If, on the other hand, the
electron orients itself in the direction of the applied magnetic field,
the spin quantum number is -1/2 and the electron is in a lower
109
magnet
switch
flashlamp
Figure 5.2.
A block diagram of the apparatus used for timeresolved EPR studies of spin labeled bR mutants. The pulse
generator triggers the data acquisition system and the flashlamp. and
in addition supplies a reference signal to the lock-in amplifier.
With
this arrangement, the spectrometer can be operated in two kinetic
modes. In mode 1. the switch at the right is closed and the EPR
signal goes directly to the computer.
The time dependence of the
EPR signal following a light flash can be recorded directly at any
fixed position of magnetic field, that is, at any fixed part of the EPR
spectrum.
Repetitive flashing of the light will provide signal
averaging. This mode is used to determine the time constant for
spectral changes. In mode 2. the switch is open and the EPR signal
goes into the lock-in amplifier that is set to amplify only signals that
have tne same phase and frequency a the reference signai. Since
the reference signal coincides with the freauencv of the pulsed-light
source. the output signal a a function of field will give only the
change in EPR signal due to light. This mode gives the spectral
chane at a!! field. but no informatiov o the time constants.
110
energy and more stable state. This situation is analogous to when a
small compass needle is placed between the two poles of a magnet
where the needle can align iteslf with its south pole pointing towards
the north pole of the magnet or vice versa. The Zeeman effect
described above, which leads to the splitting of an energy level by
the application of a magnetic field, can be shown by an energy
diagram:
Ebam
p Rem
Themy
6I
I
f
0
Iz
'U
.-I
MAGNETICFIELD
The technique of ESR is based upon making electronic
transitions between more populated lower energy states to less
populated states which are higher in energy. In the above diagram,
such a transition would be between the state with -1/2 spin to the
state with +1/2 spin and the energy required, hv, is the quantum of
energy with frequency v (h is Planck's constant).
In an EPR
instrument, this energy of transition is provided by a microwave
source. As mentioned, the unpaired electron of the free radical
travels in an orbit around a nucleus, specifically the nucleus of a
nitrogen atom in nitroxide spin-labels such as the ones used in our
studies. Similar to the orbiting electron, the quantized energy levels
of the nucleus are described by assignments of nuclear spin quantum
numbers (MI). The magnetic quantum numbers of the nitrogen
nuclear spin are -1, 0, and +1 in our case. The interaction between
the electron and the nucleus, both of which have their own magnetic
moments, causes the energy level of the electron to be split by a
small amount, giving rise to hyperfine structure.
This splitting will
depend upon how the proton spin of the nucleus aligns itself with
respect to the applied magnetic field and the electron spin. This
means that in an applied magnetic field the electron spins of -1/2
111
and +1/2 each will be split by the nuclear spins, as shown in the
energy diagram below:
Ms M
I
2
I
0
-I
2
0
The only allowed transitions between the nuclear levels shown
in the above diagram are those for which AM, =0. Consequently, only
the three transitions depicted in the diagram are allowed, each of
which gives rise to a resonance peak in the EPR spectrum. As can be
observed, the amount of energy required to make each transition
possible depends on the strength of the applied magnetic field.
In Figure 5.2, the big magnet on the sides of the spin-labeled
sample provides the external field. Omitted from the figure for
simplicity is the presence of a microwave source which is connected
to the sample chamber and provides the energy required for the
transitions discussed above. In our case, the microwave source was
at the X-band frequency of 9 GHz. The sample is also surrounded by
a coil carrying alternating current which produces an oscillating
magnetic field. Therefore, a small oscillation is present at each value
of the external field causing the output EPR signal to be AC in nature.
The advantage of having an AC output instead of a DC signal is that
the level of noise is considerably reduced as the detected output
value does not change due to the effects of environmental factors
such as temperature or electrical noise. The temperature of the
sample is controlled by a constant flow of nitrogen gas, the
temperature of which is measured by a thermocoupler positioned
right above the sample chamber. For obtaining the EPR spectra of
the spin-labeled mutants in the ground state, we placed the capillary
containing the sample in a loop-gap resonator.
Loop-gap resonators
have a filling factor of 1, so that 100% of the magnetic flux goes
112
through the sample, giving rise to better signal to noise ratios. The
EPR spectra of light-adapted samples in the ground state were
measured by sweeping the entire magnetic field.
For time-resolved
EPR measurements which were performed to monitor lightdependent conformational changes of bR, the sample holder was
placed in a rectangular cavity. The filling factor of a rectangular
cavity is two orders of magnitude smaller than that of a loop-gap
However, the openings on the side of the cavity allow the
resonator.
exposure of the sample to triggered light flashes. In addition, a much
larger amount of each sample could be used, which is important for
the ablility to observe light-dependent EPR transients.
Figure 5.2 shows that the data acquisition system and the
In addition, the pulse
flashlamp are triggered by a pulse generator.
generator supplies a reference signal to the lock-in amplifier, the
purpose of which is discussed below. This arrangement allows the
use of the EPR spectrometer in two kinetic modes. In mode 1, the
switch shown at the right side of the diagram is closed and the EPR
signal coming from the spectrometer goes directly to the computer
By keeping the external magnetic field at any
(data acquisition).
fixed position (i.e., staying fixed at any part of the EPR spectrum), the
time dependence of an EPR transient signal following excitation of bR
by light can be recorded directly, if any changes occur at the selected
field value. This mode is used to determine the time constant for
spectral changes.
Repetitive light flashes will provide signal
averaging. In mode 2, the switch is open and the light-dependent
EPR signal goes into the lock-in amplifier that is set to amplify only
signals that have the same phase and frequency as the reference
signal. As depicted in Figure 5.3, the reference signal coincides with
the frequency of the pulsed light source which, like the reference
signal, is triggered by the pulse generator. The net effect of this is
that the reference signal amplifies only those changes in EPR signal
which are due to light and subtracts out any background electrical
noise. This mode gives the spectral change at all magnetic field
values but does not provide information on the time constants.
5.2.3 Flash Spectroscopy:
410 nm and 570 ni
Monitoring
Absorbance
Changes
To correlate the light-induced EPR spectral changes with
specific intermediates of the photocycle (Figure 5.4), the M kinetics
and the kinetics for recovery of the ground state structure during the
second half of the photocycle were monitored. This was achieved by
at
113
CD-DC
oo
_.
CD
C
0
D
D.
P
CD-t
0
-11
©C
CDO
')
ro
M
,
B
C)
CD
3
O
©
CD
CD CD
c
D
0
C,
m
o2
cn
(n
p. O-p:~CD
_" ©
U)CD D
CD
0
1_
0
_F
cn
i- ,
e
-D
CD3
X
>L
DC
=
-C
_.
CD
nn
114
bR 5 70
7
-·H'
0640
H+
"
*,A.
,yLupydasrnIuc
surface
K5 9 0
r
0x
L5 5 0
Extracellular
H+f
Figure 5.4.
surface
The photocycle of bR. The sheme is that of Lozier,
Bogomolni and Stoeckenius (1975), with the addition of back
reactions. Although a single intermediate is shown for M, it has been
proposed that there are at least two distinct forms. The wavelengths
of maximal absorbance are given for each intermediate.
115
measuring light-dependent absorbance changes of the samples at
410 nm and 570 nm respectively.
were
The measurements
performed for both unlabeled and spin-labeled samples.
To perform
the measuremnts, 100 lisec light flashes at 580 nm were used for bR
excitation, with a probe beam passing through the sample. The
transmitted light from the sample passed through either a 410- or a
570-nm cut-off filter before reaching a photomultiplier detector.
5.3
Results
5.3.1
Time-Resolved
EPR
Measurements
The ground state EPR spectra of the spin-labeled mutants and
their light-dependent EPR changes are shown in Figures 5.5, 5.6, and
5.7. The shapes of the EPR spectra of the mutants (Figure 5.5, a
through d; Figure 5.6, a through c, and Figure 5.7, A through C)
reflect the internal motion of the nitroxide side chain.
determined
protein.
by the degree of interaction
This is in turn
with nearby groups in the
Specifically, the broadness of the first and last resonance
peaks (+1 and -1 resonance lines) are convenient markers for the
degree of immobilization of the spin-label. The spectra of C10R1,
C72R1 and C103R1 (Figure 5.5, a and c; Figure 5.6, b) reflect
intermediate immobilization of the nitroxides and are similar to
those reported earlier for monomeric bR (47, 207).
This suggests
that the interactions restricting the motions are intramolecular and
not between the nitroxides and adjacent bR molecules in the lattice.
In contrast, the EPR spectrum for C105R1 and C130R1 reflect strong
immobilization of the attached nitroxide (Figure 5.5, d; Figure 5.6, c),
in contrast to the relatively high degree of mobility observed in the
monomeric state (45, 47). Thus, the immobilization in the lattice
must be the result of intermolecular interactions at these sites. The
rest of the mutants have not been studied in the monomeric state
previously, and the results shown here are the first EPR
Interestingly, the maximum of the
measurements of these mutants.
+1 resonance peak is split in spin-labeled S35C and G231C (Figure
5.6, a; Figure 5.7, c) with one component more mobile than the other.
This suggests that the spin-label at each of these sites assumes two
Finally, an increase of the label mobility is
different conformations.
observed in C101R2 relative to ClO1R1 (compare Figure 5.5, a and b).
This indicates that the nitroxide in the side chain of C101R2 is
extended from the backbone relative to C101R1. Considering the
relative lengths of the spin labels R1 and R2, the degree of this
extension may be as large 3 angstroms.
116
a
C11
.
I!
e
__INC101R1
IZ
I
I
I
b
f
!
I
C101R2
1
i
c
C
I
9
!1
A
i
I
ii
C72R1
~~~
5--
-
I .~~~~~~
\ j1
~
h
II I
COI5RI
V
10G
I
Ground state EPR spectra (a through d) and difference
Figure 5.5.
spectra e Inrougi ih) for tne indicated species. The dotted vertica
lines indicate the position; of the outer hyperfine extrema that
The arrows in (a) indicate the
characterize immobilized nitroxides.
sirnal from a small amournt <15%;)of unreacted spin label that is
oresent in each sampie (a-dL.
117
Difference Spectrum
Spectrum
(a)
(d)
,/ ,
f
v
-
C35R1
'l
(b)
(e)
i
,i
,,j
C103R1
(f)
(c)
C 30R1
4Y,
Ground state EPR spectra (a through c) and difference
Figure 5.6.
spectra (d throuih f) for th indicated species. The splitting of the
lovw-field maximumI (+1 resonance ine= for C35RI indicates that the
attacnec, spin ia'.
assume, ti', differen: conformations.
One
component i. more mobile. a iudged b the broadness of the lines.
.-. similar splittin
i observed fo, the sPin labeled C-23i- mutant.
errorme
a )i 7.0) iln then stancard buffer a;
The ex>eriment. wert
room tem erature .
118
A
A
I II~
Bi
B
A
IC -I-
f
-- 0
xCCL
" i
C
4
A,
1
Ii"
ft\i J
Figure 5.7.
Ground state EPR spectra for (A) D96A/C101R1, (B)
C160RI and (C) C231R1. Arrows at selected B-field values indicate
the direction of the observed transient changes during the
photocvcle.
The conditions were as in Figure 5.6.
119
(a)
1
RI
0
(b)
1
R2
0
(c)
OD 5 7 0
1
WT
n
10-3
10-2
10-1
100
time (s)
Normalized optical densities (ODs! and EPR spectral
Figure 5.8.
function of time foliowing a iight fiasi, o-:
amplitudes recorded as
mutant (A) ClOIRI and (B) CIO1R2. Optical densities at 412 nm and
570 nm monitor the Ni intermediate and around state. respectiv el\.
For comparison. (C! the opticai density transients for the wild-tvp
nroteil
are sown .
120
Some of these results can be usefully compared with the bR
structural model shown in Figure 5.9. According to this model, the
nitroxide in ClOIR1 should interact with groups in the E-F
interhelical loop, which is consistent with the partial immobilization
noted above. The increased nitroxide mobility that is observed in
Cl10R2 relative to ClOIR1 is also consistent with the model shown in
Figure 5.9, because an increase in the length of the side chain should
move the nitroxide group away from interaction with the tip of the
E-F interhelical loop. The nitroxide side chain at C105R1 points
outward in the model, with limited intramolecular contacts. This is
consistent with the high mobility of the label in monomeric bR.
However, in the lattice, this side chain would make direct contact
with the B helix of a neighboring bR molecule, resulting in significant
immobilization, such as that which is observed. The strong
immobilization observed for C130R1 in the lattice form, in contrast to
what is observed in the monomeric state, also suggests that in the
lattice form the nitroxide side chain at position 130 is in direct
contact with a neighboring bR molecule. Reliable predictions
regarding the nature of the interactions of nitroxides at the other
sites cannot be made from the model, because of the uncertainty in
the conformotion of these bigger loop regions. This is especially the
case for C72R1, which is located within the long B-C loop. However,
the EPR data sheds light on the relative mobility of these sites in the
ground states of monomeric and lattice forms of bR. Thus,
information is provided on the local effects of protein-protein
interaction in the lattice, and a point of reference is established for
examining light-dependent conformational
changes.
The light-dependencies of the EPR spectra are shown for all the
mutants in Figures 5.5, e through h; 5.6, d through f; and 5.7, arrows
in A through C. Changes in the EPR spectra caused by photoexcitation
are observed for all the spin-labeled mutants except C72R1 and
C105R1. A quantitative comparison of the magnitude of the lightdependent spectral changes for the different mutants is difficult to
However, a qualitative inspection of the intensity of the
perform.
difference spectrum for COlRl (Figure 5.5, e) suggests that a
particularly significant conformational change occurs at this location.
The difference spectrum for C101R is unambiguously interpreted as
a decrease in nitroxide mobility, as judged by the changes in the
The structural transition in the
amplitude at the spectral extremes.
protein causing this change is not likely to be global because no
difference spectra were detected for C72R1 or C105R1 (Figure 5.5, g
and h). The mutant CO11R2 has a difference spectrum of lower
121
Figure 5.9.
The structure of the bR backbone. The helices are
identified as A-G. and the side chain R1 at 72 101 and 105 is shown
as a stick model. Coordinates of the helical segments were obtained
from the Brookhaven Protein Data Bank. Helix D was shifted by 3 A
toward the cytoplasmic side. as suggested by recent diffraction data
(R. Henderson. personal communication').
Coordinates for the
interhelical loops and nitroxide side chains were constructed by
homology and energy minimization methods (Insight II. Homology
and Discover: Biosym Technologies. San Diego. CA).
122
kl-,..IN
I
123
amplitude but of the same character (Figure 5.5, f). This is expected,
because R2 at this site has weaker tertiary interactions and should
therefore be less sensitive to changes in structure.
5.3.2 Flash Spectroscopy:
at 410 nm and 570 nm
Monitoring
of
Absorbance
Changes
To correlate the light-induced EPR spectral changes in the
mutants with a specific intermediate of the photocycle, the change in
the optical densities (ODs) at 412 nm (M intermediate) and 570 nm
(bR recovery) were recorded for both the unlabeled and spin-labeled
samples following light excitation. The kinetics of these changes for
labeled samples were compared with the time dependence of their
EPR signal measured following a light flash and with the magnetic
field fixed at the maximum of the difference signal (mode 2 in Figure
5.2). These results are presented in Figures 5.8, 5.10, 5.11, and 5.12.
Figure 5.8 shows that for both ClOIR1 and C101R2, the EPR spectral
changes appear with the decay of the M intermediate and reverse
with the recovery of the ground state species.
Therefore, the
structural changes that take place near the nitroxide at site 101
occur either during M decay or during the formation of the N and/or
O intermediate, or a combination thereof. The mutant C103R1, which
is in the C-D interhelical loop two residues away from site 101 also
shows conformational changes, with kinetics similar to the changes
detected by COIR1 and C101R2 (Figure 5.12, A). Interestingly,
however, the conformational changes observed with the
D96A/C101R1 double mutant occur after M formation, but prior to M
decay. A comparison between this result and the observations for
C1OlR1 is treated further in the Discussion.
For mutants C35R1, C130R1, C160R1, and C231R1, protein
motion near the nitroxide side chains has already occurred by the
time M reaches its maximum amplitude (Figures 5.10; 5.11, B; 5.12,
B). As in all the other cases, these conformational changes are
reversible, and disappear with the recovery of the ground state
species. For the C231R1 mutant the results should be considered
with caution, because the EPR signal was quite weak due to a low
stiochiometry of labeling.
However, it is interesting that the timecourse of the decay of the observed EPR transient for C231R1
depends on which fixed value of the magnetic field is chosen for
observing the transient.
Therefore two different decay times are
obtained (Figure 5.12, B). It may be that the reorientational
of the nitroxide side chain is restricted differently after
motion
124
®
A
0
\
·*
0o
0
0
s
I
0
0 0
0
00**
0
10' 3
.·~·
10 ' 1
10-2
Time/s
-
-
B
0
-
I~
*
*
0
S..X
\
.0
0
0
00
0
*0
4
0
0
I0
.
0
10 -2
10 - 1
Time/s
Figure 5.10.
Normalized optical densities (ODs) and EPR spectral
amplitudes recorded as a function of time following a light flash for
mutant (A) C35R1 and (B) C130R1. Optical densities at 410 nm
monitor the M intermediate. () EPR signal; (-) OD 410.
125
A
-g \
II"
'E'
eeIe
*
0.00 1 3.01
.1
I
Ied
'e'U
10
time/s
B
0.001
0.01
time/s
Figure 5.11.
Normalized optical densities (ODs) and EPR spectral
amplitudes recorded as a function of time following a light flash for
mutant (A) D96A/C101R1 and (B) C160R1. Optical densities at 410
nm monitor the M intermediate. () EPR signal; (-) OD 410.
126
10-3
10-1
10'-2
10O
Time (seconds)
B
ICI
102
10
10c
Timeis
Figure 5.12.
Normalized optical densities (ODs) and EPR spectral
amplitudes recorded as a function of time following a light flash for
mutant (.) Ci03R. and (B) C231R.
Optical densities at 410 nm
monitor the M intermediate. (..U) EPR sinal: ( ) OD 410. In the
G231C mutant. the time course of the decay of the EPR transient
signal depends on which part of the EPP spectrum is analyzed for
detection o
transients.
127
photoexcitation for each of the two separate orientations that it
appears to assume at site 231 in the ground state (Figure 5.7, C).
However, additional measurements of EPR transients at different Bfield values of the magnetic field are necessary in order to verify
this.
Except for C35R1, all the other spin-labeled mutants which
revealed light-dependent motion in bR had some effect on the
kinetics of the photocycle, as compared to the unlabeled proteins. On
the other hand, the photocycle kinetics for C72R1 and C105R1, which
do not detect any motion, were the same before and after labeling.
Both were similar to the wild-type photocycle. These effects are
expected, because structural changes are detected by means of an
interaction of the spin label with the protein. If the protein
structural transition is coupled to an optical transient, such
interaction must necessarily modify the kinetics of the photocycle in
a manner related to the strength of the interaction. This is shown
when comparing the photocycle kinetics of C1O1R1 and C101OR2,
which are slowed down by factors of 10 and 2 respectively (Figure
5.8). As observed in Figure 5.5 (e and f), these effects correlate with
the strength of the observed light-dependent EPR signals for C1O1R1
and C101R2, which are otherwise identical. Spin-labeling of V130C
and A160C slowed down the photocycle by less than a factor of 2
compared to the photocycle of unlabeled samples.
5.4
Discussion
The spin-labeling
studies described
in this chapter have
provided structural information regarding the loop regions of ground
state bR in the purple membrane lattice.
This information
complements structural data based on electron diffraction studies, in
which the structure of the loop regions cannot be modeled (35). In
addition, these studies for the first time provided information on the
motion of bR in real time, during the completion of the photocycle.
The current resolution of the technique for observing lightdependent conformational changes of bR is in the millisecond time
range. However, it should be possible to improve the resolution by
use of a laser excitation light flash having a much shorter life-time,
and/or the use of faster lock-in amplifiers. These improvements
may allow monitoring of conformational changes of bR during the
earlier steps of the photocycle without the need to have the samples
in a frozen state.
128
Previous structural studies have shown that conformational
changes of bR occur in association with M formation.
This was
confirmed by the EPR results obtained from several of the spinlabeled mutants shown here. Furthermore, we find that some
structural changes occur in a specific fashion near the C-D interhelical
loop during M decay, whereas at some other sites, the tagged
residues indicate no change. These results indicate that sequential
molecular motion occurs in separate localized regions of bR during
the formation and subsequent decay of M (e.g., compare Figure 5.8, A
with Figure 5.11, B). The nitroxide in ClOIR1 is predicted to be in
contact with the E-F interhelical loop (Figure 5.9). Therefore,
appropriate movements of helices E, F, C, D, or even B, at the
cytoplasmic surface, could theoretically be responsible for the altered
nitroxide mobility at site 101.
However, the spin label in C105R1
does not sense any changes upon photoexcitation, suggesting that the
cytoplasmic part of helix D does not move. Also, because C105R1
contacts the B helix of an adjacent bR molecule in the protein lattice
trimer, the absence of a spectral change further implies a lack of
movement at the cytoplasmic end of helix B. Therefore, we conclude
that the transient spectral change is the result of movement in the EF and/or C-D interhelical loop, which may, or may not involve helices
E, F and C, but probably does not involve helices B and D. Consistent
with this hypothesis are the time-resolved X-ray data presented in
the next chapter for bR derivatized at site 101 with a mercury label.
In the double mutant D96A/C1O1R1, which has a very longlived M, structural changes were found to occur prior to the decay of
the 412 nm species. The existence of different types of M species, all
of which absorb at or near 412 nm, was discussed previously.
This
point should be kept in mind when comparing the results of different
structural studies involving different conditions, as they may not all
trap the same M intermediate (223, 224). In comparing the results
for D96A/C101R1 with the observations made for C1OlR1, the
observed conformational changes may correspond to a late M-like
intermediate. Time-resolved FTIR studies using mutants with neutral
substitutions of residue D96 have shown the presence of an
intermediate which absorbs at 412 nm, like M, but has a protein
backbone structure similar to N and is therefore referred to as MN
(103).
The possibility exists that the observed movement at site 101
corresponds to this intermediate, which cannot be trapped in wildtype bR, but is observed in D96N (103) or D96A (Professor Akio
Maeda, personal communication) when the pH is raised.
129
Alternatively,
mutation of site D96 to alanine may give rise to the
occurrence of different types of motion during the photocycle.
Recently, reversible light-induced conformational changes were
observed with the use of a spin label in the C-D interhelical loop of
rhodopsin (169, 170).
It is significant that light-dependent changes
in bR structure are also detected by a label in the C-D interhelical
loop. As for the functional significance of the structural changes
observed in bR near site 101, Asp96, five residues from the spin
label at this site, is involved in the reprotonation of the Schiff base
during M decay and is reprotonated from solution during the decay
of N (Figure 5.4). Evidently, the EPR signal changes are coincident
with this process and thus may reflect structural changes near 101
as a result of changes in the protonation of Asp96. The x-ray data
presented in the next chapter provides evidence that the transient
change may reflect the opening and closing of a pathway from the
aqueous solution to Asp96. Recently, a change in local electrostatic
potential around residue 101 was reported to have a similar relative
time course (225).
130
Six
Chapter
Structural
III.
Containing
6.1
Studies
Time-Dependent
X-ray
of
Bacteriorhodopsin:
Measurements
of
a Cysteine-
Mutant and Its Mercury Derivative
Introduction
The EPR results presented in the previous chapter showed that
a significant light-dependent conformational change takes place near
residue 101 during the M -
N transition in the bR photocycle.
I was
therefore interested in finding out whether more information
regarding the exact nature of this conformational change could be
The X-ray and
obtained by the use of a complementary technique.
electron microscopy diffraction techniques used to obtain twodimensional projection maps of bR helices do not provide structural
information on the putative loop regions. This is due to a lack of
sufficient atomic density in these areas, otherwise available in the
transmembrane ax-helices because of the three dimensional stacking
of the repeating peptide units. The lack of discernable electron
density may also arise from high mobility in the loop regions, which
would tend to randomize the electron density in the map. X-ray
diffraction techniques can, however, be used to locate a specific site
in the protein if sufficient atomic density is introduced into that
region. One way to achieve such a goal is the introduction of cysteine
residues in the regions of interest and their derivatization with a
heavy atom label. The label diffracts X-rays strongly and thus can
be easily located. Heavy atom labeling of many soluble proteins has
been used to determine the phases of their X-ray reflections by
multiple isomorphous replacement (226). The use of specific
attachment sites for heavy-atom labeling has also been effectively
used in structural
studies of several proteins for which three-
dimensional crystals can be obtained (227, 228, 229, 230).
This chapter describes the use of heavy atom labeling and Xray diffraction in the two-dimensional crystalline lattice of the
Our ultimate aim was to monitor changes in the
purple membrane.
precisely determined location of an introduced mercury label that
are associated with a specific step of the photocycle.
Such
131
information could help to elucidate the nature of any lightdependent movement(s) at the site to which the label has been
attached. Since the most significant movement observed in the timeresolved EPR studies occurred near residue 101, the double mutant
D96A/V101C
was chosen for use in the first studies; the presence of
the neutral substitution at site 96 causes a
more than 100-fold
decrease in the rate of M decay, making trapping of this intermediate
easier. This chapter presents the X-ray results obtained for both the
ground state and the M intermediate of this double mutant. The
work was performed in collaboration with Mr. Wolfgang Behrens in
Maarten Heyn's laboratory. Having constructed, expressed, and
purified the mutant, derivatization in the purple membrane with
methylmercury benzoate was performed, and the sample was sent to
Wolfgang for performance of the X-ray measurements on the ground
state structure. Wolfgang and I then performed the measurements
on the M intermediate
together in Maarten Heyn's
laboratory.
As stated previously, available diffraction data on unlabeled
wild type and mutant bR has shown that conformational changes
occur around the cytoplasmic ends of helices E and G during the
photocycle. Thus, in addition to examining motion around the CD
loop, we were also interested in time-dependent
X-ray analysis of
mutants labeled with mercury in the E-F interhelical loop, and the
cytoplasmic extension of helix G, e.g. mutants A160C and G231C,
respectively. Quantitative labeling of these mutants was attempted,
using two different sulfhydryl-specific mercury reagents under a
variety of conditions.
The results of these experiments will be
presented.
However, due to the inability to acheive a sufficiently
high incorporation of label, it has not been possible to obtain usable
X-ray diffraction data for these sites.
6.2
6.2.1
Experimental
Derivatization
Procedures
of the
Chloromercuribenzoate
Purified
Mutant Proteins
with p-
For use in derivatization reactions, a 5 to 8 mM solution of pchloromercuribenzoate (PCMB) in 0.4 M (NH4 )2SO4, 50 mM sodium
phosphate buffer, pH 7.0, was prepared following Boyer's method
(231).
The concentration of the PCMB stock solution was determined
spectrophotometrically
reactivity
using
e232 nm= 1 .69 x 104 Ml.cm'l.
of the reagent was subsequently
The
tested at pH 7.0 by
recording the spectral shifts that accompany its reaction with
132
cysteine.
For this purpose, the label was incubated with a 10-fold
molar excess of cysteine, using a freshly prepared and argon-purged
cysteine solution. After addition of cysteine to the label, the UV
absorption spectrum of the sample was obtained, and the extent of
mercaptide formation was analyzed by measuring the increase in
extinction at 250 nm.
The obtained value was compared to that
reported in the literature (231).
For analytical labeling reactions of the purified purple
membrane suspensions, sodium phosphate buffer (50 mM, pH 7.0)
which contained the appropriate amount of PCMB was used for
background correction on the spectrophotometer (Hitachi; U-3110).
A dark-adapted suspension of the purple membrane sheets in 50
mM sodium phosphate buffer, pH 7.0, was then added to the sample
cuvette.
An equivalent amount of the same buffer with no purple
membrane present was added to the reference cuvette. UV/Visible
absorbance spectra were subsequently collected at regular intervals
at 20°C until no further hyperchromatic shift (due to mercaptide
formation) was observed. The labeling reactions were performed
both in the presence of 1.5 fold and 4 fold molar excess of PCMB.
For
each condition, wild-type purple membrane was included as control.
In addition to collecting UV/Visible absorbance spectra, the reaction
time course was followed by monitoring the absorbance change at
250 nm using the time-drive mode of a Perkin-Elmer
spectrophotometer
(model
7) fitted with a scattered transmission
accessory. In order to obtain an accurate value for the change in
extinction at 250 nm due to mercaptide formation, each of the
mutants was reacted with substoichiometric amounts of PCMB.
Assuming complete reaction of the label in the presence of excess
substrate in these reactions, the value of EC250 nm was calculated in
each case. Difference spectra were obtained by subtraction of a
UV/Visible absorbance spectrum of dark-adapted purple membrane
sample from the last UV/Visible
spectrum in the time series
collected during PCMB reaction.
The selected spectra had matching
absorbance values at the chromophore peak. Using AA 2 5 0 nm and
AE 2 50 nm values, the extent of PCMB incorporation was measured in
each case; the concentration of the protein sample was measured
using the absorbance value of the light-adapted chromophore and
the extinction value measured for wild-type bR.
The purified mutant purple membranes were labeled on a
preparative
scale for use in X-ray measurements.
For this purpose,
133
the purple membrane (10 mg) was incubated in the dark at room
temperature with a 1.5 fold molar excess of PCMB in 50 mM sodium
phophate buffer, pH 7.0, overnight, or longer.
The membranes were
subsequently pelleted and washed by resuspension in 60 ml of 5 mM
NaCl and centrifugation at 24,000 rpm in a Beckmann 45Ti rotor at
4°C. This washing procedure was repeated twice to effect complete
removal of free label.
6.2.2
Derivatization
of A160C and G231C Mutants with
4 C-
Labeled Methylmercury iodide (' 4C-CH 3H g I)
Stock solutions of
4 C-CH
3HgI
were prepared by weighing out
the appropriate amount of reagent and dissolving it in ethanol
(100%). The specific activity of the label was measured by subjecting
a known volume of the stock solution to scintillation counting on an
LKB scintillation counter, using a fully open window for the
measurement of 14C emissions. The efficiency of the instrument for
4C was measured beforehand, in order to obtain the factor of
conversion from CPM to DPM. The obtained specific activity value
for the label matched closely the value given by the manufacturer
(Amersham).
For both mutants, analytical reactions were set up in
which the purple membrane suspensions were incubated in the dark
with a
4-fold molar excess of the label in 50 mM Tris-HC1 (pH 8.2)
under argon.
As usual, an identical tube containing wild-type purple
membrane was prepared as a negative control.
The excess label was
removed from the protein by column chromatography.
For this
purpose, Econo-Pack
O1DGcolumns (Biorad) pre-equilibrated with
the reaction buffer were used. Fractions (0.5 ml) were collected after
application of the incubated sample to the column, and equal aliquots
of each fraction were subjected to scintillation counting in order to
obtain the separation profile.
The fractions containing the protein
were pooled together and an aliquot of the pool was subjected to
scintillation counting. Using the measured specific activity of the
label, the concentration of the bound label in the pool was thus
calculated. The stoichiometry of labeling was then determined using
the measured concentration of the protein present in the pool. In
addition to the above conditions, the mutants were also incubated in
the presence of 50- andlOO-fold molar excess of the label, and the
levels of incorporation determined.
Again, wild-type bR was
included under the same conditions as a control.
In the latter
experiments, however, the removal of excess free label was achieved
by successive washing of the membranes and, in a separate set, by
using Hi-Trap Sephadex columns (Pharmacia) which were pre-
134
equilibrated with buffer. The use of the Sephadex matrix prevented
considerable binding of the purple membrane to the column resin, a
problem encountered with IODG columns containing polyacrylamidebased resin. The labeling reactions in the presence of a 50- or a 100fold molar excess of the label were performed with incubation times
of as long as one week to see whether very long incubation periods
affected the stiochiometry of labeling. For the A160C mutant,
labeling in the presence of a 100-fold molar excess of the label was
also tested in 50 mM Tris-HCl (pH 9.5).
The effect of incubating the mutant proteins with a reducing
agent prior to subjecting them to labeling reactions
was also tested,
with wild-type purple membrane treated under identical conditions
for control.
A 20 mM stock solution of the reducing agent
tributylphosphine (TBP) was prepared in ethanol under a constant
stream of argon. The mutant proteins were preincubated with a 50fold molar excess of TBP in 50 mM Tris-HCl (pH 8.2) in the dark and
under argon. For A160C, preincubation times of 5, 18, and 24 hours
were checked. For G231C, a preincubation time of 24 hours was
tested. Subsequent to TBP treatment, 4 C-CH 3HgI label was added to
each tube in a 2.2 fold molar excess over TBP for overnight
incubation.
The excess free label was removed by successive
washings of the membranes and the stiochiometry of labeling was
measured as described above.
6.2.3 X-ray
Diffraction
Measurements
Figure 6.1 shows a block diagram of the X-ray instrument.
The
bombardment of a rotating copper anode with accelerated electrons
causes the generation of X-ray beams by copper atoms (an ElliottMarconi GX21 X-ray generator was used). The X-ray beams are
passed through a narrow nozzle, in order to reduce scattering by air
prior to reaching the sample. The beams diffracted by the mounted
sample are collected in an air-free cone which is continually flushed
with a mild flow of helium to keep air out of the chamber and reduce
background scattering. The powder diffraction patterns are then
collected by a position-sensitive multichannel detector. To prevent
damage to the detector by undiffracted
placed at the back of the cone.
beams, a beam stopper is
135
CD
-
O
CD P:
3
>
S-
Q
CJD
CD 5nc
C:(D
CD
C_
x
.9
0
I<
-44
U)
0
0
-
O
CD,
Ct,
C3 D
CD
-
D
I
I
M
if
CP
.A.
pL
0CD
_
(D
0
1
9
)
6.
,<
-
CD
6U
C
CD oX
Oa
_.
iD
-*
D>
3'
ID
.
_
_
>l
-°
_
O
_
-)
0
&
*r
: <>
CD
-
3CD
D U)
Y
V)
(S
CD CD
C
_
d
_
VC
v
CD
C C01
Vre
r
_
ce
cn; Bv,
.
Co_
OCD
r
5.
_
:D~
0,
_
^-
-c
_
_ *
_
-X
-rCD
=
_
_2
.
C)
~
0
I
!
Co
- a
0 Mn
Ct,
CD
C~
~~-)r
~~
I
I
i~~~~~~~c
0
_
CL =
CD
136
Hydrated oriented membrane samples were prepared by
drying concentrated drops of the purple membrane suspension,
containing 2 mg of protein, onto a 5 tm thick mylar support at 86%
relative humidity. The samples were subsequently sealed in a
sample holder and were equilibrated for 24 hours under constant
humidity. In order to minimize differences in treatment between
the labeled (D96A/V101C-MB) and the unlabeled (D96A/V101C)
samples, all operations such as centrifugation, film preparation, and
drying were performed in parallel. The oriented membrane films
were mounted with the planes of the membranes perpendicular to
the X-ray beam in a chamber that was kept at 100% relative
humidity. X-ray data were collected for more than 40 hours, in 1
hour runs, in order to obtain diffraction patterns of the ground state
structures of D96A/V101C and its mercury-labeled derivative
D96A/V101C-MB. Two independent diffraction data sets were
obtained for each case.
By comparing the data from individual
1
hour runs, systematic errors and drift in the detection electronics
could be controlled.
For each of the four sets, 40 one hour spectra
were added for further analysis. The data collection for obtaining
the ground state diffraction patterns were performed at room
temperature.
For the measurements in the M state, 30 one hour spectra were
added together for further analysis.
In order to trap the protein in
the M state, the sample was cooled to -70°C immediately after
illumination and was kept at that temperature within the sample
chamber for the duration of data collection. The fluctuation in
temperature of the sample was +±2C during the time of data
collection.
6.2.4
Data
Analysis
Background subtractions were performed by the fitting of
polynomial functions to the data. The integrated intensities were
calculated by fitting the intensity profiles to Gaussian line shapes and
were corrected by a Lorenz factor of (h2 + hk + k2 )'/1 2; the (1,0)
reflection was not included in data analysis. The sum of the
intensities did not differ much between the labeled and the
unlabeled samples. The factor required to scale the two data sets
was accordingly close to one for the two measurements (1.38 for data
set 1 and 1.43 for data set 2). Fourier difference maps were
calculated from the intensity differences using the phases and the
137
intensity ratios for reflections with the same value of h2 + hk + k 2
from electron microscopy (37). The justification for this procedure
has been discussed in detail (232).
Experimentally,
this
approximation is supported by the excellent agreement of the
structural information obtained by X-ray or neutron diffraction with
that obtained by electron microscopy in those cases where a direct
comparison is possible.
These include determination of the in-plane
position of the cyclohexane ring of the chromophore (35, 233) and
measurement of the change in protein density in the bR to M
transition (125, 126). The label position determined in this way
from the Fourier difference map was used as the starting point for
the refinement procedure (234). The refinement of the label position
was performed for the ground state structures.
However, the
refinement of the label position in the M state remains to be
completed.
6.3
Results
6.3.1 Derivatization of the Purified Mutant Proteins with
PCMB and
Figure
14 C-CH
3 HgI
6.2 shows the hyperchromatic
shift in the absorbance
spectrum of PCMB following mercaptide formation.
The largest
increase occurs at about 250 nm. Under the buffer conditions used,
the average value of increase in the extinction coefficient at this
wavelength, obtained from four independent measurements, was
found to be 163 + 10%. This value matches the literature value of
164% (231). Figure 6.3 shows the results of the derivatization of
D96A/V1O1C with PCMB. The positive peak of the difference
spectrum in the UV region indicates the incorporation of the label
into the purple membrane.
Because no incorporation of PCMB into
the wild-type bR was observed under similar conditions, and
considering
that a single mercury peak is seen per D96A/V101C
monomer in the diffraction analysis (see below), the reaction of
PCMB was specific to cysteine in this case.
This result was expected
since mercury compounds form complexes preferentially with
protein sulfhydryl and imidazole groups (235). (Wild-type bR has no
cysteines or histidines.) In order to calculate the stoichiometry of
labeling, the change in extinction coefficient at 250 nm was
measured by performing the PCMB reaction using substoichiometric
amounts of the mercury label.
The results are shown in Figure 6.4.
The calculated change in the molar extinction coefficient at 250 nm
138
cD -
<
Absorbance
_
c,
-
&:
C
N
z
ON
CD
= CMco
s
CD
ED
CD
PD
Q)
Q1
O--
tll
C
c
IU
Lcm
=O
OD
C
:3
CD
C1
"3
_ .
-,0D
0
VO
3
CI
_~
_
c:
~-
I,
cm
.
(D
;
0-e
£ °
(3
Ct .cS N
:>
'co-.
_~O
CD
:
c,
2-
-oF
tctr
oo
3
"i tiM I CDt
II-
0t
Ct)
_
-
O
.
wr
C;
_
_
_r
C
-
_
03
_
_
_
C
C;
Ct N
i
_
-
k
-clc
w
LTn
co
,
o
I
" 0Si
_
_~
_
-
S
Nj
wa
")
s
139
a)
C
-0.
0
<n
Q0
'
.28
0
a>
~
V)
.2
B
l
\
.04-
<an
-
~~-1-^4...
-. 12i
I
2.0
I
350
i
450
55O
65C
750
Wavelength (nm)
Figure 6.3.
Derivatization
of D96A/V101C mutant bR with excess
PCMB. (A) Superimposition of the UV/Visible absorption spectra of
dark-adapted D96A/V101C bR before and after PCMB derivatization.
The spectrum shown for the PCMB-treated
protein is the final one
obtained following the collection of a series of spectra during the
time course of the reaction. Note that the PCMB-reacted sample
shows greater absorbance in the ultraviolet region than does the
unreacted species. (B) The difference spectrum obtained by
subtraction of the UV/Visible absorption spectrum of unlabeled
D96A/V101C bR from that of PCMB-reacted D96A/V101C which was
exhaustively
washed to remove unbound label.
This was the
material used in subsequent X-ray studies (see below).
140
Absorbance
zc zrV,D <> ,
- C
-
-· J:ct
(P
O
CN
Ct _
t
e
0
o,
C-
1=r· Q.-~
t
O
0
(D o
a
x C-1
;l
-
uDO
(:
'
_
_.
CD
_
N
,
5 ~
Dh
~0-·
30
Cr
0
--
Cr
_rO 3o
_.
3
Cir,. _,,
-
-
ts;
c
~
o
O~
=
Cr C
-
3 o
C-
r>3 CrV
_
_
"l
~
_
=
_~
= c
CY
,K,
_·
-
_
e
_
_s
_
_
_
__
_
_
_E
_'
w
e
P(
Cr
c, _
_o
_
"3
_·
_
_e
y
r
_
_
C
_I
o
"
_.
_
CD
D
(D
- -
A Absorbance
_
-
C
:Y
bI
0
141
upon mercaptide formation was found to be 9.75 x 103. This value is
about 22% higher than the value obtained for the formation of
mercuribenzoylcysteine (231; Figure 6.2). Considering the
experimental error, a similar difference was observed when we
previously performed the measurements for A103C-MB (the value of
the extinction coefficient at 236 nm was then found to be 18% higher
for the mercaptide form of A103C as compared to the formation of
the mercaptide with cysteine). The results of the X-ray
measurements for the ground state structure of A103C-MB has been
published (236) and will be discussed briefly later in this chapter.
Using the calculated A 2 5 0nm value for mercaptide formation in
D96A/V1O01C-MB,the level of incorporation of mercuribenzoate was
found to be nearly quantitative, corresponding to 0.93 mol of
mercuribenzoate per mol of D96A/V1O1C. This allowed the
derivatized protein to be used for X-ray diffraction.
For
determination of the precise location of the mercury label, it is
extremely advantageous to have a high stiochiometry of labeling.
Not only does a high stoichiometry provide for better signal to noise
in the intensities of the X-ray scattering by mercury, it also
facilitates data analysis, because the distribution of the label within
the unit cell of the lattice structure is then known.
Unlike the case with V101C, high stiochiometries of labeling
could not be obtained for the mutants A160C and G231C.
Attempts
to obtain a high degree of derivatization were unsuccessful with both
PCMB and 4 C-CH 3HgI. The labeling stiochiometry obtained after
removal of the excess free '4C-CH 3HgI by successive washing, or by
gel filtration was consistently about 33% for the A160C mutant. This
value is close to that obtained by pilot experiments for the reaction
of A160C with PCMB (40%).
The t, 2 of incorporation is about 15
minutes, as shown in Figure 6.5. In comparison, the t/2 of
incorporation for the D96A/V10lC mutant is about 2 minutes under
similar conditions. Incubation of the mutant A160C with the label
for much longer periods, up to a week, did not improve the labeling
efficiency.
The treatment of the mutant with the reducing agent TBP
prior to the start of the labeling reaction also did not improve the
labeling stoichiometry. Similar results were obtained for the G231C
mutant, where the best stiochiometry that could be achieved after
removal of the excess free label was only about 13%. One reason for
the poor labeling of these mutants may be the inaccessibility of these
sites to the mercury reagents.
Alternatively, studies of water-
142
0 ~
-°
Absorbance
0
CJ
-4 C
O
,,4 '
0
oL 3>O(PO CD
o
C
O
-
CD)
C1
:
-- I
CD
C',
0CD
,C
D
·
C
0.
C
C',
O <p
CD CD
D'-
or
Ab
nC
g°
CD
C0
0
C
C&
sz
0
0
° t.OD
O. .,
t ->
-t CC
_
:
CD
Absorbance
0
C,
c
.LI
oCD Co
-
·
:
-4~
r),~
'
o~D~tCDC
~r~
r~o
C ~C
D'
f
(.4
I
>
POC
(p 0(
C
3CD
O)
00(n
O_
O0
w
143
soluble proteins have indicated that highly exposed cysteines react
poorly with mercury compounds (227, 228), and this might be the
reason for the observed low labeling efficiency at sites 160 and 231.
Another plausible explanation is that the incorporation of the
mercury labels at these sites beyond a certain level may become
sterically unfavorable for the rigid arrangement of the trimeric uints
within the purple membrane lattice. Interestingly, precedent in the
literature exists for the reversible binding of p-mercuribenzoate to
cysteine residues. Spontaneous reactivation of the enzyme
homoserine
dehydrogenase
whose SH groups were blocked with p-
mercuribenzoate has been reported by Epstein and Datta (237).
These investigators demonstrated that transmercaptidation could not
explain the results, and that the reactivation
was simply due to the
dissociation of the mercury label from the cysteine residues required
for activity. The cause of this easy dissociation is not clear but it
may be due to oxidation of the derivatized cysteine residues. In our
case, such a possibility should also be given very serious
consideration as the pilot experiments for labeling G231C with PCMB
which do not include any washing steps gave a much higher labeling
stoichiometry of about 80%. Epstein and Datta's experiments also
suggested that the rate of the dissociation of the label is influenced
by a conformational change in the enzyme. This point bears on the
relationship between steric hindrance and labeling efficiency alluded
to above. As discussed previously, treatment of the mutant proteins
under stringent reducing conditions did not improve the labeling
efficiency. However, if intermolecular disulfide bonds form which
are inaccessible to reducing agents, the labeling efficiency will not
improve by the addition of reductants.
6.3.2
X-ray
Diffraction
of D96A/V101C
and D96A/VlO1C-
MB
The diffraction patterns of labeled and unlabeled D96A/V101C
samples in the ground state are shown in Figure 6.6. Figure 6.7
shows the diffraction patterns of these samples in the M state. As
shown, the diffraction peaks were observed out to the (7,1)
reflection, corresponding to a resolution of 7.2 angstroms for the
protein structure in the plane normal to the membrane.
The
diffraction patterns were indexed on an hexagonal P3 lattice with
unit cell dimensions that are the same, within experimental error, to
the value of 62.4 established for the wild-type purple membrane
(238).
D96A/VO11C and D96A/V1O1C-MB form hexagonal lattices
with the same unit cell dimensions, and the derivatization of site 101
144
\O
Intensity (Arbitrary Units)
0o
B-)
FO
00
00
m
O
-1
0
·r
CD
CD
CD
<D
o
-
__
-.
O'
a
Pz
CT
> Q3
:r
II
3r
S
,O,
-
P:
Y:r
0
0 :
R)
o.
{:
0 o' 0
F
-
0aa
_
0'
n CD
CO
.(
W
·
.
:3
oT,
e 0
-
I
cn
00
'
Q
N-XO
C
C
C
0
4`
o
r
n
U,
,
0e
:,
5
0
o
:C
.r
.
:......
,
...
._ _~~~~.
~o
W
_~~~h
C
C
m
C
: ~~~~0
0
©a m
.
C
g
2
·_
f_
0
I
>
oj
Xd
&D~f
C"
I
.,
C
C
i
-
3
.... .
.
3
C
I-
o~~~C
003 w C0
-
00
0(D
C)
o~~~C
-- 4
cnCaC
;.:
/
.
I:---
'
·- _-L_
·
I
L
---
__
cn
C)
0
0
0
0
0
145
Intensity (Arbitrary Units)
a 5 ", Cr
( -5
2x
-
00
0
CD P, C) c
D
CD
_.
0 f
~
0·
C
O
,,
ri" (p
O a* x
R.
o
cn
-o
rnC
CD k
Po
O
r
(D
QD Ut
°
3
r
;3
CD c'
CD
~=5O0t7 - C>
"
cn
_ <:
mCD r
=C
CDo0
1
x , IC° 'n
Os
t~
3 a3
IC)
(D a
>'·
3
:5
P
\9
_t
O
r-'
00
00
0
F.
O
C0
0
0
0
00
0CD
0rs
146
with mercuribenzoate does not appear to perturb the lattice
structure of the purple membrane. Positive and negative intensity
differences
due to the mercuribenzoate
group can be observed. The
data analysis was facilitated by the fact that virtually identical
background scattering was obtained from the unlabeled and
derivatized
samples (Figures 6.6 and 6.7).
Using the data shown in Figures 6.6 and 6.7, the electron
density of the label was calculated as described earlier for the
ground state and the M state. The resulting electron density map for
the ground state after refinement of the label position is shown in
Figure 6.8A. The three difference maxima correspond to the three
label positions.
Fractional coordinates are defined with the origin in
the center of the unit cell and with x and y increasing along the
oblique and horizontal axes from top to bottom and from left to right,
For the two independent sets of experiments
respectively.
performed for the ground state structure, the fractional coordinates
(x, y) corresponding to the refined
The
0.1730) and (0.3343, 0.1779).
between these two positions is less
angstrom for the Fourier difference
label position were (0.3455,
difference of 0.9 angstrom
than the internal error of 1
method and indicates good
experimental reproducibility.
The refined label location was independent
of the position
chosen for starting refinement within an 8 angstrom radius of the
final position. Several of the highest contour lines of the label
on the density map of
density are drawn superimposed
D96A/V101C in Figure 6.8B. The projected position of the mercury
label attached to site 101 is close to the C helix between
-helices C and D. This is in agreement with the
transmembrane
results of EPR collision studies with oxygen and CROX, which indicate
that VallOl is positioned at the interface between the membrane
layer and the aqueous phase and marks the beginning of the C-D
interhelical loop (47). A comparison of the label position at site 101
with the position of the same label attached to site 103 can be made
by comparing Figures 6.8 and 6.9. Both of the labels are positioned
within the interhelical C-D loop. However, as expected from previous
models, the label position for site 103 is more toward the middle of
the gap between C and D helix electron densities. An advantage of
introducing a mercury label in order to locate a cysteine residue is
that the mercury atom is only three bonds away from the main chain
Caatom. Therefore, the C atom of the derivatized cysteine lies no
147
-c
,
..
-
_c:
Ci
I
-Cl)
S)
C)
CD
0E
_
C,) C)
i
.,
!
XX
",\
\
-
e
2.
-0 0
(D
-
crQ
&,-
--
m6
04
C
_6 .3, O
Cl)C t
-'
_ O
-
O
7'
CO
<D
(;
K
-\
_-i
C
c
Cl)
0
)
CZ.
-:
S
_
,
_
C)
C
E
_
';.
OG LC ;s
_) C)
~I3·
_
-
~
::.
CS
o.
~_~
Cc.,
~ .
Cc_
_
C,
"
O
., k
Q
C)
z
T
C3
c3~~~C
n
o',
,
Ct
O.. _,
_;
....
r
_ _
-
C) -:,C #3l) _Ni
,
-.
C
m*
·
- iS
zJ
-8t
l)
C
p
-d
3
Or
",
,' ,S',I ///
r
,- -
<
__
1:~
cc
N
's
/,/ \II,
~~~~~~~~~~~~~~~~h~~~~~~~~~~~~~~
O-
~.
-.
'"I>
-,~
I
*
0
_
C
c
_
3.
,fi
>
cr
_
3
1
·____
148
Figure 6.9.
Location of the mercuribenzoate
label in A103C-MB.
(A) Two dimensional difference density map showing the in-plane
position of the mercuribenzoate label bound at Cys 103. Negative
contour lines were omitted. The six solid contour lines range from 50
to 99% of the positive difference density. (B) Superimposition of the
label position from (A) (the four highest contour lines representing
70, 80, 90 and 99% of the positive density; bold lines) on the
projected structure of A103C determined from the X-ray intensities
of A103C and A103C-MB.
labeled A-G.
The seven transmembrane a-helices
are
149
further than 4.5
1 angstroms from the peak maximum for the label
in the Fourier difference map.
A superimposition of the mercury label positions in the ground
and M states of D96A/V101C onto the projected protein structure is
shown in Figure 6.10. From this, it is clear that a movement of the
label has occurred by the time of M intermediate formation. What
can be seen from the projection map is that the density of the
mercury label has moved from its ground state position at the top of
helix C, above D96, to a new position, closer to the helix G density.
The M state diffraction data has yet to be refined. Thus, only a rough
estimate of the distance traversed by the label parallel to the plane
of the membrane can be made at this time. However, as a very
crude estimate, it appears that, in the direction parallel to the
membrane, the mercury label has moved about 5 to 7 angstroms by
the time that the 412 nm M-like species has been formed. Any
movement of the label in the third dimension cannot be observed in
our two-dimensional projection maps without tilting experiments.
6.4
Discussion
The initial X-ray diffraction results on the ground and M states
of D96A/V101C-MB both confirm and extend the EPR data discussed
in the previous chapter. Although we cannot be certain that the
observed conformational changes correspond to the same M species
in both cases, it is nevertheless significant that the diffraction results
show movement of the C-D interhelical
loop at the cytoplasmic
surface, consistent with our previous prediction based on timeresolved EPR studies of the spin-labeled VOlC
and D96A/Vl01C.
A
possible functional significance of such a structural transition in
regard to the protonation changes at residue D96 was suggested in
Chapter 4. To elaborate, the diffraction data clearly shows a
movement of the mercury label away from the C helix and toward
the G helix in making the transition from the ground state to the Mlike 412 nm species, be it M or MN. A simple hypothesis that is
consistent with this result is to propose the opening of a channel into
the interior of the protein. In this context, the opening of the
cytoplasmic region of bR would facilitate the reprotonation of D96 in
the M to N transition. The previously noted change in local
electrostatic potential around V101 (223), as well as differences in
the mobility of its spin-label between the ground state species and
the light-activated form (Chapter 5), complement the diffraction data
in suggesting such a mechanism.
Evidence for, or against this theory,
150
Figure 6.10.
A conformational change in bR upon M formation is
detected by X-ray diffraction of the mercuribenzoate-labeled
D96A/V101C mutant protein.
Shown
electron densitv of the mercury label
on the projected protein structure. A
between the round and M states is
6.7B).
is the superimposition of the
in the ground and the M state
shift in the position of the label
apparent (compare with Figure
151
may be obtained in future studies by making more extensive use of
the approaches described herein with mutants that can be suitably
modified.
The strategy described in this chapter greatly complements the
available techniques for obtaining structural information on bR. In
particular, this strategy allows mapping of the putative interhelical
loop regions of bR, information which cannot be obtained from
projection maps of the unlabeled protein structure alone. The
described approach is also attractive for refining the structure of
other integral membrane proteins for which three-dimensional
crystals are not available. The results presented for analysis of the
D96A/V1O1C-MB M state further indicate the promise of this
technique for monitoring localized conformational changes occurring
during the bR photocycle.
It is not possible to obtain such
information with other currently used approaches to analyze twodimensional
diffraction
patterns of bR.
In combination
with time-
resolved EPR studies which provide real-time resolution, the X-ray
diffraction strategy should prove particularly powerful.
152
Reference:
1. Danon, A. and Stoeckenius, W. (1974) Proc. Natl. Acad. Sci. USA
71, 1234-1238.
3.
(1990) Biochim. Biophys. Acta 1016, 293-327.
Stoeckenius, W., and Bogomolni, R. A. (1982) Annu. Rev. Biochem.
4.
Stoeckenius, W. Lozier, R., and Bogomolni, R. (1979) Biochim.
5.
6.
Khorana, H. G. (1988) J. Biol. Chem. 263, 7439-7442.
Oesterhelt, D., and Stoeckenius, W. (1971) Nature New Biol. 233,
Birge, R. R.
2.
51, 587-616.
Biophys. Acta 565, 215-278.
149-152.
Henderson,
7.
R. (1977) Annu.
Rexv. Biophys. Bioeng. 6, 87-109.
8. Ottolenghi, M. (1980) Advances in photochemistry 12, 97-200.
Wiley: New York, NY.
9. Mathies, R. A., Lin, S. W., Ames, J. B., and Pollard, T. (1991) Annu.
Rev. Biophys. Biophys. Chem. 20, 491-518.
10. Lanyi, J. K. (1993) Biochim. Biophys. Acta 1183, 241-261.
11.
Rothschild, K. J. (1992) J. Bioenerg. Biomembr. 24, No. 1, 147-
167.
12. Lanyi, J. K. (1992) J. Bioenerg. Biomembr. 24, No. 2, 169-79.
13. Smith, S. O., Lugtenburg, J., and Mathies, R. A. (1985) J. Membr.
Biol. 85, 95-109.
14. Packer, L., editor. (1982) Methods in Enzvmology,
Biomembranes, 1 88. Academic Press; New York, NY.
15. Zheng, L., and Herzfeld, J. (1992) J. Bioenerg. Biomembr. 24, 139-
146.
149-152.
16.
Oesterhelt, D., and Stoeckenius, W. (1971) Nature 233,
17.
Blaurock,
18.
Oesterhelt, D. and Stoeckenius, W. (1974) Methods
19.
Lechner, J., and Sumper, M. (1987) J. Biol. Chem. 262, 9724-
20.
Mescher, M. F., and Strominger, J. L. (1976) Proc. Natl. Acad. Sci.
21.
Sumper, M., Reitmeier, H., and Oesterhelt, D. (1976) Angew.
22.
Hwang, S.-B., Tseng, Y.-W., and Stoeckenius, W. (1981)
152-155.
A. E., and Stoeckenius,
W. (1971) Nature New Biol. 233,
667-678.
Enzymol. 31,
9729.
USA 78, 2687-2691.
Chem. Int. Ed. Engl. 15, 187-194.
Photochem. Photobio.: 33, 419-427.
153
23.
Hiraki, K., Hamanaka, T., Mitsui, T., and Kito. Y. (1981)
24.
Usukura, J., Yamada. E.. and Mukohata, Y. (1981) Photochem.
25.
Kushwaha, S. C., Kates, M., and Martin, S. G. (1974) Can. J.
26.
Neugebauer, D.-CH., Zingsheim, H. P., and Oesterhelt, D. (1978) J.
27.
Neugebauer, D.-CH., and Zingsheim, H. P. (1978) J. Mol. Biol. 123,
Photochem. Photobiol. 33, 429-433.
Phobiol. 33, 475-481.
Biochem. 53, 284-292.
Mol. Biol. 123, 247-257.
235-246.
28. Dencher, N. A., and Heyn, M. P. (1982) Methods Enzymol. 88, 510.
29. Huang, K.-S., Bayley, H., and Khorana, H. G. (1980) Proc. Natl.
Acad. Sci. USA 77, 323-327.
30. Dencher, N. A., and Heyn, M. P. (1979) FEBS Lett. 108, 307-310.
31.
Khorana, H. G., Gerber, G. E., Herlihy, W. C., Gray, C. P., Anderegg,
R. J., Nihei, K., and Biemann, K. (1979) Proc. Natl. Acad. Sci. USA 76,
5046-5050.
32. Dunn, R., McCoy, J., Simsek. M., Majumdar. A., Chang, S. H.,
RaiBhandary, U. L., and Khorana, H. G. (1981) Proc. Natl. Acad. Sci.
USA 78, 6744-6748.
33.
Seehra, J. S., and Khorana, H. G. (1984) J. Biol. Chem. 259, 4187-
34.
Rees, D. C., DeAntonio, L., and Eisenberg, D. (1989) Science 245,
4193.
510-513.
35. Henderson? R., Baldwin, J. M., Ceska, T. A., Zemlin F., Beckmann,
E., and Downing, K. H. (1990) J. Mol. Biol. 213, 899-§29.
36. Baldwin, J., Henderson, R., Beckmann, E., and Zemlin, F. (1988) J.
Mol. Biol. 202, 585-591.
37. Henderson, R., Baldwin, J. M., Downing, K. H., Lepault, J., and
Zemlin, F. (1986) Ultramicroscopy 19, 147-178.
38. Downing, K. H., and Glaeser, R. M. (1986) Ultramicroscopy 20,
269-278.
39.
40.
Mao, D., and Wallace, B. A. (1984) Biochem. 23, 2667-2673.
Nabedryk, N., Bardin, A. M.. and Breton, J. (1985) Biophys. J. 48,
41.
42.
43.
44.
Wallace, B.
Henderson,
Rothschild,
Lee, D. C.,
45.
Altenbach, C., Marti, T., Khorana, H. G., and Hubbell, W. L. (1990)
873 -876.
5783.
Science 248,
A., and Teeters, C. L. (1987) Biochem. 26, 65-70.
R., and Unwin, P. N. T. (1975) Nature 257, 28-32.
K. J., and Clark. N. A. (1979) Science 204, 311-312.
Herzyk, E., and Chapman, D. (1987) Biochem. 26, 5775-
1088-1092.
46. Altenbach, C., Greenhalgh, D. A., Khorana, H. G., and Hubbell, W. L.
(1994) Proc. Natl. Acad. •ci. USA 91, 1667-1671.
154
47.
Greenhalgh,
48.
Engelman,
D. A., Altenbach, C., Hubbell. W. L., and Khorana, H. G.
(1991) Proc. Natl. Acad. Sci. USA 88, 8626-8630.
D. M., Henderson,
R., McLachlan.
A. D., and Wallace, B.
A. (19780) Proc. Natl. Acad. Sci. USA 77, 2023-2027.
49. Popot, J., Engelman, D. M., Gurel, O., and Zaccai, G. (1989) J. Mol.
Biol. 210, 829-847.
50.
Fimmel, S., Choli, T., Dencher, N. A., Buldt. G., and Wittmann-
Liebold, B. (1989) Biochim. Biophys. Acta 978, 231-240.
Ovchinikov, Y., et al. (1985) FEBS Lett. 179, 343-350.
Engelman, D. M., Steitz T. A., and Goldman, A. (1986) Annu. Rev.
Biophys. Biophys. Chem. 15, 321-353.
5 3. Bayley, H., Huan, K.-S., Radhakrishnan, R., Ross, A. H., Takagaki,
Y., and Khorana, H. G. (1981) Proc. Natl. Acad. Sci. USA 78, 222-
51.
52.
2229.
54.
Uegaki, K., Suoiyama, Y., and Mukahata, Y. (1991) Arch. Biochem.
Biophys. 286, 10'7-110.
55. Heyn, M. P., Westerhausen, J., Wallat, I., and Seiff, F. (1988) Proc.
Natl. Acad. Sci. USA 85, 2146-2150.
56. Huang, K.-S., Radhakrishnan, R., Bayley, H., and Khorana, H. G.
(1982) J. Biol. Chem. 257, 13616-13623.
57.
Huang, J. Y., and Lewis, A. (1989) Biophys. J. 55, 835-842.
58.
Leder, R. O., Helgerson, S. L., and Thomas, D. D. (1989) J. Mol. Biol.
59.
Hauss, T. Grzesiek, S., Otto, H., Westerhausen, J., and Heyn, M. P.
60.
61.
Lin, S. W., and Mathies, R. A. (1989) Biophys. J. 56, 653-660.
Earnest, T. N., Roepe, P., Braiman, M. S., Gillespie, J., and
62.
Fahmy, K., Siebert, F., Grossjean, M. F., and Tavan, P. (1989) J.
63.
Heyn, M. P., Cherry, R. J., and Muller, U. (1977) J. Mol. Biol. 117,
64.
Bogomolni,
209, 683-701.
(1990) Biochemistry
29, 4904-4913.
Rothschild, K. J. (1986) Biochemistry 25, 7793-7798.
Mol. Struct. 214, 257-288.
607 -620.
R. A., Hwang, S. B., Tseng, Y. W., King, G. I., and
Stoeckenius, W. (1977) Bophys. J. 17, 89a.
65. Nakanishi, K., Balogh-Nair, V., Arnaboldi, M., Tsujimoto, K., and
Honig, B. (1980) J. Am. Chem. Soc. 102, 7945-7947.
66. Blatz, P. E., and Mohler, J. H. (1975) Biochemistry 14, 2304-2309.
67. Spudich, J. L., McCain, D. A., Nakanishi, K., Okabe, M., Shimizu, N.,
Rodman, H., Honig, B., and Bogomolni, R. A. (1986) Biophys J. 49,
479-483.
68.
Lugtenburg, J., Muradin-Szweykowska,
Pardoen, J. A.,Harbison
M., Heeremans, C.,
G. S., Herzfeld, J., Griffin, R. G., Smith, S. O.,
and Mathies, R. A. (186) .J. Am. Chem. Soc. 108, 3104-3105.
69. Ottolenghi, M., and Sheves, M. (1989) J. Membr. Biol. 112, 193-
212.
155
70. de Groot, H. J. M., Smith, S. O., Courtin. J.. van der Berg, E., Winkel,
C., Lugtenburg, J., Griffin, R. G., and Herzfeld, J. (1990) Biochemistry
29,
6873-6883.
71.
de Groot, H. J. M., Harbison, G. S., Herzfeld, J., and Griffin, R. G.
72.
Birge, R. R., and Zhang, C. (1990) J. Chem. Phys. 92, 7178-7195.
73.
Lussier, L. S., Dion, A., Sandorfy, C.. Le-Thanh,
(1989) Biochemistry
(1986) Photochem.
28, 3346-3353.
H., and Vocelle, D.
Photobiol. 44, 629-639.
74. Harbison, G. S., Smith, S. O., Pardoen, J. A., Courtin, J. M. L.,
Lugtenburg, J., Herzfeld, J., Mathies, R. A., and Griffin, R. G. (1985)
Biochemistry
24, 6955-6962.
75.
Scherrer, P., Mathew, M. K., Sperling, W., and Stoeckenius, W.
76.
Braiman, M. S., and Mathies, R. (1982) Proc. Natl. Acad. Sci. USA
77.
Kouyama , T., Nasuda-Kouyama, A., Ikegami, A., Mathew, M. K.,
78.
Lozier, R. H., Bogomolni, R. A., and Stoeckenius,
79.
Dencher, N., and Wilms, M. (1975) Biophys. Struct. Mech. 1, 259-
80.
Kung, M. C., Devault, D., Hess, B., and Oesterhelt,
81.
Slikfin, M. A., and Caplan, S. R. (1975) Nature 253, 56-58.
(1989) Biochemistry
28, 829-834.
79, 403-407.
and Stoeckenius, W. (1988) Biochemistry 27, 5855-5863.
Biophys. J. 15, 955-T)62.
W. (1975)
271.
Biophys. J. 15, 907-911.
D. (1975)
82. Fodor, S. P. A., Ames, J. B., Gebhard. R., van der Berg, E. M. M.,
J., and Mathies, R. A. (1988)
Stoeckenius, W., Luftenberg
Biochemistry
27,
097-7f61.
83. Nuss, M. C., Zinth, W., Kaiser, W., Kolling, E., Oesterhelt, D. (1985)
Chem. Phys. Lett. 117, 1-8.
84.
Sharkov, A., Pakulev, A., Chekalin, S., Matveetz, Y. (1985)
Biochim. Biophys. Acta 808, 94-102.
85. Ort, D. R., and Parson, W. W. (1979) Biophys. J. 25, 341-353.
86. Varo, G., and Lanyi, J. K. (1991) Biophys. J. 59, 313-322.
87. Hanamoto, J., Dupuis, P., and El-Sayed, M. (1984) Proc. Natl. Acad.
Sci. USA 81, 7083-7087.
88. Dancshazy, Z., Govindjee. R., and Ebrey, T. G. (1988) Proc. Natl.
Acad. Sci. USA 85, 6358-6361.
89. Diller, R., and Stockburger, M. (1988) Biochemistry 27, 7641-
7651.
90.
Balashov, S. P., Govindjee, R., and Ebrey, T. G. (1991) Biophys. J.
91.
Sherman, W. V., Slifkin, M. A., and Caplan, S. R. (1976) Biochim.
92.
Lozier, R. H., Niederberger, W., Ottolenghi. M., Sivorinovsky, G.,
60, 475-490.
Biophys. Acta 423, 238-248.
and Stoeckenius, W. (1978) Energetics and Structure of Halophilic
156
Microorganisms
pp. 123-139 (Caplan, S. R., and Ginzburg, M. eds.).
Amsterdam: Elsevier/North
93.
Ames, J. B., and Mathies,
7190.
Holland.
R. A. (1990) Biochemistry
29, 7181-
94. Xie, A. H., Nagle, J. F., and Lozier, R. H. (1987) Biophys. J. 51, 627-
635.
95. Otto, H. Marti. T. Holz, M., Moi, T., Lindau, M., Khorana, H. G.,
and Heyn, M. P. (1989) Proc. NatE Acad. Sci. USA 86, 9228-9232.
96. Hess, B., Kuschmitz, D. (1977) FEBS Lett. 74, 20-24.
97. Hurley, J. B., Becher, B., and Ebrey, T. G. (1978) Nature 272, 87-
88.
98.
Kalisky, O., Lachish, U., Ottolenghi, M. (1978) Photochem.
Photobiol. 28, 261-263.
99. Gerwert, K. Souvignier, G., and Hess, B. (1990) Proc. Natl. Acad.
Sci. USA 87, 9774-9778.
100. Henry, E. R., and Hofrichter, J. (1992) Methods Enzymol. 210,
129-192.
101.
Heling,
1 929-194.
B., Souvignier, G., and Gerwert, K. (1993) Biophys. J. 65,
102.
Ohno, K., Takeuchi, Y., and Yoshida, M. (1981) Photochem.
103.
Sasaki, J., Shichida, Y., Lanyi, J. K., and Maeda, A. (1992) J. Biol.
104.
Friedman, N., Gat, Y., Sheves, M., and Ottolenghi, M. (1994)
Photobiol. 33, 573-578.
Chem. 267,
20782-20786.
Biochemistry 33, 14758-14767.
105. Varo, G., and Lanyi, J. K. (1990) Biochemistry 19, 2241-2250.
106. Varo, G., and Lanyi, J. K. (1991a) Biochemistry 30, 5008-5015.
107. Varo, G., and Lanyi, J. K. (1991b) Biochemistry 30, 5016-5022.
108. Varo, G., Zimanyi, L., Chang, M., Ni, B., Needleman, R., and Lanyi,
J. K. (1992) Biophys. J. 61, 82 -826.
109. Druckmann, S., Friedman, N., Lanyi J. K., Needleman, R.,
Ottolenghi, M., and Sheves, M. (1992) Photochem. Photobiol. 56,
1041-1047.
110. Lewis, A., Spoonhower, J., Bogomolni, R., Lozier, R., Stoeckenius,
W. (1974) Proc. Natl. Acad. Sci. USA 71, 4462-4466.
111. Aton, B., Doukas, A. G., Callender, R. H., Becher, B., and Ebrey, T.
G. (1977) Biochemistry 16, 2995-2999.
112. Pettei, M. J., Yudd. A. P., Nakanishi, K., Henselman, R., and
Stoeckenius, W. (1977) Biochemistry 16, 1955-1959.
113.
Braiman, M., and Mathies, R. (1980) Biochemistry
114.
Smith, S. O., Pardoen, J. A., Mulder, P. P. J., Curry B Lugtenburg,
22, 6141-6 14.
5428.
J., and Mathies, R. A. (1983) Biochemistry
19, 5421-
157
115.
Rothschild
K. J., Roepe, P. Ahl, P. A
Earnest, T., Bogomolni,
R.
A., DasGupta, S. K., Mulliken, C. M., and herzfeld, J. 1986) Proc. Natl.
Acad. Sci. USA 83, 347-351.
116. Fodor, S. P. A., Pollard, W. T., Gebhard, R., van den Berg, E. M. M.,
Lugtenburg, J., and Mathies, R. A. (1988) Proc. Natl. Acad. ci. USA
85,
2156-2160.
Photobiol.
117.
Alshuth, T., and Stockburger, M. (1986) Photochem.
1 1 8.
Smith, S. O., Myers, A. B., Pardoen, J. A., Winkel, C., Mulder, P. P.
43, 55-66.
G. Lugtenburg, J., and Mathies, R. A. (1984) Proc. Natl. Acad. Sci.
USA 81, 205-259.
R. H., Ebrey, T. G. (1985)
119.
Deng, H., Pande, C., Callender
120.
Schertler, G. F. X., Lozier, R., Michel, H., and Oesterhelt, D. (1991)
Photochem. Photobiol. 40, 467-470.
EMBO J. 10, 2353-2361.
121. Holz, M., Drachev, L. A., Mogi, T., Otto, H., Kaulen, A. D., Heyn, M.
P., Skulachev, V. P., and Khorana, H. G. (1989) Proc. Natl. Acad. Sci.
USA 86, 2167-2171.
122. Marinetti, T., Subramaniam S., Mogi, T. Marti T., and Khorana,
H. G. (1989) Proc. Natl. Acad. ci. USA 86, 529-53 .
123. Tittor, J. Soell, C., Oesterhelt, D., Butt, H.-J., and Bamberg, E.
(1989) EMBO J. 8, 3477-3482.
124. Subramaniam S., Greenhalgh, D. A., Rath, P., Rothschild, K. J
and Khorana, H. I. (1991) Proc. Natl. Acad. Sci. USA 88, 6 8 7 3 -68 7 7 .
125. Subramaniam, S Gerstein, M., Oesterhelt, D., and Henderson, R.
(1993) EMBO J. 12, 1-8.
126. Koch, M. H. J., Dencher, N. A., Oesterhelt, D., Plohn, H.-J., Rapp, G.,
and Buldt, G. (1991) EMBO J. 10, 521-526.
127. Nakasako, M., Kataoka, M., Amemiya, Y., and Tokunaga, F.
(1991) FEBS Lett. 292, 73-75.
128. Dencher N. A., Dresselhaus, D., Zaccai, G., and Buldt, G. (1989)
Proc. Natl. Acad. Sci. USA 86, 7876-7879.
129. Ort, D. R., and Parson, W. W. (1979) Biophys. J. 25, 355-364.
130. Braiman, M. S., Ahl, P. L., and Rothschild, K. J. (1987) Proc. Natl.
Acad. Sci. USA 84, 5221-5225.
13 1. Rothschild, K. J., He, Y., Mogi, T., Marti, T., Stern, L. J., and
Khorana, H. G. (1990) Biochemistry 29, 5954-5960.
132. Souvignier, G., and Gerwert, K. (1992) Biophys. J. 63, 1393-
1405.
133. Steinhoff, H.-J., Mollaaghababa, R., Altenbach, C. Hideg, K.,
Krebs, M. P., Khorana, H. ., and Hubbell, W. L. (1994) Science 266,
105-107.
134. Alexiev, U. Mollaaghababa, R., Scherrer P Khorana H. G., and
Heyn, M. P. (1995) Proc. Natl. Acad. Sci. USA 92, 372-3'6.
135. Khorana, H. G. (1993) Proc. Natl. Acad. Sci. USA 90, 1166-1171.
158
136.
Krebs, M. P., and Khorana, H. G. (1993) J. Bacteriol. 175, 1555-
137.
Huang, K.-S., Bayley, H., Liao M.-J., London, E., and Khorana, H.
138.
London, E., and Khorana, H. G. (1982) J. Biol. Chem. 257, 7003-
139.
Liao, M.-J., London, E., and Khorana, H. G. (1983) J. Biol. Chem.
140.
Dunn, R. J., Hackett, N. R., McCoy, J. M., Chao, B. H. Kimura, K.,
141.
Karnik, S. S., Nassal, M., Doi, T., Jay E. Sgaramella, V., and
142.
Braiman, M. S., Stem, L. J., Chao, B. H., and Khorana, H. G. (1987)
143.
Nassal, M. Mogi, T., Karnik, S. S., and Khorana, H. G. (1987) J.
1560.
G. (1981) J. Biol. Chem. 256, 3802-3809.
7011.
258, 9949-9955.
and Khorana, H. G. (1987) J. Biol. Chem. 262, 9246-9254.
Khorana, H. G. (1987) J. Biol. Chem. 262, 255-9263.
J. Biol. Chem. 262, 9271-9276.
Biol. Chem. 22,
9264-9270.
144. Mogi, T., Stern, L. J., Marti, T., Chao, B. H., and Khorana, H. G.
(1988) Proc. Natl. Acad. Sci. USA 85, 4148-4152.
145. Otto, H., Marti, T., Holz, M., Mogi T Stern, L. J. Engel F.,
Khorana, H. G., and Heyn, M. P. (199)
1018-1022.
146.
P5roc. Natl. Acacd Sci. USA 87,
Braiman, M. S., Mogi, T., Marti, T., Stern, L. J., Khorana, H. G., and
Rothschild, K. J. (1988) Biochemistry 27, 8516-8520.
147. Braiman M. S., Bousche 0., and Rothschild, K. J. (1991) Proc.
Natl. Acad. ci. USA 88, 2388-2392.
148.
Marti, T., Otto, H., Rosselet, S. J., Heyn, M. P., and Khorana, H. G.
(1992) J. Biol. Chem. 267, 16922-16927.
149. Subramaniam, S., Marti T., and Khorana, H. G. (1990) Proc. Natl.
Acad. Sci. USA 87, 1013-1617.
150. Greenhalgh, D. A., Subramaniam, S., Alexiev, U Otto, H., Heyn,
M. P., and Khorana, H. G. (1992) J. Biol. Chem. 267, 25734-25738.
151. Doukas, A. G., Pande, A., Suzuki, T., Callender, R. H., Honig, B.,
and Ottolenghi, M. (1981) Biophys. J. 33, 275-279.
152.
Maeda, A. Sasaki, J., Shichida, Y., and Yoshizawa, T. (1992)
153.
154.
Grzesiek, S., and Dencher, N. A. (1986) FEBS Lett. 208, 337-342.
Scherrer, P., Alexiev U., Otto, H., Heyn, M. P., Marti, T., and
Biochemistry
31, 462-467.
Khorana, H. G. (1992) Structures and Functions of Retinal Proteins
(Rigaud, J. L., ed.), pp. 205-211, John Libbey Eurotext: Montrouge.
155. Scherrer, P., Alexiev, U., Marti, T., Khorana, H. G., and Heyn, M.
P. (1994) Biochemistry
33, 13684-13692.
156. Heberle, J., Riesle, J., Thiedemann, G., Oesterhelt, D., and
Dencher, N. A. (1994) Nature 370, 379-382.
157. Zimanyi, L., Varo, G., Chang, M., Ni, B., Needleman, R., and Lanyi,
J. K. (1992) Biochemistry 31, 8535-8543.
159
158.
Stern, L. J., and Khorana, H. G. (1989) J. Biol. Chem. 264, 14202-
159.
Needleman,
14208.
R., Chang, M., Ni. B., Varo. G., Fornes, J., White, S. H.,
and Lanyi, J. K. (1991) J. Biol. Chem. 266, 11478-11484.
160. Krebs, M. P., Mollaaghababa, R.. and Khorana, H. G. (1993) Proc.
Natl. Acad. Sci. USA 90, 1987-1991.
161. Gerwert K., Hess B., Soppa, J., and Oesterhelt, D. (1989) Proc.
NatI. Acad. Sci. USA 86, 4943-4947.
162.
Pfefferle, J.-M., Maeda, A., Sasaki, J., and Yoshizawa, T. (1991)
Biochemistry
30, 6548-6556.
163. Maeda, A., Sasaki, J., Shichida, Y.. Yoshizawa, T., Chang, M., Ni B.,
Needleman, R., and Lanyi, J. K. (1992) Biochemistry 31, 4684-4690.
164. Bousche O., Braiman, M. He, Y.-W., Marti, T., Khorana, H. G., and
Rothschild,
K. J. (1991) J. Biol. Chem. 266, 11063-11067.
Muller, K. H., Butt, H. J., Bamberg, E., Fendler, K., Hess, B., Siebert,
F., and Engelhard, M. (1991) Eur. Biophys. J. 19, 241-251.
166. Papadopoulos, G., Dencher, N. A., Zaccai, G., and Buldt, G. (1990)
165.
J. Mol. Bio. 214, 15-19.
167.
Cao, Y., Varo, G., Chang, M., Ni, B., Needleman, R., and Lanyi, J. K.
168.
Rothschild, K. J., He, Y.-W., Sonar, S., Marti, T., and Khorana, H. G.
169.
Resek, J. F., Farahbakhsh,
170.
Farahbakhsh, Z. T., Hideg, K., and Hubbell, W. L. (1993) Science
(1991) Biochemistry
30, 10972-10979.
(1992) J. Biol. Chem. 267, 1615-1622.
(1993) Biochemistry
262,
Z. T., Hubbell, W. L., Khorana, H. G.
32, 12025.
1416.
Simon, R. D. (1978) Nature 273. 314-317.
172. Weidinger, G., Klotz, G., and Goebel, W. (1979) Plasmid 2, 377-
171.
386
17 3. Pfeifer. F., Weidinger, G., and Goebel, W. (1981) J. Bacteriol.
145,
174.
375-381.
Pfeifer, F., and Ghahraman, P. (1993) Mol. Gen. Genet. 238,
193-200.
175. DasSarma, S., RajBhandary, U. L., and Khorana, H. G. (1984) Proc.
Natl. Acad. Sci. USA 81, 125-129.
176. Leong, D., Boyer, H., and Betlach, M. (1988) J. Bacteriol. 170,
4910-4915.
177.
Betlach, M. C., Shand, R. F., and Leong, D. M. (1989) Can. J.
178.
Betlach, M. C., and Gropp, F. (1994) System. Appl. Microbiol. 16,
179.
Soppa, J., and Oesterhelt, D. (1989) J. Biol. Chem. 264, 13043-
Microbiol. 35, 134-140.
712-715.
13048.
180. Soppa, J., Otomo, J., Straub, J., Tittor, J.. Meessen, S., and
Oesterheft, D. (1989) J. Biol. Chem. 264, 13049-13056.
160
181.
Cline, S. W., and Doolittle, W. F. (1987) J. Bacteriol. 169, 1341-
182.
Krebs, M. P., Hauss, T., Heyn, M. P., RajBhandarv,
1344.
U. L., and
Khorana, H. G. (1991) Proc. Natl. Acad. Sci. USA 88, 859-863.
183. Ebert, K., Goebel, W., and Pfeifer, F. (1984) Mol. Gen. Genet. 194,
91-97.
U. L., and Khorana, H. G. (1990)
184.
Krebs, M. P., RaiBhandary,
185.
Dunn, R. J., Hackett, N. R., McCoy, J. M., Chao, B. H
Nucleic Acids Res. 18, 6699.
Kimura, K.,
and Khorana, H. G. (1987) J. Biol. Chem. 262, 9246-954.
186. Hackett, N. R., Krebs, M. P., DasSarma, S., Goebel, W.,
RajBhandary, U. L., and Khorana, H. G. (1990) Nucleic Acids Res. 18,
3408.
187.
Ni, B., Chang, M., Duschl, A., Lanyi, J., and Needleman, R. (1990)
188.
Blaseio, U., and Pfeifer, F. (1990) Proc. Natl. Acad. Sci. USA 87,
Gene 90, 160-172.
6772-6776.
189. Alberts, A. W., Chen, J., Kuron, G., Hunt, V., Huff, J Hoffman, C.,
Rothrock, J. Lopez, M., Joshua, H., Harris, E. et al. (1986) Proc. Natl.
Acad. Sci. USA 77, 3957-3961.
190.
86,
Lam, W. L., and Doolittle, W. F. (1989) Proc. Natl. Acad. Sci. USA
5478-5482.
191. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular
Cloning A laboratory Manual. Cold Spring Harbor Laboratory: Cold
Spring Harbor, New York.
192. Cline, S. W., Lam, W. L., Charlebois, R. L., Schalkwyk, L. C., and
Doolittle, W. F. (1989) Can. J. Microbiol. 35, 148-152.
193. Kita, T., Brown, M. S., and Goldstein, J. L. (1980) J. Clin. Invest.
66,
1094-1100.
194. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman J.
G., Smith, J. A., and Struhl K. (1989) Current Protocols in Molecular
Biology, pp. 2.4.1-2.4.2. Wiley: New York, NY.
195. Feinberg, A. P., and Vogelstein, B. (1984) Anal. Biochem. 137,
266-267.
196. Racker, E. (1973) Biochem. Biophys. Res. Commun. 55, 224-230.
197.
Kagawa, Y., and Racker, E. (1971) J. Biol. Chem. 246, 5477-
198.
Bayley, H., Radhakrishnan, R., Huang, K.-S., and Khorana, H. G.
199.
Holz, M., Lindau, M., and Heyn, M. P. (1988) Biophys. J. 53, 623-
200.
Rosenberg, M., and Court, D. (1979) Annu. Rev. Genet. 13, 319-
201.
202.
Fischer, U., and Oesterhelt, D. (1979) Biophys J. 28, 211-230.
Marti, T, Rosselet, S. J., Otto, H., Heyn, M. P., and Khorana, H. G.
5487.
(1981) J. Biol. Chem. 256, 3797-3801.
633.
353.
(1991) J. Biol. Chem. 266, 18674-186g3.
161
203.
Cline, S. W., Schalkwvk, L. C., and Doolittle, W. F. (1989) J.
204.
Conover, R. K., and Doolittle, W. F. (1990) J. Bacteriol. 172, 3244-
205.
Thorgeirsson, T. E., Milder, S. J., Miercke, L. J. W., Betlach, M. C.,
Bacteriol. 171, 4987-4991.
3249.
D. S. (1991) Biochemistry
Shand, R. F., Stroud, R. M., and Kliger,
9133-9142.
30,
28, 7800-
206.
Flitsch, S. L., and Khorana, H. G. (1989) Biochemistry
207.
Altenbach, C., Flitsch. S. L., Khorana, H. G., and Hubbell, W. L.
208.
Rehorek, M., and Heyn, M. P. (1979) Biochemistry
209.
Kano, K., and Fendler, J. H. (1978) Biochim. Biophys. Acta 509,
7805.
(1989) Biochemistry
28, 7806-7812.
4983.
18, 4977-
289-299.
210. Fothergill, J. E. (1964) Fluorescent Protein Tracing (Narin, R. C.,
ed.), pp. 3 -59, Livingstone: Edinburg.
211. Nagle, J. F., and Dilley, R. A. (1986) J. Bioenerg. Biomembr. 18,
55-68.
212.
Teissie J., Gabriel, B., and Prats, M. (1993) Trends Biochem. Sci.
213.
214.
Haines, T. H. (1983) Proc. Natl. Acad. Sci. USA 80, 160-164.
Gutman, M., and Nachliel, E. (1990) Biochim. Biophys. Acta
215.
Lozier, R. H
216.
Drachev, L. A., Kaulen, A. D.. and Skulachev, V. P. (1984) FEBS
217.
Heberle, J., and Dencher, N. A. (1992) Proc. Natl. Acad. Sci. USA
218.
Alexiev U., Marti, T., Heyn, M. P., Khorana, H. G., and Scherrer,
219.
Poll, A., and Junge, W. (1989) Biophys J. 56, 27-31.
220.
Kates, M. (1978) Prog. Chem. Fats Other Lipids 15, 301-342.
221.
Lind, C., Hoieberg, B., and Khorana, H. G. (1981) J. Biol. Chem.
222.
Roee, P., Ahl, P. L., Das Gupta S. K., Hertzfeld, J., and Rothschild,
18, 243-246.
1015,
391-414.
Niederberger,
W
Bogomolni,
R. A., Hwan,
S., and
Stoeckenius, W. 1976) Biochim. Biophys. Acta 440, 545-556.
Lett. 178, 331-335.
89,
5996-6000.
P. (1994) Biochemistry
33,
13693-13699.
256, 8298-8305.
K. J. (1981) Biochemistry 26, 6696-6707.
223. Vonck, J., Han, B.-G., Burkard, F., Perkins, G. A., and Glaeser, R.
M. (1994) Biophys. J. 67, 1173-1178.
224. Han, B.-G., Vonck, J., and Glaeser, R. M. (1994) Biophys. J. 67,
1179-1186.
225. Alexiev, U., Marti, T., Heyn, M. P., Khorana, H. G., and Scherrer,
P. (1994) Biophys. J. 66, A44.
162
226. Blundell, T. L. and Johnson, L. N. (1976) Protein
Crystallography.
Academic Press: New York, NY.
227. Dao-Pin, S., Alber, T., Bell. J. A., Weaver, L. H., and Mathews, B.
W. (1987) Protein Eng. 1, 115-123.
228. Tucker, A. D., Batv D., Parker, M. W., Pattus, F., Lazdunski, C.,
and Tsernoglou, D. (199) Protein Eng. 2, 399-405.
229.
Stock A. M., Mottonen, J. M., Stock, J. B., and Schutt, C. E. (1989)
745-749.
Nature 37,
230. Klein, C., Vogel, W., Bender, H., and Schulz, G. E. (1990) Protein
Eng. 4, 65-67.
231. Boyer, P. D. (1954) J. Am. Chem. Soc. 76, 4331-4337.
232. Plohn, H.-J., and Buldt, G. (1986) J. Appl. Crystallogr. 19, 255-
261.
and Heyn, M. P. (1986)
233.
Seiff, F., Westerhausen, J., Wallat, I
234.
Seiff, F., Wallat, I., Westerhausen, J., and Heyn, M. P. (1986)
235.
Petsko, G. A. (1985) Methods
236.
Krebs
237.
Epstein,
238.
Glaeser, R. M., Jubb, J. S., and Henderson, R. (1985) Biophys. J.
Proc. Natl. Acad. Sci. USA 83, 7746-7756.
Biophys. J. 50, 629-635.
M. P., Behrens, W., Mollaaghababa,
Heyn, M. P. (1993) Biochemistry
4862-4866.
48,
Enzymol. 114, 147-156.
32,
R
Khorana, H. G., and
12830-12' 84.
C. C., and Datta, P. (1977) Proc. Natl. Acad. Sci. USA 74,
775-780.
163
Table 3.1.
Spectrai characteristics and proton
and mutant bR in the .e-mbrane form.
Absorption
umping of w id-type
Maximum
(nm)
Strain
Proton Pumping
DA
LA
WT(MPK1)
WT(AB)
WT(SN)
562
570
R82A
597
580
R82Q
D85A
DA
561
561
6 11
D85E
D85N
D96A
D96E
D96N
D212E
D212N
613
617
A
...
c7rof WT
569
569
597
580
611
616
615
562
562
562
567
575
573
572
582
571
571
·
100
91
110
10
13
ND
15
ND
8
56
9
6
Values for proton pumping reported are the average of six to
nine cosonication experiments.
The wild-type (WT) value
[WT(MPK1)] corresponds to an initial rate of 0.18 +/- 0.02 H+ per bR
per sec. ND, not detected: DA. dark adapted: LA. light adapted.
164
Time constants and amplitudes (; in parentheses below
Table 4.1.
the corresponding time) for the kinetics of the NI intermediate and
proton release, with fluorescein attached to various positions on bR
and with pyranine in the bulk phase.
wild-type
H+-release
M-rise
M-decay
(gs)
(PS)
(ms)
Flu
Pyr
ti
t2
-
820
2.2
49
(6)
(44) (50)
1.1
45
(6)
(42) (52)
2.0
43
(7)
3
147
3.4
6.8
(15) (53)
(32)
4.0
9.0
2.2
extraceilUar
G72C-MF
71
850
152
2.4
(34) (45)
(21)
i~~~~~~~~~~~~~~~~~~~~~~~
ctoiasmi
VIO1C-MF
Q105C-MF
A160C-MF
77
76
74
950
920
850
1.3
4.6
(58) (35)
(67)
(31)
1.0
(5)
160
30
(42) (53)
2.4
6.6
1.4
47
(41)
175
(54)
2.8
24
101
(33)
(60)
(5)
G231C-IMF
60
820
1.4
(7)
141
(20) (51)
17.7
(29)
9.9
(44) (52)
64.0
(4)
2.2 11 130.0
(77) (16) (7)
165
CURRICULUM VITAE
Ramin Mollaazhababa
EDUCATION
POSITIONS
HONORS AND
State University of New York
at Stony Brook
:i85-i989
Chemistry
B.S.
Magna Cum Laude
Massachusetts Institute of
Technology
1989-1994
Chemistry
Ph.D.
Candidate
1990-1994
Research Associate. Professor H. G. Khorana
Massachusetts Institute of Technology
Departments of Biology and Chemistry
Cambridge, MA
1989-1990
Teaching Assistant
Massachusetts Institute of Technology
Department of Chemistry
Cambridge, MA
1987-1989
Research Assistant, Professor Glenn D. Prestwich
State University of New York at Stony Brook
Department of Chemistry
Stony Brook, NY
1987-1988
Teaching Assistant, Professor Glenn D. Prestwich
State University of New York at Stony Brook
Department of Chemistry
Stony Brook, NY
Phi Beta Kappa, 1989
AWARDS
Phi Lambda Upsilon, National Honorary Chemical Society, Member,
1989
American Institute of Chemists Outstanding Senior Award, 1989
Outstanding Leadership and Service Award, State University of
New York at Stony Brook, 1989
Undergraduate Excellence Award, State University of New York at
Stony Brook, 1988 and 1989
George Emerson Outstanding Junior Award, Department of Chemistry
at the State University of New York at Stony Brook, 1988
Alpha Epsilon Delta, National Premedical Honor Society, Member,
1988
President, Undergraduate Chemistry Society, State University of
New York at Stony Brook, 1987-1989
Sigma Beta, Freshman Honor Society, State University of New York
at Stony Brook, Member, 1986
166
Ramin Mollaaghababa
PUBLICATIONS
Physiological Influence of Fenoxycarb Pro-Insecticides and Soldier Head Extracts of
Various Termite Species on Soldier Differentiation in Reticulitermes flavipes
(Isoptera). B. M. Okot-Kotber. Istvan Ujvary, Ramin Mollaaghababa, F. Szurdoki, G.
Matolcsy, and Glenn D. Prestwich (1991) Sociobiology 19, 77-90.
Gene Replacement in Halobacterium halobium and Expression of Bacteriorhodopsin Mutants.
M. P. Krebs, R. Mollaaghababa and H. G. Khorana (1993) Proc. Natl. Acad. Sci. USA 90,
1987-1991.
X-Ray Diffraction of a Cysteine-containing Bacteriorhodopsin Mutant and Its Mercury
Derivative. Localization of an Amino Acid Residue in the Loop of an Integral Membrane
Protein. M. P. Krebs, W. Behrens, R. Mollaaghababa, H. G. Khorana and M. P. Heyn
(1993) Biochemistry 32, 12830-12834.
Time-resolved Detection of Structural Changes During the Photocycle of Spin Labeled
Bacteriorhodopsin. H.-J. Steinhoff, R. Mollaaghababa, C. Altenbach, K. Hideg, M. P.
Krebs, H. G. Khorana and W. L. Hubbell (1994) Science 266, 105-107.
Site-directed Spin Labeling Studies of Structure and Dynamics in Bacteriorhodopsin.
H.-J. Steinhoff, R. Mollaaghababa, C. Altenbach, H. G. Khorana, and W. L. Hubbell
(1994), manuscript submitted to Biophysical Chemistry.
Rapid Long-Range Proton Diffusion Along the Surface of the Purple Membrane and Delayed
Proton
Transfer
into the Bulk.
U. Alexiev,
R. Mollaaghababa,
P. Scherrer,
H. G.
Khorana and M. P. Heyn (1995) Proc. Natl. Acad. Sci. USA 92, 372-376.
Intramolecular Charge Transfer in the Bacteriorhodopsin Mutants Asp85-Asn and
Asp212-Asn: Effects of pH and Anions. Stephan Molke, Mark P. Krebs, Ramin
Mollaaghababa, H. Gobind Khorana and Maarten P. Heyn, manuscript in preparation.
Expression of Mammalian Photoreceptor Rhodopsin in Yeast Saccharomyces cerevisiae. R.
Mollaaghababa, F. F. Davidson, C. Kaiser and H. G. Khorana, manuscript in preparation.
Mapping the Structural Transitions in Bacteriorhodopsin Photocycle by Use of Timeresolved Electron Paramagnetic Resonance. R. Mollaaghababa, H.-J. Steinhoff, W. L.
Hubbell and H. G. Khorana, manuscript in preparation.
Time-resolved X-ray Analysis of a Bacteriorhodopsin Mutant Derivatized with a Mercury
Label. An Analysis of the Conformation of Bacteriorhodopsin in the M State. R.
Mollaaghababa, W. Behrens, U. Alexiev, H. G. Khorana and M. P. Heyn, manuscript in
preparation.