University of Southern California Dissertations and Theses

Structural studies of nicotinic acetylcholine receptors and their regulatory complexes

(USC Thesis Other)

Structural studies of nicotinic acetylcholine receptors and their regulatory complexes

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

STRUCTURAL STUDIES OF NICOTINIC ACETYLCHOLINE
RECEPTORS AND THEIR REGULATORY COMPLEXES

by

Kaori Noridomi

_________________________________________________________________

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CHEMISTRY)

August 2015

Copyright 2015 Kaori Noridom

iii

This thesis is dedicated
To my grandfather who taught me not to give up
and
To my grandmother who taught me to live strong
and
To my mother who provides me priceless love

iv

Acknowledgements

This dissertation would not have been possible without the help of so
many people both in the United States and Japan.
First, and most of all, I would like to express the deepest appreciation
to my advisor Dr. Lin Chen for allowing me to work on challenging projects,
for patiently waiting and supporting me for past eight years, for providing us
freedom to be independent scientists, and for teaching me crystallography step
by step. It was the most enjoyable time to discuss science with him. I will
never forget the day when I finally obtained the first crystal after working on
it for five years. I will never forget the day when I finally solved the first
crystal structure. Not only have I learned so many things in his lab, but I also
had the opportunity to meet many lifelong friends.
It is a genuine pleasure to express my deep sense of thanks and
gratitude to my committee members. Dr. Don Arnold provided me with
constant support and encouragement through many difficult times during my
Ph.D. career. I appreciate Dr. Vadim Cherezov for his guidance and inputs for
my Ph.D. dissertation. Dr. Richard Roberts was a member of my dissertation
committee as well as a collaborator. His scholarly advice and scientific
approach have helped me to a very great extent to accomplish this work. I am
also grateful for the opportunity to work with his students, Aaron Nichols and
v

Lan Huong Lai. I enjoyed the cohesive teamwork and communication. I truly
hope that in my next position, I will be able to work individuals as talented as
they are. I would also like to express my gratitude to Dr. Michael Lieber and
Dr. Chih-Lin Hsieh for their support, advice, guidance, valuable comments,
and suggestions. They were not just professors but also scientists who opened
my eyes as to what kind of scientist I would like to be in the future.
Next, I would like to thank the wonderful people who I met at USC.
They are my best friends, great colleagues, and often times feel like brothers
or sisters to me. First, I would like to acknowledge Dr. Go Watanabe. I would
not be here without him who provided me with tremendous support
throughout my Ph.D. career. He helped me to establish my life as a Ph.D.
student at USC and has kept telling me to trust my abilities as a scientist. He
patiently taught me and helped me to build up strong basic skills to work in a
biology lab. Dr. Nimanthi Jayathilaka and Dr. Reza Kalhor also offered
tremendous support from day one and they are one of reasons I decided to join
Dr. Lin Chen’s lab. Reza always gave me amazing lectures and Nimanthi gave
me hands-on trainings on experiments. I am pleased that I had such
outstanding senior graduate students and was able to work with them over the
years. Our scientific discussions did not just stay at the lab, but also at home
over wine. These memories will not fade with time and I shall cherish them
forever. Dr. Keriann Oertell is the first friend I made at USC. She was always
available for me and provided support in both science and life. She also
helped me to improve and continuously work on my English. We were always
together and she made sure that I was moving forward.
I would also thank to all of the Lin Chen lab members. Katherine
Daugherty and Dr. Michael Philips were a big part of my Ph.D. life. It was the
best time for all of us; Go, Nimanthi, Reza, Katie and Michael were working
together in the lab. Even though my experiments were tough at the time, they
vi

made my life fun and special. Their smiles gave me a lot of encouragement
and motivation. Dr. Cosma Dellisanti and Dr. Aki Uchida introduced me to
Lin Chen’s lab, and it was the starting point of my Ph.D. career. I am also
grateful to have worked with Kevin Cheng, who cheered me up and made me
laugh whenever I was stressed out. I enjoyed chatting with Xiao Lei and
Haochen Li about not only science but also our future plans. Lastly, I cannot
forget to mention members of our receptor group. Melissa Hansen has an
amazing passion and toughness in pushing our challenging project forward.
She brightened up our difficult and depressing period with her positive
attitude. Kevin Wu, Fiona Obiez, and Stella Chen were also great additions to
the group. I learned how enjoyable it is to work as a team and they gave me a
happiness to see them becoming a scientist. I also had a fun time talking with
previous lab members, Dr. Yongqing Wu and Dr. Yongheng Chen.
During my Ph.D., I could meet so many amazing people who are very
considerate. Zhiwei Chen, Dr. Meng Xia, Victoria Duan, and Dr. Yiming
Xing were a mom and/or a big sister to me. They watched me all the time and
provided me with emotional support. It was nice to work with Linlin Ma,
Stephanie Chu, and Carolina Dantas in the same lab space. They were very
helpful and I really enjoyed girls talk with them. Especially, I could not
survive without a weekly coffee time with Stephanie. It did not take a long
time for me to realize that she would become another lifelong friend. It was a
huge encouragement to have Damian, who also worked on crystallography
and went through the similar difficulties. I appreciate Varuzhan Balasanyan
for checking on me time to time and having scientific discussions. He
provided me insightful inputs and motivation. I could share my frustration and
worries about being a Ph.D. student with Taehyun Ryu and Tuba Koker. We
encouraged each other and I finally could achieve the goal. Mimi Chun
vii

showed me to be a strong but a loveable woman. She provided me smiles and
motivation.
I owe a deep sense of gratitude to staff of GM/CA at APS as well as
Dr. Shuxing Li at the USC NanoBiophysics Core Facility for their support. I
am also grateful to Chemistry and MCB departments at USC. Michele Dea at
Chemistry department supported and assisted me through my long Ph.D.
period. Amazing staff at MCB department, Rokas Oginskis, Laura Cajero,
Hayley Peltz, Kathleen Walker, Eleni Yokas, and Christina Tasulis provided
me the help and guidance. I always had a place to ask for help and I could
focus on my research because of their physical and mental support. It is my
privilege to thank my undergraduate advisor Dr. Levente Fabry-Asztalos for
his constant encouragement throughout my Ph.D. career. Christopher Dayley,
a scientist at Merck provided me with courage and confidence to continue my
career as a scientist. He also introduced me to potential industry options
outside of scientific academia.
I would like to thank members of Wakahisakai, especially Hisame
Sensei and Mr. Okada who are my parental figures in the United States. They
cheered me up and pushed me through the long Ph.D. journey. I also had a
great experience working with instructors of Japanese classes at USC. It was a
nice and unique opportunity to work as an EALC TA, and I was encouraged
by their smiles and warm hearts. A lot of my friends in Japan have supported
me since I came to the United States. I could not achieve my goal without
their love, support, and encouragement. They are my lifetime treasure. In
addition, I am grateful to have lifelong mentors, Ms. Kazuko Tadakuma who
is my teacher at a junior high school and Ms. Noriko Nakamura who is my
teacher at a high school. My best friend Chie Sukeyasu supported me
throughout my life since I met her. She somehow could sense when I am in a
viii

difficult situation and she always sent me the warmest messages. I do not
know how many times I was encouraged by her.
Last of all, I would like to acknowledge my family. My grandparents
raised me with lots of love while providing me with amazing life lessons. My
grandfather taught me not to give up and my grandmother taught me to live
strong. My mother, full of endless love, is forever my role model. Lastly, Ku-
ta, my puppy supported me as a counselor. His smiles cheered me up a lot,
and I really appreciate his 24/7 support. What I learned during my Ph.D. was
not only science but I also learned how to appreciate people and how to love
people, and. my Ph.D. is the fruit of love and support, which I was received
from the tremendous people I met.

Kaori Noridomi
June 2015

ix

Table of Contents

Dedication iii
Acknowledgements iv
List of Tables xv
List of Figures xvi
Abbreviations xxi
Attributions xxiii
Abstract xxiv

Chapter 1 Introduction 1
1.1 Nicotinic acetylcholine receptors ...…………………………..…..……….... 1
1.1.1 Stoichiometries of nAChRs ……………………..…………………. 2
1.1.2 Structures of nAChRs .……….…………………………………..… 5
1.1.3 Ligand binding site ….….…….……………………………………. 6
1.2 Early structural studies of nAChRs ……….….………………………...…... 8
1.2.1 Acetylcholine binding protein ………….…..………………………. 9
1.2.2 Muscle-type nAChRs ……………………...…………………..…. 10
1.2.3 Prokaryotic LGICs ………………………....………………..……. 12
1.2.4 GABA
A
and serotoni 5-HT
3
receptors ………..………………..…. 13
1.2.5 Neuronal nAChRs ……………………….....…………………..…. 14
1.4 This work ……………………………….……………………………...….. 14

x

Chapter 2 Structural Studies of nAChR α1 Subunit and Fab35 Complex to
Understand Autoimmune Disease of Myasthenia Gravis 18
2.1 Abstract ……………………………………………………………….….... 18
2.2 Introduction ……………..………………………………………………..... 19
2.3 Results …………………………………………………………….............. 24
2.3.1 α211 purification …………….……………………..………...….. 24
2.3.2 mAb35 purification …...…….……………………………………. 26
2.3.3 mAb35 digestion and Fab35 purification ……………………...…. 28
2.3.4 Gel shift assay …………………………….……………………… 30
2.3.5 The α211/Fab35/α-bungarotoxin complex purification …..…….... 33
2.3.6 Crystallization of the α211/Fab35/α-bungarotoxin ternary
complex ……………………….………………….........…………. 35
2.3.7 Structure determination of the α211/Fab35/α-bungarotoxin
ternary complex …………............................................................... 37
2.3.8 Structure guided analysis of the α211/Fab35 interactions….….…. 40
2.4 Discussion ………………………..…………………………..............…… 60
2.5 Materials and Methods…..………………………………………………… 63
2.5.1 Construction of α211 …………………………….…………...….. 63
2.5.2 α211 expression …………………………………..……...………. 63
2.5.3 α211 purification …………….……………………………….….. 64
2.5.4 Cell culture and reagents for mAb35 ………………………....….. 65
2.5.5 mAb35 purification …...…….……………………………...…….. 65
2.5.6 mAb35 digestion and Fab35 purification ………………...………. 66
2.5.7 Gel shift assay …………………………….……………...………. 66
2.5.8 The α211/Fab35/α-bungarotoxin complex purification …........….. 67
2.5.9 Crystallization and data collection ………….………................…. 67

Chapter 3 Structural Studies of nAChR α1 Subunit and Fab210 Complex 69
3.1 Abstract …………………………………………..………………….……. 69
3.2 Introduction ……………..…………………………………………….…... 70
3.3 Results ………………………………………………………...…………... 71
xi

3.3.1 Fab210 preparation ….....…….……………………..…………….. 71
3.3.2 Gel shift assay …………………………….………………………. 73
3.3.3 The α211/Fab210/α-bungarotoxin complex purification ……......... 75
3.3.4 Crystallization of the α211/Fab210/α-bungarotoxin terynary
complex …………………………………………...……............…. 77
3.3.5 Structure determination of the α211/Fab210/α-bungarotoxin
ternary complex …………………………………………............... 81
3.3.6 Structure guided analysis of the α211/Fab210 interactions …......... 84
3.4 Discussion ………………………..…………………....….……...……..… 97
3.5 Materials and Methods…..…………………………………………..…… 100
3.5.1 α211 expression and purification …….…………..…………..…. 100
3.5.2 mAb210 and Fab210 production …………………………..……. 100
3.5.3 Gel shift assay …………………………….…………….....……. 100
3.5.4 The α211/Fab210/α-bungarotoxin complex purification ….....…. 100
3.5.5 Crystallization and data collection …………..……….............…. 101

Chapter 4 Development of a MG Antibody Detection Method
Using ELISA 103
4.1 Abstract ……………………………………….……..………………...…. 103
4.2 Introduction ……………..…………………….……..……………...…..... 104
4.3 Results …………………………………………….…...…………............. 108
4.3.1 Establishment of control assays …………..………..……………. 108
4.3.2 mAb35 detection in rat serum …………………….……….….…. 118
4.3.3 mAb35 titration in rat serum …….…………..….………….……. 121
4.3.4 Improvements of detection limit .…..………..…….…………….. 123
4.3.5 Human α211 test ……………..….…………..…….…………….. 125
4.3.6 MG patient serum dilution test ….…………..…….…………….. 125
4.3.7 Screening MG patient serum samples …...…..………………….. 128
4.4 Discussion ………………………..………….………………...…………. 130
4.5 Materials and Methods …..…………………………………...….………. 132
4.5.1 Samples preparation for ELISA .….….……..…….....….………. 132
xii

4.5.2 ELISA for cotrol experiments …….………………….…………. 133
4.5.3 Washing test …………………...….……….……….……………. 133
4.5.4 mAb35 detection in rat serum …….…….……………………….. 133
4.5.5 mAb35 titration in rat serum …….…………...………………….. 134
4.5.6 Improvements of detection limit .…..………...………………….. 134
4.5.7 MG patient serum dilution test ...…..…….…..……………….…. 135
4.5.8 MG patient serum screening …....…..………..…………….……. 135

Chapter 5 Crystallographic Studies of Novel α211 Targeting Proteins
Selected by mRNA Display Technique 137
5.1 Abstract …………………………………………………………………... 137
5.2 Introduction ……………..………………….…………………………….. 138
5.3 Results ……………………………………………….…………................ 143
5.3.1 Clone 4.2 expression and purification ………...………………… 143
5.3.2 Clone 4.2 gel shift assay ………………….……………………... 146
5.3.3 α211/Clone 4.2 complex purification and crystallization .............. 150
5.3.4 Clone 4.2 mutant expression and purification …………………... 152
5.3.5 Clone 4.2 mutant gel shift assay ……….……….……………….. 153
5.3.6 α211/Clone 4.2 mutant complex purification and crystallization 154
5.3.7 Competitive binding assay of Clone 4.2 mutant with
α-bungarotoxin …………………………...…………………….... 154
5.4 Discussion ………………………..…………………………...….………. 156
5.5 Materials and Methods ....………………………………………………... 159
5.5.1 Construction of the Clone 4.2 and Clone 4.2 mutants ……...….... 159
5.5.2 Clone 4.2 expression …………………...……………………….. 159
5.5.3 Clone 4.2 purification …..……………...………………………... 160
5.5.4 Gel shift assay …………………………….……………………... 161
5.5.5 α211/Clone 4.2 complex purification and crystallization .............. 161
5.5.6 Clone 4.2 mutant expression and purification ……….………….. 162
5.5.7 Gel shift assay for specificity ………….….…………………….. 162
xiii

5.5.8 α211/Clone 4.2 mutant complex purification and
crystallization ................................................................................. 163
Chapter 6 Structural Determination of nAChR α9 Subunit 164
6.1 Abstract ……………………………………………………………..……. 164
6.2 Introduction ……………..……………………………………………….. 165
6.3 Results ……………………………………………………………............. 167
6.3.1 α9 purification …………….……………………..………............ 167
6.3.2 Attempts of α9 crystallization …...……….……………………… 170
6.3.3 Gel shift assay …………………………….……………………... 170
6.3.4 Rat Flag α9 mutants/α-bungarotoxin complex purification ……... 172
6.3.5 Crystallization and data collection of rat Flag α9 mutants
with α-bungarotoxin ……………………….…………….…...…. 173
6.3.6 Endo Hf digestion …...….……………….…………….……….... 177
6.3.7 Endo Hf treated rat Flag α9 3Mut/α-bungarotoxin
complex purification …….……………………………………..... 179
6.3.8 Crystallization and data collection of Endo Hf treated
rat Flag α9 3Mut/α-bungarotoxin …….…………………………. 180
6.3.9 Structure guided analysis of α9 ......…………………..……....…. 185
6.4 Discussion ………………………..…………………………...…………. 193
6.5 Materials and Methods …...………………………………………...……. 196
6.5.1 Construction of α9 expression vector ...……….……….………... 196
6.5.2 α9 expression and purification ……..…………..………….……. 196
6.5.3 Gel shift assay …………………………….…………………….. 197
6.5.4 Crystallization of rat Flag α9 mutants with α-bungarotoxin ……. 197
6.5.5 EndoHf digestion ……………………….…………...…………... 198
6.5.6 Endo Hf treated rat Flag α9 3Mut/α-bungarotoxin
complex purification …….…………………………….…............ 199
6.5.7 Crystallization of Endo Hf treated rat Flag α9 3Mut/
α-bungarotoxin complex …………..….………..…………........... 199
6.5.8 Data collection ………………………...…….….……....……….. 200

xiv

Chapter 7 Expression and Purification of nAChR α7/AChBP
Chimera 201
7.1 Abstract …………………………………………………….……………. 201
7.2 Introduction ……………..………………………………….……………. 202
7.3 Results ……………………………………………………….……........... 205
7.3.1 α7/AChBP chimera expression and purification ....….………….. 205
7.3.2 Comparisons of α7/AChBP chimera proteins .……….…………. 207
7.3.3 Endo Hf digestion on α7/AChBP chimera …….…….………….. 208
7.4 Discussion ………………………..……………….…………...…………. 209
7.5 Materials and Methods …..………………………………………………. 211
7.5.1 Construction of α7/AChBP chimera …..………………..………. 211
7.5.2 α7/AChBP chimera expression …………..………….………….. 211
7.5.3 α7/AChBP chimera purification ….………..……………............. 211
7.5.4 Endo Hf digestion ………………...………..……………............ 212

Chapter 8 Conclusion 213

Bibliography 223

Appendices 241
Appendix A: Figures for structural studies …………………………………… 241
Appendix B: Sequence alignments ………………………………………...…. 246

xv

List of Tables

Table 2.1 Crystallographic analysis of human or mouse α211/Fab35/α-
bungarotoxin complexes ………………………………...…............ 39
Table 2.2 List of contacting residue pairs between α1 subunit and the heavy
or light chain of Fab35 …….…………………………………..…... 44
Table 3.1 Cryo solution test for phase separation ……………..…….…...…... 80
Table 3.2 Summary of unit cell dimensions of each complex crystal ….…...... 81
Table 3.3 Crystallographic analysis of human or mouse α211/Fab210/α-
bungarotoxin complexes ………………..……………….……….... 83
Table 3.4 List of contacting residue pairs between α1 subunit and heavy
or light chain of Fab210 ..……………………...…………………... 86
Table 3.5 Summary and comparison of contacting residue pairs between α1
subunit and heavy or light chain of Fab35 and Fab210 ..…..….…... 87
Table 4.1 Washing test of stability of mouse α211 and antibodies ……..…... 113
Table 6.1 Crystallographic analysis of Endo Hf treated rat Flag α9 3Mut
complex ……….............................................................................. 184

xvi

List of Figures

Figure 1.1 Subtypes and major stoichiometries of nAChRs ………..………… 4
Figure 1.2 Schematic of subunit comparison …………….……………............ 7
Figure 1.3 Schematic representation of the ACh binding site of α7 nAChR …. 8
Figure 1.4 Electron microscopy structure of the Torpedo nAChR …..……..... 11
Figure 1.5 X-ray crystal structures of GABA
A
and serotonin
5-HT
3
receptors ………………………………………….……….. 13
Figure 2.1 Mouse α211 purification ……………………….……………….... 25
Figure 2.2 Purification of mAb35 from hybridoma cells ...……….…………. 27
Figure 2.3 Fab35 purification …..………………………………….……….... 29
Figure 2.4 Binding site of α-bungarotoxin and expected binding site of
Fab35 …………………………………………………………….. 30
Figure 2.5 6% Native-PAGE gel showing binding of Fab35 to mouse α211 .. 31
Figure 2.6 10% Native-PAGE gel showing specific binding of Fab35 to
mouse α211 …………………………...………….………….…… 32
Figure 2.7 The mouse α211/Fab35/α-bungarotoxin complex purification ..… 34
Figure 2.8 Crystals grown as bundle with some precipitations …………….... 36
Figure 2.9 X-ray crystal structure of the human α211/Fab35/α-bungarotoxin
complex ...……………………………………………………….... 41
Figure 2.10 Interactions of Fabs with human α1 subunit ….……….………...... 42
Figure 2.11 Surface model of α1 subunit and Fab35 showing contacting
residues on each protein ...…….……………………….…..……... 45
xvii

Figure 2.12 Water molecules mediating interactions at the interface of α1
subunit and Fab35 ...… ……........................................................... 47
Figure 2.13 Binding interactions of α1 subunit and Fab35 at the MIR ……..… 48
Figure 2.14 Electron density maps of binding interactions of α1 subunit and
Fab35 at the MIR …………………………………………………. 49
Figure 2.15 Binding interactions with insertion of CDR-H3 ....……………...... 52
Figure 2.16 Additional interactions involving the N-terminal α-helix ……..…. 53
Figure 2.17 Water mediated interactions of α1 subunit and Fab35 .................... 54
Figure 2.18 A model of Fab35 bound to a pentamer receptor,
α7/AChBP chimera …...………………………………….....…… 56
Figure 2.19 A model of a full-length antibody bound to nAChR ……....…...... 57
Figure 2.20 A model of an antibody cross-link two nAChRs ………...…….… 59
Figure 3.1 Fab210 purification …………….…………………..…….............. 72
Figure 3.2 10% Native-PAGE gel showing binding of Fab210 to
mouse α211 ………..………………………………...…………… 74
Figure 3.3 The α211/Fab210/α-bungarotoxin complex purification .….…...... 76
Figure 3.4 Preliminary crystals of mouse α211/Fab210/α-bungarotoxin …..... 77
Figure 3.5 Optimized crystals of the α211/Fab210/α-bungarotoxin …...…..... 78
Figure 3.6 Phase separation problem of cryo solution ……………...……..… 80
Figure 3.7 Superimposed view of the human Fab35 ternary structure and
the human Fab210 ternary structure ............................................... 85
Figure 3.8 Interactions at the N-terminal α-helix mediated by interfacial
water molecules ……………………………………………..….... 89
Figure 3.9 Comparison of the interfaces (Fab35 vs. Fab210) at the
N-terminal α-helix …………………………………………..…… 90
Figure 3.10 Electron density map of interactions at the N-terminal α-helix ….. 90
Figure 3.11 Binding interactions of α1 subunit and Fab210 at the MIR .……... 93
Figure 3.12 Comparisons of electron density maps of binding interactions
at the MIR (Fab35 ternary vs. Fab210 ternary) …..……...………. 94
Figure 3.13 Unique interaction site of the Fab210 ternary structure .……..…... 95
Figure 3.14 Binding interactions with insertion of CDR-H3 .…………….…... 96
xviii

Figure 3.15 Binding interaction with aromatic residues ………………….…... 99
Figure 4.1 Schematic model of ELISA using a Ni coated ELISA plate ...….. 108
Figure 4.2 Control experiments to determine non-specific binding ……....... 109
Figure 4.3 Non-specific biding test of 2°Ab to a plate and/or mouse α211 ... 111
Figure 4.4 1°Ab titration test ………………………………………..…..….. 115
Figure 4.5 Additional 2°Ab test .………..……………………………....…... 117
Figure 4.6 Serum dilution test ………………...………………………..…... 120
Figure 4.7 mAb35 titration test in rat serum ..………………………….…… 122
Figure 4.8 Detection limit optimization with 1:50 serum dilution ................. 124
Figure 4.9 Mouse α211 vs. Human α211 ...…………………….…..……..... 126
Figure 4.10 MG patient dilution test ...………………………………………. 127
Figure 4.11 Human MG samples screening .…………………………….…... 129
Figure 5.1 Structural comparison of VHH and the 10FnIII …….…..….......... 140
Figure 5.2 Selection cycle of a typical mRNA-display ...…………………… 142
Figure 5.3 Small scale expression test for Clone 4.2 ...………….................... 143
Figure 5.4 Clone 4.2 purification over Mono Q HR 5/5 and Superdex 75
10/300 ....…………………………………………………..…….. 145
Figure 5.5 Gel shift assay showing binding of Clone 4.2 to mouse α211 …... 148
Figure 5.6 Gel shift assay showing Clone 4.2 is temperature stable
as well as competitive with α-bungarotoxin ………..…………… 149
Figure 5.7 Mouse α211/Clone 4.2 complex purification …………………… 151
Figure 5.8 Gel shift assay showing Clone 4.2 mutant 5b binds specifically
to α211 …………………………………..………………….…… 153
Figure 5.9 Competition assay of Clone 4.2 mutant 5b with α-bungarotoxin 155
Figure 6.1 Rat Flag α9 3Mut purification …..………………………………. 169
Figure 6.2 10% Native-PAGE gels showing α-bungarotoxin binding to
rat Flag α9 3Mut and rat Flag α9 single Mut ……........................ 171
Figure 6.3 Rat Flag α9 3Mut/α-bungarotoxin purification .………………… 172
Figure 6.4 Rat Flag α9 3Mut/α-bungarotoxin crystals ….………………….. 174
Figure 6.5 Rat Flag α9 single Mut/α-bungarotoxin crystals …...………........ 175
xix

Figure 6.6 Crystals of α9 complexes in different crystallization condition
…………………………………….……........................................ 176
Figure 6.7 Removal of N32 glycan by Endo Hf ……………………..………178
Figure 6.8 Endo Hf treated rat Flag α9 3Mut/α-bungarotoxin
purification .................................................................................... 179
Figure 6.9 Crystals of Endo Hf treated rat Flag α9 3Mut complex. ............... 183
Figure 6.10 Electron density maps at glycosylation sites …………….……… 187
Figure 6.11 Introducing a correct and a false mutation to test model-bias …... 188
Figure 6.12 Electron density map after building in a sugar chain ………….… 189
Figure 6.13 Comparison of sugar chains between α211 complex and
α9 complex ……............................................................................ 190
Figure 6.14 Interactions of three mutations introduced to enhance
α-bungarotoxin binding ………………………………….….…... 192
Figure 7.1 Sequence information of α7/AChBP chimera ….………….....…. 204
Figure 7.2 α7/AChBP purification …….......................................................... 206
Figure 7.3 Comparison of α7/AChBP purified with different constructs .…. 208
Figure 8.1 Hypothesized schematic view of the immobilized synthetic
peptide which corresponds to the MIR loop on a poly-D-Lys
coated well with glutaraldehyde ….…...……………..………..... 222
Figure A.1 Superimposed view of the human Fab35 ternary structure and
the mouse Fab35 ternary structure ...……...………..………......... 242
Figure A.2 Ramachandran plot of the human Fab35 ternary complex
structure ………………...…….…….………………………......... 243
Figure A.3 Ramachandran plot of the human Fab210 ternary complex
structure ………………………………………..……………........ 244
Figure A.4 Ramachandran plot of Endo Hf treated α9/α-bungarotoxin
complex structure …………………………………..….……….... 245
Figure B.1 Sequence alignments of human α subunits and other related
subunits to determine cross-reactivity as well as critical
residues for different mAb bindings (PISA results based) ............. 247

xx

Figure B.2 Sequence alignments of human α subunits and other related
subunits to determine cross-reactivity as well as critical
residues for different mAb bindings
(CCP4 contact results based) ………………………….…..….... 249
Figure B.3 Sequence alignments of Fabs (Light Chain) …....……….……..... 251
Figure B.4 Sequence alignments of Fabs (Heavy Chain) ……….................... 252

xxi

Abbreviations

ACh Acetylcholine
AChBP Acetylcholine binding protein
ALS Amyotrophic lateral sclerosis
APS Advanced photon source
cDNA Complementary DNA(deoxyribonucleic acid)
CDR Complementarity determining region
CDR-H1 Complementarity determining region 1 of the heavy chain
CDR-L1 Complementarity determining region 1 of the light chain
C
H
Constant domain, Heavy chain
C
L
Constant domain, Light chain
CMT Charcot-Marie-Tooth disease
CTRL Control
dAb Domain antibody
dsDNA Double stranded DNA
ECD Extracellular domain
ELISA Enzyme-linked immunosorbent assay
Em Emission filter
Ex Excitation filter
e10FnIII Modified fibronectin type III domain
Fab Fragment, antigen-binding
Fc Fragment, crystallizable
FPLC Fast protein liquid chromatography
GABA
A
Type-A γ-aminobutyric acid (receptors)
GluCl Glutamate-gated chloride channel
IPTG Isopropyl β-D-thiogalactopyranoside
xxii

IVIg Intravenous immunoglobulin
LGIC Ligand gated ion channel
LRP4 Low-density lipoprotein receptor-related protein 4
MG Myasthenia Gravis
mAb Monoclonal antibody
MIR Main immunogenic region
mRNA Messenger RNA (ribonucleic acid)
MuSK Muscle specific kinase
nAChR Nicotinic acetylcholine receptor
PBS Phosphate-buffered saline
PCR Polymerase chain reaction
RIA Radioimmunoassay
scFv Single-chain antibody
SDS-PAGE Sodium dodecyl sulfate – polyacrylamide gel electrophoresis
VH (V
H
) Variable domain, Heavy chain
VHH Variable domain, Heavy chain of Camelidae antibodies
VL (V
L
) Variable domain, Light chain
WT Wild type
α-Btx α-bungarotoxin
α211 Extracellular domain of nAChR α1 subunit
1°Ab Primary antibody
2°Ab Secondary antibody
10FnIII Fibronectin type III domain

xxiii

Attributions

I performed all of the work presented in this dissertation with the
exceptions detailed here.
My advisor Dr. Lin Chen contributed to structural analysis by program
O in Chapters 2, 3, and 6. Go Watanabe contributed to harvesting crystals and
collecting diffraction data in Chapters 2, 3, and 6. Melissa Hansen contributed
to purifying mouse α211, mAb35, Fab35, and the complex of α211/Fab35/α-
bungarotoxin in Chapter 2. Aaron Nichols performed a selection of binding
proteins with mRNA display in Chapter 5. Kevin Wu and Fiona Obiezu
contributed to purification of Clone 4.2 and Clone 4.2 mutant 5b in Chapter 5.

xxiv

Abstract

Nicotinic acetylcholine receptors (nAChRs) play a key role in
neuronal communication by sending electric signals upon binding of
neurotransmitters, acetylcholines. nAChRs relate closely to various neuronal
diseases including Alzheimer’s disease and schizophrenia. Understanding the
mechanism of nAChRs would lead to better understanding of those diseases
and eventually to new treatments. This dissertation covers six different
projects, four of which are related to a nAChRs α1 subunit. The other two
projects involve nAChR subunits, α7 and α9. For many of our studies, the X-
ray crystallography technique is utilized for structural determination, which
provides useful information to develop drugs or deeper understandings of
molecular mechanisms.
The main objective of this dissertation is to understand how a disease
called myasthenia gravis (MG) is triggered at molecular level. It is an
autoimmune neuromuscular disease, and it is known to be caused by
autoantibodies targeting nAChRs. Tremendous researches have been done, but
it is still not known how the autoantibodies actually bind the receptors. The
first part of discussion is the structural determination of nAChR α1 subunit
bound by one of MG mAbs named mAb35, which is a prototypical MG mAb.
The first X-ray crystal structure of the complex was determined at 2.6 Å, and
xxv

detailed interactions of the interface were revealed. The study was then
extended to the determination of another complex structure of nAChR
α1/mAb210. Various types of MG mAbs have been generated, and they have
shown different characteristics. Thus, it is important to learn the differences of
these mAbs, specifically mAb35 and mAb210, to understand MG better.
Furthermore, this knowledge could be applied for drug developments in the
future. The crystal structures of these two distinct complexes provide, for the
first time, molecular details for understanding the disease mechanisms and for
structure-guided drug design of MG. As the third aim, a new detection method
of MG antibody was developed by utilizing all materials and knowledge
obtained in the first section of MG study. The current diagnosis of MG has
several limitations due to the usage of radiolabeled materials. Even though
additional experiments are required for further improvement, our newly
developed assay system is more time efficient and cost effective and could
contribute to a MG diagnosis in the feature. The last aim related to MG is to
discover proteins which can specifically target the nAChR α1 subunit at the
interface between the receptor and the autoantibody, and also in near the
future, those which target the binding site of MG antibodies. This project is a
collaborative work with Dr. Rorberts’ group at USC and is in progress. It has
already been achieved to develop a binder protein which targets the specific
subunit. Structural information would enable us to design a variety of binder
proteins (e.g. weaker binders or stronger binders), which could be used as
therapeutic molecules. Then, the topic is moved onto other nAChR subunits.
The 5th project is the structural determination of a nAChR α9 subunit. The
focus of this project is a sugar chain on the α9, which is highly conserved on
the cys-loop of nAChR α subunit and might have important roles in the gating
mechanism of nAChRs. The X-ray crystal structure of nAChR α9 bound by α-
bungarotoxin was solved at 3.0 Å. Our structure provides more insights into
the possible role of the sugar chain. Lastly, this dissertation ends with
xxvi

purification of another nAChR subunit, α7. Due to the difficulty in expressing
of the wild type α7 subunit, a chimera of α7/AChBP was produced and
successfully purified using purification techniques and knowledge from the
nAChR α1 study. The structure determination of the α7 bound with
therapeutic compounds would lead for further drug development.
In summary, our structural studies on nAChRs related to MG revealed
the detailed information of the binding interface between the receptor and
autoantibodies, resulting in further understanding of the disease. In addition,
the new diagnostic method has been developed as well as possible therapeutic
protein molecules for MG. The X-ray crystal structure of another nAChR
subunit, α9 was successfully solved with a clear picture of a sugar chain,
which might play an important role in functions of nAChRs. Finally, a robust
purification system for the nAChR subunit α7 was developed and optimized
for additional structural studies in order to develop therapeutic drugs.

1

Chapter 1

Introduction

1.1 Nicotinic acetylcholine receptors
Nicotinic acetylcholine receptors (nAChRs) are one of the ligand-
gated ion channels (LGICs), which are activated by the binding of
neurotransmitters such as acetylcholine (ACh). Ion channels were evolved to
transport ions across the cell membrane, and opening of the LGICs is
controlled by the ligand binding whereas other gated ion channels open with
electrical, chemical or mechanical signals
1
. The LGIC superfamily includes
anion selective channels: glycine and γ-aminobutyric acid (GABA) receptors
and cation selective channel: the 5-HT
3
serotonin receptors
2
. nAChRs are the
prototype members of the LGICs superfamily and are known as non-selective
cation channels allowing cations such as Na
+
, K
+
and Ca
2+
to cross the
membrane by activation. The nAChRs were the first receptors to be
recognized and named in the neuromuscular junction as the ‘nicotinic
receptive substance’ in 1907
3
. ACh was identified as the very first
neurotransmitter in 1914
4
. nAChRs have very important function and roles in
both the central and peripheral nervous systems, and have been extensively
2

studied since they are pharmaceutical targets for many human diseases and
substance addiction
5
. Myasthenia gravis, epilepsy, schizophrenia, and
depression are well known nAChRs-related diseases, and recent discoveries
show that nAChRs are correlated to some cancers
6
. For instance, α7 nAChRs
are related to lung cancers
7, 8
and overexpression of α9 nAChRs were
observed in breast cancer
9, 10
. As an example of a nAChRs function, the ACh
activates muscles by binding to the nAChRs on skeletal muscle fibers, which
allows sodium ions to go through the membrane, resulting in muscle
contraction. The motions of our bodies are resulted from the activities of
nAChRs.
To date, seventeen different subunits have been identified in nAChRs
(α1 – α10, β1 – β4, γ, δ, and ε), and they share a similar topology. There are
two types of nAChRs: the muscle-type and the neuronal type
11
. The α1
nAChR is known as the muscle-type nAChR and controls muscle contraction.
Most nAChRs, however, are found in the central and peripheral nervous
systems. Those neuronal type nAChRs are closely related to diseases such as
Alzheimer’s disease and schizophrenia
1, 2, 12, 13
. Recent studies have also
shown that some neuronal nAChRs can also be found in non-neuronal
tissues
14
. For example, α3 nAChRs and α7 nAChRs are found in skin
keratinocytes and lung cells
2
. α9 nAChRs is also known to have an important
role in epithelialization
15
.
1.1.1 Stoichiometries of nAChRs
nAChRs are composed of five subunits to work as a functional
receptor; some form a homopentamer with only α subunits (α7, α8 and α9),
whereas the others form a heteropentamer that may include different α
subunits and non-α subunits. Stoichiometries vary, but at least two α-subunits
are contained in each receptor to form the pentamer structure (Figure 1.1). In
3

the fetal form of muscle-type nAChRs, a pentamer receptor is formed by four
different subunits (α, β, γ, and δ), and the γ subunit is replaced by the ε
subunit in the adult form. Heteropentamer receptors can be further divided
into two classes: heterodimeric αβ nAChRs and complex nAChRs.
Heterodimeric αβ nAChRs are composed of one type α subunit (α2, α3 or α4)
and one type β subunit (β2 or β4). On the other hand, complex nAChRs
contain more than one type of α or β subunit such as (α3β4)
2
α5. Neither the α5
subunit nor the β3 subunit can form receptors without at least two other
subunits (e.g. α3/β4 for α5 or α6/β2 for β3)
16-19
. The α7 subunit was known as
a homopentamer receptor, but recent studies show that it could form a
heteropentamer with the β2 subunit
20-23
. The α8 subunit has been found in
chickens, and α9 subunits are expressed better together with α10 subunit
12
.
There are various stoichiometries of the nAChRs; however, the number of
different combinations is unknown. Even though each subunit has high
homologies, functions of nAChRs are different depending on subunits or
stoichiometries. It is important to develop therapeutic molecules which target
specific subunit; therefore, functions of other nAChRs would not be affected.
Localization and function of nAChRs are affected by the composition
of subunits
13, 24
. Homopentamers, for example, are known to bind to α-
bungarotoxin (krait snake toxin), whereas heteropentamers, with the exception
of muscle-type nAChRs do not
25
. The α-bungarotoxin from Bungarus
multicinctus is the first toxin used to characterize nAChRs and demonstrated
the blockage of Torpedo
26
and Electrophorus
27
nAChRs. Subunits which have
a loop C are called α subunits
28
, and the remaining given Greek letters based
on their molecular weights as resolved on polyacrylamide gels. A loop C has
adjuscent Cys residues to form an unusual disulfide bond
1
. This loop
contributes to ligand binding by forming acetylcholine binding sites with other
loops.
4

Figure 1.1: Subtypes and major stoichiometries of nAChRs
This figure was adapted from Lindstrom, L. (2010) eLS.
12
. Seventeen subunits have
been identified and they form either a homopentamer or a heteropentamer. The exact
number of stoichiometry is not known. Important arrangements of nAChRs are
shown.α9 subunit could form functional homopentamer receptors, but expression
level is higher with α10 subunit.
5

1.1.2 Structures of nAChRs
nAChRs consist of three different domains: the extracellular domain
(ECD), the transmembrane domain which has four transmembrane helices
(TM1 – 4), and a small intracellular domain
29
. The large ECD, which consists
of about 210 – 220 amino acids, is located at the N-terminus of nAChR, and is
important for ligand binding, with subunits being homologous. nAChRs are
also known as members of the Cys-loop LGICs, which have the Cys-loop
located at the extracellular N-terminal domain. The Cys-loop has the signature
sequence of thirteen highly conserved amino acid residues which have a
cysteine at both sides of those residues, forming a disulfide bond. Figure 1.2
shows a unique Cys-loop is located in all subunits at a similar position (C128
– C142 for α1 subunit). It also shows that all α subunits contain adjacent
cysteine residues, which are at C192 and C193 in loop C of α1 subunit. In
addition, the N-glycosylation site on the Cys-loop (141 in α1 subunit) is
highly conserved except for α7, α8, and α9. The glycosylation has been shown
to be important for toxin binding or the gating mechanism
30, 31
. The sugar
chain of α9 is at the site adjacent to the Cys-loop, and it still could interact
similary with the receptor to mediate the communication between loop C and
the Cys-loop. Those features imply the importance of post-translational
modifications for the structure and function of the nAChRs. The
transmembrane domains (M1 – M4) consist of four short α helices of
hydrophobic segments (15 – 20 amino acids each) with small linking regions
of M1 – M2 and M2 – M3. These transmembrane segments penetrate the cell
membranes in an alternating fashion to form a channel of the nAChR. The M2
segment together with a third portion of the M1 segment is known to be a
lining of the channel. The segment between M3 and M4 is large (100 – 150
amino acids), and forms the intracellular domain. This region is flexible;
6

therefore, detailed crystal structure has not been obtained
12
. Finally, the C-
terminus segment (4 – 28 amino acids) ends at the extracellular domain.
1.1.3 Ligand binding site
The ligand binding site can be found at interface of an α-subunit and a
neighboring subunit
32
. In muscle-type nAChRs, the binding sites are located
between the α1 subunit and an adjacent γ/ε or δ subunit, which results in non-
equivalent ligand-binding sites
33-35
. On the other hand, neuronal nAChRs have
binding sites between an α subunit (α2, α3, α4, or α6) and a β subunit (β2 or
β4). Most nAChRs carry two binding sites, but homopentameric nAChRs
contain five identical binding sites though the binding of two ACh molecules
is sufficient to activate channel opening
36-38
. Among the seventeen subunits,
the subunit, which carries the primary agonist binding site, has two adjacent
cysteine residues (Cys192 and Cys193 in α1 subunit), which are unique
among the subunits. Those residues are believed to be important for agonist
binding, and two α subunits are included in all heteropentameric nAChRs.
Three subunits, β1, β3, and α5 are known as “accessory subunits” which do
not participate in the formation of ligand-binding sites. However, they still
contribute to conformational changes, localization, activation, and other
properties of the receptors
12, 39, 40
. In addition, the α5 and the α10 subunits
cannot form functional nAChRs unless they co-assemble with other α subunits
despite being α subunits. It has been shown that the α5 subunit is most closely
related to the β3 subunit based on sequence similarity compared to the other
nAChR subunits
41
. One important residue in α1 subunit which is believed to
be important for the agonist binding is Tyr 190, and the α5 subunit is missing
this residue though the α10 subunit retains it.
7

Figure 1.2: Schematic of subunit comparison.
This figure was adapted from Lindstrom, L. (2000) Muscle & Nerve
42
. Each subunit
was shown as a primary structure. Transmembrane domains are shown as orange and
labeled M1 – M4. General locations of disulfide-linked loops and glycosylation sites
can be compared between different subunits.

8

The binding site consists of several loops from both an α subunit and a
neighboring subunit. Loops A
43
, B
44
and C
44-46
from the the α subunit
contribute the principal binding site, whereas Loops D
47
, E
47-49
and F
50
from
the neighboring subunit contribute the rest (Figure 1.3). In homopentamer
nAChRs, α subunit provides the equivalent loops to form the binding site. The
opening of gates is initiated by ligands binding to the receptors. It triggers a
closure of Loop C (also called C-loop) over the binding site, and the electrical
signaling is produced by ions flowing through the cell membrane.

Figure 1.3: Schematic representation of
the ACh binding site of α7 nAChR.
This figure was adapted from Taly, L.
(2009) Nature reviews
51
. The ACh binding
site is illustrated. Three loops from α
subunit (Loop A, B and C) contribute to
the primary binding site, and the other
three loops (Loop D, E, and F) from the
adjacent subunit (γ/ε, δ, or β in
heteropentamer and α in homopentamer)
contribute to the complementary binding
site.

1.2 Early structural studies of nAChRs
It has been more than a century since nAChRs were discovered and
have been studied by many scientists in the decades since. It was not easy to
study nAChRs, which are the membrane proteins. However, several
breakthroughs were achieved from tremendous efforts particularly, structural
information of the receptors which were obtained in recent years and provided
significant insights to understand the nAChRs further.
9

1.2.1 Acetylcholine Binding Protein
The X-ray crystal structures of acetylcholine binding protein (AChBP)
were one of major breakthroughs in studies of nAChRs. The structures were
solved with agonist and antagonist by several groups
38, 52-58
. The first crystal
structure of AChBP was solved at 2.7 Å resolution in 2001
38
, and an even
higher resolution structure at 1.75 Å is now available, providing detailed
information at the atomic level
59
.
Acetylcholine binding protein (AChBP), which is found in the snail
Lymnaea stagnalis
60
, has been intensively studied as a model of nAChRs
because of its unique characteristics
18, 61-64
. It aligns well with all LGICs, and
the sequence identities are 15% – 20% with GABA
A
, glycine, and 5HT
3

serotonin receptors. It most resembles the α subunits of nAChRs, with which
it shares 20% – 25% sequence identity
65
. Even though sequence similarity is
not particularly high, important residues, such as those in the ligand binding
site, are well conserved
65
. The AChBP, which consists of 210 amino acids,
assembles into a homopentamer and is a homolog of the extracellular domain
of homopentameric receptors such as the neuronal α7 nAChR
66
. Since it does
not have the transmembrane region, it is a soluble protein and easy to obtain
for structural studies. It is also known to bind most nAChR agonists and
competitive antagonists such as ACh, nicotine, d-tubocurarine and α-
bungarotoxin though it binds nicotine with ten-fold higher affinity than ACh
60,
67
. It functioned as a receptor with ACh binding when it was linked with the
pore domain of the 5-HT
3A
receptor
68
. The AChBP is a great model to study
nAChRs due to those characteristics; however, there is a difference in the
Cys-loop. The Cys-loop is a characteristic feature of nAChRs, and is defined
by a conserved sequence of 13 amino acid residues between two cysteines.
However, AChBP has only 12 residues, and there is not sequence
10

conservation with LGICs. Although there was a slight difference, the crystal
structure of AChBP revealed detailed structural information of the
extracellular domain, especially the ligand-binding sites.
1.2.2 Muscle-type nAChRs
The electron microscopy structure of the Torpedo nAChR was solved
at 4 Å, and structural information of not only the extracellular domain but also
the transmembrane domain became available
69, 70
. Studies of nAChRs in
vertebrate skeletal muscle have been conducted for more than one hundred
years. In 1907, Lengley noted a “receptive substance”, and tonic contraction
of certain muscles in fowls, frogs and toads with treatment of nicotine was
observed. Upon the addition of curare, which is known to block nAChRs, no
muscle contraction is observed, proving the receptor is important in that
function
3
. Upon studies of the receptors, the electric organs of the electric fish
such as the electric ray Torpedo californica and eel Electrophorus electricus
were very useful since those tissues were nAChR rich. In particular, Torpedo
has been heavily studied as a model of muscle nAChRs because of a high
degree of homology and the same stoichiometry as muscle nAChRs
71
. Several
three-dimensional structures were reported
37, 72, 73
, but one major breakthrough
in this field was the high resolution structure of the Torpedo electric organ
nAChR at 4 Å using a cryo-electron microscope
69, 70
. The structures of
Torpedo nAChRs provided detailed information on the extracellular domain
and also the transmembrane domain in both the open and closed states of the
receptors. Figure 1.4 shows a pentameric structure of full-length Torpedo
nAChR. The main immunogenic region (MIR), which has a high correlation
with a disease called myasthenia gravis (discussed in Chapter 2 and Chapter
3), was first determined in this structure as well as ACh binding sites. The
11

dimensions of the ECD and AChBP were consistent with ~ 60 Å in height, ~
80 Å diameter, and ~ 20 Å diameter channel
38, 69, 73
.
Figure 1.4: Electron microscopy structure
of the Torpedo nAChR.
Adapted from Unwin (2005) J. Mol. Biol.
70
.
Structures are shown in ribbon diagrams. (A)
The whole receptor viewed from the synaptic
cleft. (B) Viewed parallel with the membrane
plane. (C) A single subunit is shown with
detailed labels of structural details. The MIR
is located at the top of the receptor adjacent to
the N-terminal α-helix of the receptor. ACh
binding sites can be found at the interfaces of
α subunit and γ/δ. Trp149 of the B loop is a
conserved residue in nAChR α subunits, and is
believed to be important as well as residues in
the C loop.
A B
C
12

Another breakthrough was the first atomic resolution view of a
nAChR. The crystal structure of the extracellular domain of the mouse
nAChRα1 subunit bound to α-bungarotoxin was solved at 1.94 Å resolution
30
.
Although it was only the ECD of the α1 subunit, this atomic level structure
enabled the study in more detail of a number of receptor-specific elements
such as the MIR, the N-linked glycosylation and the Cys-loop. A sugar chain
has been shown to be important for toxin binding as well as for receptor
function
30, 31, 74, 75
. This structure also confirmed that the MIR was located in
an exposed area of the receptor; therefore, it is highly accessible to
monoclonal antibodies causing autoimmune disease. It was also shown that
the MIR interacts with the N-terminal α-helix and the β5 – β6 turn of the
receptor. Those details contributed to further studies of the disease.
1.2.3 Prokaryotic LGICs
X-ray crystal structures of two prokaryotic LGICs were solved: the
bacterium Erwinia chrysanthemi and the bacterium Gleobacter violaceus each
at ~ 3Å resolution. Even though those are prokaryotic receptors, they have
considerable homology with eukaryotic LGICs (16% sequence identity with
the nAChR α1 subunit), and they were solved as full-length proteins providing
more detailed information of the transmembrane domain
76-78
. The major
difference from the nAChRs was the lack of the N-terminal α-helix, the Cys-
loop, and the cytoplasmic loop. Those missing pieces of information would
limit the study of nAChRs; however, the first high resolution of a pentameric
receptor in full-length was hoped to be an important model of LGICs to learn
gating mechanisms within the family.

13

1.2.4 GABA
A
and serotonin 5-HT
3
receptors
Some Cys-loop receptors have been solved as pentamer: the
glutamate-gated chloride channel (GluCl)
79
, the type-A γ-aminobutyric
receptor (GABA
A
) receptor
31
, and the serotonin-gated 5-HT
3
receptor
80
. Those
studies revealed detailed structural information of transmembrain domains and
channel gating mechanisms. Two structures of eukaryotic ligan-gated ion
channels of the Cys-loop receptor, GABA
A
and serotonin 5-HT
3
receptors are
shown in Figure 1.5. The structure of the GABA
A
receptor receptor contained
sugar chains and it showed the importance of those sugar chains for gating
mechanisms which was also discussed in a previous study. The serotonin 5-
HT
3
receptor was crystallized with nanobodies, and it suggested that they
could be used as diagnostic and therapeutic tools in future.

Figure 1.5: X-ray crystal structures of GABA
A
and serotonin 5-HT
3
receptors.
The left figure is a structure of the human GABA
A
receptor which is created using
PDB: 4COF. It contains sugar chains. The right figure is a structure of the mouse
serotonin 5-HT
3
receotor stabilized with nanobodies. It is created using PDB: 4PIR.
Both receptors are solved as full-length receptors.
14

1.2.5 Neuronal nAChRs
In 1986, the first neuronal nAChR subunit, α3, was cloned
81
, eleven
nAChR subunits, α2 – α7, α9, α10, β2 – β4, in mammals
82, 83
and one, α8, in
avian species
84
have been identified as the neuronal nAChRs since then.
The first major discovery in neuronal nAChRs was the X-ray crystal
structure of chimeric human nAChR α7 subunit with Lymnaea stagnalis
AChBP
85
. It was again only the extracellular domain of the receptor, but it
was purified and crystallized as a pentamer. The chimera shared 64%
sequence identity and 71% sequence similarity with the human α7 subunit.
Crystals of the receptor bound by agonists and antagonists were also solved,
and those crystals would be used for structure guided drug design.
Lastly, structures of the human α9 subunit with several agonists and
antagonists were also solved at high resolution
86
. The α9 subunit (ECD) is
known to be soluble in native form, and it was the first neuronal nAChR
structure of the wild type receptor. If more structures of receptors would be
available, it would be possible to learn the functionality and characteristics of
each subunit. Also, it would be possible to design drugs specific for each
subunit/nAChR.
1.3 This work
Understanding of the nAChR would lead to treatments or cures of
receptor-related diseases. Among them, myasthenia gravis (MG) is a well-
known autoimmune disease in which the nAChR α1 subunit is a target of
autoantibodies. The crystal structure of Torpedo nAChR and mouse nAChR
α1 subunit revealed the structure and location of the main immunogenic
region which were exposed to the outside of the receptor where autoantibodies
have easy access. Even though the structural details are available, it has not
15

known how autoantibodies actually bind to the receptor. The complex
structure of the nAChR α1 subunit and MG autoantibodies would provide us
with tremendous insights to understand the pathogenicity of MG, and could
also lead to finding a treatment for the disease. In Chapter 2, the first X-ray
crystal structure of the nAChR α1 subunit complexed with Fab35 (the Fab
portion of mAb35) is solved. mAb35 is one of the monoclonal antibodies
(mAbs) known to target the MIR to cause MG. It is a prototypical
autoantibody for the disease, and has been intensively studied. The structural
details of the nAChR α1 subunit/Fab35 complex are revealed, and it is shown,
for the first time, how the autoantibody recognizes the receptor to trigger MG.
A next question raised is if all autoantibodies bind to the receptor in
the same way as mAb35. Many biochemical studies have shown that binding
characteristics vary among mAbs though they target the same region of the
receptor, the MIR. For example, mAb35 is known to bind the receptor in a
highly conformational-dependent manner. To understand the difference,
another structure of nAChR α1 subunit complex with Fab210 is solved.
mAb210 shows different binding properties from the receptor when compared
to mAb35. mAb210 can bind a denatured nAChR α1 subunit or peptides of
the receptor while mAb35 cannot. Both mAbs are known to cross-react with
other subunits (α3, α5, and β3), but only mAb210 can be used in laboratory
settings to detect such subunits on Western blots
87-90
. In Chapter 3, the
differences between those two mAbs are discussed using newly solved
structures of complexes of Fab210.
Next our challenge is to develop a new detection method to diagnose
MG using molecular techniques learned during crystallization studies. One of
major difficulties in receptor studies is to obtain soluble receptor proteins in
large quantity. Even though there has been some success with the purification
of Torpedo nAChR and AChBP, they are still different from the human α1
16

subunit. Another issue in the current detection method is the use of radioactive
material which is not only expensive but also requires higher safety precaution.
In Chapter 4, a novel detection method for MG using a well-known molecular
biology technique, ELISA, is discussed.
In Chapter 5, the development of a possible treatment for MG with
collaboration with Dr. Richard Roberts (USC, Los Angeles, CA) and his
group is initiated. Dr. Roberts invented a novel technique called mRNA
display
91
. This technique enables the creation of a binder protein which targets
the protein of interest. It also allows us to select a binder which would bind a
specific area of the protein. That is, it is possible to create a binder protein
which could compete with MG autoantibodies. A challenge in this study is the
high homology of each subunit of nAChRs. First our aim is to develop
subunit-specific ligands, and then to perform structural studies for further
understanding to design better binders. As a future plan, this study hopes to
contribute to a new treatment of MG.
The focus of Chapters 2 – 5 is nAChR α1 subunit and an autoimmune
disease, MG. In Chapter 6, the knowledge and skills learned from the α1
subunit are utilized to crystallize the nAChR α9 subunit which is known to be
soluble as a wild type protein. A structure of the human α9 subunit was solved
in 2014 without a glycan
86
. Our belief is that glycosylation has an important
role in the receptor function, which was also discussed in previous studies
30, 31,
74, 75
. The structure of neuronal-type nAChR focusing the glycosylation and
comparison with other known structures of nAChRs is discussed.
Lastly, our study is expanded to the nAChR α7 subunit. Previously,
the structure of the chimeric human nAChR α7 subunit with Lymnaea
stagnalis AChBP was solved by our lab collaborating with Dr. Steven Sine
(Mayo Clinic, Rochester, MN)
85
. Structures with some agonists and
17

antagonists were also solved. Our next focus is to obtain structural
information of therapeutic drugs bound to the α7 subunit including
compounds which are currently in clinical trials. A protein expression system
had not previously been worked out in our lab; therefore, the protein
purification of α7/AChBP chimera in Chapter 7 is presented, which will lead
to crystallographic studies later.

18

Chapter 2

Structural Studies of nAChR α1
Subunit and Fab35 Complex to
Understand Autoimmune Disease of
Myasthenia Gravis

2.1 Abstract
Myasthenia gravis (MG) is an autoimmune neuromuscular disease that
causes severe muscle weakness and fatigue which can be fatal if left
untreated. Currently, there is neither a cure nor an effective treatment for this
disease. It has been known that a nicotinic acetylcholine receptor (nAChR) α1
subunit is strongly related to MG. This chapter discusses the crystal structure
of the nAChR α211 bound by the Fab fragment of mAb35 (Fab35) which was
solved at 2.6 Å. This structure reveals, for the first time, the detailed
interactions between the nAChR α1 subunit and the Fab35, providing a
framework for further studies of the disease mechanisms of MG and
therapeutic development.
19

2.2 Introduction
Myasthenia gravis (MG) is an autoimmune neuromuscular disease,
with 200 per million people are affected in the United States
92
. The actual
number may be much higher since the diagnosis of MG remains difficult. The
name, myasthenia gravis, is based on three different words; in Greek, mayo
and asthenia, meaning “muscle” and “weakness” respectively, and in Latin,
gravis means “serious”
93
. As the name describes, the neuromuscular junction
in MG patients is affected causing severe muscular weakness. The major
cause of MG is the binding of autoantibodies to nicotinic acetylcholine
receptors (nAChRs) in muscles, as they recognize the receptors as foreign
antigens, thus impairing the motor function of the patient. The first discovery
of the autoimmune character of MG was a demonstration of an experimental
autoimmune myasthenia gravis (EAMG). After immunization of the nAChRs
purified from Electrophorus electric organ to rabbits, they showed muscle
weakness and eventually died
94
. Autoantibodies could directly block the
function of nAChRs,
95-98
or cross-link the receptors resulting in internalization
and degradation
99-102
. They also could destruct the postsynaptic membrane
curvature which contains the nAChRs
103
.
There are two main types: ocular MG and generalized MG. In ocular
MG, ocular muscles are often affected, causing drooping eyelids and double
vision. On the other hand, many patients with generalized MG suffer from
fatigue and muscular weakness. Because of the unique characteristics of the
symptom, which often are misunderstood as the laziness, it is hard to
definitively diagnose MG early. If not properly treated, the respiratory system
is eventually affected, leading to death. Women are predominantly affected by
early-onset MG (before the age of 50), while very late-onset MG (after the age
of 60) is higher in males
104, 105
. Juvenile MG in Europe and North America is
20

low, accounting for 10% to 15% of cases; however, it is more common in
Asia. It has been reported that 50% of MG patients with ocular symptoms in
China are below 15 years of age
106
.
Currently, there is no cure for MG. However, there are several
effective but very invasive treatments. Thymoma, a type of tumor originating
from epithelial cells in the thymus, is often found in MG patients, and the
symptoms could be dramatically improved by surgical removal. Moreover,
acetylcholinesterase inhibitors, corticosteroids, and immunosuppressant
medications are commonly used to control the symptoms
107, 108
. MG patients
tend to have a decreased number of nAChRs, impairing a signal transmission.
Thus, inhibiting the activity of acetylcholinesterases, which break down
neurotransmitters, acetylcholines (AChs), and modulate neurotransmission,
more AChs become available for the activation of nAChRs to compensate the
low level of nAChRs. Furthermore, both corticosteroids and
immunosuppressants are used to treat many autoimmune diseases to inhibit
the immune system. This approach could result in some side effects including
nausea and vomiting as well as an increased risk of infection and liver damage.
As for MG therapies, plasmapheresis and intravenous immunoglobulin (IVIg)
are commonly performed. Normal antibodies from health people are injected
into MG patients to reset their immune system. Even though it has a lower
risk of side effects, however this treatmen would take a longer time (about a
week) to see an effect and it would not last long (3 – 6 weeks). Another
therapy, plasmapheresis works similar to dialysis. A patient might require
implanting a catheter into his/her chest for this procedure and autoantibodies
causing MG are removed from patient’s blood going through a machine as
well as other necessary antibodies. Therefore, important antibodies need to be
replaced before blood is returned back to a patient’s body. It has higher risks
such as heart rhythm problems, but it has shown a rapid effect
109
.
21

Immunotherapy has been studied as a new approach for MG treatment, which
showed antigen-specific immunosuppression of MG by immunization of
purified nAChRs
110-113
.
Among several possible targets for autoantibodies, nAChR α1 subunit
is known to be a major target for autoimmune response. Approximately 80% –
90% of MG patients carry anti-nAChR antibodies, but the rest of are
seronegative for nAChR-antibodies. Autoantibodies targeting muscle specific
kinase (MuSK) have been identified in 5% - 10% of MG patients
114-116
.
Recent studies have also shown that a small population of MG patients have
autoantibodies against low-density lipoprotein receptor-related protein 4
(LRP4)
117, 118
. Even though there are several targets for autoantibodies which
cause MG, it could be still possible to treat many MG patients just by
blocking the binding of the anti-nAChR antibodies to the nAChR α1 subunit.
Hundreds of monoclonal antibodies against the nAChR were
previously created and extensively studied
119-121
. It has been identified that
autoantibodies bind to a specific region of nAChR α1 subunit called the main
immunogenic region (MIR). It is located at the extracellular domain (ECD) of
the α1 subunit and strongly correlates with the disease severity
122-126
. Many
studies were performed to localize the MIR in 1980s using synthesized
peptides: 46 – 120 of Torpedo
127
, 6 – 85 and 37 – 85 of mouse
128
, 61 – 76 of
Torpedo
129
, and 67 – 76 of both human and Torpedo α-subunits
130, 131
, and
α67 – 76 is currently the most accepted region as the MIR. More recently, an
importance of N-terminal α-helix was shown for binding of autoantibodies
132
,
and three regions within the nAChR α1 subunit were also identified as
important recognition sites for autoantibodies: α(1 – 12), α(65 – 79), and
α(110 – 115)
133
. Some mutational studies were also performed to identify
especially important residues among the MIR for autoantibodies bindings
131,
22

134-137
. In addition, the location of the MIR was identified on the ECD of
Torpedo nAChR α subunit by electron microscopy studies
138
. Continuous
efforts of structural studies revealed more details of interactions of
autoantibodies and nAChRs. A structure of cryo-electron microscopy showed
scFv35 or Fab35 binds to the top and periphery of the nAChR
139
. More than
50 % of autoantibodies are known to target the MIR
42, 140, 141
and higher levels
of autoantibodies are detected in severe MG patients. However, very little is
known about how autoantibodies actually bind to the nAChR α1 subunit at the
molecular level. Crystal structures of the complex will provide detailed
information of the interface between nAChR α1 subunit and autoantibodies,
which could lead the way for structure-guided drug design for MG. We used
X-ray crystallography to determine the structure of the antibody/nAChR α1
subunit complex.
One challenge in structural studies of nAChRs is to obtain large
amounts of the receptor proteins which contain the hydrophobic region of the
transmembrane domain. Moreover, membrane proteins such as nAChRs
contain glycosylation needed for proper trafficking and folding. Therefore, a
prokaryotic expression system is not well-suited even though it is often used
to produce large quantities of cytosolic proteins for structure studies. In a
previous study, an engineered nAChR α1 subunit was successfully stabilized
and purified in yeast. It was truncated at the 211
th
amino acid from the N-
terminal and 3 mutations (Val8Glu, Trp49Arg, and Val155Ala) were
introduced to stabilize the protein and increase solubility, and the engineered
protein was designated as α211
142, 143
. In 2007, the first atomic-resolution
structure of the extracellular domain of the nAChR subunit α211 was solved
30
.
The purification protocol was optimized in the previous study
144
; however,
there was difficulty in obtaining protein for our experiments. Therefore, a
23

novel method was developed for our experiments and it describes here to
obtain 0.5 mg/1L culture of the recombinant nAChR α211 in yeast.
As for a MG autoantibody, mAb35 was chosen to study the binding
interface with both human and mouse nAChR α1 subunit (α211). It was raised
against Electrophorus electricus in rats, and mAb35 is a prototypical mAb
known to recognize nAChR α1 subunit as an antigen and trigger MG in
experimental animal models
120, 141, 145
. One of the interesting features of MG
mAbs is that the binding affinity for a denatured form of the α1 subunit is
much lower than for the native form
119, 120, 146, 147
. Among them, mAb35 is a
well-known autoantibody which binds to the receptor in a conformational
dependent manner. It cannot bind to denatured α1 subunit or synthetic
peptides
131, 135, 148
. Studies also have shown that the binding of some MG
mAbs such as mAb35 did not show a significant difference in channel
activation
149
nor α-bungarotoxin binding
150
. X-ray crystal structures of two
Fab fragments of MG mAbs were solved (Fab192 and Fab198)
151, 152
. Even
though the detailed structural information of binding sites of mAbs was
revealed, it was still unclear how exactly mAbs interact with the nAChR α1
subunit. Therefore, the direct X-ray crystal structure of the complex of
nAChR bound by an autoantibody is needed to provide clear information to
understand MG better. However, one question which must be asked is how
relevant the information obtained from the use of rat MG antibodies is to
human MG. Several studies have shown cross-reactivity of autoantibodies
with nAChRs from various species
94, 119, 120, 141, 153-156
. Anti-nAChR antibodies
which are raised from one species can bind to nAChRs of different species
due to the highly conserved sequence of the MIR
130, 134, 137
. Symptoms of
myasthenia gravis can be seen in immunized animals with purified nAChRs
from difference species. It is also seen that passive transferring of EAMG to
animals by injection of serum from MG patients or EAMG animals
89, 154, 157
.
24

Therefore, our structure will be useful to study the interface of the nAChR and
the autoantibody even though the autoantibody is not from human MG
patients.
2.3 Results
2.3.1 α211 purification
Both mouse and human α211 proteins were expressed in P. pastoris
and purified (Refer to Materials and Methods for details). The expression
level was extremely low following the typical yeast expression protocol. To
optimize the expression, several factors were tested including OD level,
induction time, pH of media, methanol concentration, and temperature.
Obvious differences were not observed with most variables, but temperature
clearly affected expression. Yeast is usually cultured at 30 °C including the
protein expression period. By lowering the temperature to under 20 °C,
expression level dramatically increased (≥5-fold). There was not a large
difference between 15 °C and 20 °C. Since 30 °C is a favorable temperature
for yeast, yeast was inoculated at 30 °C to obtain enough biomass, and then
induced at 20 °C. After Ni-NTA purification, the protein was subjected to
Superdex 75 10/300. Proteins were relatively pure after Ni-NTA purification.
Two peaks were observed on the chromatogram of Superdex 75 10/300
(Figure 2.1). The first peak was within the void volume and was likely
aggregates of α211 with some contaminations of proteins. The second peak
fractions were pooled and used for the complex formation with Fab and α-
bungarotoxin. Protein expression level in the yeast expression system tend to
fluctuate widely by environmental factors compared to other expression
systems such as E.coli., but a minimum of ~ 0.5 mg of protein was obtained
from the 1L starting culture.
25

Figure 2.1: Mouse α211 purification.
(A) Chromatogram of mouse α211 purification with Superdex 75 10/300. The second
peak indicates α211 protein. The first peak is the aggregation of the protein and some
other protein contaminants. A similar chromatogram was obtained for human α211.
(B) 15% SDS-PAGE gels of α211 purification (16 μL of sample + 4 μL of 5x loading
dye except Ni beads: 2 μL of beads + 6 μL of water + 2 μL of 5x loading dye and
Injection: 1 μL of sample + 7 μL of water + 2 μL of 5x loading dye). A faint protein
band was observed in a media sample. This purification was the minimum expression
level, and twice more expression was achieved. Ni elution sample was already at
crystallography level of purity. Fraction 17 shows some contaminants proteins
including aggregated mouse α211. Fraction samples of the main peak were clean.
Fractions 22 – 27 were pooled and concentrated down for future experiments.

A
B
26

2.3.2 mAb35 purification
Hybridoma cells of mAb35 were cultured for 7 to 10 days, and the
supernatant harvested for mAb35 purification. Many antibody purifications
are usually performed using Protein A beads or sometimes Protein G beads.
However, the mAb35 antibody is a rat IgG
1
, and the binding affinity to both
Protein A and Protein G beads were none or very limited. The harvested
media, containing secreted mAb35, was incubated with Protein G beads which
have some binding affinity. Due to the binding limitation, large amounts of
Protein G beads were used (approximately 10 mL culture/1 mL Protein G
beads). Despite the high amount of beads used, much of the mAb35 was lost
in the initial flow though. Therefore, the second and the third round of
purification using the flow through were necessary to recover as much mAb35
as possible. After purifying with Protein G beads, the eluted sample was run
over Superdex 200 10/300 (Refer to Materials and Methods for details). The
chromatogram showed a single peak with no indication of the formation of
aggregates (Figure 2.2). An SDS-PAGE gel showed reasonably clean bands of
mAb35 though there were some contaminants of a high molecular weight.
Since further purification was required before crystallization, the impurity was
not a concern at this step. Fractions of the peak were collected and proceeded
to Fab digestion and purification. Although the binding affinity to the beads
was weak, ~ 8 mg of mAb35 were obtained from 100 mL of media culture
with several rounds of Protein G beads purification from the same cultured
media.

27

Figure 2.2: Purification of mAb35 from hybridoma cells.
(A) Chromatogram of mAb35 purification with Superdex 200 10/300. After Protein
G beads purification, the elution sample was subjected to Superdex 200 10/300
column. There was no obvious void volume peak, and only a single peak of mAb35
was obtained. (B) 15% SDS-PAGE gel of mAb35 purification. Higher bands indicate
the heavy chain of mAb35, and lower bands indicate the light chain of mAb35.
Protein G beads elution with 1 mL fraction volume. After 8 fractions, mAb35 was
eluted with 5 mL elution buffer to collect all mAb35 (16 μL of sample + 4 μL of 5x
loading dye). Elution samples were pooled and concentrated to further purify with
Superdex 200 10/300. There were some bands indicating impurity with some higher
molecular weight proteins, which would not affect future experiments (16 μL of
sample + 4 μL of 5x loading dye except Injection: 1 μL of sample + 7 μL of water +
2 μL ofnn 5x loading dye). Fractions 23 – 28 were pooled and concentrated down for
the following Fab35 purification.
A
B
28

2.3.3 mAb35 digestion and Fab35 purification
It would be challenging to crystallize with an entire mAb35 due to the
size and flexibility of the protein. The interface of mAb35 and α211 is the
focus point in our study, and it is known that Fab portion of antibodies
(variable regions of both heavy and light chains) contains the antigen-binding
site. Therefore, mAb35 digestion was performed to obtain Fab35 for
crystallization.
The purified mAb35 was incubated with immobilized papain beads to
obtain the Fab portion. Digestion efficiency was low, and the protocol had to
be optimized. Papain digestion is usually performed at 37 °C for wide range
of incubation time, from 2 hours to 9 hours, depending on the type of species
and concentration of the antibodies. mAb35 required a longer incubation
period, and was incubated at 30 °C for ~ 20 hours. Even longer incubation
times were also tried, but the antibodies seemed to start being degraded. After
the enzyme treatment, the digested mAb35 sample was subjected to Mono Q
HR 5/5 column. In a previous study, Fab was isolated from Fc (constant
region of antibody) using DEAE-cellulose column which is an anion exchange
column and equivalent to Mono Q column
158

152
. Both mAb and Fc are often
neutral; however, Fab has basic pI
159
. As was expected, undigested mAb35
and Fc were bound to the column, and Fab came out in the flow through
(Figure 2.3). A Mono S HR 5/5 column, a cation exchange column, was also
tried to purify Fab as a bound peak, but it failed to separate Fab from the rest.
Therefore, a negative run of Mono Q was performed to purify Fab. Flow
through and peak 1 were combined and concentrated for complex purification
with α211 and α-bungarotoxin. The recovery rate of Fab was still low (~ 50%),
and there could be room to optimize it. However, sufficient Fab was obtained for the
study, so further optimization was not done.
29

Figure 2.3: Fab35 Purification.
(A) 15% SDS-PAGE gel of Fab digestion. Not all mAb35 was digested, but digestion
can clearly been shown. The sample was subjected onto Mono Q column. (B)
Chromatogram of Mono Q HR 5/5 of Fab35 purification. Peak 1: Fab in flow
through. Peak 2: Undigested mAb35. Peak 3: Fc. (C) 15% SDS-PAGE gel of Mono
Q fraction samples (8 μL of sample + 2 μL of 5x loading dye except Injection: 1 μL
of sample + 7 μL of water + 2 μL of 5x loading dye). Mono Q peaks were confirmed
with the gel.
A
B
C
1
2
3
30

2.3.4 Gel shift assay
Before moving to a complex crystallization, a gel shift assay was
performed to confirm the binding of Fab35 to α211. α-bungarotoxin was
included in the assay since it would help crystallization by stabilizing α211
30
.
The α-bungarotoxin binding site is far away from the MIR which mAb35
known to bind (Figure 2.4); however, it still needs to be confirmed that α-
bungarotoxin is not interfering mAb35 from binding to α211.

Figure 2.4: Binding site of α-bungarotoxin and expected binding site of Fab35.
A crystal structure of α211 (shown in green) bound by α-bungarotoxin (shown in
orange). The pink indicates a glycosylation on α211. PDB:2QC1. The position with
the yellow arrow indicates MIR which has been predicted as a binding site of
autoantibodes including mAb35.
MIR
α211
α-Btx
31

Mouse α211 and Fab35 were successfully purified as described above.
Each sample was mixed in equimolar ratios and incubated on ice for an hour
before loading onto a native-PAGE gel. Lane 1 shows α211 alone, and is then
compared to each of the mixture lanes (4 – 6). The native-PAGE gel showed
clear band shifts indicating Toxin binding to α211 (Lane 1 vs. Lane 4) and
Fab35 binding to α211 (Lane 1 vs. Lane 5). Also, Lane 6 (α211, Fab35 and
Toxin mixture) proved the ternary complex was formed containing all three
components: α211, mAb35, and α-bungarotoxin (Figure 2.5).

Figure 2.5: 6% Native-PAGE gel showing binding of Fab35 to mouse α211.
Both α-bungarotoxin and Fab35 are basic, so they migrated upward not showing any
bands in the gel (Lane 2 and 3). Binding of α –bungarotoxin and Fab35 individually
to α 211 were tested (Lane 4 and Lane 5). Also, the ternary complex formation was
also observed (Lane 6). Lane 6 proves that Fab35 binds to mouse α211 without being
interfered by α-bungarotoxin.

32

To show Fab35 binds specifically to α211, a similar gel shift assay was
performed with rat α9, which is another subunit of nAChR and belongs to a
neuronal type (Figure 2.6). Three mutations were introduced in the wild type
rat α9 for better toxin binding (See details in Chapter 6). Lane 7 shows
binding of Fab35 to α211 (Lane 1 vs. Lane 7), whereas Lane 8 shows that
Fab35 is not binding to α211 (Lane 2 vs. Lane 8). This experiment showed
that Fab35 binding to α211 is specific even though α211 and α9 have high
sequence homology (approximately 40% of identity and 60% of similarity).

Figure 2.6: 10% Native-PAGE gel showing specific binding of Fab35 to mouse
α211.
Binding specificity of Fab35 to mouse α211 was shown using another nAChR
subunit, rat Flag α9 3Mut. Lane 5 shows that α-bungarotoxin binding to rat Flag α9
3Mut though it is a slight shift and difficult to identify (Lane 2 vs. Lane 5). Lane 8
indicates no binding of α-bungarotoxin to rat Flag α9 3Mut. Similar results were also
obtained with human α211.

33

2.3.5 The α211/Fab35/α-bungarotoxin complex purification
The ternary complex formation was demonstrated by the gel shift
assay in the previous section, and finally, a large amount of the ternary
complex was purified. α211, α-bungarotoxin and Fab35 were mixed at a
1:1.5:1.5 molar ratio to form the complex, followed by Ni-NTA purification to
remove excess α-bungarotoxin and Fab35. Higher molar ratio of Fab35 and α-
bungarotoxin (e.g. 1:2:2) was ideal, but the ratio worked fine when there was
not enough purified Fab35. For further purification and buffer exchange, the
elution was run over Superdex 200 10/300. The ternary complex was stable
throughout the entire purification process. Two peaks were observed on the
Superdex 200 10/300 chromatogram (Figure 2.7). Since separation by
Superdex column is based on molecular size, the second, larger peak
corresponds to the monomer ternary complex even though on a denaturing
SDS-PAGE gel, both looked similar. The left peak could be the result of
oligomers (possibly a dimer) of the ternary complex. Since there is no clear
answer for the quality of left peak protein at this point, fractions from each
peak were pooled separately and concentrated for crystallization. Proteins
could crystallize as oligomers, so both samples were screened separately. The
oligomerization could result from oligomerization of Fab35. There is no clear
proof, but a peak indicating a oligomer or a dimer was observed in Fab35
Superdex chromatogram additionally to a monomer Fab35 peak. Even though
the oligomer peak was separated in the first Superdex purification, it was
observed again when the purified and stored Fab35 sample was run over
Superdex column. From these observations, it is possible that Fab35 is easy to
oligomerize or dimerize. A concern was that two peaks were merged together
too closely, and crystallization would be affected by contamination from each
other.
34

Figure 2.7: The mouse α211/Fab35/α-bungarotoxin complex purification.
(A) Chromatogram of Superdex 200 10/300 of the complex. Two peaks were
obtained, and each peak (Fractions 23 – 28 and Fractions 29 – 33) was separately
concentrated to set crystallization trays. (B) 15% SDS-PAGE gel of summary ternary
complex purification. Both left and right peaks show the right ratio of the ternary
complex formation. (C) Fab35 purification over Superdex 200 10/300 right after
Mono Q HR 5/5 purification. There is very small peak of oligomer Fab35 (left peak).
(D) Chromatogram of Superdex S200 10/300 of Fab35 (already purified once) after a
few weeks of storage at 4 °C. The oligomer formation was observed by the right peak.
A
B
C
D
35

2.3.6 Crystallization of the α211/Fab35/α-bungarotoxin ternary
complex
Crystallization conditions for the α211/Fab35/α-bungarotoxin ternary
complex were screened with 2.5 mg/mL concentration of the purified complex
protein by the hanging drop method at room temperature. Initials crystals of
mouse α211 ternary complex were small and grew as clusters (Figure 2.8. Left
drop). After optimization of crystallization conditions, rod-like shaped crystals
were grown as bundle with the size of 10 μm x 100 μm – 20 μm x 200 μm in a
crystal condition of 0.1M sodium cacodylate trihydrate pH 6.5, 0.1 M – 0.15
M calcium acetate hydrate and 18% – 20% PEG8K (Figure2.8, Middle drop).
Precipitations were observed right after setting drops. Some crystals were
formed about 3 days later with background precipitations, and continued to
grow for 7 days. Obvious crystal degradations were not observed and crystals
were harvested in 2 – 4 weeks. Crystals of human α211/Fab35/α-bungarotoxin
ternary complex were also produced in a similar condition with ones of mouse
α211 ternary complex. It was harder to obtain crystals for the human complex,
but crystals quality was better than crystals of mouse α211 ternary complex.
Morphologies of human α211 ternary complex were also rod-like cluster with
slightly larger in size. The complex proteins of both mouse and human α211
were fairly stable and could be stored at 4° C for some time. However, the
quality of crystals dramatically decreased with time (approximately 3 weeks
later).

36

Figure 2.8: Crystals grown as bundle with some precipitations.
Left: Preliminary crystals from the initial screening. Middle: Ternary complex of
mouse α211/Fab35/α-bungarotoxin. Right: Ternary complex of human
α211/Fab35/α-bungarotoxin. Crystals were grown as bundle at room temperature
with the hanging drop method. Some precipitations were observed in the background.
The reproducibility of the crystals was moderate, but the quality of crystals was
highly affected by the protein quality.

37

2.3.7 Structure determination of the α211/Fab35/α-
bungarotoxin ternary complex
The diffraction data were collected at the Advanced Photon Source
(APS) beamline 23-ID-B at Argone National Laboratory. The data were
scaled and merged using HKL2000 package
160
. Both crystals of human α211
and mouse α211 ternary complex belong to the space group C2 with unit cell
dimensions a = 159.9 Å, b = 42.1 Å, c = 136.5 Å, β = 117.1°; and a = 160.0 Å,
b = 42.0 Å, c = 137.6 Å, β = 116.5°, respectively. The scale files were
converted to mtz files using Scalepack2mtz in CCP4 package
161
. Structures of
both human and mouse ternary complexes were solved using programs in
CCP package
161
. Here, the human ternary complex is used as an example to
explain the process of structural determination. A preliminary structure was
solved at 3.5 Å, and a higher resolution structure was solved at 2.6 Å using the
preliminary crystal structure.
First, it was determined that there was one set of ternary complex
molecule in an asymmetric unit using the Matthews_coef program. Molecular
replacement was then performed with Phaser MR
162, 163
. Three different
structures were used as ensembles for the search models: α211 (from PDB:
2QC1), Fab search model (a poly-Ala model derived from Fab198), and α-
bungarotoxin (from PDB: 2QC1). The α211 and α-bungarotoxin of PDB:
2QC1 were separated using Pdbset. The Fab search model was created from
Fab198 (PDB: 1FN4) by changing it to a poly-Ala model. Each search model
was added into the Phaser MR program
162, 163
in the order of α211, the Fab
search model, and α-bungarotoxin. Searching the α211/α-bungarotoxin
complex (PDB: 2QC1) as one ensemble was attempted, but using each
molecule as an individual ensemble provided a better result. After one round
38

of refinement with Refmac5
164, 165
, R
work
/R
free
was 39.9/48.7 at 3.5 Å cut-off.
Side chains were then replaced with that of Fab35.
Then, this Fab35 model was used as a search model for better data
sets. Molecular replacement was similarly performed with three ensembles:
α211 (from PDB: 2QC1), Fab35 model, and α-bungarotoxin (from PDB:
2QC1), yielding a single solution (RFZ=2.14 TFZ=21.7 PAK=0 LLG=718
RFZ=8.0 TFZ=18.6 PAK=8 LLG=259 RFZ=4.3 TFZ=12.2 PAK=2 LLG=6
LLG=1733). After one round of refinement, R
work
/R
free
was 32.9/39.0 at 2.6 Å
cut-off. This higher resolution structure clearly showed the extra density
indicating the CDR3 loop of the heavy chain. The loop of Fab35 is much
longer than one of Fab198; therefore, several amino acids were needed to be
inserted based on the electron density and sequence alignments of two Fabs.
The CDR3 loop of the heavy chain was rebuilt using O
166
. Also, since the
α211 in the PDB: 2QC1 was mouse α1 subunit, some side chains were needed
to be replaced for the sequence of human α1. The side chains were fitted into
electron densities using Coot
167
, and water molecules were added using O
166
.
The structure was analyzed using Procheck, and further modifications were
performed. Ramachandran Plot can be found in Appendix A. After multiple
refinements, the final R
work
/R
free
was 19.9/27.7.
The final resolution of the complex with human α211 was 2.6 Å and
the one with mouse α211 was 2.65 Å. The resolution cut-off was determined
at I/σ I = 1. Values of crystallographic analysis are shown in Table 2.1.

39

Table 2.1: Crystallographic analysis of human or mouse α211/Fab35/α-
bungarotoxin complexes.
Data was processed with HKL2000 package, and structure was solved by molecular
replacement using Phaser MR with partial search models of PDB: 2QC1 and 1FN4
followed by refinement with Refmac 5. Highest resolution shell is shown in
parenthesis.

Human α211/Fab35/α-Btx Mouse α211/Fab35/α-Btx
Data collection

Space group C2 C2
Cell dimensions

a, b, c (Å) 159.9 160.0

42.1 42.0

136.5 137.6
α, β, γ (
o
) 90.0, 117.1, 90.0 90.0, 116.5, 90.0
Resolution (Å) 45.0 - 2.60 45.0 - 2.65

(2.64 – 2.60) (2.70 – 2.65)
R
sym
(%) 14.5 (0.00) 14.4 (0.00)
I/σI 10.1 (0.8) 10.2 (0.8)
Completeness (%) 99.4 (94.9) 94.5 (53.7)
Redundancy 3.6 (3.0) 3.5 (3.0)

Refinement

Resolution (Å) 45.0 – 2.60 45.0 – 2.65
No. reflections 25,125 23,363
R
work
/R
free
19.9/27.7 22.9/32.8
No. atoms

Receptor 1852 1861
αBtx 553 553
Light chain 1641 1650
Heavy chain 1677 1721
Sugar

Water 119 55
B-factors (Å
2
)

Receptor 47.5 50.7
α-Btx 76.9 60.2
Light chain 52.5 67.5
Heavy chain 57.3 89.5
Sugar

Water 47.2 47.1
R.m.s deviations

Bond length (Å) 0.011 n/a
Bond angles (
o
) 1.605 n/a
40

2.3.8 Structure guided analysis of the α211/Fab35 interactions
The first X-ray crystal structure of nAChR α1 subunit bound by
Fab35 was solved, and detailed structure information was revealed. The
overall structure of the complex is shown with a cartoon model in Figure 2.9
(Top left), and a surface model of the complex structure is also shown in the
figure (Top right). As expected, Fab35 targeted the MIR of α211 which is an
opposite side of ligand binding site. The close view of the interface is shown
at the bottom of the figure. The previous study using scFv198, it was
suggested that CDR2 and CDR3 of heavy chain (CDR-H2 and CDR-H3) and
CDR3 of light chain (CDR-L3) would be mainly interacting with the
receptor
168
. Generally, CDR-H3 and CDR-L3 are known to be involved in
binding, and CDR-L2 has the least interaction with antigens
169, 170
. Fab35
seems to follow the general characteristics of antibodies. Three loops, CDR-
H2, CDR-H3 and CDR-L3 interact with the MIR as if they are pinching the
region. Figure 2.10 shows some key interactions between Fab35 and human
α1. Fab35 is represented as surface model in Figure 2.10, A, and it clearly
shows that it is grabbing the MIR motif (shown in orange) of α1 subunit. On
the other hand, human α1 is shown in surface model in Figure 2.10, B, and a
loop of CDR-H3 inserts into a deep groove on the surface of α1. It is an
interesting finding since many antibodies recognize antigen by capturing an
epitope. The feature is indeed observed at the MIR motif, but an additional
interaction of the loop of CDR-H3 insertion contributes tighter and unique
binding of Fab35 to α1. Previously, X-ray crystal structures of two different
Fabs that were from MG mAbs (mAb192 and mAb198) were solved
151, 152
,
and their structures were compared with our structure. Figure 210, C and D
show superimposed views of Fab192 vs. Fab35 and Fab198 vs. Fab35.
Interestingly, the important CDR-H3 loop in Fab35 binding is much shorter in
those two Fabs; therefore, there is no insertion in the groove of α1.
41

Figure 2.9: X-ray crystal structure of the human α211/Fab35/α-bungarotoxin
complex.
Top left: The overall structure of the human α211 (cyan, shown as α1 in the figure)
bound by α-bungarotoxin (green, α-Btx) and Fab35 (yellow: heavy chain, magenta:
light chain). The variable domains of heavy and light chains are labeled as V
H
and
V
L
, respectively. The constant domains heavy and light chains are indicated as C
H

and C
L
, respectively. A cartoon model is shown. Top right: A surface model of the
complex structure. Bottom: A close view of the interface of human α211 and Fab35.
Fab35 interacts with the MIR of the receptor.

42

Figure 2.10: Interactions of Fabs with human α1 subunit.
(A) Fab35 is shown in surface model (the variable domain of heavy chain, V
H
shown
in yellow and the variable domain of light chain, V
L
shown in magenta).The α1
subunit (α211 ECD) is shown in cyan, and the MIR is colored in orange.Fab35
recognizes the MIR motif, and it grabs the region. (B) CDR3 of the heavy chain is
inserted into a deep groove of the α1 subunit. (C) Fab198 (PDB: 1FN4) is
superimposed on the ternary complex using the C-alpha backbone of Fab35 V
H
.
Alignment is similar using the C-alpha backbone of Fab35 V
L
. Blue indicates V
H
of
Fab198 and green indicates V
L
of Fab198. Fab198 structure aligns well with Fab35,
but a major difference can be seen at CDR3 of the heavy chain. The loop of Fab35
interacts with α1 by inserting into a groove. However, the loop of Fab198 is much
shorter, and it does not interact with α1. (D) Fab192 (PDB: 1C5D) is superimposed
on the ternary complex using the C-alpha backbone of Fab35 V
H
(blue: V
H
of Fab192
and green: V
L
of Fab192). A similar feature is observed on CDR3 of the heavy chain
of Fab192. It is shorter than one of Fab35 resulting in no interaction with α1.
Superimpose using the C-alpha backbone of Fab35 V
L
aligns two light chains fairly
well, and the heavy chain slightly shifted clockwise, which created more distance
from Fab35 V
H
at many locations except the CDR-H3 loop. The loop is not long
enough to form interactions as seen in the Fab35 ternary structure.
A B
C D
43

Some obvious interactions of Fab35 and α1 could already be observed
from the overall structure of the complex. In order to obtain more detailed
interaction information, contact program in CCP4 was used to analyze protein
molecular contacts, and also Protein Interfaces Surfaces and Assemblies
(PISA) program was run to analyze protein interfaces and surfaces. Those
programs provided information of residues which have close contacts with
residues from a binding protein partner. PISA also calculated interface area,
which was 557.7 Å
2
. Table 2.2 lists contacting residue pairs between α1
subunit and Fab35, and then interacting residues are visualized in colors on
surface model of the structure (Figure 2.11). The figure provides a brief idea
of how many residues from either the heavy chain (shown in yellow) or the
light chain (shown in magenta) contribute in binding to the α1 subunit. It is
clear that the heavy chain of Fab35 has more interactions than the light chain.
Figure 2.11, B is a side view of Fab35, and it is obvious that the unique CDR-
H3 loop, which inserts into a groove of α1, sticks out from the rest of
structure. Residues which interact with both the heavy and light chains are
colored in orange, and those residues belong to the MIR which is known to be
an antibody recognition site. Furthermore, these contacting analyses revealed
that N-terminal α-helix is indeed important for Fab35 binding. The heavy
chain had main interactions with the α-helix through hydrogen bonds and salt
bridges. It was an interesting finding since MIR has been heavily studied as
the target of autoantibodies, but the importance of α-helix in binding has not
been discussed until recently
132, 133
. Our structure revealed that α-helix had a
significant role in binding of autoantibodies. Fab192 and Fab198 structures
align well with Fab35 at the α-helix interaction site, so it would be highly
possible that many of MG antibodies also carry the feature.

44

Fab35 Ternary Interaction Table

Cyan: Hydrogen bonds
Magenta: Both hydrogen bonds and salt bridges

Table 2.2: List of contacting residue pairs between α1 subunit and the heavy or
the light chain of Fab35.
Two different programs, contact and PISA were run to determine contacting residues
at the interface of α1 subunit and Fab35 within 4.5Å. The table shows are combined
results of the two programs.

α211 (B) Light Chain (C) Heavy Chain (D)
H3 T57, V58
R6 W52, D54, G56, T57
L7 A103
K10 W52, D53, D54, R100, R102, A103
L11 A103
D14 R102
Y15 R102
Y63 Y32, K50
N64 R102
K66 Y32 A103, I104
W67 I92 A103, I104
N68 Y91, I92, N93, G94, Y95
P69 I92, N93
D70 G94, Y95 W47, V58
D71 Y95 W47, R50, W52, V58, A103, N105
Y72 W52, V58, A103, N105
G73 V58
45

Figure 2.11: Surface model of α1 subunit and Fab35 showing contacting
residues on each protein.
(A) Side view of α1 subunit with α-bungarotoxin. (B) Side view of Fab35. (C)
Rotated (A) with 90° downward. Top view of α1 subunit with α-bungarotoxin. (D)
Rateded (B) with 90° upward. Bottom view of Fab35. Yellow indicates interaction of
the heavy chain and α1 subunit. Magenta indicates interaction of the light chain and
α1 subunit. Orange color on α1 subunit shows the residues contacting with both
heavy and light chain of Fab35. Heavy chain has more interaction with α1 subunit.
The residues with orange color belong to the MIR. It indicates that MIR is indeed a
target of autoantibodies by interacting with both chains. In (B), the unique CDR-H3
loop is clearly showed.
α211
α-Btx
Fab35
A
B
C
D
α211 α-Btx
Fab35
90° 90°
46

The structure was analyzed further, and some interesting interactions
were observed. Four important water molecules were also detected between
the interface of α1 and Fab35 (Figure 2.12). Those water molecules show
important roles in binding, and details are discussed later. The first discussion
point is the interactions at the MIR. Previous studies have pointed out the
importance of the region, especially residue Asn68 and Asp71
134-136
α1.
Figure 2.13, A is a zoomed out view of interface at the MIR, and Figure 2.13,
B is a zoomed in view showing detailed interactions. The structure revealed
that Asp71 is indeed significant for the binding by forming the interaction
network with surrounding residues. It forms a salt bridge with Arg50 of V
H

and a hydrogen bond with Tyr95 of V
L
. Figure 2.13, C is a view from a
different orientation and it shows how residues in the interaction network
further interact with their surrounding residues. Arg50 of V
H
packs face-to-
edge with Trp47 and Trp52. Phe107 of V
H
also interacts with Tyr95 of V
L
by
face-to-edge. Asp71 is located where its side chain interacts with Asn105 of
V
H
; however, the orientations of the side chains of both residues are not
optimal to form a hydrogen bond. They are still within a distance where they
can make van der Waals contact (3.4Å). There are two critical interfacial
water molecules involved in this interaction network, H4 and H5. Asp71
forms hydrogen bonds to both of the water molecules. H5 in turn forms
hydrogen bonds to the main chain carbonyl of Tyr91 and the main chain
amide of Asn105, respectively. Another interfacial water molecule, H4 also
forms two more hydrogen bonds in addition to the one with Asp71. One of
them is with the main chain carbonyl of A103, which in turn forms another
hydrogen bond with Y72. The other one is with the main chain amide of N68
which is believed to be an important residue among the MIR. Asn68 interacts
with the main chain of light chain, Figure 2.13, D shows the interactions with
Asn68. It forms hydrogen bonds with the main chain carbonyls of Tyr91 and
Gly94 of V
L
. It also forms another hydrogen bond with the main chain amide
47

of Gly94 of V
L
. Figure 2.14 shows electron densities of those two residues,
D71 and N68. Overall, Asp71 of α1 is located at a critical position forming
interactions to both the heavy and light chains. It also interacts with
surrounding residues through water molecules. A similar phenomenon was
observed with N68 of α1, which has also been heavily studied as a critical
residue. As a result, Fab35 interacts extensively with the MIR. These
structural observations are consistent with biochemical data and also indicate
the importance of the conformation of the MIR. Certain residues need to be at
certain orientations for Fab35 to be bound. It is one of the reasons for Fab35
to be identified as a conformation dependent antibody.

Figure 2.12: Water molecules mediating interactions at the interface of α1
subunit and Fab35.
Some water molecules mediate interactions between α1 (cyan) and Fab35 (heavy
chain: yellow and light chain: pink). Four water molecules are shown in a green
sphere (H2, H4, H5, and H82). Interfacial water molecules, H4 and H5 are important
for binding at the MIR. H2 water molecule mediates interactions of α1 and V
H
. The
last water molecule H82 is related to interactions with the CDR-H3 loop. Detailed
interactions are discussed later of the section.
48

Figure 2.13: Binding interactions of α1 subunit and Fab35 at the MIR.
(A) The overview of the interactions at the atomic center of the MIR (cyan: α1,
orange: MIR, yellow: V
H
, magenta: V
L,
and green sphere: water). (B) A zoomed in
view of the MIR. D71 of α1 forms a salt bridge with R50 of V
H
and a hydrogen bond
with Y95 of V
L
. Two interfacial water molecules, H4 and H5 forms hydrogen bonds
with surrounding residues (H4: D71 and N68-NH of α1 and A103-CO of V
H
, H5:
D71 of α1, N105-NH of V
H
, and Y91-CO) (C) R50 of V
H
and Y95 of V
L
further
interact with surrounding residues. R50 of packs face-to-edge with W47 and W52 of
V
H
, and a similar interaction is seen with F107 of V
H
and Y95 of V
L
. (D) N68 of α1
additionally forms hydrogen bonds with the main chain carbonyls of Y91 and G94 of
V
L
. It also forms another hydrogen bond with the main chain amide of G94 of V
L
.
Those interactions would require a certain orientation of amino acid residues which
would explain why mAb35 is a conformational dependent antibody.

A
B
C
D
49

Figure 2.14: Electron density maps of binding interactions of α1 subunit and
Fab35 at the MIR.
(A) Detailed binding interactions at D71 of α1 (cyan: α1, orange: MIR, yellow: V
H
,
magenta: V
L,
and green sphere: water). (B) The electron density map at D71of α1
which interacts heavily with R50 of V
H
and Y95 of V
L (
yellow: α1, blue: V
H
,
magenta: V
L,
and magenta cross: water). (C) Detailed binding interactions at N68 of
α1. (D) The electron density map at N68 α1 which forms hydrogen bonds with the
main chain carbonyls of Y91 and G94 of V
L
. It also forms another hydrogen bond
with the main chain amide of G94 of V
L
.
A
C
B
D
50

Our structure revealed that the MIR is indeed an antigenic site of MG
antibodies, at least for mAb35. Next finding was interactions at the groove of
α1 where CDR-H3 is inserted in. It was briefly mentioned in Figure 12.11, but
the loop can be recognized easily on a surface model of the structure.
Especially, R102 is located at the tip of the loop, and it has multiple
interactions with surrounding residues and backbones. Figure 2.15, A is an
overall view of the interaction site, which is seen from the top of α1. Figure
2.15, B shows how R102 is oriented at the groove, and it nicely aligns with
N64 and the main chain carbonyl of Asp14 forming polar-charge interactions.
In addition, it forms hydrogen bond with H82 water molecule which also
forms two more hydrogen bonds with the main chain carbonyls of Leu11 and
N64 of α1.
At the same location, another important interaction was observed.
Figure 2.15, C describes an important interaction of A103 of V
H
. It engages
extensive interactions with the MIR and the N-terminal α-helix of α1. The
interaction of the main chain carbonyl of A103 of α1 was discussed in Figure
2.13. It interacts with the MIR directly and indirectly through a water
molecule (H4). On the other hand, the side chain of A103 of α1 makes van der
Waals contacts to a number of residues from the N-terminal α-helix (Leu7,
Leu11, and the aliphatic portion of the side chain of Lys10) as well as to
Trp67 from the MIR. In a structural sense, those side chains of residues from
α1 creates a hydrophobic pocket where the side chain of A103 can insert.
In terms of the N-terminal α-helix, more interactions were found to be
important. A network of interactions at the binding site is shown in Figure
2.16, A. Asp53 of V
H
forms a salt bridge with Lys10 of α1, and Asp54 of
heavy chain forms bidendate hydrogen bonds with Arg6 of α1. The Arg6 then
forms a cation-π stack interaction with Trp52 V
H
which also packs face-to-
edge interaction with Tyr72 of α1. The last interaction at the N-terminal α-
helix is mediated by a water molecule, H2 (Figure 2.16, B). The water
51

molecule was located at close proximity to the heavy chain of Fab35. His3 of
α1 is stretched out to the water molecule forming a hydrogen bond. K64 and
the main chain carbonyl of Y59 of V
H
also form a hydrogen bond with H2
water molecule. The MIR was the first targeting site to be recognized for MG
autoantibodies, but recent studies suggested that the N-terminal α-helix would
be involved in binding
132, 133
. Our structure revealed that it indeed contributed
to the interaction with the heavy chain of Fab35. It also showed that the
unique loop of CDR-H3 play an important role by interacting with both the N-
terminal α-helix and the MIR.
Lastly, an interaction was found at the opposite side of the MIR or the
N-terminal α-helix (Figure 2.17). Tyr63 of α1 and Lys50 of V
L
form a
hydrogen bond, and those residues are supported by surrounding residues and
water molecules. Lys50 of V
L
forms a cation-π stack interaction with the
above residue, Tyr32 of V
L
. On the other hand, Tyr63 of α1 is supported by
Glu23 of α1 from the bottom. The residue Glu23 forms a hydrogen bond with
the water molecule H73, which in turn forms hydrogen bonds with N52 and
the main chain carbonyl of K50 of V
L
. It would help to hold Glu23 to be a
certain orientation. Additionally, the loop of Glu23 is held by the water
molecule H7, which forms hydrogen bonds with Asn64 and the main chain
amide of Glu23of α1. The main chain carbonyl of Ans64 of α1 contributes to
the interaction with R102 of V
H
as discussed in Figure 2.15. In this regard,
Asn64 of α1 would also be an important residue for MG autoantibody binding
though it is not a part of the MIR.

52

Figure 2.15: Binding interactions with insertion of CDR-H3.
(A) The overview of the interaction site (cyan: α1, orange: MIR, yellow: V
H
,
magenta: V
L,
and green sphere: water). The view is from the top side of α1. (B) R102
stacks well with N64 and the main chain carbonyl of D14 of α1 forming polar-charge
interaction. (C) R102 forms a hydrogen bond with H82, which in turn forms
hydrogen bonds with the main chain carbonyls of L11 and N64. Another residue of
CDR-H3, A103 engages extensive interactions with the MIR and the N-terminal α-
helix. The main chain carbonyl of A103 forms a hydrogen bond with Y72 of α1 as
well as with the interfacial water molecule H4. Its side chain then inserts into a
hydrophobic pocket which is defined by L7, K10 (the aliphatic side chain), L11, and
W67 of α1.
A B
C
53

Figure 2.16: Additional interactions involving the N-terminal α-helix.
(A) A network of interactions at the binding interface between the N-terminal α-helix
and the heavy chain of Fab35. (cyan: α1, orange: MIR, yellow: V
H
, and magenta: V
L
).
D53 of V
H
forms a salt bridge with K10 of α1. D54 of V
H
and R6 of α1 are interacted
with two hydrogen bonds. And R6 then forms a cation-π stack interaction with W52
of V
H
, which in turn packs face-to-edge interaction with Y72 of α1. (B) H2 water
molecule mediates an interaction between the N-terminal α-helix and V
H
through
hydrogen bonds. H3 of α1 forms a hydrogen bond with H2 water molecule, and K64
and the main chain carbonyl of Y59 of V
H
form additional hydrogen bonds with the
water molecule.

A B
54

Figure 2.17: Water mediated interactions of α1 subunit and Fab35.
(A) The overview interaction at Y63 of α1 (cyan: α1, orange: MIR, yellow: heavy
chain, magenta: light chain
,
and green: α-bungarotoxin). It is located at the other side
of the MIR or the N-terminal α-helix. (B) The original structure (A) is turned
counterclockwise by 90°. Y63 of α1 and K50 of V
L
form a hydrogen bond, and
surrounding residues and water molecules provide supports to the two residues. K50
of V
L
interacts with Y32 of V
L
with cation-π stacking. Y63 of α1 is supported by E23
of α1, which is also supported by hydrogen bonds with water molecules (H7 and
H73). The water molecule H7 holds a loop of E23 of α1 by forming a hydrogen bond
with N64 of α1 which is located at a neighboring loop.

A B
55

Detailed interface interactions were determined at the atomic level,
and new insights were obtained to understand how Fab35 interacts with α1
subunit. Next question would be whether mAbs interacts with neighboring
subunits. In order to answer the question, our Fab35 complex structure was
superimposed with α7/AChBP chimera structure (PDB: 3SQ9) which was
solved as a pentamer. Even though it is not a muscle-type nAChR, it is still a
good model since nAChRs have high homology between subunits and
receptors. The structure of α1/Fab35/α-bungarotoxin was superimposed on
one of α7/AChBP chimera subunit using the C-alpha backbone of α1. Figure
2.18, A shows the overall view of the model showing how Fab35 interacts
with a pentermer nAChR. Figure 2.18, B is a zoomed in view of the interface.
Fab35 seems to interact only with α1 subunit (blue of cyan), not neighboring
subunits (shown in white and dark gray). Therefore, the ternary complex
structure would provide sufficient information to understand the interaction of
α1 and mAb35 (Fab35) in the context of a native pentamer nAChRs.
Since only Fab portion was used for our structural study, another
superimposed model was created to determine how an entire antibody would
interact with the nAChR. For better understanding, the Torpedo nAChR was
used since it was solved as a full-length pentamer (PDB: 2BG9). For a full-
length antibody, PDB: 1IGY, which is IgG1 subtype was picked because
Fab35 belongs to the same subtype. First, the ternary complex structure was
superimposed on the Torpedo nAChR structure using the C-alpha backbone of
α1, and then the full-length antibody was superimposed on Fab35 heavy chain.
For a clear view, the original ternary complex structure was removed from the
model structure. Figure 2.19 shows that the antibody is projected away from
the central pore of the receptor. It provides an orientation of the antibody
when it binds to the receptor.

56

Figure 2.18: A model of Fab35
bound to a pentamer receptor,
α7/AChBP chimera.
(A) The ternary complex is
superimposed on the pentamer
α7/AChBP chimera structure
(PDB: 3SQ9) using the C-alpha
backbone of α1. Magenta indicates
light chain and yellow indicates
heavy chain of Fab35. Cyan shows
α1 and blue represents one of
α7/AChBP chimera subunits.
Other four subunits including
neighboring subunits of the α1
colored in either white or gray. (B)
Zoomed in view of the interface. It
clearly shows that Fab35 interacts
with only α1 subunit, not
neighboring subunits.
A
B
57

Figure 2.19: A model of a full-length antibody bound to nAChR.
A model is created by superimpose the ternary complex on the Torpedo nAChR
pentamer structure (PDB: 2BG9) using the C-alpha backbone of α1. A full-length
IgG1 antibody (PDB: 1IGY) is then superimposed on heavy chain of Fab35. The
original ternary structure is removed for clarity. The antibody is shown as a surface
model (blue and cyan: heavy chain, orange and yellow: light chain). Fab and Fc are
as labeled. The full-length Torpedo nAChR is shown in light brown and one α1 ECD
is colored in magenta. It shows an orientation of the entire antibody when it binds to
the nAChR.

58

The ternary complex structure as well as models created based on the
structure suggests that mAb35 does not interact with other subunits from the
same receptor. Fab35 binds to the receptor by pointing away from the central
axis; therefore, it is unlikely that two Fabs from the single antibody bind to
two α1 subunits within the same receptor. The last question would be whether
a single antibody can target two α1 subunits which are from different
receptors. If it is possible, it would be interesting to determine what kind of
orientation they would interact. In order to answer the question, additional
model, were created. Another ternary complex was superimposed on heavy
chain which was available for binding. And then, another Torpedo nAChR
was superimposed on α1 of the ternary complex structure using the C-alpha
backbone. Once again, the original ternary complex structure was removed for
better view.
Figure 2.20, A shows how the antibody interacts with two nAChRs,
and Figure 2.20, B is a view from the bottom, which is obtained by rotating
the original view with 90° upward. Antibodies have a twisted structure;
therefore, the receptors have two fold symmetries. The receptors are pointed
outward, it looks difficult for a single antibody to target two receptors. Thus,
antibody binding to two nAChR pemtamers may induce membrane curvature.
This is an interesting poin considering that MG autoantibodies cause
internalization of nAChRs. In addition, the hinges of antibodies could provide
enough flexibility for antibodies to adjust angles to bind targets. Based on
those possibilities, a single antibody could bind to two α1 subunits from
different receptors though it cannot cross-link the two α1 subunits in the same
receptor, which agrees with a previous EM study
139
.

59

Figure 2.20: A model of an antibody cross-link two nAChRs.
(A) A front view of one antibody cross-links two α1 subunits from different nAChRs.
Those receptors are projected away from the antibody. (B) A view from the bottom.
The original model is rotated on horizontal axis by 90°. It shows that it has two fold
symmetries since antibodies have a twisted structure. In actual situation on cells, the
orientation of receptors would be different due to an uneven cell surface and/or
flexible hinges of antibodies. It shows a possibility of a cross-linking of two α1
subunits though there would be some limitations. Color distributions are same with
Figure 2.16.
90°
A
B
60

2.4 Discussion
It was challenging to obtain large quantity of α1 subunit which is a
membrane protein. The soluble portion of ECD of α1 was purified, and
purification protocol was optimized to obtain ~ 0.5 mg from 1L of starting
culture. The mAb35 was purified from hybridoma cells, and Fab35 was
further purified by digesting mAb35with papain. The X-ray crystal structure
of α1 bound by Fab35 was solved for the first time and revealed detailed
information of binding interface. It was achieved by adding α-bungarotoxin to
stabilize proteins. Ternary complex structures using both human and mouse
were solved. The differences of those two structures were not discussed in this
chapter, but superimposed structures can be found in Appendix (Figure A.1).
A slight shift of Fab35 was observed between human and mouse structures
with a large view, but there was not significant shift at interfaces. Therefore,
the binding patterns would be similar between different species of α1 subunits.
The importance of the MIR has been discussed by many scientists, and it has
been known as a target of MG autoantibodies. Our structure confirmed that
the MIR was indeed important for Fab35 targeting. The high resolution
structure also enabled to cross-validation with previous biochemical studies.
In 1989, Das and Lindstrom found an important difference of MIR
sequences between human muscle and Xenopus. They used one of MG mAbs
called mAb210 (discussed details in Chapter 3) and tested bindings with
synthetic peptides similar to the MIR (α66 – 76). In the study, it was shown
that N68 and D71 (D68 and K71 in corresponding Xenopus sequence) were
critical residues for MG mAb binding
135
. Additional studies were performed
using intact receptors expressed in Xenopus. Importance of those residues
were examined by introducing mutations
136
. Their results correlated with
another mutational study in which inhibitions of antibodies were observed by
61

substitution of either residues to glycine
134
. Here, our structure revealed that
those residues have significant roles in binding (Figure 2.13). The MIR
contributed significantly to the binding of Fab35. Especially D71 of α1 was at
the center of the interaction network forming hydrogen bonding with
surrounding residues. Two important interfacial water molecules were also
identified at the interface of the MIR and Fab35, and they mediated
interactions through hydrogen bonds. N68 of α1, which has been also found as
a critical residue involved in the network interaction of the MIR interface.
These structural observations of bindings at the MIR indicate that a
conformation of the MIR is important for those residues to be able to interact
with Fab35. It would be one of reasons why Fab35 would not bind to a
denatured α1 subunit Additional interaction sites other than the MIR were also
found. One of them was the N-terminal α-helix of α1 subunit which was
actually discussed in recent papers
132, 133
. The heavy chain of Fab35 interacted
with the α-helix directly and indirectly mediated by water molecules (Figure
2.14). Another important and interesting finding was an interaction of CDR-
H3 which inserts into a groove of α1 subunit. It also engaged interactions
between the N-terminal helix and the MIR. Those interactions would be
another reason for mAb35 to be a conformation dependent antibody. Many
important residues for Fab35 binding are located on different structural
elements of nAChR α1 such as the N-terminal α-helix and MIR. Moreovere,
Fab35 does not have a tight binding to a localized motif like Fab210 (Refer to
Chapter 3). This is why Fab35 (mAb35) binding is conformation dependent.
Overall, Fab35 interacted with several regions of α1 in a conformational
dependent manner; therefore, mAb35 cannot bind α1 subunits unless they are
properly folded to have enough interactions.
A model of a pentamer nAChR bound by Fab35 was created using
α7/AChBP chimera structure (PDB: 3SQ9) in order to determine if Fab35
62

(mAb35) interacts with other neighboring subunits (Figure 2.18). The answer
was that it does not interact with them; it binds to the receptor only through α1
subunit. It means that our structure would provide sufficient information to
develop therapeutic drugs for MG. Another model was created to analyze how
a full-length antibody binds to nAChR and how it binds to two different α1
subunits if it is possible. The Torpedo nAChR (2BG9) and mouse IgG1
(1IGY) were used to create the model. Previously, three dimentional structural
studies of cryo-EM showed that the variable domains of mAh35 were away
from the channel; therefore, it concluded that a single mAb35 could not bind
to two α1 subunits within the same receptor. However, it would be possible to
cross-link two α1 subunits with neighboring receptors
139
. Indeed, our structure
showed a binding of Fab35 away from the central pore of the receptor. The
model of an antibody bound to two receptors showed that a single antibody
could cross-link two α1 subunits. Two nAChRs were pointed out when they
were bound by the antibody. It could have a restriction based on the angle of
those receptors; however, it is still possible considering deformation of cell
membranes and flexibility of antibody hinges. The membrane curvature
induced by MG autoantibody binding to two nAChR pentamers could be
related to receptor internalization observed in MG.
Now, detailed crystal structure is available and it agrees with many of
biochemical studies previously performed. The mutagenesis studies of
replacing two critical amino acids (N68 and D71) showed a possible
development of therapeutic drugs. If those residues are blocked, the binding of
antibodies would be dramatically reduced; that is, another treatment would be
available for MG patients. Our structure will help to understand how the
autoantibodies recognize the nAChR α1 subunit, which in turn can aid in
understanding the complexities that cause MG and also contribute in
developing a more effective treatment of the disease. As one of possible drug
63

developments, a technique called mRNA (Chapter 5) display could be
utilized. In fact, the idea has been worked on by selecting molecules which
would target Fab35 binding site with collaboration with Dr. Roberts’ group.
The detailed structure and interaction information raised another idea as a
future work, which would be discussed in Chapter 8.
2.5 Materials and Methods
2.5.1 Construction of α211
The α211 mouse construct was provided by Dr. Zuo-Zhong Wang,
Zilkha Neurogenetic Institute, Department of Cell and Neurobiology, Keck
School of Medicine, University of Southern California. The detailed
construction information is as previously published paper
144
. Briefly, Flag-tag
and His-tag were added at the N-terminus and at the C-terminus, respectively,
for higher expression and purification purposes. The construct was truncated
at the 211
th
amino acid from the N-terminal of nAChR α1 subunit without
signal sequence, and 3 mutations (Val8Glu, Trp49Arg, and Val155Ala) were
introduced for better solubility and stability. The α211 human construct was
designed based on the α211 mouse construct, and the synthesized DNA was
ordered from GenScript. The gene was codon optimized for yeast, and it was
cloned into a pPICZαA vector using EcoRI and XbaI sites.
2.5.2 α211 expression
Both mouse and human α211 constructs were linearized by digesting
with Sac I restriction enzyme and transformed into KM71H of P. pastoris
(Invitrogen) by electroporation. The transformants were plated on YPDS
plates which contained 100 µg/mL Zeocin. Plates were incubated at 30 °C for
3 – 5 days until colonies formed, and several colonies were picked and
64

restreaked on fresh YPDS plates. After another 3 – 5 days of incubation,
colonies formed to be used for protein expression. Pre-inoculation was made
seeding a single colony in 30 mL BMGY medium. The culture was incubated
at 30 °C with shaking (180 rpm) overnight. This 5 – 7 mL culture was used to
inoculate 500 mL of BMGY in a 2L baffled flask at 30 °C with shaking (180
rpm) to OD
600
6. Invitrogen recommends OD
600
to be 2 – 6 which is log phase
growth, but expression issues were not observed with higher OD
600
. More
protein was obtained with higher OD
600
rather than lower OD
600
, therefore, it
was cultured to hit an OD
600
of 6 – 8. Cells were harvested by centrifugation at
3,000 x g for 15 minutes at room temperature using sterile bottles. The
supernatant was discarded, and cell pellets were re-suspended in 400 mL of
BMMY medium for induction. Aeration was important in the Pichia
expression system especially during methanol induction. Therefore, the media
volume was kept between 10 – 30% of the total flask volume all the time and
baffled flasks were used. The re-suspended culture was divided between two 2
L baffled flasks (200 mL media each) and incubated at 20 °C with shaking
(220 rpm) for 72 hours. 100% methanol was added every 24 hours to a final
concentration of 0.5% to induce protein expression. After 72 hours of
induction, cells were harvested by centrifuging at 6,000 x g for 20 minutes at
room temperature. The cell suspension was filtered using 0.22 µm filter to
remove any cell debris. Protein purification proceeded with the supernatant as
the protein was secreted.
2.5.3 α211 purification
Ni-NTA agarose beads (QIAGEN) were incubated with the
supernatant at 4 °C overnight with end-over-end rotation. The protein was
eluted with elution buffer (50 mM NaH
2
PO
4
; pH 7.8, 0.5 M KCl, 10%
glycerol, and 500 mM imidazole) after washing with washing buffer
65

containing a lower amount of imidazole (20 mM) and 0.1% Triton X-100 to
remove loosely bound proteins. The eluted protein was concentrated and ran
over a size exclusion column (Superdex 75 10/300 GL, GE Healthcare) with
20 mM HEPES; pH 7.5 and 150 mM NaCl buffer for further purification.
After each protein fractions was identified and confirmed by OD
280
and SDS-
PAGE gels, protein peak fraction were pooled and concentrated for future
experiments.
2.5.4 Cell culture and reagents for mAb35
Hybridoma cells of mAb35 were purchased from American Type
Culture Collection (ATCC). The cells were maintained in DMEM medium
containing 1.97 g/L NaHCO
3
and 10% fetal bovine serum (FBS). The cells
were cultured in a 37 °C incubator with 5% CO
2
and subcultured every 2 to 3
days with cell density of between 1 x 10
5
and 1 x 10
6
cells/mL. For protein
production, cell culture was kept for several days until medium color changed
to yellow.
2.5.5 mAb35 purification
After 7 to 10 days of incubation at 37 °C, cell culture was harvested by
centrifuging at 6,000 x g for 15 minutes. The supernatant was filtered with
0.22 µm filter, and affinity purification was performed using Protein G
Sepharose
TM
4 Fast Flow (GE Healthcare). The supernatant and beads were
incubated at room temperature for 2 hours with rotation, and the protein was
eluted with elution buffer (0.1 M glycine-HCl, pH 2.7) after washing with
washing buffer (20 mM sodium phosphate, pH 7.0). Due to the low pH of the
elution buffer, a neutralizing buffer (1 M Tris-HCl, pH 9.0) was added to the
collection tubes (60 to 200 µlL/mL elute) prior to collection. After checking
the expression of the protein with SDS-PAGE gels, the protein elution was
66

concentrated to run over a size exclusion column (Superdex 200 10/300 GL,
GE Healthcare). The fraction of protein peak was pooled and concentrated for
further study.
2.5.6 mAb35 digestion and Fab35 purification
mAb35 was exchanged into digestion buffer (20 mM sodium
phosphate, 10 mM EDTA; pH 7.0 and 20 mM cysteine-HCl; adjust pH to 7.0
right before use) using Zeba Spin columns (Thermo Scientific), and resulting
sample incubated with immobilized papain beads (Thermo Scientific). Papain
beads limitation is 10 mg per 250 µL beads based on a company protocol. The
efficiency of the digestion was low for mAb35, so double the amount was
used. Due to the low recovery rate, approximately 10 mg of mAb35 was used
for a round of Fab35 purification. The sample was rotated overnight at 30 °C.
The beads were transferred into a gravity column, and the flow through was
collected. For the maximum recovery of the protein, the beads were washed
with PBS buffer, and the wash flow through was collected in fractions. The
protein fractions were combined after identifying with SDS-PAGE gels, and
then the sample was concentrated down to ~ 500 µL while exchanging buffer
to Mono Q Buffer A (20 mM sodium phosphate, pH 7.0). An affinity column
(Mono Q HR 5/5, GE Healthcare) was used to separate the Fab portion from
the rest using a salt gradient of 0 – 100% Buffer B (Mono Q Buffer B contains
1 M NaCl). Fab fractions were collected and concentrated for α211 complex
purification.
2.5.7 Gel shift assay
α211, α-bungarotoxin and Fab35 were mixed in an equimolar ratio
(0.5 µg of α211 was used), and the mixture incubated on ice for 1 hour. For
the binding specificity experiment, α9 3Mut (Refer to Chapter 6 for details)
67

was prepared in the same manner. Either 6% or 10% native-PAGE was
prepared and stored at 4 °C. Just before samples were loaded into a gel, 5X
loading buffer (containing no dye) was added to each sample tube. The gel
was run for 3 – 3.5 hours at 100 – 120 V, ~ 15 mA in 4 °C room with TBE
buffer. Bands were detected with Coomassie staining.
2.5.8 The α211/Fab35/α-bungarotoxin complex purification
α211, α-bungarotoxin and Fab35 were mixed at a 1:1.5:1.5 molar
ratio, and the mixture incubated on ice for 1 – 2 hours. Ni-NTA purification
was performed to remove excess α-bungarotoxin and Fab35. The elution was
run over a size exclusion column (Superdex 200 10/300 GL, GE Healthcare)
with 20 mM HEPES; pH 7.5 and 150 mM NaCl buffer. Two peaks were
obtained, and each peak fractions were pooled separately to concentrate for
crystallization.
2.5.9 Crystallization and data collection
The purified and concentrated ternary complex of α211 (both mouse
and human)/Fab35/α-bungarotoxin was diluted to 2.5 mg/mL. Crystals were
screened using Crystal Screen, Crystal Screen 2 and Index kits (Hampton
Research) by the hanging drop method at room temperature. Each drop
contained 0.5 μL of protein and 0.5 μL of reservoir solution, and 0.5 mL of
screening solution was added in each well. Three conditions gave initial
crystals, and one condition was picked to optimize a crystallizing condition.
The optimized condition was 0.1 M sodium cacodylate trihydrate pH 6.5, 0.1
M – 0.15 M calcium acetate hydrate and 18% – 20% PEG 8K. Rod-like
shaped crystals were grown as bundle with the size of 10 μm x 100 μm – 20
μm x 200 μm. Crystals were harvested using harvest solution and cryo
solution which contained 0.2 M calcium acetate hydrate, 0.1 M sodium
68

cacodylate trihydrate pH 6.5 and 30% PEG 8K (additional 20% glycerol for
cryo solution). Crystals were transferred to different concentrations of
harvest/cryo mixture solution to protect crystals from osmotic shock (100%
harvest solution, 1:3 harvest/cryo solution, 1:1 harvest/cryo solution, 3:1
harvest/cryo solution, 100% cryo solution). Each incubation time was
approximately 5 – 10 minutes. Longer incubation period, 30 minutes, was also
tried, but there was no obvious difference.
Data were collected at the Advanced Photon Source (APS) beamline
23-ID-B at Argone National Laboratory using a 10 x 10 beam (λ = 1.0332 Å,
12.000 keV) with the attenuation factor of 5.0 and MARmosaic 300 CCD
detector. The detector distance was 300.0 mm. The oscillation range and the
exposure time per frame were 0.5° and 2.0 sec, respectively. Data were
processed and scaled using HKL2000 package
160
. The phases were
determined by molecular replacement using the coordinates of α211 and α-
bungarotoxin complex (PDB: 2QC1) and Fab198 (PDB: 1FN4, poly Ala).
Refmac was used to refine the structures. The sequence of mAb35 was
obtained from an antibody sequencing company, MCLAB Molecular
Laboratories (San Francisco, CA), and Fab35 side chains were added using
Coot. Some extra amino acids in Fab35 were inserted with O, and further
structural model buildings were also carried out in O.

69

Chapter 3

Structural Studies of nAChR α1
Subunit and Fab210 Complex

3.1 Abstract
mAb210 is another MG mAbs which has an unique feature. The
binding of mAb35 is highly conformational dependent, whereas mAb210 can
bind to denatured nAChRs or peptides. It would be interesting to determine
the differences of these mAbs in binding at the atomic level. The structure of
Fab210 ternary complex is successfully solved at 2.6 Å resolution, and
structures are compared. The Fab210 ternary structure once again shows an
importance of the MIR loop, but it also shows different binding patterns from
Fab35. Another structure of MG mAbs binding to α1 would help to
understand the disease further and help to develop molecules and strategies
that can block a wide range of MG autoantibodies.
70

3.2 Introduction
In the previous chapter, the structure of nAChR α1 subunit bound by
Fab35 was successfully solved at 2.6 Å. The structure revealed the importance
of the MIR upon bindings of autoantibodies. There are hundreds of
monoclonal antibodies that were created against various species of nAChRs,
and they show different features on bindings
119-121
. Among them, mAb35 is
the prototype MG mAb, which is known as a conformational dependent
binding antibody. That is, it can bind to native α1, but it cannot bind to
denatured α1
89, 135
. For better understanding of the disease, it will be desirable
to crystallize another complex with other well-defined MG mAb, which
carries different characteristics from mAb35. mAb210 was picked as a next
MG mAb to be studied since mAb210 binds to synthetic peptides or denatured
nAChR
135
, where as mAb35 binds to the nAChR with a highly conformational
dependent manner. Both of mAbs belong to IgG
1
class, and mAb35 was raised
by immunizing Electrophorus AChR to rats
120, 132
, whereas mAb210 was
raised against mammalian nAChR (Bovine and mouse) by immunizing rats
89,
132
. Another interesting feature of mAb210 is that it can be used as a research
tool since it can bind to denatured receptors or peptides. It has cross-reactivity
with some nAChR subunits such as α3, α5, and β3; therefore, it is sometimes
used to detect those subunits on Western blots
87-90
. Even though mAb35 also
cross-reacts with them, it cannot bind when they are denatured.
The structure of the complex with Fab210, along with that of nAChR
α1 subunit/Fab35, would provide us a better idea on how auto-antibodies with
different features bind to nAChR α1 subunit and cause the same autoimmune
disease, MG. If common features of bindings among a variety of antibodies
are found, we could efficiently block autoantibodies from binding to nAChR
α1 subunit.
71

3.3 Results
3.3.1 Fab210 preparation
mAb210 was obtained from cultured hybridoma cells. Cells were
cultured in the same way as mAb35 and purified using Protein G beads
followed by Superdex 200 10/300 (Refer to Chapter 2 for details). A single,
sharp peak was obtained with a small peak within the void volume (Figure 3.1,
A). The pooled samples were subjected to Mono Q HR 5/5. A similar Mono Q
chromatogram was obtained, but bound peaks were slightly different from
those for the Fab35 purification. Rather than the sharp, slightly fused peaks
seen in Mono Q trace for Fab35 (Chapter 2, Figure 2.3, B), the two peaks
(peaks 3 and 4) for Fab210 were completely fused, and appeared to be shifted
to the right (Figure 3.1,B). An SDS-PAGE gel showed that both peaks 3 and
4 contain Fc (Figure 3.1, D). Since the efficiency of papain digestion was low
in the mAb35 experiment, more papain beads were used to increase a
recovery rate of Fab210. As a result, most of mAb210 was digested with
papain. The left broad shoulder next to the fused peak on Mono Q
chromatogram (peak 2) could be intact mAb210 though it was too small of an
amount to detect on an SDS-PAGE gel. Fab210 was obtained in the flow
through as with Fab35, and the sample was run over Superdex 200 10/300.
The resolution of the SDS-PAGE was not great, and it was hard to distinguish
the light and heavy chains of Fab210. However, it clearly showed that papain
digestion was successfully performed and Fab210 was purified for further
study.

72

Figure 3.1: Fab210 purification.
(A) mAb210 was purified over Superdex 200 10/300. A nice single peak was
obtained. Fractions 23 – 27 were pooled and concentrated down for the following
Fab210 purification. (B) Chromatogram of Mono Q HR 5/5 of Fab210 purification.
The bound peak was shifted to the right compared to Fab35 purification, and is Fc.
(C) Superdex 200 10/300 chromatogram of Fab210 purification. The main peak
fractions 22 – 26 were pooled for the ternary complex purification. (D) 15% SDS-
PAGE gel of Mono Q and S200 fraction samples (16 μL of sample + 4 μL of 5x
loading dye except Injection: 1 μL of sample + 7 μL of water + 2 μL of 5x loading
dye). MonoQ injection sample shows almost complete digestion of mAb210. The
bound peak (F23 and F25) was indeed Fc portion of mAb210.
D
A
B
C
1
2
3
4
73

3.3.2 Gel shift assay
The ternary complex formation was tested before the ternary complex
purification was attempted though Fab210 should bind to α211 as well as
Fab35 does. The gel shift assay was performed (Figure 3.2) in the same way
as with Fab35 (Refer to Chapter 2). Both α-bungarotoxin and Fab210
migrated upward due to basic pI of the proteins (Lanes 2 and 3). Fab35 was
used as a positive control for the assay system. The experiment shows the
binding of α-bungarotoxin, Fab35, and Fab35/α-bungarotoxin to mouse α211
as expected (Lane 4, Lane 7, and Lane 8 respectively).
Lane 5 shows the binding of Fab210 to mouse α211 and Lane 6 shows
the ternary complex formation with Fab210 and α-bungarotoxin. However,
faint bands were observed in both Lane 5 and 6 indicating there were free
mouse α211 (Lane 5) and mouse α211/α-bungarotoxin (Lane 6) which Fab210
was not bound to. It is possible that it could be the result of from an inaccurate
concentration measurement of Fab210. There is also the possibility that
Fab210 binding is not as strong as Fab35 binding, and the complex could be
dissociating during the assay. There were no obvious differences between the
Fab35 and the Fab210 complexes; however, a slight difference was seen in
migration (Lane 5 vs. Lane 7 and Lane 6 vs. Lane 8). The migration of Fab35
complex was faster than the one of Fab210. It could be due to charge
differences between two Fabs resulting from their unique amino acid
sequences or conformational differences of the ternary complexes. Since
major shifts, which indicate the ternary complex formation, were observed,
the Fab210 ternary complex was carried through purification for
crystallization.

74

Figure 3.2: 10% Native-PAGE gel showing binding of Fab210 to mouse α211.
Fab35 was used to test the assay system. Bindings of Fab35 and Fab35/α–
bungarotoxin were observed in Lane 7 and Lane 8, respectively. Basic pI proteins of
α-bungarotoxin and Fab35 migrated upward not showing any bands on the gel (Lane
2 and 3). Fab210 binding was also observed both with and without α–bungarotoxin
(Lane 5 and Lane 6). Faint bands were seen in both Lane 5 and Lane 6 indicating free
α211 (or α211/α–bungarotoxin) without Fab210. It could be caused by not enough
Fab210 by a concentration measurement error or dissociation of Fab210 during the
assay.

75

3.3.3 The α211/Fab210/α-bungarotoxin complex purification
The ternary complex of both mouse α211/Fab210/α-bungarotoxin and
human α211/Fab210/α-bungarotoxin were purified a similar way as the Fab35
ternary complex (Refer to Chapter 2 for details). Briefly, α211, α-
bungarotoxin and Fab35 were mixed at a 1:2:2 molar ratio to form the
complex, and excess α-bungarotoxin and Fab35 were removed by Ni-NTA
purification before Superdex 200 10/300 purification.
The chromatograms of Superdex column were different from those of
the Fab35 ternary complex. There was a gradual slope on the left side of the
main peak in Fab210 ternary complex purification (Figure 3.3), while a
distinct peak was observed to the left side of the major peak in the purification
of Fab35 ternary complex, which indicated oligomer forms of the complex
(Chapter 2, Figure 2.7). Superdex chromatograms of mouse α211 and human
α211complex were similar as seen in Figure 3.3, A and B, respectively. Figure
3.3 C shows the overlap chromatograms of Fab35 ternary complex (in blue)
and Fab210 ternary complex (in red). The major peaks were observed at
similar positions, but the left shoulder looked different for the different Fab
complexes. An SDS-PAGE showed the ternary complex formation in all
Superdex fractions (Figure3.3, D). Generally, homogeneity is important for
crystallization, but the required purity is different depending on proteins. If
homogeneity is critical for crystallization for the Fab210 ternary complex, it
would be more challenging than for the Fab35 ternary complex crystallization
since it is hard to separate the left shoulder from the main peak. This result
indicated that there are no clear number of species in the sample such as a
monomer and dimer. It is also possible that Fab210 oligomer complex is in a
dynamic equilibration with the monomer complex. Thus, it is difficult to
separate, but it may not affect crystallization as these complexes will
interchange and crystallize in one form. There was a limited amount for the
76

preliminary purification of mouse α211/Fab210/α-bungarotoxin; therefore,
more fractions were collected. However, more selected fractions of the main
peak were pooled and concentrated down to set trays in later experiments.

Figure 3.3: The α211/Fab210/α-bungarotoxin complex purification.
(A) Chromatogram of Superdex 200 10/300 of the mouse α211 ternary complex with
Fab210. Fractions 23 – 28 were pooled and concentrated down for crystallization. (B)
Chromatogram of Superdex 200 10/300 of the human α211 ternary complex with
Fab210. Fractions 26 – 30 pooled and concentrated down for crystallization. (C)
Overlap chromatograms of Fab35 (in blue) and Fab210 α211 (in red) ternary
complex purification over Superdex 200 10/300. (D) 15% SDS-PAGE gel of mouse
α211/Fab210/α-bungarotoxin ternary complex purification. (16 μL of sample + 4 μL
of 5x loading dye except Injection: 1 μL of sample + 7 μL of water + 2 μL of 5x
loading dye). All fractions show the right ratio of the ternary complex formation.
A B
C
D
77

3.3.4 Crystallization of the α211/Fab210/α-bungarotoxin
ternary complex
First, crystallization of the ternary complex of mouse α211/Fab210/α-
bungarotoxin was attempted by screening around the condition (0.1M sodium
cacodylate trihydrate pH 6.5, 0.1 M – 0.15 M calcium acetate hydrate and
18% – 20% PEG 8K) in which α211/Fab35/α-bungarotoxin was crystallized
instead of screening with ordinary screening kits. However, no crystals were
obtained in those screening conditions; therefore, crystallization conditions
were re-screened using the screening kits with 2.5 mg/mL concentration of the
purified complex protein by the hanging drop method at room temperature
(total 192 conditions). Among them, crystals grew in only one condition, 0.1
M sodium citrate pH 5.6, 20% isopropanol, and 20% PEG 4K. Initial
preliminary crystals were small and morphology was not as desired (Figure
3.4). Even though they were not the best crystals, they were birefringent and
diffracted at the APS 23-ID-B with a limited diffraction.

Figure 3.4: Preliminary crystals of the mouse α211/Fab210 /α-bungarotoxin.
One of the crystal screen kits gave initial crystals of the Fab210 ternary complex.
Left: Showing an entire drop. Right: Zoomed in view of a drop. It is focused on some
crystals. Diffraction was tested using these crystals at the APS 23-ID-B.
78

Next, the crystallization condition was optimized by narrowing
screening around the initial condition. Rod-like shaped crystals were again
obtained with the size of 10 μm x 100 μm – 40 μm x 300 μm. Some of the
crystals were similar to the ones for the Fab35 ternary complex, but some
were wider and thicker (Figure 3.5)

Figure 3.5: Optimized crystals of the α211/Fab210/α-bungarotoxin.
Top two drops: Crystals of mouse α211/Fab210/α-bungarotoxin. Bottom two drops:
Crystals of human α211/Fab210/α-bungarotoxin. Some crystals were wider and
thicker than the ones of the Fab35 ternary complex. Background precipitations were
also observed.

Crystals grew with background precipitations as with Fab35 ternary
crystals. Crystals formed within 3 days and continued to grow for a week.
They were fairly stable, and crystal degradation was not observed within a
Mouse α211
Ternary Complex
Human α211
Ternary Complex
100 µm
79

month or even longer. Protein freshness (approximately 3 weeks of storage at
4 °C) and protein quality definitely affected crystallization, and it was slightly
harder to obtain crystals compared to Fab35 ternary crystals. Human
α211/Fab210/α-bungarotoxin crystals were difficult to obtain, and crystals
tended to grow as a bulk. Some crystals had measurable width, and looked
more plate-like than bar-like crystals.
There was difficulty in harvesting crystals due to cryo solutions. Since
isopropanol and PEG 4K works for cryo, glycerol was not added in the
original cryo solution. Crystals were gradually transferred to cryo solution
(1:1 reservoir/cryo solution and 100% cryo solution) with 5 – 10 minutes
incubation time at each condition. While crystals were harvested, unusual
phenomena were observed. Bubbles formed around the harvesting well
solution, and eventually the bubbles trapped crystals (Figure 3.6). Crystals
were degraded in the bubble, and eventually disappeared, a sign of the phase
separation in the cryo solution. After realizing the problem, the cryo solution
was left to sit some time, and a clear phase separation was observed in the
tube (Figure 3.1). It was important to determine a good ratio of isopropanol to
PEG 4K. It probably occurred by PEG 4K pulling water resulting in
isopropanol being repelled away from aqueous solution. Since sodium
chloride also could affect phase separation, the concentration of sodium
chloride was decreased to 0.1 M or removed entirely. Four different
conditions were newly tested, and two conditions (Cryo 1 and Cryo 2) did not
show phase separation (Table 3.1). One major difference between Cryo 1 and
2 was the addition of 10% glycerol to replace the cryo protection which was
lost by lowering the concentrations of isopropanol and PEG 4K. Crystals were
harvested using both cryo solutions to see if any difference would be observed
in diffractions. Although both cryo solutions were good, phase separation had
80

started by the time of harvesting. Therefore, it was important to harvest
crystals within a certain time period to minimize the phase separation.

Figure 3.6: Phase separation problem of cryo solution.
Left: Right after transferred into a cryo solution. Intact crystals can be seen. Middle:
Some bubbles were observed during an incubation time. Eventually, crystals were
trapped in the bubble and disappeared. Right: Phase separation was observed in a
cryo solution tube. The top layer is aqueous solution and the bottom layer is
isopropanol.

Table 3.1: Cryo solution test for phase separation.
Several cryo solutions were made and tested for phase separation. Concentrations of
isopropanol and PEG 4K were mainly tested. The concentration of sodium chloride
was lowered since it could contribute to the phase separation. Two conditions were
found to be good, and a major difference was glycerol content.

Original Cryo 1 Cryo 2 Cryo 3 Cryo 4
Sodium citrate, pH5.6 0.1 M 0.1 M 0.1 M 0.1 M 0.1 M
Isopropanol 30% 20% 15% 20% 25%
PEG 4K 25% 20% 20% 30% 20%
Sodium chloride 0.15 M 0.1 M 0.1 M ─ 0.1 M
Glycerol ─ ─ 10% ─ ─
Result Separation Good Good Separation Separation
Aqueous
Isopropanol
81

3.3.5 Structure determination of the α211/Fab35/α-
bungarotoxin ternary complex
The diffraction data were collected at the Advanced Photon Source
(APS) beamline 23-ID-B at Argone National Laboratory. The Data were
scaled and merged using HKL2000 package
160
. Interestingly, structures were
solved with two different space groups, P1 and C2. It was consistent that both
crystals of mouse α211 and human α211 ternary complex harvested with Cryo
1 solution belong to the space group P1, and those harvested with Cryo 2
solution belong to the space group C2. Human α211/Fab210/α-bungarotoxin
crystals in Cryo 2 were solved at the best resolution. Two difference space
groups could be due to the different cyo solutions. Cryo 1 could have affected
the crystal packing; therefore crystals lost 2-fold symmetry. Preliminary data
sets are summarized in Table 3.2.
human α211 mouse α211 human α211 mouse α211
Cryo solution Cryo 1 Cryo 1 Cryo 2 Cryo 2
Space group P1 P1 C2 C2
a 41.4 Å 42.0 Å 178.3 Å 177.4 Å
b 39.9 Å 89.6 Å 41.4 Å 41.9 Å
c 125.0 Å 125.9 Å 123.9 Å 125.8 Å
α 104.2° 75.0° 90° 90°
β 91.2° 90.0° 105.2° 104.6°
γ 100.0° 79.1° 90° 90°
Resolution 2.9 Å 3.7 Å 2.6 Å 3.2 Å

Table 3.2: Summary of unit cell dimensions of each complex crystal.
Crystals of both mouse α211 and human α211 (Cryo 1 solution) were solved as P1
space group, and ones (Cryo 2 solution) were solved as C2 space group. Cryo 2 gave
better diffraction/resolution in both mouse and human α211 complex crystals. Those
numbers are from preliminary data set, not from the final refined data.
82

The scale files were converted to mtz files using Scalepack2mtz in
CCP4 package
161
Structures of both human and mouse ternary complexes
were solved in a similar manner as discussed in the determination of the
Fab35 ternary structures (Chapter 2, section 2.3.7). The human Fab210 ternary
structure is discussed here as an example.
One set of ternary complex molecule was detected in an asymmetric
unit with the Matthews_coef program. In order to select the best search model
for Fab210, previously solved three Fab structures (Fab192, Fab198, and
Fab35) were tested. Among them, Fab35 provided the best R values.
Therefore, molecular replacement was performed using the Fab35 ternary
complex structure, as a search model, which was solved earlier as described in
Chapter 2. The α1/toxin complex and the Fab35 of the human Fab35 ternary
structure were separated as ensembles using Pdbset, and two of them were
added into the Phaser MR program
162, 163
(Ensemble #1: α1/toxin complex and
Ensemble #2: Fab35). A single solution was obtained with the molecular
replacement (RFZ=16.5 TFZ=13.4 PAK=0 LLG=301 RFZ=13.3 TFZ=29.3
PAK=0 LLG=210 LLG1448). After one round of refinement with Refmac5
164,
165
, R
work
/R
free
was 27.4/34.2 at 2.6 Å cut-off. Even though the CDR3 loop of
the heavy chain was similar between Fab35 and Fab210 compared to that of
Fab192 or Fab198, reconstruction was still necessary. The amino acid
sequence alignment of Fab35 and Fab210 showed that Fab210 had two extra
amino acids at the loop. Thus, these two amino acids were inserted using O
166
.
The side chains were replaced based on the Fab210 sequences using Coot
167
,
and the water molecules were added using O
166
. The structure was analyzed
using Procheck. Ramachandran Plot can be found in Appendix A. After
multiple refinements, the final R
work
/R
free
was 19.9/27.7. The resolution cut-off
was determined at I/σ I = 1, and values of crystallographic analysis are shown
in Table 3.3
83

Human α211/Fab210/α-Btx Mouse α211/Fab210/α-Btx
Data collection

Space group C2 C2
Cell dimensions

a, b, c (Å) 178.3 177.4

41.4 41.9

123.9 125.8
α, β, γ (
o
) 90.0, 105.2, 90.0 90.0, 104.6, 90.0
Resolution (Å) 45.0 – 2.60 45.0 – 3.20

(2.64 – 2.60) (3.26 – 3.20)
R
sym
(%) 7.3 (92.5) 12.5 (93.9)
I/σI 14.9 (0.8) 8.9 (0.8)
Completeness (%) 99.8 (99.6) 99.9 (99.9)
Redundancy 3.3 (3.1) 3.2 (3.1)

Refinement

Resolution (Å) 45.0 – 2.60 45.0 – 3.20
No. reflections 26,697 14,837
R
work
/R
free
19.9/27.7 24.6/34.6
No. atoms

Receptor 1852 1861
αBtx 553 553
Light chain 1631 1622
Heavy chain 1656 1653
Sugar

Water 67 n/a
B-factors (Å
2
)

Receptor 81.8 95.0
α-Btx 129.1 136.5
Light chain 73.3 103.4
Heavy chain 73.8 106.9
Sugar

Water 75.6 n/a
R.m.s deviations

Bond length (Å) 0.011 n/a
Bond angles (
o
) 1.617 n/a

Table 3.3: Crystallographic analysis of human or mouse α211/Fab210/α-
bungarotoxin complexes.
Data was processed with HKL2000 package, and structure was solved by molecular
replacement using the previously solved Fab35 ternary structure followed by
refinement with Refmac 5. Highest resolution shell is shown in parenthesis.
84

3.3.6 Structure guided analysis of α211/Fab210 interactions
The X-ray crystal structure of nAChR of α1/Fab210/α-bungarotoxin
was solved at 2.6 Å, and detailed interactions of the interface were determined.
Similar interactions were observed at some parts, but differences were also
found. The structure of the Fab210 ternary complex was superimposed with
the one of the Fab35 ternary complex (Figure 3.7). It showed that the main
targeting site of both mAbs was the MIR as expected. The MIR motif was
inserted into a surface pocket between the light chain and the heavy chain of
Fab210. Other binding features were also observed, which were also
discovered in the Fab35 ternary structure. The N-terminal α-helix of α1 also
contributed to the binding as well as the CDR3 of the heavy chain which is
inserted into a groove of α1. Although the orientations of Fab35 and Fab210
were slightly different, the overall structures were similar. The CDR-H3 loop
showed some differences (Figure 3.7, B), but it was not as big difference as
seen in Fab192 or Fab198 models (Figure 2.10). Therefore, Fab35 and Fab210
could be categorized in the same group in terms of a targeting epitope or
binding patterns. However, further detailed analyses are necessary to explain
why Fab210 is different than Fab35.
In order to determine detailed interactions, two programs, CCP4
contact and PISA were again used as with the Fab35 ternary structure. The
distance cut off was set at 4.5 Å for the CCP4 contact program. Table 3.4 lists
contacting residue pairs between α1 subunit and Fab210, and Table 3.5 has
combined lists of both Fab35 and Fab210 for comparison. The Fab210 ternary
structure was analyzed according to those results from the programs though
those are not absolute answers. It was necessary to consider bond angles or
surrounding interfacial water molecules. Detailed interactions at the interface
were determined for the Fab210 ternary structure, and any differences from
Fab35 binding were discussed.
85

Figure 3.7: Superimposed view of the human Fab35 ternary structure and the
human Fab210 ternary structure.
(A) The structure of human α211/Fab210/α-bungarotoxin (purple: α211, dark green:
α-bungarotoxin, red: Fab210: light chain, orange: Fab210 heavy chain) is
superimposed with the structure of human α211/Fab35/α-bungarotoxin (cyan: α211,
green: α-bungarotoxin, magenta: Fab35 light chain, yellow: Fab35 heavy chain)
using the C-alpha backbone of human α1 of Fab35 ternary complex. Fab35 portion is
shifted to left-down side, and the gap is larger at the constant regions of Fab (C
H
and
C
L
). The overall structures are similar; Fab210 also interacts heavily at the MIR. (B)
Differences at the CDR-H3 interaction site. There is a little shift of the loop though it
is not a major difference.

A B
86

Fab210 Ternary Interaction Table

Light Blue: Hydrogen bonds
Pink: Both hydrogen bonds and salt bridges

Table 3.4: List of contacting residue pairs between α1 subunit and heavy or
light chain of Fab210.
Two different programs, contact and PISA were run to determine contacting residues
at the interface of α1 subunit and Fab210 within 4.5Å. The table shows are combined
results of the two programs.

α211 (B) Light Chain (C) Heavy Chain (D)
H3 Y56
R6 W52, D54, Y56
L7 M104
K10 W52, D54, R100, R103, M104
L11 R103, M104
K13 R103
D14 R103
N64 R103
K66 M104, Y106
W67 D92 M104, Y106
N68 Y91, D92, N93, G94, Y95 Y106
P69 D92, N93
D70 N1, N93, G94, Y95 A58, Y59
D71 Y95 R50, W52, Y56, M104, G105, Y106, W107
Y72 Y56, M104
G73 Y56
Y112 Q27, N93
87

Fab35 vs. Fab210 Ternary Interaction Table

Light Blue: Hydrogen bonds
Pink: Both hydrogen bonds and salt bridges
Table 3.5: Summary and comparison of contacting residue pairs between α1
subunit and heavy or light chain of Fab35 and Fab210.
Two different programs, contact and PISA were run to determine contacting residues
at the interface of α1 subunit and Fab210 within 4.5Å. The table shows are combined
results of the two programs.

α211 (B)
Light Chain (C)
Fab35
Light Chain (C)
Fab210
Heavy Chain (D)
Fab35
Heavy Chain (D)
Fab210
H3 T57, V58 Y56
R6
W52, D54, G56,
T57
W52, D54, Y56
L7 A103 M104
K10
W52, D53, D54,
R100, R102,
A103
W52, D54, R100,
R103, M104
L11 A103 R103, M104
K13 R103
D14 R102 R103
Y15 R102
Y63 Y32, K50
N64 R102 R103
K66 Y32 A103, I104 M104, Y106
W67 I92 D92 A103, I104 M104, Y106
N68
Y91, I92, N93,
G94, Y95
Y91, D92, N93,
G94, Y95

Y106
P69 I92, N93 D92, N93
D70 G94, Y95
N1, N93, G94,
Y95
W47, V58 A58, Y59
D71 Y95
Y95 W47, R50, W52,
V58, A103, N105
R50, W52, Y56,
M104, G105,
Y106, W107
Y72
W52, V58,
A103, N105
Y56, M104
G73 V58 Y56
Y112 Q27, N93
88

The major interactions sites were the same, and the Fab210 ternary
structure also showed interactions at three main sites, the MIR, N-terminal α-
helix, and CDR-H3. In addition, interfacial water molecules were also found
in this structure though locations or binding patterns of those waters were
different except one (Figure 3.8, A) Figure 3.8, B shows a overall view of the
interacting site at the N-terminal α-helix of α1, and Figure 3.8, C is a zoomed-
in view of the site. Arg6 forms a cation-π stacking with Tyr56 of V
H,
which
forms hydrogen bonds with His3 of α1 and the main chain amide of Gly73 of
α1. These interactions are then further stabilized by an interfacial water
molecule H2 which forms hydrogen bonds with His3 and Arg6 of α1. The
guanidinium group of Arg6 of α1 also packs face-to-edge to the aromatic ring
of Trp52 of V
H
, which in turn in backed up by Trp107 of V
H
. Trp52 of V
H

forms a hydrogen bond with an interfacial water molecule H5, which is also
make another hydrogen bond with the main chain carbonyl of Asp71 of α1.
Finally, Asp54 of V
H
forms salt bridges with Arg6 and Lys10 of α1. Here,
interesting findings were the orientation and interaction patterns of Arg6 of α1
and its interaction pattern. Several aromatic residues are packed creating a
strong interaction (Figure 3.8, D). Such aromatic packing interactions were
seen at the interface of α-bungarotoxin and α1
30
. This interaction was
completely different from that of Fab35. Arg6 of α1 forms a cation-π stack
interaction with W52 of V
H
in the Fab35 ternary structure, whereas Arg6 of
α1 packs face-to-edge interaction with W52 of V
H
in the Fab210 ternary
structure. It is a due to two amino acid residues in Fab210, Y56 and W107.
Those aromatic rings create an extremely strong binding at the site. The
differences in those binding patterns can be clearly identified by comparing
two structures (Figure 3.9). Figure 3.10 shows the electron density map of the
aromatic packing interactions.

89

Figure 3.8: Interactions at the N-terminal α-helix mediated by interfacial water
molecules.
(A) Three important interfacial water molecules are identified between α1 (purple,
MIR is colored in orange) and Fab210 (heavy chain: orange and light chain: red) and
are shown in a cyan sphere (H2, H5, and H49). (B) The overview of the interaction
site at the N-terminal α-helix. (B) R6 of α1 forms a cation-π stacking with W56,
which forms a hydrogen bond to H3 of α1 and the main chain amide of G73. This
interaction is further stabilized by a hydrogen bond with interfacial water molecule
H2 which forms another hydrogen bond with R6 of α1. R6 then packs face-to-edge to
the aromatic ring of W52 V
H
, which in turn backed up by W107 of V
H
. The
interaction between W53 of V
H
and the main chain amide of D71 of α1 is mediated
by another interfacial water molecule H5 with hydrogen bonds. Finally, D54 of V
H

forms salt bridges with K10 and R6 of α1.
A B
C D
90

Figure 3.9: Comparison of the interfaces (Fab35 vs. Fab210) at the N-terminal
α-helix.
(A) The interface of Fab35 and the N-terminal of α-helix of α1 MIR (cyan: α1,
orange: MIR, yellow). R6 of α1 forms a cation-π stack interaction with W52 of V
H
.
Surrounding residues also contribute to the binding. (B) The interface of Fab35 and
the N-terminal of α-helix of α1 (purple: α1, green: MIR, orange: V
H
). R6 forms
cation-π stacking with Y56 of V
H
, and fate-to-edge with W52 of V
H
which is
supported by another aromatic ring of W107 of V
H
. These aromatic packing
interactions provide tight binding, and they can be seen in the binding interface of
toxins.

Figure 3.10: Electron density
map of interactions at the N-
terminal α-helix.

The electron density map at the
N-terminal α-helix of α1
(
yellow:
α1, green: V
H
, and magenta
cross: water). The orientation is
the same with Figure 3.9, B for
comparison. Aromatic packing
interactions are clearly shown.

A B
91

The next binding interface to be discussed is at the MIR which has
been believed to be a target of MG autoantibodies. Similar binding patterns
were observed in some interactions, but differences were also found (Figure
3.11). First, a similar interaction through interfacial water molecule was found
(Figure 3.11, B). A water molecule H49 forms hydrogen bond with Asp71 of
α1 and the main chain carbonyl of M104 of V
H
, which in turn forms a
hydrogen bond with Tyr72 of α1. A difference was a binding pattern of Asp71
of α1 and Arg50 of V
H
. In Fab35 structure, they were directly interacted with
a salt bridge; however, the distance between two residues was far, and they
interacted through an interfacial water molecule, H5. The water molecule then
forms hydrogen bonds with Trp52 of V
H
and the main chain carbonyl of
Asp70 of α1. On the other hand, Asp71 of α1 interacted directly with the main
chain amide of Tyr106 of V
H
though the interaction was mediated by a water
molecule in the Fab35 structure. The interaction between Asp71 of α1 and
Tyr95 of V
L
was consistent. Lastly, Tyr106 of V
H
interacts with a loop of α1
around the MIR and forms a van der Waals contact. The side chain interacts
by sticking into a cave of α1. This interaction was unique since there is Trp at
the position in Fab35 and the orientation of the side chain is different.
Therefore, such an interaction could not be observed in the Fab35 ternary
structure.
Another important residue at the MIR, Asn68 also showed the same
pattern of interaction (Figure 3.11, C). Asn68 forms hydrogen bonds to the
main chain carbonyl of Tyr91 and Gly94 of V
L
as well as to the main chain
amide of Gly94 of V
L
. Electron density maps of binding interactions at Asp71
and Ans68 of α1 subunit are also shown in Figure 3.12, which also indicates
differences of interaction patterns at Asp71. A different feature of interaction
was found around the site, and it was not observed in the Fab35 ternary
structure (Figure 3.11, D). Asp70 of α1 forms a hydrogen bond to Asn1 of V
L

92

and van der Waals contact to Gly94 of V
L
. This interaction could not be
observed, since a residue at position 1 in Fab35 is Asp1. Another unique
interaction at the area was that the adjacent Pro69 of α1 makes a van der
Waals contact Asn93 of V
L
. Those residues are also found in the Fab35
ternary structure; however the distance of those two residues were far to form
a van der Waal contact. As a unique interaction of Fab210, Figure 3.13, A
shows that Tyr112 of α1 forms hydrogen bonds with Gln27 and Asn93. It is
located at under the MIR loop, and those residues are conserved in Fab35.
However, those residues are again not in close proximity to form hydrogen
bonds in the Fab35 structure (Figure 3.13, B).
Lastly, the CDR3 of the heavy chain strongly contributes to the
binding of Fab210. As observed in the Fab35 ternary structure, the loop of
CDR-H3 is inserted in a groove of α1 (Figure 3.14). M104 inserts deeply into
a hydrophobic pocked formed by the aliplatic side chain of Lys10, Leu7,
Leu11, Trp67, and Tyr72 of α1. This interaction was observed in the Fab35
ternary structure (Figure 2.14), but the inserting amino acid was Ala which is
much shorter than Met. It was also observed an interaction with Arg104 of
V
H
. There was Arg at the tip of CDR-H3 loop of Fab35, but interacting patters
are different between two. Arg103 forms hydrogen bonds with the main chain
carbonyl of Lys10, Leu11, Asp14, and Asn64 of α1. Its side chain is also in
van der Waals contact with Lys10, Lys13, and Asp14 of α1. In addition, R103
of α1 can form the electrostatic interaction with Asp14 of α1.
93

Figure 3.11: Binding interactions of α1 subunit and Fab210 at the MIR.
(A) The overview of the interactions at the atomic cage of the MIR (purple: α1,
green: MIR, orange: V
H
, red: V
L,
and cyan sphere: water). (B) A zoomed in view of
the MIR. Two interfacial water molecules, H49 and H5 forms hydrogen bonds with
surrounding residues (H49: D71 and N68-NH of α1 and M104-CO of V
H
, H5: D71
and D70-CO of α1, R50 and W52 of V
H
). The water molecule H5 is unique in
Fab210 ternary structure. R50 of V
H
and D71 of α1 interacts through a salt bridge in
the Fab35 structure, but water mediated interactions are seen in the Fab210 structure.
(C) N68 of α1 additionally forms hydrogen bonds with the main chain carbonyls of
Y91 and G94 of V
L
. It also forms another hydrogen bond with the main chain amide
of G94 of V
L
. These interactions are consistent with that of Fab35. (D) Unique
interactions of Fab210 are found. D70 of α1 forms a hydrogen bond with N1 of V
L
,
and D70 of α1 also makes van der Waals contact to Gly94 of V
L
. Pro69 of α1
interacts to N93 of V
L
through van der Waal as well.

A B
C D
94

Figure 3.12: Comparisons of electron density maps of binding interactions at
the MIR (Fab35 ternary vs. Fab210 ternary).
Interactions at D71 and N68 are compared between the Fab35 ternary and the Fab210
ternary structures (Fab35 - yellow: α1, green: V
H
, blue: V
L,
and magenta cross: water
and Fab210 - yellow: α1, blue: V
H
, green: V
L,
and magenta cross: water). Top two
figures show differences of interaction patterns at D71. D71 of the Fab210 ternary
complex forms hydrogen bonds with Y95 of V
L
and the main chain amide of Y106 of
V
H
. D71 then interacts with R50 of V
H
through a water molecule. Interaction patterns
at N68 are very similar. The electron density between N68 and the main chain
carbonyl of Y91 can be clearly seen in the Fab210 ternary structure.
D71
N68
Fab35 Ternary
Ternary
Fab210 Ternary
Ternary
95

Figure 3.13: Unique interaction site of the Fab210 ternary structure.
(A) Y112 of α1 forms hydrogen bonds with Q27 and N93 of V
L
(purple: α1, green:
MIR, orange: V
H
). This site is located at the beneath of the MIR loop. (B) Those
three residues are conserved in Fab35; however such interactions cannot be formed in
the Fab35 structure. Superimposed view shows that Q27 and N93 of V
L
are far from
Y112 of α1 to form contacts (Fab210 - purple: α1, green: MIR, orange: V
H
and
Fab35 - cyan: α1, orange: MIR, yellow).

A
B
96

Figure 3.14: Binding interactions with insertion of CDR-H3.
(A) The overview of the interaction site (purple: α1, green: MIR, orange: V
H
). (B)
M104 of α1 is inserted into a hydrophobic pocked defined by L7, L11, W67, Y72,
and the aliphatic side chain of K10 of α1. A similar interaction is observed in the
Fab35 ternary structure, but the side chain inserted is Ala which is much shorter than
Met. (C) R103 of α1 forms hydrogen bonds with the main chain carbonyl of K10,
L11, D14 and N64 of α1. The side chain of R103 also forms van der Waals contact
with K10, K13, and D14 of α1. R103 and D14 of α1 are close enough for
electrostatic interaction.

A
B
C
97

3.4 Discussion
The X-ray crystal structure of Fab210/α1 ECD/α-bungarotoxin was
successfully solved. It was the second crystal structure which is a MG mAb
bound to α1. As the second complex structure, mAb210 was picked since it
showed a major difference in binding to α1. Both mAb35 and mAb210 can
bind to intact nAChRs; however, mAb35 cannot bind to neither synthesized
peptides nor denatured nAChRs, whereas mAb210 can. To answer the
question, the complex structure of Fab210 was attempted to solve using X-ray
crystallography. All preparations were done in the same manner with the
Fab35 ternary structure. Some differences were observed during the ternary
complex purification and crystallization. In the purification step, a distinct
peak which could be resulted from oligomerizaion was not observed in the
Fab210 ternary complex (Figure 3.3). Crystallization conditions were also
different from that of the Fab35 ternary complex. However, similar
morphologies of crystals were obtained within a week, and they were
diffracted at the Advanced Photon Source (APS) beamline 23-ID-B at Argone
National Laboratory. Both human and mouse Fab210 ternary structures were
solved, and differences of Fab35 and Fab210 were determined using the
human ternary complexes of Fab35 and Fab210.
The overall structures of two were similar, and main binding sites were
also similar. Fab210 interacted extensively with the MIR loop, and some
interaction patters were similar with that of Fab35. However, Fab210 had
much more extensive interactions through highly inter-connected interactions
in a small area. This would be one reason that mAb210 can bind to peptides.
mAb210 can target a focused area with multiple binding interactions. As a
result, peptides could form a confirmation at some degree for mAb210 to bind
to them. Next, a major difference was seen at the N-terminal of α-helix of α1
98

(Figure 3.8 and 3.9). Aromatic packing interactions were observed, and such
interactions can be seen in a toxin binding to the receptor
30
. Those interactions
were not found in the Fab35 ternary structure since important aromatic
residues, Y56 and W107 are missing in Fab35. This rigid structure formed by
those aromatic rings could also enable mAb210 to bind to peptides. It would
facilitate to fold the peptides to the right confirmation; therefore mAb210
easily can bind. It works as a framework of peptides, and even liner peptides
could have a confirmation in the pocket. More interactions were observed at
the interface such as the loop of CDR-H3. Interaction patterns of the loop
were similar between two Fabs, but intensities would be different. For
example, Fab210 inserts Met into a hydrophobic pocket of α1 while it is Ala
in Fab35. Even though many differences were seen, a critical difference
would be that Fab210 CDRs contain multiple aromatic residues. And, those
aromatic rings pack tightly with each other as well as with residues on α1.
Figure 3.15 shows aromatic residues which significantly contributed to
binding to α1. Left represents the Fab35 ternary structure and the right
represents the Fab210 ternary structure. It is clear that so many aromatic rings
are clustered at a focused area of α1 in the Fab210 structure. mAb210 can bind
to peptides or denatured protein due to this extremely focused and strong
interactions. It was a very interesting finding, and finally the answer is here.
As a next target MG mAb, mAb192 would be interesting since it also
shows unique results in previous studies
132, 171
. It showed that its binding did
not require the MIR loop though it could bind at a close enough distance since
it competes binding with mAb35
132
. The X-ray crystal structures could give an
answer which could not be obtained by biochemical experiments. Also, those
detailed insights to structural information would lead the therapeutic
developments.
99

Figure 3.15: Binding interaction with aromatic residues.
The left is the Fab35 ternary structure (cyan: α1, orange: MIR, yellow), and the right
is the Fab210 ternary structure (purple: α1, green: MIR, orange: V
H
). Aromatic
residues which are showed critical interactions are shown with sticks. It is obvious
that Fab210 has clustered aromatic rings at a focused area of the binding site. Even
though Fab35 also has many aromatic rings for interactions, they are spread out.
Important residues of CDR-H3 are also shown with sticks. It is clearly shown that
they insert into a groove of α1.

100

3.5 Materials and Methods
3.5.1 α211 expression and purification
The construct information of α211 mouse and α211 human can be
found in Chapter 2 materials and methods (section 2.5.1) as well as protocols
of α211 expression and α211 purification (section 2.5.2 and 2.5.3).
3.5.2 mAb210 and Fab210 production
Hybridoma cells of mAb210 were generously provided by Dr. Jon
Lindstrom, Department of Neuroscience, Perelman School of Medicine,
University of Pennsylvania. The cells were cultured in a similar manner as for
mAb35. mAb210 was purified and papain digestion was performed to obtain
Fab210. Refer to Chpter2 materials and methods (sections 2.5.4, 2.5.5, and
2.5.6) for detailed purification protocols.
3.5.3 Gel shift assay
The gel shift assay was performed following the same protocol for
mAb35 gel shift assay which can be found in Chapter 2 materials and methods
(section 2.5.7).
3.5.4 The α211/Fab210/α-bungarotoxin complex purification
α211, α-bungarotoxin and Fab210 were mixed at a 1:2:2 molar ratio,
and the mixture incubated on ice for 1 – 2 hours. Ni-NTA purification was
performed to remove excess α-bungarotoxin and Fab210. The elution was run
over a size exclusion column (Superdex 200 10/300 GL, GE Healthcare) with
150 mM NaCl and 20 mM HEPES, pH 7.5 buffer. A broad left shoulder was
101

observed, and the main peak fractions were pooled and concentrated for
crystallization.
3.5.5 Crystallization and data collection
The purified and concentrated ternary complex of α211 (both mouse
and human)/Fab210/α-bungarotoxin was diluted to 2.5 mg/mL. Initially,
crystallization was attempted by screening around the condition in which the
Fab35 ternary complex was crystallized (0.1 M sodium cacodylate trihydrate
pH 6.5, 0.1 M – 0.15 M calcium acetate hydrate and 18% – 20% PEG 8K).
Since crystals didn’t form in those conditions, crystals were re-screened using
Crystal Screen, Crystal Screen 2 and Index kits (Hampton Research) by the
hanging drop method at room temperature. Each drop contained 0.5 μL of
protein and 0.5 μL of reservoir solution, and 0.5 mL of screening solution was
added in each well. Only one condition gave initial crystals, which was 0.1 M
sodium citrate pH 5.6, 20% PEG 4K, and 20% isopropanol. The condition was
optimized and diffracting crystals were obtained in 0.1 M sodium citrate pH
5.6, 13% – 14% PEG 4K, and 18% – 26% isopropanol. Several concentrations
of the protein were tested (2.5 mg/mL – 5.7 mg/mL). Rod-like shaped crystals
were grown as bundle with the size of 10 μm x 100 μm – 40 μm x 300 μm.
Crystals were harvested using two different cryo solutions: Cryo 1 containing
0.1 M sodium citrate pH 5.6, 20% PEG 4K, 20% isopropanol, and 0.1 M
sodium chloride, Cryo 2 containing 0.1 M sodium citrate pH 5.6, 15% PEG
4K, 20% isopropanol, 0.1 M sodium chloride, and 10% glycerol. Since PEG
4K and isopropanol work as a cryo, crystals were directly transferred to a
solution which was 1:1 mixture of reservoir solution and Cryo 1 (Cryo 2)
solution to prevent from osmotic shock. There was a phase separation issue in
this crystallization/cryo condition; therefore, it was critical to harvest in a
102

certain time range. Crystals were incubated in each solution for less than 5
minutes and harvested as quickly as possible.
Data were collected at the Advanced Photon Source (APS) beamline
23-ID-B at Argone National Laboratory using a 20 x 20 beam (λ = 1.0332 Å,
12.000 keV) with the attenuation factor of 2.0 and a Pilatus 3 6M detector
(423.6 x 434.6 mm
2
sensitive area, 172 x 172 µm
2
pixel size) (DECTRIS Ltd.,
Baden, Switzerland). The detector distance was 350.0 mm. The oscillation
range and the exposure time per frame were 0.5° and 0.5 sec, respectively.
Data were processed and scaled using HKL2000 package. The phases were
determined by molecular replacement using the coordinates of two α211/α-
bungarotoxin complex; mouse α211 (PDB: 2QC1) and human α211 (solved
structure in Chapter 2). Also, three Fabs, Fab198 (PDB: 1FN4), Fab192
(PDB: 1C5D), and Fab35 (solved structure in Chapter 2), were compared for
the best R value. As a result, Fab198 was used to solve four different Fab210
ternary structures (mouse/human α211 with two difference space groups).
Refmac was used to refine the structures. The sequence of mAb210 was
obtained from an antibody sequencing company, MCLAB Molecular
Laboratories (San Francisco, CA), and Fab210 side chains were added using
Coot. Extra amino acids in Fab35 were inserted with O, and further structural
model buildings were also carried out in O.

103

Chapter 4

Development of a MG Antibody
Detection Method Using ELISA

4.1 Abstract
Diagnosis of MG is challenging since most of symptoms share similar
characteristics with other neuromuscular diseases. A main cause of MG is
autoantibodies targeting MIR, and concentration of the autoantibodies related
to MIR binding has a correlation to disease severities. It is important to
measure the concentration accurately with high sensitivity. Current method
mostly used is radioiimunoassay which requires radioactive materials. The
aim of this study is to develop more time efficient, cost effective and easy-to-
use method; therefore, it could be widely used not only to a diagnosis but also
follow up of the disease. This chapter discusses a development of a novel
method to detect autoantibodies causing MG by taking advantage of the
highly purified nAChR ECD that was generated from our crystallography
studies.

104

4.2 Introduction
There are several tests to diagnose myasthenia gravis, MG, including
electrophysiological tests, the tensilon (edrophonium chloride) test which
checks the effect of cholinesterase inhibitors, and autoantibodies test
108
.
Giving some stress by physical force or electrical stimulation to a muscle on
MG patients, the strength of that muscle would decrease due to the damaged
neuromuscular junction. On the other hand, the injection of cholinesterase
inhibitors would improve some MG symptoms such as eyelid ptosis or limb
weakness since they block the activity of the enzyme breaking down the
neurotransmitter, acetylcholine. Diagnosis of MG is sometimes difficult
especially at early stages because some symptoms are similar to other
neuromuscular disorders
108
. The detection of autoantibodies to nAChR is one
common test used for MG diagnosis. However, some patients are seronegative
for anti-nAChR. In that case, other autoantibodies targeting different proteins
are tested while other diseases such as amyotrophic lateral sclerosis (ALS)
and polymyositis are suspected to affect neurotransmission
107
. Thus, a series
of tests are usually necessary for an accurate diagnosis.
The most powerful and sensitive method used for diagnosing MG is
radioimmunoassay (RIA). Though it is commonly used in a clinical setting to
measure the concentration of anti-nAChR antibodies, there are some
disadvantages. It often uses human muscles which are obtained from
amputated limbs as a receptor source. It would not be easy to have a constant
supply of human muscles. Another disadvantage is a usage of radioactive
material. For detection,
125
I-labeled α-bungarotoxin is used, and it requires
extra safety caution. The radioactive material has a limited half-life, so it
would not be cost effective. Although there are some drawbacks in RIA, it is
useful since it is able to measure concentration of anti-nAChR with high
105

sensitivity. It has been suggested that the autoantibody concentration and the
disease severity are correlated. Some studies have been performed to
determine a correlation of autoantibodies and disease severities. A correlation
was not seen in some studies
95, 99, 154, 172-174
, but some other studies saw a
decreasing in concentration of autoantibodies with clinical improvement
122, 123,
175
. Decrease of autoantibodies was seen with some MG therapies.
Plasmapheresis is one of therapies for MG to remove antibodies including
autoantibodies from patients’ blood. Improvement of symptoms with the
therapy indicates the correlation of antibody concentration and the severities
of MG. A study also suggested that there was no correlation with specificity
of autoantibodies and disease severities
141
. Most of patients, both ocular MG
and generalized MG, have autoantibodies mostly targeting the MIR.
Antibodies specificity would not determine a type of MG or severities of the
disease. Recently, a study showed the correlation of autoantibody
concentration and MG severity, and it was shown a difference between ocular
and generalized MG
124-126
. It would be useful if autoantiboies causing MG
could be accurately measured with more simple assays since there are some
difficulties to diagnose MG with current methods. It would be also important
not only to determine the severity but also actual pathogenesis (e.g. anti-
nAChR antibody or anti-MuSK) for proper treatments. Some neuromuscular
diseases could show similar symptoms; therefore, it is necessary to distinguish
them by an accurate diagnosis.
As a new detection method for MG diagnosis, Enzyme-linked
immunosorbent assay (ELISA) is developed, which would be more efficient
and cost effective. Previously, some ELISA based detection methods were
proposed in both the direct and indirect assays
176-180
, and some ELISA kits are
also commercially available (e.g. BioVendor). In some methods, muscle-
delivered nAChRs are coated on polystyrene wells without any controlled
106

orders. It still has a problem for a supply of the receptors. To overcome the
issue, usage of the human rhabdomyosarcoma cell line TE671 was proposed
for ELISA
181, 182
since the cell line is known to express nAChR on its
surface
183
The cell line has also been applied to develop RIA
184-187
. In other
methods, anti-nAChR autoantibodies or α-bungarotoxin are immobilized on
wells to capture receptors; however, they could inhibit bindings of MG
autoantiboes resulting in inaccurate measurements. With a commercially
available kit, concentration of autoantiboies is measured by an indirect
method competing with antibodies known to bind the receptors. Receptors and
a sample are mixed first and applied to wells. Those wells are coated with
anti-nAChR antibodies which target a different site from that of MG
autoantibodies. Receptors are immobilized on wells through the anti-nAChR
antibodies, and labeled autoantibodies known to cause MG are added. If the
sample contains MG antibodies, the labeled autoantibodies cannot bind to the
immobilized receptors. Thus, the lower signal indicates more MG
autoantibodies in samples.
In our new detecting method, nAChR (ECD of α1 subunit) will be
immobilized on an ELISA plate, and MG serum will be directly measured by
binding the receptor. It became possible due to the success of α211
purification in large quantities. Previous studies attempted to coat plates with
receptors in random orientation, but in our study, it is achieved to immobilize
receptors in a certain orientation mimicking the cell surface. That is, the MIR
region is exposed to the outside for autoantibodies to bind them easily. It was
not possible with purified receptors from muscle. The α211 is a recombinantly
expressed protein, thus the protein can be engineered depending on
experiment purposes. An orientation of nAChRs can be controlled by
immobilizing them using His-tag which is inserted at C-terminus of the
receptors. Upon development, specificity and sensitivity would be important
107

aspects. It could be imagined that there would be various components in
patient serum samples which could interact with wells and/or α211 resulting
in false positive signals. It would be critical, therefore, to minimize such non-
specific bindings. The other challenge would be sensitivity of detection. The
lower limit would be especially important to be able to diagnose MG in the
early stages, when the number of autoantibodies is small. To achieve the goal,
a series of control experiments and optimizations was necessary using purified
autoantibody, mAb35. Because mAb35 is also successfully purified in our lab,
it can be used for those experiments. Since it was difficult to obtain human
MG samples, the pseudo sample was prepared by mixing known amount of
purified mAb35 with rat serum. Optimizations of the assay were then
achieved using the pseudo rat MG samples. In order to apply for human
samples, human serum of healthy people are obtained and compared with rat
serum. Mouse α211 is initially used to develop the assay, but human α211 is
purified for further accurate measurements for human MG antibodies. After a
series of control experiments, a protocol of a possible new detection method is
achieved. Due to the difficulty of obtaining well analyzed MG patients’
samples, it still requires to test the system using actual MG samples. It would
probably need further optimizations, but promising results are obtained.
If this assay is completely established, it would help a diagnosis and a
treatment of MG. An affordable clinical test can be used not only for a
diagnostic purpose but also for monitoring the disease. This chapter describes
how a novel assay system is designed and optimized. The beginning of
chapter covers an experimental set up with a series of control experiments,
and then it moves into a detection test with pseudo rat MG samples. Finally,
the assay system is tested with human MG samples which unfortunately are
not fresh and lack RIA data for us to determine the new method.
108

4.3 Results
4.3.1 Establishment of control assays
Some optimizations are necessary to establish ELISA in order to
measure actual samples. First of all, it is important to choose the correct type
of ELISA plate. A Ni coated plate was chosen to immobilize α211 protein
using His-tag, which was an ideal setting since the His-tag was introduced at
the C-terminus of α211 away from the expected binding site of autoantibodies
(Figure 4.1). Additionally, a Flag-tag that could be used as a control primary
antibody (1°Ab) was introduced at the N-terminal of α211, where 1°Ab can
easily recognize it. The only concern was that Ni tends to have higher non-
specific binding which could cause false positive readings by a non-specific
binding of other components such as a secondary antibody (2°Ab).

Figure 4.1: Schematic model of ELISA
using a Ni coated ELISA plate.
nAChR α211 is immobilized on a plate
through the His-tag which is at the C-
terminus of the receptor. The primary
antibody recognizes α211, and the
secondary antibody carrying a
fluorochrome binds to the primary
antibody resulting in a signal. A Flag-tag
was introduced at N-terminus and
autoantibody binding sites are also known
at N-terminal side, which would be ideal
for this assay setting. Any non-specific
binding of 2°Ab to the well would give a
false positive signal.
109

A series of control experiments was necessary to determine the extent
of non-specific binding as shown in Figure 4.2. Some non-specific bindings
were expected, and it was important to minimize them for accurate
measurements. mAb35 (rat IgG) was used as an autoantibody, and rat anti-
Flag antibody was used as a positive control of the 1°Ab. Goat anti-rat IgG
H&L (Alexa Fluor®488) was chosen as the 2°Ab, and was pre-adsorbed, and,
therefore, should not react with antibodies of human, mouse, chicken, cow,
rabbit or sheep.

Figure 4.2: Control experiments to determine non-specific binding.
CTRL 1 and CTRL 2 determine a background reading with and without the presence
of mouse α211 (yellow ovals). CTRL 3 and CTRL 4 confirm if the 2°Ab (blue Y)
does not bind to a well or mouse α211 non-specifically. CTRL 5 tests a non-specific
binding of 1°Ab (red Y) to a well. Finally, CTRL 6 determines non-specific binding
which could rise from any components in serum such as a variety of antibodies,
proteins and hormones (purple Y and multiple color circles).
110

The first control experiment was performed to test any non-specific
bindings of the 2°Ab to a well or mouse α211 (Figure 4.2, CTRL 3 and CTRL
4) as well as a background reading with and without mouse α211 (Figure 4.2,
CTRL 1 and CTRL 2). In this set of experiments, non-specific binding of the
1°Ab to a well was determined using two different 1°Abs, mAb35 and anti-
Flag antibody (Figure 4.2, CTRL 5). The maximum binding capacity of
mouse α211 to a well was also tested. Based on the manufacture’s protocol,
the binding capacity of each well is ~ 9 pmol His-tagged protein (27 kDa).
The molecular weight of mouse α211 should be 30 kDa including a glycan, so
9 pmol of mouse α211 (~ 0.27 µg) needs to be added to each well for
saturation. The experiment was designed by immobilizing 6 different amounts
of mouse α211 (0 µg, 0.1 µg, 0.2 µg, 0.3 µg, 0.4 µg, and 0.6 µg) on each well.
As for the amount of antibodies, the minimum required amount was calculated
as following. Since antibodies are divalent, theoretically half of the molar
equivalent amount of antibodies (~ 4.5 pmol) would be enough to bind all
mouse α211. However, because not all immobilized α211 is oriented to bind
the two arms of one antibody, the actual antibody required may be between
4.5 pmol to 9.0 pmol. Antibodies are high in cost, therefore, 5 pmol of
antibodies were used for the first round of control experiment though excess
was ideal. CTRL 6 would be performed after several more control
experiments.
The control experiment worked nicely, and no non-specific binding of
the 2°Ab to neither the wells nor to mouse α211 was observed (Figure 4.3,
CTRL 3 and CTRL 4). The saturation point (maximum binding capacity of
mouse α211) was around 0.2 µg – 0.3 µg based on mAb35 readings even
though it was not conclusive with anti-Flag readings since the signal was still
slightly increasing. Higher signal detection was obtained with anti-Flag
antibody compared to mAb35, which could be due to several reasons, one of
111

which would be tightness of antibody binding. mAb35 could have a weaker
binding and be washed off more during washing steps.

Figure 4.3: Non-specific biding test of 2°Ab to a plate and/or mouse α211.
(A) The top table shows actual values of fluorescent intensity, and the lower table
represents the experimental set up for each control assay. There was no non-specific
binding of the 2°Ab binding to the plate (CTRL 3) or α211 (CTRL 4). Non-specific
binding of the 1°Abs was not observed (CTRL 5). (B) The PBS curve is not visible
on the graph since it is overlapped with the 2°Ab curve, indicating no non-specific
binding of the 2°Ab. The signal due to the anti-Flag was stronger than that of mAb35.
The saturation point is around 0.2 µg – 0.3 µg of α211.
A
B
0
50
100
150
200
250
300
350
400
450
0 0.2 0.4 0.6
Fluorescence Intensity
α211 (μg)
Control Experiment
PBS
2° Ab Only
anti-Flag
mAb35
Point of saturation
0.3 μg

112

In order to further optimize the protocol, a washing test was performed
to measure how much of the antibodies could be washed off at each wash
(Table 4.1). It could also determine how efficient each washing step was, and
how many washing was necessary to minimize background readings. An
ELISA plate was prepared similarly to the first control experiment. Briefly,
different amounts of mouse α211 were added into each well, and either anti-
Flag antibody or mAb35 was added. The first measurement was done right
after secondary antibodies against anti-Flag and mAb35 were removed from
wells as a reference point (Wash 0). Five washes (5 minutes each) followed,
and fluorescent intensity was measured after each wash (Wash 1 – Wash 5).
Higher signals were detected as the amount of mouse α211 increased
per well, which was consistent with the previous experiment. Surprisingly,
after a single wash (Wash 1), readings for the negative control wells (no
mouse α211) dropped down to background levels (Table 4.1), implying that
non-specific binding of anti-Flag antibody and mAb35 was very limited if any.
This experiment also showed that extensive washing (5 washes) did not wash
off antibodies nor strip off mouse α211 from a well. Based on these results,
three washes for 5 minutes would be reasonable for the assay even though a
single wash was sufficient. After the wash test, the plate was left for several
hours to check bleaching effect of fluorescent signal, and significant
difference was not observed. However, it was noticed that fluorescent
readings were different time to time, which could be caused by environmental
factors such as pH changes and temperature changes. From these results, it
would be important to establish a method to normalize the fluorescent singal
in future analysis of the assay. The washing test was successful, but a strange
phenomenon was again observed. Anti-Flag antibody showed stronger signals
compared to mAb35 throughout wells, which was seen in the control
experiment. A hypothesis of higher reading of anti-Flag antibody was a
113

difference in antibody binding formation. mAb35 is known to bind its antigen,
nAChR, in a conformation-dependent manner. An antibody binding could
inhibit another antibody binding depending on the orientation of an antigen
(mouse α211). On the other hand, anti-Flag antibody is a polyclonal antibody,
which would enhance the antibody binding to mouse α211. Another
hypothesis was that the 2°Ab binds better to anti-Flag antibody than mAb35
for some reason. Some additional experiments would be necessary to solve the
puzzle. Since mAb35 had been behaving as expected, further experiments
were continued to develop the assay system.

Table 4.1: Washing test of stability of mouse α211 and antibodies.
Washing did not interfere antibody binding (1°Ab to mouse α211 or 2°Ab to 1°Ab).
It did not strip off mouse α211 from a Ni coated plate. A single wash was sufficient
to remove any unbound antibodies from wells. Anti-Flag antibody showed higher
fluorescent signals, which could indicate higher binding of the binding.

114

Next, it was determined the minimum amount of 1°Ab required to
saturate the antibody binding to mouse α211. Once the amount was obtained,
the required amount of 2°Ab could be determined. Based on the previous
experiment result, 0.3 µg (10 pmol) of mouse α211 was used to saturate each
well, and an antibody titration experiment was performed. In order to
determine the minimum amount of 1°Ab, a different amount of 1°Ab was
added into each well, and a fixed amount (5 pmol) of 2°Ab was used to detect
signals. Additionally, rat Flag α9 (Refer to Chapter 6 for details about subunit
α9) was used as a negative control. mAb35 is an autoantibody which targets
subunit α1; therefore, it should not bind to subunit α9. However, the construct
design of rat Flag α9 is similar to mouse α211, and it should be detected by
anti-Flag antibody as well as mouse α211. The results show that mAb35 was
specifically binding to mouse α211 and not to rat Flag α9 (Figure 4.4). The
result was consistent with a gel shift assay in Chapter 2, Section 2.3.4 (Flag α9
3Mut was used in the gel shift assay, and should not affect the result of
mAb35 binding).
The control, anti-Flag antibody, bound to both mouse α211 and rat
Flag α9 with similar fluorescent intensity curve until the possible saturation
point (2 pmol). However, the intensity of anti-Flag/mouse α211 gradually
increased (Figure 4.4). Both mAb35/mouse α211 and anti-Flag/rat α9 reached
saturation around 2 pmol of 1°Ab though the intensities were different. This
indicated that the saturation point was around 2 pmol of 1°Abs, which was
much less than expected (4 – 5pmol: if antibody binding is divalent, 8 – 10
pmol: if antibody binding is univalent). One explanation would be 0.3 µg (10
pmol) of mouse α211 was necessary to saturate each well, but it would not be
necessarily all mouse α211 bound to the well. In other words, actual saturation
amount of mouse α211 on a well would be much less than the amount added
into the well.
115

Figure 4.4: 1°Ab titration test.
The negative control worked well, proving mAb35 binding is mouse α211 specific.
The anti-Flag antibody bound to both mouse α211 and rat Flag α9; however, a slight
increase in intensity was observed in the binding to mouse α211. Even though the
intensity levels of fluorescence were different between mAb35 to mouse α211 and
anti-Flag to rat Flag α9, the shapes of curves were similar. Both hit a saturation point
around 2 pmol of antibodies.
0
100
200
300
400
500
0 2 4 6 8 10
Fluorescent Intensity
Antibody Amount (pmol)
Antibody Titration
anti-Flag (a211)
mAb35 (a211)
anti-Flag (a9)
mAb35 (a9)
116

Finally, the amount of 2°Ab was tested. Throughout the control
experiments, a fixed amount of 2°Ab (5 pmol) was used, but it was necessary
to test if that amount was enough to saturate. To determine this, the same
ELISA plate, which the 1°Ab titration experiment was performed with, was
used by adding an additional 5 pmol of 2°Ab. If the original amount of 2°Ab
(5 pmol) was not enough, an increase of signal intensity should have been
observed with the additional 2°Ab. The result was clear that mAb35 sample
indeed required higher amount of 2°Ab to achieve maximum readings (Figure
4.5, A). A linear regression line was fitted of the data points between 0 pmol
and 2 pmol, and R
2
value was > 0.99, which indicated the accurate
measurement of antibodies with this system (Figure 4.5, B). Further studies
with more data points will be needed to establish a more accurate relationship.
With the additional 5 pmol of 2°Ab, a titration curve of mAb35/mouse
α211 became similar to the one of anti-Flag/rat α9 WT; however, an unusual
phenomenon was again observed in anti-Flag/mouse α211. A signal of anti-
Flag/mouse α211 kept increasing further, and it did not saturate. This was
seen previously (Figure 4.4), but the increment was more severe. The curve
was clearly different from those of mAb35/mouse α211 and anti-Flag/rat Flag
α9. Even though there was no clear explanation as to what would be causing
the gradual increase of anti-Flag/rat Flag α9 signal, experiments were
continued to optimize detection of mAb35 in serum. Based on the result from
this experiment, it was concluded that 10 pmol of 2°Ab was necessary for
future experiments. A series of control experiments were performed, and
necessary information was obtained to do the last control experiment, CTRL 6
(non-specific binding of components in serum to a well/mouse α211 or non-
specific binding of 2°Ab to serum components).

117

Figure 4.5: Additional 2°Ab test.
(A) An additional 5 pmol of 2°Ab was added to determine the minimum required
amount of 2°Ab needed to maximize the reading. The mAb35 reading showed a
dramatic difference. Higher amounts of 2°Ab would be necessary for mAb35
samples. (B) The gradual increase of anti-Flag/mouse α211after the saturation point
(~ 2 pmol of 2°Ab) cannot be explained; however, a linear curve was obtained for all
samples before the saturation point with. R
2
values of > 0.99.
0
100
200
300
400
500
600
700
800
0 2 4 6 8 10
Fluorescent Intensity
Antibody Amount (pmol)
Antibody Titration (Additional 5 pmol 2°Ab)
anti-Flag (a211)
mAb35 (a211)
anti-Flag (a9)
mAb35 (a9)
0
100
200
300
400
500
600
0 0.5 1 1.5 2
Fluorescent Intensity
Amount of Antibodies (pmol)
Antibody Titration with additional 5 pmol 2°Ab (0 pmol – 2 pmol)
anti-Flag (a211)
mAb35 (a 211)
anti-Flag (a9)
mAb35 (a9)
A
B
118

4.3.2 mAb35 detection in rat serum
In Section 4.3.1, an autoantibody detection method was established
using antibody samples (mAb35) prepared with PBS. Linear curves with high
R
2
values were obtained, which indicated a new perfect linear relationship
between antibody concentration and fluorescent intensity (Figure 4.5, B). A
next challenge was to detect mAb35 autoantibodies in serum, with the final
goal of detecting autoantibodies in human patient serum. Therefore, the
protocol needed to be optimized against serum samples. It would be more
difficult since there would be a lot of other components in serum including
many antibodies produced by the immune system. It was expected to have
high background readings caused by non-specific binding of those
components. ELISA is often used to in a medical setting, and patient serum
samples are usually diluted for detection
188-192
. The dilution factor varies
based on detecting targets. The first experiment was to determine the best
dilution factor of serum samples for the optimal MG autoantibody detection.
There was not a human MG sample with known concentration, so a pseudo
MG sample was made with purified mAb35 and rat serum. The rat serum was
used because mAb35 is a rat antibody. Anti-rat IgG was needed to detect
mAb35; therefore, it was reasonable to use the rat serum containing other rat
antibodies to learn about non-specific binding.
A previous study
181
found that 120 nM of autoantibodies in serum was
a high level among MG patients. Based on this, 100 nM of pseudo MG rat
serum was prepared by adding purified mAb35 into rat serum. Using PBS,
samples of several dilution factors (no dilution, 1:2, 1:5, 1:10, 1:100, and
1:200), were made. Since 100 µL of each sample was added to a well, the
amount of mAb35 to be detected in each well was 10 pmol, 5 pmol, 2 pmol, 1
pmol, 0.1 pmol, and 0.05 pmol. Background readings of human serum were
119

tested as well as rat serum to determine any problems that could be faced. For
human serum detection, anti-human IgG was used as the 2°Ab since anti-rat
IgG would not detect human antibodies in human serum. It was expected to
have higher background readings in samples which were diluted with lower
dilution factors because the samples would have higher rate non-specific
bindings of other contents in serum. A main question was how much dilution
was necessary to see a significant difference between the serum only sample
(background control) and the mAb35 doped serum sample. If the background
is too high, signal from autoantibodies would not be clear, with samples of
low concentration antibodies difficult to detect.
Rat serum and human serum behaved completely differently. Rat
serum had a U-shaped curve of intensity (Figure 4.6, pink line), while human
serum intensity continuously increased as dilution factor was lowered (Figure
4.6, yellow line). This result implied that information obtained with rat serum
would not be applicable to human MG sample, and additional optimization
experiments would be necessary. mAb35 doped serum samples showed a
similar curve to that of the rat serum samples (background control). Although
fairly high intensity for both doped and undoped rat sera was detected at high
dilution factors (1:100 and 1:200), those dilution factors would not be the best
choice since readings of background and mAb35 sample reading (rat serum vs.
mAb35 doped rat serum) were similar. The 1:10 dilution showed the biggest
difference between the background and sample readings (Figure 4.6, red line
box), and should be the dilution factor to be used for further optimization.
Even though background readings of rat and human showed differences, it
would not be a major problem as long as a dilution factor which has a large
separation from a background reading can be found in human MG sample.
Next focus is on how accurately mAb35 can be detected with background
readings resulted from non-specific binding of contents in serum. Since
120

human MG sample was not available, the pseudo rat MG sample was used to
develop the assay further. Homologies of nAChRs are extremely high
between species, and most of MG mAbs can cross react
130, 134
. Therefore, it is
still valid to optimize the assay system using rat samples.

Figure 4.6: Serum dilution test.
Serum and mAb35 doped serum were diluted (total 6 dilutions) with PBS to
determine how serum affects signal detection. The original concentration was 100
nM, and 100 µL of each diluted sample was added to each well. The blue line
indicates the mAb35 doped serum sample, pink indicates rat serum only (background
readings of rat serum), and yellow indicates human serum only (background readings
of human serum). Rat serum and human serum showed obvious differences in
background readings. Dilution of 1:10 showed the most difference from background
reading (red line box).

0
50
100
150
200
250
300
350
400
0 2 4 6 8 10 12
Fluorescent Intensity
Serum Dilution/ Amount of Antibody (pmol)
Serum Control Experiment (mAb35 Detection)
mAb35 in Rat
Serum
Rat Serum
Human Serum
121

4.3.3 mAb35 titration in rat serum
As discovered in the previous section, the best dilution factor was
determined as 1:10 for the pseudo rat MG sample. MG patients would have
various concentrations of MG autoantibodies in their serum, and thus it is
important to be able detect a wide range of autoantibody concentrations in
patient samples. Therefore, a determination of the lowest and highest sample
concentrations was attempted using pseudo rat MG serum sample (mAb35
doped rat serum) and using this dilution factor. Five different pseudo samples
were prepared (1 nM, 10 nM, 25 nM, 50 nM, and 100 nM) together with a
reference sample which did not contain any mAb35. After a 1:10 dilution, 100
μL of each sample was added to wells; therefore each well contains 0.01 pmol,
0.1 pmol, 0.25 pmol, 0.5 pmol and 1 pmol of mAb35.
The result showed a linear relationship between mAb35 concentration
and fluorescent intensity (Figure 4.7). The highest concentration was 100 nM
in this experiment, which saturated the signals. Once the highest point (100
nM) was removed, a fitted curve was obtained with R
2
> 0.99 as expected.
The background reading was high compared to PBS, but it proved that this
assay system could measure accurate autoantibody concentrations within a
certain range. It was necessary to optimize the system further to be able to
detect a higher concentration range of serum sample since it could not
accurately measure the 100 nM sample. Since patients could have even higher
amounts of autoantibodies
181
, higher antibody concentration samples may be
read if the background could be minimized. It is also important to focus on a
lower concentration range so that this assay would be able to diagnose MG at
the early stage, which is currently very difficult.

122

Figure 4.7: mAb35 titration test in rat serum.
Five different concentrations of mAb35 were prepared using rat serum. Samples were
diluted 1:10 before adding into wells. The highest concentration sample (100 nM)
could not be accurately detected, but a linear curve with a high R
2
value was obtained
proving that this system worked at a limited concentration range.
y = 251.43x + 89.075
R² = 0.996
0
50
100
150
200
250
0 0.1 0.2 0.3 0.4 0.5 0.6
Fluorescent Intensity
Amount of mAb35 (pmol)
mAb35 in Rat Serum Titration
(1:10 Dilution, 1 nM - 50 nM: 0 pmol - 1 pmol)
y = 172.53x + 99.318
R² = 0.9381
0
50
100
150
200
250
300
0 0.2 0.4 0.6 0.8 1 1.2
Fluorescent Intensity
Amount of mAb35 (pmol)
mAb35 in Rat Serum Titration (1:10 Dilution)
123

4.3.4 Improvements of detection limit
In the last experiment, the lowest detection limit was around 1 nM, and
the highest was somewhere between 50 nM and 100 nM. Several things were
attempted to improve a detection limit, starting with the incubation of serum
samples on wells upon which rat Flag α9 was immobilized on. The idea was
to remove any antibodies and components, which would non-specifically bind
to either mouse α211 or a well, by incubating with rat Flag α9 prior to mouse
α211 since it has high homology to mouse α211. The sensitivity of detection
slightly increased, but 1 nM was still a limitation. Another attempt was adding
milk to a final concentration of 5% in both the serum sample and the 2°Ab to
block any non-specific bindings. Even though Ni-coated ELISA plates were
pre-blocked with BSA, it could reduce non-specific bindings, but
unfortunately, no obvious differences were observed. Next, different dilution
factors of serum sample were tried based on the serum control experiment.
Three dilution factors were tested; 1:5, 1:50, and 1:100. There were no
changes with 1:5 dilution compared to 1:10 dilution, and readings of 1:100
dilution did not give good R
2
values. However, some improvements were seen
with 1:50 dilution (Figure 4.8). Even though R
2
value of 1 nM – 200 nM (R
2
=
0.9782) was not as high as was obtained previously, it was reasonable, and it
achieved a reading of the highest concentration (200 nM). By removing higher
points (150 nM and 100 nM), R
2
values were improved. It was successfully
measured 100 nM and 150 nM with R
2
= 0.9985 and R
2
= 0.9833, respectively.
Problems with lower concentration measurements are possible with this
dilution (the lowest was 25 nM in this experiment), but readings can be
performed with two different diluted samples if necessary (ex. 1:10 dilution
and 1:50 dilution). Using 1:50 dilution, the highest detection limit was
expanded to 200 nM.
124

Figure 4.8: Detection limit optimization with 1:50 serum dilution.
Serum was diluted 1:50, and 100 µL of diluted sample was added to each well. A
reasonable linear curve was obtained including all readings though removing 150nM
and 200nM readings gave the best R
2
value. It showed that 1:50 dilution would work
to read high autoantibody concentration of MG sample.
y = 411.71x + 132.25
R² = 0.9782
0
50
100
150
200
250
300
350
0 0.1 0.2 0.3 0.4 0.5
Fluorescent Intensity
Amount of mAb35 (pmol)
mAb35 in Rat Serum
(1:50 Dilution, 0 nM - 200 nM : 0 pmol - 0.4 pmol)
y = 457.95x + 128.23
R² = 0.9833
0
50
100
150
200
250
300
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
Fluorescent Intensity
Amount of mAb35 (pmol)
mAb35 in Rat Serum
(1:50 Dilution, 0 nM - 150 nM: 0 pmol - 0.3 pmol)
y = 532.86x + 123.6
R² = 0.9985
0
50
100
150
200
250
0 0.05 0.1 0.15 0.2 0.25
Fluorescent Intensity
Amount of mAb35 (pmol)
mAb35 in Rat Serum
(1:50 Dilution, 0 nM - 100 nM, 0 pmol - 0.2pmol)
125

4.3.5 Human α211 test
Control experiments had been done using mouse α211, but human
α211 was also purified for more accurate detections since the main aim of this
assay system was to detect autoantibodies in human serum (Refer to Chapter 2
for human α211 preparation). Even though mouse α211 and human α211 have
high homology, it was necessary to confirm that all optimized conditions,
which were obtained using mouse α211, could be applied to an assay system
with human α211. The titration assay of mAb35 was performed to compare
the two different proteins. Six different mAb35 concentration samples in rat
serum were prepared (0 nM, 1 nM, 10 nM, 25 nM, 50 nM, and 100nM) and
they were diluted to 1:10 using PBS. Very similar readings were obtained
(Figure 4.9) indicating human α211 could be used without further
optimization to adapt the current setting.
4.3.6 MG patient serum dilution test
A sample of MG patient with an unknown antibody concentration was
tested to determine a dilution factor since rat serum and human serum behaved
differently in the previous experiment (Section 4.3.2, Figure 4.6) The sample
was diluted with six different dilution factors (1:500, 1:200, 1:100, 1:50, 1:10,
and 1:5) using PBS. Due to sample limitation, lower dilution samples such as
1:2 and no dilutions could not be tested, but the 1:5 dilution seemed to work
well (Figure 4.10). The antibody signal was clearly separated from
background. Readings of other dilution factor samples were very similar to the
background reading; therefore, it would be difficult to detect MG
autoantibodies in those dilutions. Once MG patient serums are obtained, 1:5
dilution factor could be used for the detection.

126

Figure 4.9: Mouse α211 vs. Human α211.
Human α211 was purified for more accurate assay development. Since the assay
system was optimized using mouse α211, the mAb35 titration experiment was
performed to compare them. The results showed that they behaved similarly, and all
information obtained with mouse α211 could be applicable to a system with human
α211.

0
50
100
150
200
250
300
0 0.2 0.4 0.6 0.8 1 1.2
Fluorescent Intensity
Amount of mAb35 (pmol)
Mouse α211 vs. Human α211
a211 Mouse
a211 Human
y = 213.69x + 97.871
R² = 0.9814
y = 219.41x + 98.17
R² = 0.9821
0
50
100
150
200
250
0 0.1 0.2 0.3 0.4 0.5 0.6
Fluorescent Intensity
Amount of mAb35 (pmol)
Mouse α211 vs. Human α211 (0 nM - 100 nM)
a211 Mouse
a211 Human
127

Figure 4.10: MG patient dilution test.
MG patient serum was diluted to 6 different dilution factors using PBS. Human
serum behaved differently from rat serum (Refer to Section 4.3.2). Most dilutions
show similar readings with background, but 1:5 dilution shows an obvious difference.
Due to the sample limitation, lower dilution factors could not be tested in this
experiment. However, it could be tried if it would give better results in future.

0
20
40
60
80
100
120
140
160
0 2 4 6 8
Fluorescent Intensity
Serum Dilution
Human Patient Sample Dilution Test
Patient Serum
Control Serum
128

4.3.7 Screening MG patient serum samples
A group of MG patients serum samples were tested if our newly
established system works or not. Unfortunately, samples were not fresh and
they had been stored at -80 °C for years. There were 79 samples including
controls, MG patient samples, and samples of other neuromuscular diseases
such as Amyotrophic Lateral Sclerosis (ALS) and Charcot-Marie-Tooth
disease (CMT). Each sample was diluted to 1:5 using PBS.
The results were inconclusive because some serum samples of MG
patients showed negative results (Figure 4.11). Since the samples were old, it
is possible that something happened to the autoantibodies affecting detection.
It is also possible that those MG patients had different types of autoantibodies,
not targeting nAChRs that, therefore, would not be detected by this assay. Of
more concern is the false positive result. It would be a major issue if our
system cannot distinguish MG from other neuromuscular diseases which
would share similar symptoms. It is necessary to determine the problem in
depth, but limited sample information was obtained; only disease diagnostic
information of each sample was available. Radioimmunoprecipitation assay
(RIA) results were not provided and the assay could not be performed in our
facility. Therefore, it could not be confirmed whether the false positives were
caused by a defect in our assay system, sample quality, or the accuracy of
sample information. Fresh samples were requested to collect from patients.
Once the fresh samples are obtained with RIA results, the assay system can be
tested within a better controlled setting.

129

Figure 4.11: Human MG samples screening.
Human MG samples were obtained and they were screened using the assay system developed. Each sample was diluted to 1:5
with PBS. The control intensity was subtracted from each reading for analyzing purpose. Red indicates MG samples, and blue
indicates other neuromuscular diseases. There were two control samples colored as green. Results were inconclusive. There were
false positives and some MG samples showed negative results. Samples were old and not enough information was provided for
further analysis.

130

4.4 Discussion
A novel detection method was developed using solid-phase
immunoenzymatic method, ELISA. A series of control experiments was
performed to set up the assay system. The nAChRs (mouse α211) were
successfully immobilized on wells though His-tag which was introduced at the
C-terminus of the receptors. Non-specific bindings of 1°Ab (to wells) and
2°Ab (to wells or mouse α211) was tested using PBS, and it showed no such
bindings at all. Maximum binding capacity of mouse α211 was determined
followed by a determination of a detection limit of autoantibodies using
purified mAb35. Titration of mAb35 in PBS was attempted, and it showed a
clear linear relationship. It also proved a specificity of MG autoantibodies
targeting to α1 subunit; they did not bind to α9 subunit at all. Next, pseudo rat
MG samples were prepared diluting with rat serum. Different dilution factors
were tested to obtain the best readings. Further optimizations for detection
limits were performed and it was concluded the detection range of 1 nM to
200 nM (1:10 dilution for the lower limit and 1:50 dilution for the higher
limit). For more accurate measurement, human α211 was purified and
compared to mouse α211 results. They did not show obvious differences, but
human α211 would be more suitable for this assay system detecting human
MG autoantibodies. Even though fresh human MG samples could not be
obtained, some preliminary experiments have been performed with serum
samples which were stored for years. Background of human serum was
checked, and human serum behaved differently from rat serum. However, 1:5
dilution seemed to be able to detect human MG samples. Unfortunately, RIA
results were not available to cross-validate our ELISA results. Disease
diagnosis was available for each sample at least; therefore, some analysis was
possible. Some MG samples gave negative values. It could be caused by a
storage condition of samples. It is also possible that those could be from
131

patients who are seronegative for anti-nAChR. More concern was false
negative results. There were two control samples, and one of them gave
slightly higher reading. Because of this, it cannot be concluded anything from
this result. Further experiments need to be repeated using more controlled
samples with RIA data in future. It is also necessary to standardize readings
since absolute values changes to time to time. Although the results were
inconclusive, it is promising to develop a complete assay system with further
optimizations.
It was challenging to improve the sensitivity to detect low
concentration of MG autoantibodies. However, it was achieved to obtain a
low detection limit of 1 nmol/L, which was better than a previous ELISA
study using TE671 cells (i.e. 2 nmol/L)
181
. Though the detection limit might
not be as sensitive as RIA, it was fairly close. A commercially available RIA
kit (RSR Diagnostic for Autoimmunity) sets a cut off of positive sample at 0.5
nM. Our system requires further optimization to detect a lower concentration
of autoantibodies comparable with RIA. RIA would be more suitable to detect
the early stage of MG at this point, but our method would be useful for a
quick initial diagnosis and clinical evaluation of patients. It is important for
patients to monitor concentration of autoantibodies. Even though there is a
reference range to determine disease severities, a definition of high
concentration would be different depending on individuals. Some might not
show a high reading in a test, but they could show sever symptoms. In this
regard, it would be necessary to compare readings to previous test results, and
monitoring of antibody concentration would help to provide the best treatment
for each patient. RIA would be difficult to be performed as a routine test to
check clinical status. However, our method could be easily used to monitor
patients since it is a simple and cost effective.
132

One drawback of our system is that the use of a single subunit α1, as
compared to a pentamer receptor in other methods. Critical autoantibodies are
known to target MIR of α1 subunit, but it is still unknown how other subunits
contribute to the bindings. Based on our first crystal structures of α1/Fab35
complex and α1/Fab210 complexes, only α1 subunit would be enough to
detect the anti-AChR antibodies targeting MIR which are the major cause of
MG. The sensitivity of an assay would help to diagnosis MG at the early
stage; therefore, patients can start proper treatments. RIA would meet the
criteria better, but it is not a simple test to run. The status of symptoms would
change from time to time in MG, and it is important to monitor it. An
affordable test would be useful for follow up of the disease. Our system
requires the final step of optimization, which is a sample screening of actual
human MG samples. Once it is established, this assay will be a useful
detection method in a clinical setting.
4.5 Materials and Methods
4.5.1 Sample preparation for ELISA
Mouse α211, human α211, and mAb35 were prepared as explained in
Chapter 2, Materials and Methods. Briefly, mouse α211 and human α211 were
expressed in P. pastoris and purified. mAb35 was purified from hybridoma
cells, purchased from ATCC. Pierce® Nickel Coated Plates (white, 96-well)
were used to develop the detection assay. The primary antibody of rat-anti
Flag was purchased from Agilent Technologies. Pre-adsorbed secondary
antibody of goat anti-rat IgG H&L (Alexa Fluor®488) was purchased from
Abcam. The signal was measured using a plate reader (Thermo Fluoroskan
Ascent FL) with following filters; Ex: 485 and Em: 527 or 538.

133

4.5.2 ELISA for control experiments
Pierce® Nickel Coated Plates (white, 96-well) were pre-blocked with
bovine serum albumin; therefore, an additional blocking step was not
necessary. An appropriate amount of α211 (100 µL) was added in a well and
incubated with shaking for 1 hour at room temperature. Then, wells were
washed three times using 200 µL of Washing Buffer (PBST containing 0.05%
Tween 20) for 5 minutes each. The primary antibody was diluted to the
preferred concentration with PBS, and 100 µL was added. After 1 hour of
incubation at room temperature, wells were again washed three times using
200 µL of Washing Buffer. The secondary antibody diluted to the proper
concentration with PBS was added (100 µL per well) and incubated for 1 hour
at room temperature. The final washing was done three times with 200 µL of
Washing Buffer (5 minutes each). Fluorescent intensity was measured using a
plate reader using filters Ex: 485 and Em: 527 or 538.
4.5.3 Washing test
Mouse α211was immobilized on a well and incubated with primary
antibodies, anti-Flag and mAb35, followed by a secondary antibody (Refer to
4.5.2 for details). Signals were measured right after the secondary antibody
was removed from wells (Wash 0). Wells were washed with 200 µL of
Washing Buffer for 5 minutes, and then the second measurement was
performed (Wash 1). It was repeated four more times (Wash 2 – Wash 5).
Fluorescent intensity was measured after each wash.
4.5.4 mAb35 detection in rat serum
Rat (sprague dawley rat) and human serum (pooled normal human)
were ordered from Innovative Research. Alexa Fluor®488 goat anti-rat IgG
134

H&L (Life Technologies) was used for human serum sample detection. A
pseudo MG rat serum (100 nM) was made by doping rat serum with purified
mAb35. The pseudo sample was diluted with different dilution factors using
PBS (no dilution, 1:2, 1:5, 1:10, 1:100, and 1:200). Since serum samples
would be expected to have high background due to other components, five
washes (5 minutes each) were performed instead of three washes after serum
sample incubation (before the 2°Ab incubation). The main protocol was
followed as in 4.5.3 except the amounts of mouse α211 and 2°Ab were
changed. After optimization with control experiments, 0.3 µg (10 pmol) of
mouse α211 was added to each well and 10 pmol of 2°Ab per well was used
for detection.
4.5.5 mAb35 titration in rat serum
Five different concentrations (1 nM, 10 nM, 25 nM, 50 nM, and 100
nM) of mAb35 doped rat serum were prepared. These pseudo MG serum
samples were diluted to 1:10 with PBS, and then added to a well upon which
0.3 µg (10 pmol) of mouse α211 was immobilized. The general ELISA
protocol of this experiment was same as control experiments except for the
number of washes. Five washes (5 minutes each) were performed after serum
incubation and signals were detected using 10 pmol of 2°Ab.
4.5.6 Improvements of a detection limit
Rat Flag α9 (0.3 µg, 10 pmol) was immobilized on wells as in the
same manner as mouse α211. Desired concentration of mAb35 doped serum
was added in the wells and incubated for an hour before being transferred to
wells which mouse α211 was immobilized on. The subsequent steps were
identical to the general ELISA protocol (modification: 5 washes). Testing
points of each dilution factors were following; for 1:5 dilution: 0 nM, 1 nM,
135

10 nM, 25 nM, 50 nM, and 100 nM, for 1:50 dilution: 0 nM, 25 nM, 50 nM,
100 nM, 150 nM, and 200 nM, for 1:100 dilution: 0 nM, 25 nM, 50 nM, 100
nM, 150 nM, and 150 nM.
4.5.7 MG patient serum dilution test
A group of MG positive serum sample was obtained by MG patient
serum samples were obtained from Dr. Iadarola Michael, Anesthesiology
Research Laboratories, Department of Perioperative Medicine, National
Institutes of Health, and were collected by his collaborator, Dr. Andrea M.
Corse, Department of Neurology and Neurosurgery, Johns Hopkins University
School of Medicine. Human α211 was immobilized on wells (0.3 µg, 10
pmol), and diluted MG patient serum samples were added. Six different
diluted samples were prepared (1:500, 1:200, 1:100, 1:50, 1:10, and 1:5), and
human serum was used as a background reading with the same dilutions.
4.5.8 MG patient serum screening
Human α211 was immobilized on wells (0.3 µg, 10 pmol), and MG
patient serum samples, which were diluted to 1:5, were added to each well.
After 1 hour of incubation, wells were washed with PBST (5 washes, 5
minutes each), and the 2°Ab added. In addition to actual sample wells, some
wells were prepared to obtain standard curves. Six different concentration of
samples (0 nM, 1 nM, 10 nM, 25 nM, 50 nM, and 100 nM) in either PBS or
rat serum were prepared to create two different standard curves to determine
which one would be better for future analysis. For the standard curve samples,
different 2°Ab was used from control assays. Anti-rat IgG Alexa Fluoro 488
from Abcam had been used throughout optimization experiments, was
switched to one from Invitrogen. It was a lower cost, and also it would be
better to use antibodies from the same company for better consistency since
136

anti-human IgG was purchased from Invitrogen. Before the change, the
performance of the new antibody was checked and compared with the
previous one.

137

Chapter 5

Crystallographic Studies of Novel
α211 Targeting Proteins Selected by
mRNA Display Technique

5.1 Abstract
Therapeutic antibodies have been extensively studied and some
successes have been achieved with their unique and powerful features: high
binding specificity and binding affinity. As a new-generation protein binder,
usage of alternative protein scaffolds has been proposed mimicking
immunoglobulins. The mRNA display works as a powerful in vitro selection
system of proteins for binding to a target of interests. In this chapter, it is
attempted to generate a specific binding protein targeting nAChR α1 subunit.
These studies cover a selection of a binding protein as well as purification
aiming to obtain a X-ray crystallography structure for further development of
therapeutic molecules.

138

5.2 Introduction
The project of this chapter was a collaborative work with Dr. Richard
Roberts’ group, Mork Family Department of Chemical Engineering and
Materials Science, Viterbi School of Engineering, University of Southern
California. A technique called mRNA display, which was invented by Dr.
Roberts, was utilized to achieve protein selections in vitro
91
. Target-specific
binding proteins are extremely useful for therapy, diagnostics, drug target
analysis, and laboratory research
193, 194
. The main goal of this project is to
develop molecular tools to study the functions of nAChR in neuromuscular
junction and potential treatment for MG using mRNA display technique.
Antibodies have been extensively studied for therapeutic drugs and
treatment. Antibodies are a great tool since they have high affinity and
specificity to bind a target protein. In addition, they are naturally occurring
human proteins; therefore, they have relatively low immunogenicity and low
toxicity. Drawbacks were a time consuming and complicated process to
develop therapeutic antibodies. Due to the success of therapeutic antibodies, a
next focus was moved to something similar but better; small engineered
antibody fragments. Single-chain antibodies (scFvs) are one of examples of
those fragments. They are composed of two variable domains (VH and VL)
linked with a short length of peptides
195
. Compared to the full-length
antibodies (~ 150 kDa), those are only ~ 27 kDa
196
. Another example is
domain antibodies (dAbs) which are either VH or VL domains with size of 11
– 15 kDa. They still keep binding specificities to target proteins as with
conventional antibodies
196-198
. One example of dAbs is nanobody which is
single variable domain (VHH) of antibodies found in camels or llamas
199, 200
.
Finally, advanced techniques of protein engineering with selection libraries
increased the possibility to generate novel binding proteins using alternative
139

protein scaffolds
193, 194
. Several protein scaffolds have been proposed as the
next-generation antibody therapeutics, and the 10th fibronectin type III
domain of human fibronectin (10FnIII) is one of them as a useful tool for
protein recognition
91, 201-203
.
The wild-type 10FnIII, which is a size of ~10 kDa, was discovered to
be a potential scaffold for novel binding proteins
204
, and a new scaffold
e10FnIII was developed for optimized expression
205-207
. The fibronectin type
II domain is a naturally occurring protein in human, and it has a role for the
formation of the extracellular matrix and cell-cell interactions
204
. Thus, it is
expected to have a minimized immunogenic reaction and toxicity as with
antibodies. It is a large protein which has 15 repeating units consisted from
three types of small domains (I, II and III). The 10th

unit was developed as a
alternative scaffold of binding protein
204
. Structures of the 10FnIII are
determined by NMR and X-ray crystallography and revels interesting
characteristics of the protein
208-210
. The 10FnIII has an antibody-like structure
though primary sequence is different from conventional antibodies. It contains
three diversified loops which is similar to VH (or VHH) of antibodies. It also
has a structure of β-sandwich which resemble to that of the antibody VH (or
VHH) domain (Figure 5.1). Previous studies demonstrated a successful
generation of libraries and selection of binder proteins using those variable
loops
88, 205, 206, 211-213
. The 10FnIII looks similar, but the antibody has nine β-
strands whereas 10FnIII has seven. It could bind specifically to a target
protein with high affinity like antibodies
88, 204-206, 211-214
. A major difference is
that the 10FnIII does not have disulfide bonds, which enables us to purify the
protein using bacteria expression system. This is a significant advantage since
bacteria expression system is generally the simplest and least expensive
purification method.
140

Figure 5.1: Structural comparison of VHH and the 10FnIII.
This figure is adapted from Takahashi, R. et.al. TRENDS in Biochemical Sciences
(2003)
202
. (A) Llama variable domain of heavy chain (VHH). (B) The 10
th
fibronectin
type III domain (10FnIII). Both share similar features: three diversified loops (CDR-
H1, CDR-H2, and CDR-H3 for Llama VHH vs. BC, DE, and FG loops for the
10FnIII) and β-sandwich structure.

In our experiments, modified scaffold e10FnIII is used, which contains
five single mutations and deletion of seven amino acids at the N-terminus.
Three loops, BC, DE, and FG, are often main candidates of diversification
since they are analogue to CDRs of antibodies (H1, H2, and H3,
respectively)
215, 216
. For our studies, two loops (BC = 7 residues and FG = 10
residues) are diversified, which generated specific binders with high affinity
in previous studies
205, 212
. A typical mRNA selection cycle of mRNA display
is shown in Figure 5.2. This technique is a powerful selection system of
peptides and proteins and enables us to generate a target-specific binding
protein from diverse libraries of up to 10
14
molecules. Specific binders of
nAChR α1 subunit are selected using mRNA display. Further selections are
performed for binders which are resistant to temperature and protease. In
A B
141

order to select the temperature stable binders, the library of binders are
incubated at 65 °C for 10 minutes. It is also challenged by selecting protease
resistant binders by treating selections with proteinase K. In addition, α-
bungarotoxin is added to the library to select binders which compete with α-
bungarotoxin.
The entire selection experiments are designed and performed by Dr.
Roberts’ group; therefore, details of selection process are not discussed in this
Chapter. The discussion is focused on purification and crystallization of novel
binding proteins targeting nAChR α1 subunit. This Chapter covers the
optimization of binder protein purification as well as determination of binding
affinity in our setting. Selected binding proteins are supposed to have a certain
level of binding affinity defined by selection process; however, it is re-
examined by a gel shift assay with purified proteins. The specificity of the
binding protein is also examed to confirm it would not have any cross reaction
to other subunits of nAChRs due to the high homogeneity. Finally,
crystallizations of a binder protein, by itself and as a complex with mouse
α211, are discussed. The crystal structure of a binder protein in complex with
mouse α211 will be extremely useful to learn interaction of those two proteins.
This project is still ongoing, and our final aim of solving structures has not
been achieved. However, this Chapter follows through a selection and
development of a possible therapeutic molecule targeting nAChR α1 subunit.

142

Figure 5.2: Selection cycle of a typical mRNA-display.
This figure is taken from Takahashi, R. et.al. TRENDS in Biochemical Sciences
(2003)
202
. A library of dsDNA is generated from synthetic oligonucleotides. After
transcription of dsDNA sequences for generation of mRNA templates (a), puromycin
oligonucleotide (blue circle) is covalently ligated to mRNA templates followed by in
vitro translation (c). cDNA synthesis is performed after purification (d,e), and
resulted products of the cDNA/mRNA-protein fusion is subjected to the target
protein which is immobilized on beads. PCR is performed to regenerate an enriched
dsDNA pool. This cycle is repeated for 4 – 10 times to converge.

143

5.3 Results
5.3.1 Clone 4.2 expression and purification
After several rounds of selections in mRNA display, a binder named
Clone 4.2 showed a temperature and proteinase K resistance as well as
competition with α-bungarotoxin. Several similar Clones were provided by
the Roberts group, but Clone 4.2 was the best carrying characteristics of high
protein expression with tight binding to mouse α211. The construct of Clone
4.2 was transformed into Rosetta
TM
2(DE3)pLysS competent cells with the
heat shock method. Four different colonies were screened to test expression
with a small scale inducing with isopropyl β-D-1-thiogalactopyranoside
(IPTG). The expression was determined with an SDS-PAGE gel, and all
colonies showed protein expression with high level (Figure5.3).

Figure 5.3: Small scale expression test for Clone 4.2.
Whole cell samples of Clone 4.2 were run over 15% SDS-PAGE gels (8 μL of
sample + 2 μL of 5x loading dye for Before samples and 6.4 μL of sample + 1.6 μL
of 5x loading dye for After sample). Four colonies were tested, and high expression
of Clone 4.2 was observed in all samples after 3 hours of induction at 37 °C.
144

Once the expression of the protein was confirmed, it was scaled up for
protein purification. Since His-tag was introduced into the construct, protein
was purified using Ni-NTA beads after lysing cells by sonication. After Ni-
NTA purification, an ion exchange chromatography was performed by
subjecting Ni-NTA elution samples to an anion exchange column, Mono Q
HR 5/5. It was necessary to exchange buffer to lower salt concentration prior
to running the sample over the column. Most of Clones had a serious
precipitation problem at this stage. Although some precipitation was observed
in Clone 4.2, it was clearly less than other Clones. The sample of Clone 4.2
was centrifuged at high speed to remove any precipitations and was run over
Mono Q HR 5/5 column. A nice bound peak was obtained (Figure5.4, A, Peak
2), and it was further purified with a size exclusion column, Superdex 75
10/300. The chromatogram of the size exclusion columns (Figure 5.4, B) was
unexpected. The observed protein peak (Figure 5.4, B, Peak 2) was much
smaller, and a large peak was observed in a void volume (Figure 5.4, B, Peak
1). The flow through sample of Mono Q HR 5/5 was also run over Superdex
75 10/300 (Figure.4, C), and it was surprising to obtain a nice sharp peak
representing monomer Clone 4.2 (Figure 5.4, C, Peak 4). One concern was
that there were some contaminants which had higher molecular weights based
on an SDS-PAGE analysis. It is not clear what proteins with large molecular
weight migrate together with Clone 4.2 which has small molecular weight.
The migration of Clone 4.2 was slower than expected; it was eluted at the
volume where salt would be. It is possible that Clone 4.2 would have some
interaction with beads of Superdex 75 10/300. There were several additional
peaks (Figure 5.4, C, Peak 1 – 3), and an SDS-PAGE showed the same protein
component in all peaks. Those peaks would be indicating oligmer states of
Clone 4.2. Further analysis of those proteins was necessary; therefore,
fractions of each peak were pooled and concentrated down.
145

Figure 5.4: Clone 4.2 purification over Mono Q HR 5/5 and Superdex 75 10/300.
(A) Chromatogram of Clone 4.2 purification with Mono Q HR 5/5 with 15% SDS-
PAGE gel showing samples from each peak. A nice bound peak was obtained (Peak
2) though an SDS-PAGE shows some contaminants in the sample (16 μL of sample +
4 μL of 5x loading dye except precipitation sample). The flow through sample also
contained a lot of Clone 4.2 protein (Peak 1). (B) Chromatogram of bound peak of
Mono Q run purified over Superdex 75 10/300. 15% SDS-PAGE gel is shown for
each peak (16 μL of sample + 4 μL of 5x loading dye except Injection: 1 μL of
sample + 7 μL of water + 2 μL of 5x loading dye). Fractions 17 – 21 were pooled and
concentrated down for future experiments. (C) Chromatogram of Mono Q flow
through purified over Superdex 75 10/300. Multiple peaks were obtained, and 15%
SDS-PAGE gels shows that all peaks contain Clone 4.2. Fractions 38 – 44 were
pooled and concentrated down.
A B
C
146

5.3.2 Clone 4.2 gel shift assay
Binding of Clone 4.2 to mouse α211 was tested with a gel shift assay.
Two different Clone 4.2 samples were prepared: possible oligomer sample
(could be aggregated proteins) and monomer sample. The oligomer sample
was obtained from the peak which bound to a Mono Q HR 5/5 column (Figure
5.5, B), and the monomer sample was from the flow through sample of the
Mono Q run. Both samples were mixed with mouse α211 in equimolar ratio
and incubated on ice for an hour. Since Clone 4.2 was selected as a
temperature stable binder, it was tested if the Clone 4.2 treated with heat could
still bind to mouse α211. In order to test, the purified Clone 4.2 was heated for
10 minutes at 65 °C, which was the same condition the library was treated
with. As a positive control for the system, the sample of mouse α211 with α-
bungarotoxin was prepared.
A native-PAGE gel in Figure 5.5 showed clear shifts of bands
indicating binding of Clone 4.2 to mouse α211 (α211 vs. α211 + Clone 4.2,
monomer). An obvious difference was observed between oligomer and
monomer of Clone 4.2. It was shown that the possible oligomer sample indeed
had a large molecular weight, and it was stuck at the well. There was not
enough information to conclude if it was an oligomer or aggregated protein,
but it was shown that there were possibly some proteins could bind to mouse
α211. Faints band were observed in the oligomer sample (α211 + Clone 4.2,
oligomer) with mouse α211 though most of mouse α211 did not shift. Heated
Clone 4.2 lost its function to bind to α211 (α211 + Heated Clone 4.2,
monomer). Even though it was challenged with a high temperature during the
selection, it could not have been stable enough in our system. It is possible
that Clone 4.2 could tolerate a little lower temperature such as 37 °C, which is
significant enough regarding protein stability. Thus, another heating test was
147

performed at 37 °C for 30 minutes as well as a competitive assay with α-
bungarotoxin since Clone 4.2 was a selected against α-bungarotoxin binding
site. It is not known if Clone 4.2 binds at the same place with α-bungarotoxin
or its binding indirectly inhibits binding of α-bungarotoxin. A crystal structure
of the complex would answer how Clone 4.2 competes with α-bungarotoxoin.
The results of the second gel shift assay were interesting. First, the
heating test was successful, and heated Clone 4.2 bound to mouse α211 (α211
vs. α211 + Heated Clone 4.2). Even though it could not tolerate the expected
temperature treatment, 65 °C for 10 minutes, it showed stability at 37 °C for
30 minutes (Figure 5.6). It was a substantial breakthrough considering many
proteins are sensitive to temperature. Next, a competitive assay with α-
bungarotoxoin showed not only Clone 4.2 competed with α-bungarotoxoin but
also it had a fairly strong binding affinity to mouse α211. It was surprising
that Clone 4.2 showed a similar binding affinity with α-bungarotoxoin, which
binds to the receptor extremely strong (α211 + Toxin + Clone 4.2 with
equimolar ratio of Toxin and Clone 4.2). It was not expected to bind so tightly,
so the experiment was designed in a way to increase the amount of α-
bungarotoxoin hoping to observe a competition with excess amount of Clone
4.2. Bands in excess amount of Clone 4.2 were sharper indicating most of
mouse α211 was bound by Clone 4.2. A specific value of binding affinity
could not be obtained by this assay; however, it was shown that Clone 4.2
indeed competes with α-bungarotoxoin with high binding affinity. An
additional sample (Toxin : Clone 4.2 = 0.5:1) would provide a better idea if
Clone 4.2 binds to mouse α211 tighter, but it needs to measure a Kd value in
future. In this assay, it was shown that Clone 4.2 binds to mouse α211,
specifically around α-bungarotoxoin binding site.

148

Figure 5.5: Gel shift assay showing binding of Clone 4.2 to mouse α211.
(A) 10% Native-PAGE gel shows Clone 4.2, monomer binds to mouse α211 (α211 vs.
α211 + Clone 4.2, monomer). Oligomer Clone 4.2 nor heated Clone 4.2 do not bind
to mouse α211 (α211 + Clone 4.2, oligomer, α211 + Heated Clone 4.2, monomer).
Samples were mixed in equimolar ratio and incubated on ice for an hour. (B)
Chromatograms of Superdex 75 10/300 of oligomer or monomer Clone 4.2. Indicated
peaks were pooled and concentrated to use for the gel shift assay in A.

A
B
149

Figure 5.6: Gel shift assay showing Clone 4.2 is temperature stable as well as
competitive with α-bungarotoxin.
10% Native-PAGE gel shows temperature treated Clone 4.2 binds to mouse α211
(α211 vs. α211 + Heated Clone 4.2). Different amount of Clone 4.2 was added into a
fix amount of α-bungarotoxin (Toxin : Clone 4.2 = 1:1, 1:2, and 1:3). It clearly shows
the competition of Clone 4.2 over α-bungarotoxin. An equimolar ratio of Clone 4.2
with α-bungarotoxin was enough to bind to mouse α211 indicating Clone 4.2 would
have a similar binding affinity with α-bungarotoxin.

150

5.3.3 α211/Clone 4.2complex purification and crystallization
The gel shift assay confirmed the binding of Clone 4.2 to mouse α211;
therefore, it was moved on to a complex purification of Clone 4.2 and mouse
α211. Mouse α211 and Clone 4.2 were mixed at a 1:1.5 molar ratio and
incubated on ice for 1 – 2 hours for complex formation. The mixture was run
over Superdex 75 10/300, and three main peaks were observed (Figure 5.7,
A). The first peak would be protein aggregates, and the second peak was the
complex of mouse α211 and Clone 4.2. The excess amount of Clone 4.2 was
observed in the third peak. An SDS-PAGE gel confirmed proteins in each
peak Figure 5.7, B), and the fractions of peak 2 were concentrated down for
crystallization. A surprising finding was that most of contaminants of large
molecular weight in Clone 4.2 disappeared after the complex purification.
There were obvious contaminants showing in an SDS-PAGE gel (Figure 5.7,
C) after Superdex 75 10/300 purification of Clone 4.2. Those bands were not
detected in the inject sample of complex purification (Figure 5.7, B). It could
not be explained why they disappeared, but those proteins could be
precipitated over by the time due to instability.
Crystallization was attempted with the concentrated sample of the
complex using screening various kits. Initial crystals were obtained in two
different conditions, and they looked like sea urchin. It was further screened to
optimize the conditions; however, those crystals did not improve. It was also
tried to crystallize as Clone 4.2 alone, but no crystals were obtained. Another
attempt was performed by adding Fab35 (Refer to Chapter 2 for details) to
stabilize the complex, and it was also failed to obtain crystals. Since some
precipitations and unknown contaminations were observed during purification
steps, further optimization would be necessary to stabilize the protein and
obtain high quality crystals.
151

Figure 5.7: Mouse α211/Clone 4.2 complex purification
(A) Chromatogram of mouse α211/Clone 4.2 complex purification with Superdex 75
10/300. The first peak is the aggregation of the protein. The second peak indicates the
complex of mouse α211 and Clone 4.2, and the third peak is the excess amount of
Clone 4.2. (B) 15% SDS-PAGE gels of the complex purification (16 μL of sample +
4 μL of 5x loading dye except Injection: 1 μL of sample + 7 μL of water + 2 μL of 5x
loading dye). Fractions 22 – 25 were pooled and concentrated down for future
experiments. (C) 15% SDS-PAGE gel, a sample lane of Clone 4.2 purification over
Superdex 75 10/300. The detailed purification results can be found in Section 5.3.2
and Figure 5.5, C. Some contaminants were detected in Clone 4.2 purification, but
they disappeared in the complex purification (Figure 5.5, B). They could be
precipitated due to instability.

A
B
C
152

5.3.4 Clone 4.2 mutant expression and purification
Several mutants were designed based on a structure of e10FnIII from
human fibronectin (PDB: 1TTF) to improve the stability of the protein.
Among them, Clone 4.2 mutant 5b was the best candidate considering
expression level, binding affinity, and stability. Prior to purification of the
mutant, the purification protocol was optimized since the best behaving binder,
Clone 4.2 still had an issue with precipitation. Various buffers were tested:
different concentrations of salt, different kinds of salt (ex. NaCl vs. KCl), and
different pH. Some of them showed slight improvements, but the precipitation
problem could not be solved. Finally, it was discovered that addition of 10 %
glycerol would stabilize the protein. Several binder proteins including Clone
4.2 mutants, which had a major precipitation problem, were used to test the
finding, and most of them could be purified without precipitations.
Transformation of Clone 4.2 mutant 5b was performed as with Clone 4.2, and
expression was confirmed with a small scale prior to a large scale expression.
Clone 4.2 mutant 5b was purified following the original protocol of Clone 4.2
except for the buffer. The optimized buffer was used and no precipitation
problem was observed throughout the purification.
Clone 4.2 mutant 5b behaved similarly to Clone 4.2; monomer did not
bind to a Mono Q HR 5/5 column, and oligomer/aggregates bound to the
column. However, a significant difference was observed in protein purity.
Clone 4.2 had some contaminants which had a large molecular weight, and it
could not be removed. A sample of Clone 4.2 mutant 5b did not show any
contaminants after Superdex 75 10/300 purification. It could be a good sign of
protein stability since it was concerned that those contaminants could be heat
shock proteins stabilizing Clone 4.2 proteins and may be tightly associated
with Clone 4.2 protein.
153

5.3.5 Clone 4.2 mutant gel shift assay
A gel shift assay was performed to confirm the binding of Clone 4.2
mutant 5b to mouse α211. Samples were prepared in the same manner with
Clone 4.2 gel shift assay. In this assay, rat Flag α9 (Refer to Chapter 6 for
details) was added as a negative control. Clone 4.2 mutant 5b should bind
specific to mouse α211 (α1 subunit), not rat Flag α9 (α9 subunit). The result
clearly showed that Clone 4.2 5b mutant bound to mouse α211specifically
(Figure 5.8).

Figure 5.8: Gel shift assay showing Clone 4.2 mutant 5b binds specifically to
α211.
6% Native-PAGE gel shows that Clone 4.2 selectively binds to mouse α211, not α9
subunit. Different molar ratio of Clone 4.2 mutant 5b is mixed with α 211 (Lane 3
and Lane 4). The excess amount of Clone 4.2 mutant 5b is shown as a faint band on
Lane 4 (upper band). Twice amount of Flag α9 was used since Flag α9 band is harder
to detect on a Native-PAGE gel (ran as a smear band). Therefore, a band intensity of
unbound Clone 4.2 mutant 5b is stronger compared to the one of α211 (Lane 4 vs.
Lane 6). Excess amount Clone 4.2 mutant 5b did not shift Flag α9 indicating it binds
specific to α211 (Lane 6).

154

5.3.6 α211/Clone 4.2 mutant complex purification and
crystallization
The complex of mouse α211 and Clone 4.2 5b mutant was purified as
with the complex of wild type Clone 4.2. A similar result was obtained, and
the purity improved a little. Higher bands of contaminants which were seen in
Clone 4.2 (Figure 5.7, C) were not detected (data now shown). Crystallization
conditions were screened the same way. Unfortunately, no obvious
improvements were observed, and similar kinds of crystals, sea urchin like
crystals, were obtained in the same conditions with mouse α211/Clone 4.2.
Preliminary crystals did not improve with narrow screening of the conditions.
5.3.7 Competitive binding assay of Clone 4.2 mutant with α-
bungarotoxin
This section of experiment was performed by Kevin Wu (lab
technician) with my guidance. The binding affinity of Clone 4.2 5b mutant
was tested by competing with α-bungarotoxin. It was tested with two different
ways: titration with Clone 4.2 mutant 5b or titration with α-bungarotoxin. The
previous experiment indicated a strong binding of Clone 4.2; therefore, a
wider range of titration curve was used (ex. α211 : α-bungarotoxin : Clone 4.2
Mut 5b = 1:1:0.5 up to 1:1:10 of molar ratio). The results were interesting
(Figure 5.9). The half amount of Clone 4.2 mutant 5b was enough to compete
with α-bungarotoxin (Figure 5.9, A, Lane 4). The toxin binding was almost
completely inhibited by the same molar ratio of Clone 4.2 mutant 5b (Figure
5.9, A, Lane 5). More interesting results were observed with α-bungarotoxin
titration (Figure 5.9, B). Binding of α-bungarotoxin was not observed with
half amount of α-bungarotoxin to Clone 4.2 mutant 5b (Figure 5.9, B, Lane 4).
Shifts of bands were detected by increasing amount of α-bungarotoxin. It was
155

interesting how tight Clone 4.2 mutant 5b binds to α211 (Figure 5.9, B). Clone
4.2 mutant 5b still bound to α211 with ten-fold higher amount of α-
bungarotoxin (Figure 5.9, B Lane 8). The concentration of α-bungarotoxin
could be off since faint bands were observed in control lanes (Figure 5.9, A
and B, lane 3). Even considering the possible error range, this result showed
that Clone 4.2 mutant 5b bound to α211 tighter than α-bungarotoxin. It
demonstrated how Clone 4.2 mutant 5b bind to α211 compared to α-
bungarotoxin which is known to bind extremely tight.

Figure 5.9: Competition assay of Clone 4.2 mutant 5b with α-bungarotoxin.
6% Native-PAGE gel are used to do gel shift assays. (A) Titration with Clone 4.2
mutant 5b (labeled as Clone 5.2 5b in the figure). The amount of α211 and α-
bungarotixn is fixed, and amount of Clone 4.2 is increased. A band shift indicating
α211/Clone 4.2 mutant 5b formation is detected in the sample which contains only
half amount of Clone 4.2 mutant 5b (Lane 4). (B) Titration with α-bungarotoxin. The
amount of α-bungarotoxin is increased. Clone 4.2 mutant 5b competes well with α-
bungarotoxin. The binding of Clone 4.2 mutant 5b is observed with even addition of
10 times higher molar ration of α-bungarotoxin (Lanne 8). It indicates an extremely
tight binding of Clone 4.2 mutant 5b.

A B
156

5.4 Discussion
Binding proteins targeting nAChR α1 subunit were successfully
selected with mRNA display performed by Aaron Nichols (Dr. Roberts’
group). Several selected binders were tested for protein expression and
binding affinity. Some binders were not stable and tended to precipitate even
though it showed high binding affinity in mRNA display. It is not surprising
based on previous studies
217-220
. The wild-type of 10FnIII has been known to
be highly soluble and stable (Tm = 90 °C)
202, 214
. However, the solubility and
stability could be affected by modulation of variable loops
217-220
. Binding
affinity was also tested by a gel shift assay even though it was determined by
the Dr. Robert’s group as part of the mRNA sepection process. A binder
which showed the best solubility and binding affinity was selected for
crystallization. Among several candidates, few of binders showed high
expression with tighter binding. One of them was Clone 4.2. It showed a
specificity of binding to mouse α211 by testing with nAChR α9 subunit.
Clone 4.2 did not bind to nAChR α9 subunit, but it shifted a band by binding
to mouse α211 (Figure 5.8). The protein was also challenged for
thermostability during a selection process. It was supposed to tolerate 65 °C
heating for 10 minutes; however, it lost the binding function after the heat
treatment in our assay system. It is possible that Clone 4.2 could not tolerate
the temperature and was denatured. Although it could not show the expected
thermostability, it kept the binfing function after heating at 37 °C for 30
minutes, which is fairly good as a protein. Clone 4.2 behaved well compared
to other binder proteins, but it still had precipitation issues during purification
process. Thus, some optimization experiments were necessary, and it was
discovered that the addition of 10% glycerol in buffer improved the protein
stability.
157

After optimization of purification, crystallization was finally attempted
by itself and as a complex with mouse α211. Sea urchin-like small cluster
crystals were obtained in two different conditions. Optimization of
crystallization conditions were performed by narrow screening and additive
screening. However, crystals did not show any improvements and stayed as
small clustered crystals. Therefore, further stabilization of the protein was
achieved by introducing mutations based on available structures of 10FnIII.
Several mutants were designed, and binding affinity was tested. It is possible
that binding function would be lost by mutations introduced. Indeed, some of
them lost the function to bind α 211 or binding affinity decreased. Expression
level of proteins was then tested among candidate mutants which kept the
binding function. After all of screening processes, Clone 4.2 mutant 5b was
picked for further studies. The protein still had stability issues; it tended to
precipitate during purification. After a series of optimization with buffer, it
was found that 10% of glycerol could help to stabilize the protein. The purity
of the protein increased compared to wild-type Clone 4.2, but it still did not
give diffractable crystals. Similar morphology of crystals was obtained with
the mutant protein. It is possible that Clone 4.2 mutant 5b itself is still not
stable, or Clone 4.2 mutant 5b is not enough to stabilize α211, which required
an addition of α-bungarotoxin to be crystallized
30
. Even though it competes
with α-bungarotoxin to bind to α211, it is possible that Clone 4.2/Clone 4.2
mutant 5b bind at a different location, but it is close enough to the binding site
of α-bungarotoxin interfering binding of α-bungarotoxin. Stabilization of loop
C of α211 is important for crystallization, and Clone 4.2/Clone 4.2 mutant 5b
might not be able to do so.
Although further attempts are necessary to solve the structure of
Clone4.2/Clone 4.2 mutant 5b bound to α211, the binding affinity of the
Clones showed interesting results. It is known that α-bungarotoxin binds to
the nAChRs very tight, and the natural selection achieved such a tight binding.
158

However, our results showed that Clone 4.2/Clone 4.2 5b could bind to α211
even tighter (Figure 5.6 and Figure 5.8). Clone 4.2 5b still bound to α211
when it was mixed with ten-fold higher amount of α-bungarotoxin (Figure 5.8,
B). It showed that in vitro selection could work as efficiently as a natural
selection, or even better.
In termas of the long term objective of this study, it is possible to
create a binder which is not only protein-specific but also site-specific with a
variety of binding affinity. For example, a binder which can compete with MG
autoantibodies could be used as a therapeutic molecule to treat MG (MG is
discussed in Chapter 2 and Chapter 3). The area specific binder could be
developed by a competition with known MG mAbs such as mAb35. Once
structure information of Clone 4.2 bound to α211 becomes available, a weaker
binder could also be generated. A weaker binder competing with α-
bungarotoxin might have a potential to work as a pain killer by blocking
channels. The nAChRs are also found in insects and other invertebrates. They
could be the targets for neuroactive pesticides, and indeed many insecticides
directly or indirectly target the receptors
12, 221
. Binder proteins targeting
specific to the receptors of insects would be useful as insecticides. Moreover,
neuronal nAChRs are closely associated to the pathogenesis of several
diseases, and diseases are often related to specific nAChR subunits. For
instance, mutations in α4 or β2 could cause autosomal dominant frontal lobe
epilepsy
222, 223
. Some other diseases including schizophrenia, Tourette’s
syndrome, depression, Alzheimer’s disease, and Parkinson’s disease are all
related to the nAChR α7 subunit
13, 24
. It is important to develop drugs which
could target specific subunits to minimize side effects since all nAChR
subunits have high homology to each other. The technique, mRNA display,
has a high potential to create a subunit specific molecule which could
contribute in various fields of nAChR studies.

159

5.5 Materials and Methods
5.5.1 Construction of the Clone 4.2 and Clone 4.2 mutants
The constructs of Clone 4.2 and Clone 4.2 mutants were made and
provided by Dr. Richard Roberts Group, Mork Family Department of
Chemical Engineering and Materials Science, Viterbi School of Engineering,
University of Southern California. The genes of interest were cloned into
pET24a vector with a kanamycin resistant marker. His-tag was introduced at
the C-terminus for purification purpose. Several mutants were created, and
one of the mutants that heavily worked on had mutations at V11R and L19E
for protein stabilization. The mutations including Clone 4.2 5b were designed
by my PI, Dr. Lin Chen based on a structure of e10FnIII from human
fibronectin (PDB: 1TTF).
5.5.2 Clone 4.2 expression
The construct of Clone 4.2 was transformed into Rosetta
TM

2(DE3)pLysS competent cells with the heat shock method. The transformants
were plated on 2XYT agar plates containing 50 µg/mL of kanamycin. Plates
were incubated at 37 °C for overnight, and colonies were formed next day.
Four different colonies were screened for expression check with a small scale.
Once expression was confirmed, it was scaled up for protein purification. Pre-
inoculation was prepared seeding a single colony in 50 mL of 2XYT media
with kanamycin. The culture was incubated at 37 °C with shaking (220 rpm)
overnight. One liter of inoculat was set using 10 mL of the overnight
preinoculant. After 2.5 – 3 hours of incubation at 37 °C, OD
600
reached 0.6 –
0.8. Once a desired OD
600
was obtained, cells were induced with the final
concentration of 1mM isopropyl β-D-1-thiogalactopyranoside (IPTG). After 3
160

hours of induction at 37 °C, cells were harvested by centrifugation at 6,000 x
g for 20 minutes at 4 °C. Cell pellets were stored at -20 °C until a purification
day.
5.5.3 Clone 4.2 purification
Frozen cells were thawed on ice and resuspended with Ni-NTA
binding buffer (50 mM NaH
2
PO
4
; pH 7.8, 0.5 M KCl, 10% glycerol, 0.1%
Triton X-100, and 20 mM imidazole) containing protease inhibitors (1mM
PMSF, 1 µg/mL leupeptin, and 1 µg/mL pepstatin A). Then, cells were lysed
with sonication followed by centrifugation at 18,000 rpm for 30 minutes at 4
°C. The supernatant was incubated with Ni-NTA agarose beads (QIAGEN) at
4 °C for 2 – 3 hours with end-over-end rotation. The protein was eluted with
elution buffer (50 mM NaH
2
PO
4
; pH 7.8, 0.5 M KCl, 10% glycerol, and 500
mM imidazole) after washing with binding buffer to remove loosely bound
proteins. The eluted protein was concentrated and diluted with Mono Q Buffer
A (20 mM HEPES; pH 7.5) which did not contain any salt to lower salt
concentration for an anion exchange chromatography. The final salt
concentration was adjusted to 100 mM – 120 mM. Some precipitations were
observed, and they were removed by centrifugation at 10,000 x g for 10
minutes or by filteration using 0.22 µm filter. The sample was subjected onto
an affinity column (Mono Q HR 5/5, GE Healthcare) to purify proteins using
a salt gradient of 0 – 100% Buffer B (Mono Q Buffer B contains 1 M NaCl).
Bound peak fractions and flow through fractions were separately concentrated
down and run over a size exclusion column (Superdex 75 10/300 GL, GE
Healthcare) using 20 mM HEPES; pH 7.5 and 300 mM NaCl buffer. Each
peak was identified and confirmed by OD
280
and SDS-PAGE gels, and
fractions of the peaks were pooled and concentrated for future experiments.
The final amount of monomer Clone 4.2 was 1 – 2 mg from 1 L culture.
161

5.5.4 Gel shift assay
Mouse α211 and Clone 4.2 (oligomer, monomer, or heated sample)
were mixed in an equimolar ratio (1 µg of α211 was used), and the mixture
incubated on ice for 1 hour. Heated samples were prepared as following.
Clone 4.2 protein was heated in a heat block for 65°C for 10 minutes or 37 °C
for 30 minutes, and samples were cooled down on ice before adding mouse
α211. For the competitive assay, amount of Clone 4.2 was increased (α-
bungarotoxin : Clone 4.2 = 1:1, 1:2, and 1:3). Just before samples were loaded
into a gel, 5X loading buffer (containing no dye) was added to each sample
tube, and samples were loaded on 10% native-PAGE at 4 °C. The gel was run
for 3 – 3.5 hours at 100 – 120 V, ~ 15 mA in 4 °C room with TBE buffer.
Bands were detected with Coomassie staining.
5.5.5 α211/Clone 4.2 complex purification and crystallization
Mouse α211 was used to form a complex with Clone 4.2. Purification
method of mouse α211 can be found in Chapter 2, Section 2.5.2. Mouse
α211were mixed at a 1:1.5 molar ratio, and the mixture was incubated on ice
for 1 – 2 hours. Then, the sample was run over a size exclusion column
(Superdex 75 10/300 GL, GE Healthcare) with 20 mM HEPES; pH 7.5 and
300 mM NaCl buffer. Three peaks were obtained: Peak 1 – aggregation, Peak
2 – the complex of mouse α211 and Clone 4.2, and Peak 3 – excess Clone 4.2.
The peak 2 were pooled and concentrated down for crystallization.
Crystallization was attempted with the concentrated sample of the
complex (2.5 mg/mL). Crystal Screen, Crystal Screen 2 and Index kits
(Hampton Research) were used to screen conditions by the hanging drop
method at room temperature. Each drop contained 0.5 μL of protein and 0.5
μL of reservoir solution, and 0.5 mL of screening solution was added in each
162

well. Sea urchin like crystals were obtained in two different conditions: 0.2 M
ammonium sulfate, 0.2 M BIS-TRIS pH 5.5, 25% w/v PEG 3350 and 0.2 M
lithium sulfate monohydrate, 0.1 M BIS-TRIS pH 5.5, 25% w/v PEG 3350.
Optimizations of crystals were attempted by narrow screening and additive
screening (Hampton Research); however, no obvious improvements were
observed.
5.5.6 Clone 4.2 mutant expression and purification
Clone 4.2 mutant 5b was expressed and purified as with Clone 4.2.
The major change was made in buffer. After optimization, it was discovered
that 10 % glycerol would stabilize proteins. Therefore, 10 % glycerol was
added into Mono Q Buffer A, Buffer B, and Superdex 75 10/300 buffer.
5.5.7 Gel shift assay for specificity
Clone 4.2 mutant 5b was used for this experiment. Two different
molar ratio of Clone 4.2 mutant 5b were mixed with 1 µg of mouse α211 (1:1
and 1:2). Rat Flag α9 was used to test specificity. It migrates as smear band,
and it is hard to detect on a native-PAGE gel. Thus, twice the amount (2 µg)
was mixed with Clone 4.2 mutant 5b at 1:2 molar ratio. The excess amount of
Clone 4.2 mutant 5b was added to confirm it does not bind to rat Flag α9. The
sample mixture was incubated on ice for 1 – 2 hours and loaded on 6% native-
PAGE at 4 °C. The gel was run for 2.5 hours at 110 V, ~ 15mA in 4 °C room
with TBE buffer. Bands were detected with Coomassie staining.

163

5.5.8 α211/Clone 4.2 mutant complex purification and
crystallization
The complex of mouse α211 and Clone 4.2 mutant 5b was mixed with
1:2 molar ratio and incubated on ice for 1 – 2 hours before subjecting onto
Superdex 75 10/300 column. For stabilizing reason, 10 % of glycerol was
included to the original size exclusion buffer which was used to purify mouse
α211/Clone 4.2 complex. Fractions of peak were identified with a SDS-PAGE,
and concentrated down for crystallization. Crystallization conditions were
screened as with the complex of mouse α211 and Clone 4.2.

164

Chapter 6

Structural Determination of nAChR
α9 Subunit

6.1 Abstract
The nAChR α9 subunit belongs to neuronal nAChRs, and it has some
unique characteristics. It can be expressed as a homopentamer, but it also be
expressed as a heteropentamer together with nAChR α10 subunit. It was
originally discovered in hair cells of the inner ear, but recent studies show
relationships between α9 and some other clinical factors such as pain and
cancer. The first X-ray crystal structure of human α9 ECD was solved in
2014
86
, but a sugar chain was removed in the structure. The importance of
sugar chains have been discussed in previous studies
30, 31, 74, 75
. Crystallization
of rat α9 ECD complexed with α-bungarotoxin is attempted with a sugar chain.
This chapter discusses purification, crystallization, and the structure of α9
ECD/α-bungarotoxin complex.

165

6.2 Introduction
The nAChR α9 subunit is one of neuronal nAChRs, and it is known to
form both homopentamer
224, 225
and heteropnetamer with α10 subunit
226-230
.
Even though it can be expressed as homopentamer, it showed higher
expression of functional receptors with co-expression with α10 subunit in
Xenopus oocytes
226, 230
. α9 subunit was first found in hair cells of the inner
ear
231
, and it was shown to be located at other non-neuronal cells including
keratinocytes
232
and lymphocytes
233
. This subunit is interesting because it has
been shown correlations with not only ear disorders but also chronic pain and
breast cancers
9, 234, 235
. Some compounds which target α9 subunit have been
studied as a potential clinical application for pain
236-242
. Those compounds are
the α9 nAChRs antagonists and could suppress a pain pathway signaling.
Even though some compounds have been shown to have potential applications
for a clinical use, the specificity of binding is still a challenge due to the high
homology of nAChR subunits. Detail insights of structural information would
reveal different features of each subunit which could lead to a subunit-specific
drug discovery. Although nAChRs are known to be related to neuronal
diseases since they are often found in the CNS and PNS, overexpression of
specific nAChRs has been ovserved in various types of cancers.
Overexpression of α9 nAChRs was detected in breast cancer, and treatment of
nicotine showed the higest expression of α9 nAChR mRNA
9
. Additionally, α9
nAChRs have unique characteristics compared to other nAChRs: nicotine
works as an agonist for most of nAChRs, but it does not for the α9 subunit
224,
226
. Instead, it works as an antagonist, and other well-known agonists such as
epibatidine for other nAChRs showed antagonist effects on the α9 subunit
243
.
It is an interesting future since α9 subunit has high sequence homology with
other nAChRs (identities of 41 – 67%)
86
. Identifying the cause of differences
166

would help to understand the function mechanisms of nAChRs and design
drugs specific to α9 nAChRs.
The X-ray crystal structures of human α9 ECD were solved with free
and antagonist-bound states
86
. Even though those structures were solved at a
high resolution, sugar chains were removed for the crystallization purpose.
The importance of sugar chains were previously discussed in α1 ECD
30, 74, 75

and GABA
A
structures
31
. Glycosylation is one of the post translational
modifications, and it is well-known to be important for receptor folding and
trafficking
244
. In fact, glycosylation-defective mutants of receptors failed to
express on cell surface
30
. The previous study of α1 showed that the interaction
of a sugar chain with α1 as well as with α-bungarotoxin in α1 ECD (α211)/α-
bungarotoxin structure
30
. The sugar chain stabilized the binding of α-
bungarotoxin to the receptor. It was also discussed that a sugar chain could
have an important role in the gating mechanism of the nAChRs. Information
regarding to the sugar chain is still limited, and it is still unclear if it
contributes to the receptor function or antagonists/agonists binding. Therefore,
it is attempted to solve a structure of rat α9 ECD with the potentially
important sugar chain.
This chapter discusses structural determination focusing on the N-
linked carbohydrate chain located at 145 which is near the stem of the Cys-
loop and under loop C. Compared to α 211, the location of sugar chain in α9
subunit is shifted to the C-terminus side by two amino acids based on the
primary structure. It would be interesting to see how the location shift shows
differences in structures and/or interactions. Our structure also contains three
mutations which are introduced based on the previous structure of α211/α-
bungarotoxin. Those mutations are introduced to enhance α-bungarotoxin
binding. Binding of α9 and α–bungarotoxin is tested by a gel shift assay
167

before purifying as a complex. Our structure reveals the interactions of side
chains of those mutants as well as the sugar chain.
6.3 Results
6.3.1 α9 purification
Rat α9 was expressed in P. pastoris and purified similarly to
mouse/human α211 (Refer to Chapter 2, Section 2.3.1). Several constructs
were made to obtain α9 proteins for crystallization. Before the optimization of
the yeast expression system, which was performed with mouse α211, rat α9
had not been expressed. Two different constructs were made for better protein
expression purpose; one without Flag-tag (rat α9) and the other with Flag-tag
at the N-terminus like α211 (rat Flag α9). Originally, a construct without Flag-
tag was created for better pentamer assembly since α9 has been known to
form a homopentamer
224, 225
. Flag-tag was introduced because the previous
study showed the increase of protein expression in α211
144
. Rat Flag α9 had
the same backbone construct as mouse α211. No obvious differences were
seen between the two α9 constructs. Once the new expression system was
developed using mouse α211, proteins from both constructs were highly
expressed and purified resulting in ~ 2 – 3 mg of protein from the 1L of starter
culture.
For the crystallization purpose, two additional mutant constructs were
made: 3 mutations were introduced in one construct, and 1 mutation was
introduced in the other one. Those mutations were for tighter α-bungarotoxin
binding, which would stabilize α9 protein for crystallization. Those mutations
were designed based on mouse α211/α-bungarotoxin structure
30
. Three
mutations were Val189Trp, Ser191Phe, and Gly193Ser. Two residues, W189
and F191 were known to be important for toxin binding based on the structure
168

and Gly193Ser mutation was added since possible interaction was seen in the
α211/α-bungarotoxin structure even though the interaction may not be as
important as the other two residues. Interestingly, W189 and F191 were
different in human α1 (S187 and T189 in corresponding human α1 sequence)
suggesting that it occurred during evolution since mice are preys of snakes,
but not human. As for a single mutant, only Ser191Phe was introduced for
enhance the toxin binding. In a discussion with Dr. Steven Sine, Department
of Physiology and Biomedical Engineering, Mayo Clinic College of
Medicine, the importance of the single mutation was brought up. Most
nAChR subunits which are known not to be bound by α-bungarotoxin have
lysine at the position, and tight α-bungarotoxin binding subunits, α1 subunit
(mouse) and α7 subunit (mouse and human), have phenylalanine. It was
possible that the single mutation would be sufficient enough to enhance α-
bungarotoxin binding; therefore, the single mutant construct was made as well
as 3 mutants. Both mutant constructs were expressed well with the newly
developed protocol.
The purification protocol was similar to the one used for mouse/human
α211 (Refer to Chapter 2). Proteins, which were expressed and secreted in
media, were purified with Ni-NTA beads. The purity of α9 protein was as well
as mouse/human α211 after Ni-NTA purification though a little bit of
contaminants was observed at lower molecular weights (Figure 6.1). Since
expression level of α9 was 4 – 6 times higher than α211, samples could not be
concentrated as low volume as α211. Thus, multiple runs of Superdex 75
10/300 (3 – 4 rounds) were necessary. A larger void volume peak was
observed in α9 sample compared to α211 sample. A migration of α9 was
similar to α211 indicating that α9 was purified as a monomer form. The
additional peak was sometimes observed between peaks of the void volume
and the α9 peak. Since α9 is known to form a pentamer, the peak could be an
169

indication of oligomers. Pentamers could be eluted in the void volume with
the column; therefore, it was determined by running Superdex 200 10/300. It
was confirmed that they were aggregations of α211 with some contaminations
of other proteins. Due to the large peak of void volume, fractions of α9 were
collected away from the peak to increase purify of the sample.

Figure 6.1: Rat Flag α9 3Mut purification.
(A) Chromatogram of rat Flag α9 3Mut purification with Superdex 75 10/300. Three
to four times of Superdex purification were necessary for one batch of purification
due to high expression of the protein. The main peak indicates monomer α9 protein.
A large peak within a void volume was observed. A similar chromatograms were
obtained for rat α9, rat Flag α9, and rat Flag α9 single Mut. (B) 15% SDS-PAGE gel
of rat α9 3Mut purification (16 μL of sample + 4 μL of 5x loading dye except
Injection: 1 μL of sample + 7 μL of water + 2 μL of 5x loading dye). A descent
protein band was observed in a media sample showing a high expression of the
protein. Ni elution sample was fairly pure though some contaminations could be seen
at lower molecular weight. Fraction 18 shows a peak within a void volume. Even
though it seems to be clean, it was aggregations of α9 proteins. To increase a protein
purify, only a middle portions of the peak, Fractions 23 – 26, were pooled and
concentrated down for further experiments.
A
B
170

6.3.2 Attempts of α9 crystallization
Once high quality α9 protein was obtained, crystallization of α9 was
attempted in various ways. First, rat α9 by itself was screened for
crystallization, and hair-like small crystals were obtained in 2.8 M sodium
acetate trihydrate pH 7.0. Optimization of the crystal condition was attempted
by screening around the condition as well as using additive screening kits.
Unfortunately, crystals could not be improved to a diffracting quality. Several
more attempts were done with the protein; co-crystallize with α-bungarotoxin
as with α211, and removal of glycans using an endoglycosidase (Endo Hf).
None of those attempts produced crystal; therefore, rat Flag α9 was purified
for crystallization. Crystals did not grow with rat Flag α9 in the condition
which rat α9 crystals grew. After many failed attempts, it was decided to
introduce mutations to enhance α-bungarotoxin binding (discussed in Section
6.3.1).
6.3.3 Gel shift assay
Rat Flag α9 3Mut and rat Flag α9 single Mut were successfully
purified following the original protocol of α211 purification. Even though rat
α9 and α-bungarotoxin could have been purified as a complex, obvious
binding was not observed in a gel shift assay. It was possible that the binding
was not strong enough to hold the complex interaction during gel
electrophoresis which took ~ 3 hours. It is usually seen the opposite
phenomena: it is harder to maintain protein complexes on FPLC than in gel
shift. One explanation would be that the FPLC purification process is a half of
time length compared to a gel shift assay; thus, it could have been purified as
a complex before they dissociate. Mutations were introduced to enhance the
α-bungarotoxin binding, and binding was confirmed by a gel shift assay for
171

both mutants (Figure 6.2). It was slight shift in rat Flag α9 3Mut compared to
rat Flag α9 single Mut, but it was clear enough to confirm the binding.

Figure 6.2: 10% Native-PAGE gels showing α-bungarotoxin binding to rat Flag
α9 3Mut and rat Flag α9 single Mut.
(A) It was a slight shift; therefore, it was difficult to observe. However, Lane 3
clearly shows a partial shift of bands with half amount of α-bungarotoxin. A band of
rat Flag α9 3Mut was fully shifted by mixing with the same molar amount of α-
bungarotoxin. Mouse α211 was used as a positive control. (B) A similar assay was
performed to test rat Flag α9 single Mut. Only two amino acid residues were different
from rat Flag α9 3Mut, but dramatic difference was seen in a shift of bands. These
results confirmed the formation of the complex.
A
B
172

6.3.4 Rat Flag α9 mutants/α-bungarotoxin complex
purification
After the binding of Flag α9 mutants to α-bungarotoxin was confirmed
by a gel shift assay, the next step was purification of the complex proteins. Rat
Flag α9 3Mut (or rat Flag α9 single Mut) was mixed with α-bungarotoxin at
1:2 molar ratio, and the mixture was incubated on ice for an hour before
Superdex 75 10/300 run. There was a little bump on the left side of the main
peak, which could be the merged peak containing aggregated proteins from
the rat Flag α9 3Mut purification. However, fairly pure protein complex was
obtained, and the excess amount of α-bungarotoxin was successfully separated
(Figure 6.3). An SDS-PAGE gel confirmed the complex formation of rat Flag
α9 3Mut and α-bungarotoxin.

Figure 6.3: Rat Flag α9 3Mut/α-bungarotoxin purification.
(A) Chromatogram of rat Flag α9 3Mut/α-bungarotoxin complex purification with
Superdex 75 10/300. The second peak indicates the excess α-bungarotoxin. There
was a small bump at the left side of the main peak, which would be aggregates of
proteins. A similar chromatogram was obtained for rat Flag α9 single Mut. (B) 15%
SDS-PAGE gels of rat Flag α9 3Mut/α-bungarotoxin purification (16 μL of sample +
4 μL of 5x loading dye except Injection: 1 μL of sample + 7 μL of water + 2 μL of 5x
loading dye). Fractions of the main peak show a complex formation and fraction 31
shows unbound α-bungarotoxin. Fractions 22 – 25 were pooled and concentrated
down for crystallization.
A
B
173

6.3.5 Crystallization and data collection of rat Flag α9 mutants
with α-bungarotoxin
With the high quality of the protein complex was in hand,
crystallization of rat Flag α9 3Mut/α-bungarotoxin were attempted to obtain
by screening conditions with the hanging drop method. Screening was done
using 3 mg/mL proteins at room temperature. Initial crystals looking like sea
urchins grew in a crystal condition of 0.1 M ammonium acetate, 0.1 M BIS-
TRIS pH 5.5, and 17% w/v PEG 10K. For optimization, two different
temperatures were tested: 4°C and room temperature, and crystals grew better
at room temperature. Even though it was tested using the same crystallization
condition, larger crystals were observed as a cluster of small bar instead of sea
urchins in 3 days. Some other factors could have affected crystal growths.
Once a better morphology of crystals was obtained, crystal conditions were
further screened around the initial condition with different concentrations of
proteins (1 mg/mL, 3 mg/mL, and 6 mg/mL). Crystals were optimized in a
condition of 0.05 M – 0.1 M ammonium acetate, 0.1 M BIS-TRIS pH 5.5, and
11% – 14% w/v PEG 10K. Thick rod-like crystals were obtained with the size
of 10 μm x 100 μm – 20 μm x 200 μm, and protein concentration of 6 mg/mL
produced larger crystals compared to lower protein concentrations (Figure 6.4,
Drop 1). Some crystals had a groove (Figure 6.4, Drop 2) on the surface, and
some had sward edges (Figure 6.4, Drop 2 and 3). Crystal decay was observed
after 5 days, and the phenomena started earlier in a condition which contained
higher concentration of ammonium acetate. Crystals were shot at the APS 23-
ID-B and APS 19-ID-D, but diffraction was very limited (~ 6 Å) despite of
the size and morphology of the crystals. Even though enough data set could
not be obtained due to crystal decay, Bravais Lattice was obtained as primitive
tetragonal with the unit cell dimensions of a = 106.0410 Å, b = 106.0410 Å,
and 365.3280 Å. Since the diffraction of the crystals was poor, dehydration
174

was performed to improve a crystal packing. Unfortunately, no obvious
improvement was achieved by the dehydration, but different Bravais Lattice,
primitive orthorhombic (a = 77.094 Å, b = 127.558 Å, and 351.665 Å), was
obtained indicating a change of crystal packing by dehydration.

Figure 6.4: Rat Flag α9 3Mut/α-bungarotoxin crystals.
Crystals were grown as rod shapes. The protein concentration of 6 mg/mL was the
best to obtain bigger crystals. Drop 1 and 3 (0.5 µL protein + 0.5 µL reservoir) and
Drop 2 (1 µL protein + 1 µL reservoir). Larger drop size gave larger crystals.
Crystals in Drop1 had a square edge, whereas crystals in Drop 2 and 3 showed sharp,
sward-like edge. Drop 2 crystals also had a groove on a surface of crystals. Those
crystals had a limited diffraction though the quality of crystal seemed to be good.
Crystallization of rat Flag α9 single Mut/α-bungarotoxin was also
attempted. The single mutant of rat Flag α9 was purified in the same manner
as the complex with α9 3Mut bound to α-bungarotoxin. Rat Flag α9 3Mut and
rat Flag α9 single Mut has only two amino acids difference, and they should
be very similar. Therefore, screening of crystal conditions was performed
around the condition which was optimized to obtain crystals of rat Flag α9
175

3Mut/α-bungarotoxin. The higher concentrations of ammonium acetate and
PEG 10K were necessary for rat Flag α9 single Mut crystallization compare to
rat Flag α9 3Mut complex; 0.075 M – 0.1 M ammonium acetate, 0.1 M BIS-
TRIS pH 5.5, and 16% – 22% w/v PEG 10K. Crystals were thin and long with
a needle-like morphology (Figure 6.5). They were more stable than rat Flag
α9 3Mut crystals; no obvious crystal decays were detected even after a month.
These crystals were also shot at the APS 19-ID with an expectation of a better
diffraction given the crystal stability. Diffraction spots were fairly nice, and
the crystals diffracted better (~ 4 Å). However, it was difficult to collect good
quality of data set that is sufficient to solve a structure. Since crystals were
very thin, they easily decayed with the X-ray beam. Even though a structure
was not solved, two possible Bravais Lattices were obtained for the crystals: C
centered orthorhombic (a = 77.645 Å, b = 133.976 Å, and 56.206 Å) and
primitive hexagonal (a = 77.495 Å, b = 77.495 Å, and 56.038 Å).

Figure 6.5: Rat Flag α9 single Mut/α-bungarotoxin crystals.
Crystals were thin and long. A protein concentration was 6 mg/mL, and a drop was
set with 0.5 µL protein + 0.5 µL. Larger drop size gave larger crystals. Even though
crystals were thin (5 µm x 300 µm – 500µm), they diffracted well with clean
diffraction spots. However, it was difficult to obtain enough data set to solve a
structure due to the small size of crystals and decay in X-ray beam.
100 μm
176

The last attempt was to crystallize the complex with a different
condition; 0.2 M lithium sulfate monohydrate, 0.1 M BIS-TRIS pH 5.5, 25%
w/v PEG 3350. The preliminary crystals were obtained with rat Flag α9 3Mut,
and crystals looked very small clusters (Figure 6.6). Both complexes of 3
mutants and single mutant produced crystals as needle-like cluster after
optimization; 3Mut: 0.2 M – 0.25 M lithium sulfate monohydrate, 0.1 M BIS-
TRIS pH 5.5, 15% – 20% w/v PEG 3350 and single Mut: 0.1 M lithium
sulfate monohydrate, 0.1 M BIS-TRIS pH 5.5, and 15% w/v PEG 3350.
Crystals seemed to stable for 3 – 4 weeks, but many crystals disappeared
while harvesting, which could be caused by a wrong cryo solution or low
quality of crystals. Some crystals were still harvested and shot at the APS 23-
ID, but they did not diffract much.

Figure 6.6: Crystals of α9 complexes in different crystallization condition.
Morphology of crystals between 3Mut and single Mut was slightly different; Single
Mut crystals were much longer. A protein concentration was 3 mg/mL or 6 mg/mL,
and a drop was set with 0.5 µL protein + 0.5 µL. A larger drop size gave larger
crystals. Those crystals had limited diffraction.
177

6.3.6 EndoHf digestion
After optimizing crystal conditions of rat Flag α9 mutants/α-
bungarotoxin, complex of wild type α9 was crystallized by screening the
optimized condition of those mutants crystals, but they were not diffraction
quality. Crystals of rat Flag α9 3Mut and rat Flag α9 single Mut with α-
bungarotoxin diffracted to some degree. However, it was difficult to collect
enough data set to solve the structure due to the limited crystal quality. One
possible reason for the low quality of crystal could be non-ideal crystal
packing, which could be caused by a glycosylation in this case. Rat α9
protein has two glycosylation sites: N32 and N145, and N32 glycosylation site
is located at the top the receptor, which could interfere crystal packing.
nAChR α1 has a glycosylation site at a similar place with N145, which was
shown to be important for toxin binding
30
. Thus, a new approach to get
diffracting crystals was to remove a glycan at N32. Mutagenesis was
performed to remove the glycosylation site by changing asparagine to alanine;
however, there was a major problem in expression. Some expression was
detected, but it was 1/20 – 1/30 of expression level compared to rat Flag α9,
rat Flag α9 3Mut, or rat Flag α9 single Mut. It makes a sense since
glycosylation is one of the important post-translational modifications for
receptor folding and trafficking
245
. Therefore, it was not an option to remove a
glycan by mutagenesis. Endo Hf enzyme could cleave glycans, but not
selectively. A special experimental protocol was developed to achieve the
goal removing only one of glycans; rat Flag α9 3Mut was incubated with α-
bungarotoxin first before adding Endo Hf. A hypothesize was that the glycan
at N32 was exposed on a surface of the receptor and could be easily cleaved
off by Endo Hf, while the glycan at N145 could be covered by α-bungarotoxin
or hindered by the loops (loop C and Cys-loop) of the receptor resulting in
difficult accessibility for Endo Hf (Figure 6.7). The Endo Hf treatment is
178

usually done in a denaturing condition; however, a native protein is necessary
for crystallization. Thus, a protocol was optimized to remove only a glycan at
N32 with a native condition. Two different sample sets (with and without α-
bungarotoxin) were prepared to prove that α-bungarotoxin blocks Endo Hf
from cleaving a glycan at N145. A time course experiment was performed to
determine the best incubation time to remove a glycan N32 but not N145.
Results showed that clear difference between two samples. As expected, α-
bungarotoxin indeed inhibited Endo Hf from removing a glycan at N145.
Even though the end products were mixture of rat Flag α9 3Mut with a single
glycan and no glycan, it was moved to a crystallization step.

Figure 6.7: Removal of N32 glycan by Endo Hf.
(A) Schematic model of Endo Hf digestion on α9 bound by α-bungarotoxin. Endo Hf
would have a better access to cleave the glycan at N32 compared to the one at N145.
(B) 15% SDS-PAGE gel of time course experiment of Endo Hf digestion on rat Flag
α9 3Mut/α-bungarotoxin (8 μL of sample + 2 μL of 5x loading dye). Small scale
experiment was performed to optimize a protocol before a large scale experiment.
The sample of T = 0 indicates a quick enzymatic reaction of Endo Hf. However, it is
also possible that Endo Hf had a better access to glycans while samples were being
heated to make samples for SDS-PAGE gels. For more accurate results, Endo Hf
should have been de-activated first by heating before mixing with 5x loading dye
followed by heat treatment.
A
B
179

6.3.7 Endo Hf treated rat Flag α9 3Mut/α-bungarotoxin
complex purification
After the removal of N32 glycan was confirmed by an SDS-PAGE gel,
Ni-NTA purification was performed to remove Endo Hf. Elution fractions
were combined and run over Superdex 75 10/300 for further purification.
There was a small fused peak on the left side of the main peak (Figure 6.8, A).
It was too small amount to be detected on an SDS-PAGE gel, but it could be
oligomers since α9 had a tendency to cause an oligomerization. An SDS-
PAGE gel showed a nice complex formation of Endo Hf treated rat Flag α9
3Mut with α-bungarotoxin (Figure 6.8, B). It was hard to see, but an
additional faint band, which indicated α9 protein without both glycans, was
observed under a main α9 protein band.

Figure 6.8: Endo Hf treated rat Flag α9 3Mut/α-bungarotoxin purification.
(A) Chromatogram of Endo Hf treated rat Flag α9 3Mut/α-bungarotoxin complex
purification with Superdex 75 10/300. A small fused peak would result from an
oligomerization of α9 proteins. (B) 15% SDS-PAGE gel of Endo Hf treated rat Flag
α9 3Mut/α-bungarotoxin purification (16 μL of sample + 4 μL of 5x loading dye
except Injection: 1 μL of sample + 7 μL of water + 2 μL of 5x loading dye). Fractions
of the main peak show a complex formation. Fractions 22 – 26 were pooled and
concentrated down for crystallization.
A B
180

6.3.8 Crystallization and data collection of Endo Hf treated rat
Flag α9 3Mut/α-bungarotoxin
Crystallization of the complex ofrat Flag α9 3Mut/α-bungarotoxin
treated with Endo Hf was attempted. Instead of screening conditions using
screening kits, it was tried to screen around the condition in which Flag α9
3Mut/α-bungarotoxin was crystallized (0.1 M ammonium acetate, 0.1 M BIS-
TRIS pH 5.5, and 17% w/v PEG 10K). The optimized crystallization
condition for the complex was 0.04 M – 0.07 M ammonium acetate, 0.1 M
BIS-TRIS pH 5.5, and 15% – 16% w/v PEG 10K. Stick-like crystals were
formed within a week, and they often crystallized as a bundle as seen in
Figure 6.9, Drop 1. The size was 10 μm x 100 μm – 20 μm x 200 μm. Some
drops had few crystals which grew longer (10 μm – 20 μm x 200 μm – 400
μm), and they were more likely a single crystal rather than a bundle (Figure
6.9, Drop 2). An interesting feature of these crystals was their morphologies;
they looked like a tube or a duct (Figure 6.9, Drop 2 zoomed in view). Those
crystals were fairly stable, and they were intact even after a month.
Data were collected at the Advanced Photon Source (APS) beamlines
19-ID-D and 23-ID-D at Argone National Laboratory. Data were processed
and scaled using HKL2000
160
. Crystals had very limited diffractions at the
beamline 19-ID-D; however, several good data sets were obtained at the
beamline 23-ID-D using its special functions of the station, raster and vector.
The raster function made it possible to screen a good diffracting portion of a
crystal before collecting a data; therefore, the best quality of data set could be
collected for each crystal. The other function, vector, was suitable for a stick-
like long crystal. Those thin crystals could easily decay under strong X-ray
beam so that it is difficult to collect enough data set to solve a structure.
However, using the vector function, a crystal moves along the crystal as well
181

as it rotates like a helix to minimize the X-ray exposure on the same spot of a
crystal. After obtaining some data sets, HKL2000 was used to process the data,
and a structure solution was tried to solve using CCP4 package
161
. There were
some difficulties in the initial process of solving the structure. First, peak
search in HKL2000 was critical to indexing compared to data of other crystals.
Additional peaks needed to be selected to process the data even though it
looked as if enough peaks were selected. Second, the crystal had a twinning
issue, in which more than two crystals of the same protein are joined together
in different orientation. Fortunately, the twinning data could have been
separated by using de-twinning function in Refmac5
164, 165
. Third,
determination of a space group was challenging which could have caused by
the twinning issue. There were four possible Bravais Lattices, primitive
hexagonal, C centered orthorhombic, primitive monoclinic, and C centered
monoclinic. At the beginning, it was tried to solve a structure with primitive
hexagonal; however, several unusual outcomes were observed in possible
space groups such as P61. The electron density seemed to match well with a
structure, but the number of intensity of systematic absence, which is one of
the tools to determine a space group, did not agree with the space group. The
major problem was seen in an electron density map; there was unknown
density along a screw access. The presence of the density did not make sense
considering a size of the density and the protein.
Several data sets were collected from the same crystal, and two data
sets were merged and scaled using HKL2000 package
160
to solve the structure
followed by conversion of scale files to mtz files using Scalepack2mtz in
CCP4 package
161
. The structure was solved using several programs in
CCP4
161
. The structure was finally solved as primitive hexagonal with the unit
cell dimensions of a = 129.327 Å, b = 129.327 Å, and 53.747 Å at 3.0 Å, and
it belongs to the space group P3
1
(Table 6.1). After running the
182

Matthews_coef program, three copies of α9/α-bungarotoxin were detected in
an asymmetric unit. The human α9/α-bungarotoxin structure (PDB: 4UY2)
was used as one ensemble for a search model for molecular replacement. It
was necessary to select one set of complex beforehand using Pbdset since it
contained two sets. The molecular replacement failed when each molecule
(human α9 and α-bungarotoxin) was separately added as individual ensembles.
Molecular replacement was pefromed using Pharer MR
162, 163
, and two
solutions were obtained (selected solution: RFZ=6.1 TFZ13.7 PAK=0
LLG=262 RF++ TFZ=15.1 PAK=0 LLG=251 RFZ=3.6 TFZ=8.7 PAK=0
LLG=191 LLG=1172). The result also showed a warning of twinning crystals.
Therefore, an amplitude based twin refinement was performed using
Refmac5
164, 165
. The R
work
/R
free
was 27.7/35.8 without the twin refinement, and
the R
work
/R
free
was 24.8/30.5 with the amplitude based twin refinement at 3.0
Å cut-off. The resulted solution was tested based on twin laws in Phaser, and
the structure no longer showed a twinning issue. The possible of a model bias
due to a twinning issue was also examied by introducing correct and false
amino acids (details are discussed later). After the examination, our structure
was confirmed to be reliable. Since our structure was rat α9, the side chains
were replaced from human α9 residues using Coot
167
. Our structure showed
the density of a sugar chain, which was necessary to be built in. The sugar
chain was adapted from the previous α211/α-bungarotoxin structure
(PDB:2QC1) using O
166
followed by refinements. Sugar molecules which did
not fit into the electron density were removed. The structure was analyzed
using Procheck, and Ramachandran Plot can be found in Appendix A. The
final R
work
/R
free
was 22.2/31.6 after multiple refinements. The resolution cut-
off was determined at I /σI = 1, and values of crystallographic analysis are
shown in Table 6.1. The resolution was high enough to observe a density of a
sugar chain which was approaching toward α-bungarotoxin.
183

Figure 6.9: Crystals of Endo Hf treated rat Flag α9 3Mut complex.
Tube-like crystals were obtained, and they were fairly stable. Some grew as a bundle,
and some grew as a single crystal with larger size. The condition which gave crystals
of rat Flag α9 mutants complex was used as a starting point to optimize a
crystallization condition. A protein concentration was 6 mg/mL, and a drop size was
0.5 µL protein + 0.5 µL. Data was collected with those crystals and a structure was
solved.

184

Table 6.1: Crystallographic analysis of Endo Hf treated rat Flag α9 3Mut
complex.
Data was processed with HKL2000 package, and structure was solved by molecular
replacement using PDB: 4UY2 followed by amplitude based twin refinement with
Refmac5. Highest resolution shell is shown in parenthesis.

Data Collection
Space group P3
1
Cell Dimensions
a, b, c ( Å) 129.3

129.3
53.7
α, β, γ (°) 90.0, 90.0, 120.0
Resolution (Å) 50.0 – 3.00
(3.05 – 3.00)
R
sym
(%) 8.4 (0.00)

I/σI 24.7 (0.4)
Completeness (%) 99.7 (93.8)

Redundancy 9.8 (5.8)

Refinement
Resolution (Å) 48.46 – 3.00
No. relections 19,032
R
work
/R
free
22.2/31.6
No. atoms
Receptor 1780
α-Btx 548
Water n/a
B-factors (Å
2
)
Protein 134.8
α-Btx 148.7
Water n/a
R.m.s deviations
Bond length (Å) 0.012
Bond angles (°) 2.015
185

6.3.9 Structure guided analysis of α9
The phase was determined by molecular replacement using the
previous human α9 ECD structure bound by α-bungarotoxin (PDB: 4UY2) as
the search model. Despite crystal twinning, the structure was refined
reasonably well though some flexible loops had week density (Table 6.1). The
main focus of this study was the sugar chain which is located at the site
between the Cys loop and loop C of the α9 subunit. There are two
glycosylation sites in the subunit, and one of them (at Asn32) should have
been cleaved off with EndoHf. The search model (PDB: 4UY2) did not
contain either sugar chains; therefore, extra density should be observed in the
2Fo-Fc map if it exists in the current structure. Indeed, a mass of extra density
was observed under loop C of α9 subunit (Asn145), which could be
representing a sugar chain (Figure 6.10, A). The Fo-Fc map also confirmed a
strong positive density of the sugar chain (Figure 6.10, B). The density was
detected in all three complexes in the asymmetric unit indicating that it could
be a real density from the structure. To check the gycosylation site, Asn32
where the sugar chain should not be detected was first analyzed. As expected,
there was no extra density around the area. Instead, there was a density which
corresponds to only one glycan subunit (Figure 6.10, C). The Fo-Fc map also
did not show any positive density at the site. Based on these observations, it
was promising that our α9 structure would reveal interesting sugar chains that
interact with the receptor and/or α-bungarotoxin. It also confirmed that our
strategy to cleave only a sugar chain at Asn32 worked out very well. The
sugar chain at Asn32 was indeed exposed, while Endo Hf had a limited access
to Asn145. The density of a possible sugar chain is exciting since it would be
an indication of glycan interaction to α9 and/or α-bungarotoxin. Sugar chains
are known to be flexible, and it would be harder to obtain electron densities
unless they are stabilized by some interactions.
186

Since it was a twin crystal, there was still a concern that the structure
could have been biased by the model. In order to eliminate the possibility, a
simple test was performed by modifying side chains. Since three mutations
(V189W, S191F, G193S) were introduced into our structure, side chains were
modified at two different positions: a correct mutation (V189W) and a false
mutation (I190W). If nicely fitted electron density appears at the faulse
mutation position, the structure could be model-biased. Before modifying side
chains, 2Fo-Fc and FoFc maps were determined. Extra density in the 2Fo-Fc
map was already seen at 189 position before replacing amino acids (Figure
6.11, A). It indicates that the residue should have an amino acid which has a
larger side chain than Val. The Fo-Fc map also confirmed that there was
strong positive density at the position (Figure 6.11, B). As a negative control,
the next residue I190 was replaced by Trp as with V189. Extra density was
not observed at 190 in the neither the 2Fo-Fc map nor the Fo-Fc map. After
introducing mutations, the structure was again refined with Refmac 5. A nice
fitting electron density was observed at 189 where a residue was actually
mutated to Trp at DNA level. On the other hand, a weak density appeared at
190 where Ile is a correct amino acid (Figure 6.11, C). This means that
electron densities were not biased depending on amino residues provided.
Based on the result, it can be concluded that our structure was not biased by
the model; therefore, the sugar density observed would be reliable even
though the data is from a twin crystal. After it was confirmed that the structure
was not biased by the model, structure was further refined. Some residues
needed to be replaced to correct residues since the model used was human α9
subunit, and our structure was rat α9 subunit. There were differences in nine
amino acids. Also, side chains of three mutations which were introduced for
enhanced α-bungarotoxin binding were also modified to right amino acids.
187

Figure 6.10: Electron density maps at glycosylation sites.
The blue density indicates 2Fo-Fc map (counter at 1 sigma), and the yellow density
indicate FoFc map (counter at 2 sigma). (A) 2Fo-Fc map around N145 of α9. An
extra density can be seen at a red arrow. The original model had one sugar unit
showing with yellow sticks. (B) 2Fo-Fc (blue) and Fo-Fc (yellow) maps are shown.
The yellow density clearly shows a missing sugar chain density (red arrow). (C) 2Fo-
Fc (blue) and Fo-Fc (yellow) maps are shown at N32 of α9. A sugar chain at N32
should have been cleaved off, and it indeed does not show extra density. Only one
sugar molecule is detected (red arrow).

A
B
C
188

Figure 6.11: Introducing a correct and a false mutation to test model-bias.
(A) Electron density of 2Fo-Fc map is shown in blue. An extra density (pointed with
a red arrow) can be observed at 189 where it is Val in the original structure (wild
type). Trp mutation is introduced at a DNA level, and the extra density represents the
Trp. (B) Fo-Fc map (shown in yellow) is added. It also indicates a density of mutated
Trp (red arrow). A false mutated residue is added at 190 where the map does not
show an extra density (yellow arrow). (C) After refinement, a new 2FoFc map is
shown in purple. A nicely fit electron density is shown at 189 where Trp is a correct
residue (red arrow), whereas a poor electron density is shown at 190 where Ile is an
actual residue (yellow arrow). This test shows that the structure from a twin crystal is
not significantly biased by the model used.

V189
I190
V189
I190
V189W
I190W
A
C
B
V189W
I190W
189

After several rounds of refinements, a sugar chain was added based on
the previous structure of α211 ECD bound by α-bungarotoxin (PDB:2QC1)
30
.
The density resembled the high mannose sugar structure see in α211/α-
bungarotoxin structure. In fact, major modifications of the sugar chain were
not necessary when it was adapted to our structure. The further refinement
was performed, and sugar molecules which did not match with electron
densities were removed. A nice density of the sugar chain was observed in all
three molecules in the asymmetric unit, and interaction with α-bungarotoxin
was also seen (Figure 6.12). However, it looked different from the α211
ECD/α-bungarotoxin structure. Thus, the final structure was superimposed
with α211 ECD/α-bungarotoxin structure, and sugar chains were compared
(Figure 6.13).

Figure 6.12: Electron density map after building in a sugar chain.
A sugar chain is built in based on the one from α211 ECD/α-bungarotoxin. A clear
fitted electron density is seen after a refinement (red arrow). The sugar chain is in
contact to a bottom part of α-bungarotoxin.
190

Figure 6.13: Comparison of sugar chains between α211 complex and α9
complex.
Blue indicates a sugar chain of α211 and pink indicates a sugar chain of α9. White
blue color is α211/α-bungarotoxin complex and gray color is α9/α-bungarotoxin
complex. Sugar chains support α-bungarotoxin in different manners. A sugar chain of
α211/α-bungarotoxin complex is stretched out between α211 and α-bungarotoxin,
whereas a sugar chain of α9/α-bungarotoxin complex supports α-bungarotoxin from
the bottom.

191

The two structures of α211 ECD/α-bungarotoxin and α9 ECD/α-
bungarotoxin were superimposed based on the receptor. A glycosylation site
of α9 is shifted up by two amino acid residues compared to α211. In previous
structural study, it was shown that the sugar chain was supporting α-
bungarotoxin to bind to α211
30, 75
. Some interactions were observed in both
α211/sugar chain and α-bungarotoxin/sugar chain. The sugar chain was also
stretched out between α211 and α-bungarotoxin. It was interesting to see that
a sugar chain interacted completely different from the previous structure. It
seems that the sugar chain in α9 supports α-bungarotoxin by pushing up from
the bottom, and the sugar chain was spread to horizontally rather than
vertically. It is possible that the sugar chain was pushed down by α-
bungarotoxin binding. Without the α-bungarotoxin binding, the sugar could be
stretched out making a contact with loop C.
The interactions of three mutated residues were determined and
compared with α211 ECD/α-bungarotoxin toxin structure (Figure 6.15). In
α211 ECD/α-bungarotoxin toxin structure, W189 interacted with the sugar
chain, and it was indirectly supporting α-bungarotoxin binding. However, in
α9 ECD/α-bungarotoxin structure, it seems to be directly supporting a loop of
α211 ECD/α-bungarotoxin. The residue F191 has a similar interaction to the
one in the α211 ECD/α-bungarotoxin toxin structure thought it had two
different positions in the α9 complex structure. The residue inserts into a
groove of α-bungarotoxin. This residue would provide enough interaction to
hold α-bungarotoxin. This agrees with a result of a gel shift assay (Figure 6.2).
Even though S193 does not seem to have a strong interaction with α-
bungarotoxin, the toxin could have better interaction than Gly.

192

Figure 6.14: Interactions of three mutations introduced to enhance α-
bungarotoxin binding.
(A) Side view of α9/α-bungarotoxin complex. (B) Superimposed with α211/α-
bungarotoxin comples (shown in blue and light blue). Side view. (C) Top view of
α9/α-bungarotoxin complex. (D) Superimposed with α211/α-bungarotoxin comples
(shown in blue and light blue). Top view. (A) and (B) show a close interaction of
F191; a residue sticks into a groove of α-bungarotoxin loops. W189 supports a loop
of α-bungarotoxin, and S193 would provide an interaction with α-bungarotoxin. (C)
and (D) show that the three mutated residues are similar to restudies of α211.

A
C
B
D
W189
F191
S193
W189
F191
S193
193

6.4 Discussion
A structure of the nAChR α9 subunit was solved. The construct of
nAChR α9 subunit was made based on α211, and its expression was
optimized. It was achieved to obtain ~ 2 – 3 mg of protein from 1L of starter
culture. Crystallization was attempted, but it was challenging to obtain
diffracting quality of crystals. The key point of the α1 subunit crystallization
was addition of α-bungarotoxin to stabilize loop C of α1 which is flexible.
Even though α-bungarotoxin is known to bind to the α9 subunit, a weak
binding was observed. In order to enhance α-bungarotoxin binding, three
mutations (V189W, S191F, G193S) were introduced based on the structural
information of the α211 ECD/α-bungarotoxin. A complex single mutant
(S191F) was also made, and it was shown that the single mutation was enough
to enhance binding of α-bungarotoxin with a gel shift assay. Once tighter
binding of α-bungarotoxin was confirmed, a complex of α9 3MUT ECD/α-
bungarotoxin was purified to set crystallization trays. After screening
conditions, preliminary crystals were obtained. Crystallization conditions
were optimized, and large enough crystals were produced for data collection.
However, they did not diffract well to provide a complete data set. Our
hypothesis was that an additional sugar chain on α9 subunit could be
interfering crystal packing since it is located close to the top of the receptor. A
sugar chain which stretches out from N145 is believed to be important, and it
showed to help α-bungarotoxin binding in α1 subunit
30, 75
. Mutation was
introduced at Asn32 and it reduced protein production (data now shown).
Therefore, it was attempted to remove only the sugar chain on the top of the
receptor. In order to remove only one sugar chain, α-bungarotoxin was added
first to form a complex followed by Endo Hf treatment. The sugar chain at
N145 is under loop C; therefore it would be harder to be cleaved off compared
to the other sugar chain which is more exposed. It indeed showed differences;
194

one of the two sugar chains was removed easily, but the other was not. Even
though there was not a proof that a sugar chain which was easily removed was
the sugar located at the top of the receptor, it was more likely true based on
knowledge from the structure of α211 ECD/α-bungarotoxin.
Finally, well-diffracting crystals were obtained, and a full data set was
collected. Crystals were also more stable than crystals of fully glycosylated
proteins which showed decay within a week. In this regard, the sugar chain
located at the top of the receptor (Asn32) could have actually interfered
crystal packing. Unfortunately, crystals were twinned, and amplitude based
twin refinement was necessary using CCPP4
161
Refmac 5
164, 165
. The phase
was determined by molecular replacement, and three copies of α9 ECD/α-
bungarotoxin complex were found in the asymmetric unit. Despite crystal
twinning, reasonable electron densities were obtained to solve the structure
(Rw = 22.2 % and Rfee = 31.6 % at 3Å resolution) though electron densities
of some flexible loops were weak. Importantly, the expected sugar density
was observed. Since the search model (PDB: 4UY2) does not contain a sugar
chain, it was shown as extra density in the 2Fo-Fc map. In addition, the Fo-Fc
map showed a strong positive density indicating existence of the sugar chain.
The sugar chain at the top of the receptor was indeed cleaved off showing
only one sugar residue, whereas extensive density was observed under loop C.
Due to the crystal twinning, the possibility of the effect of the model bias
cannot be completely ruled out. However, the authentic features of the crystal
structure were observed at sites where mutations were introduced. Extra
density was observed at V189 where the residue was mutated to W189. It was
also confirmed by a strong positive density in the Fo-Fc map. A fake mutation
was also introduced to test the model bias, and a nice fitted electron density
could not be obtained. These results confirmed that our structure is reliable,
and the sugar chain is a real. The sugar density resembled the high mannose
195

sugar structure seen in α211 ECD/α-bungarotoxin complex structure. A sugar
chain from α1 was introduced into our structure, and further refinement was
performed. Amazingly, it was not necessary to make major modifications.
Branching patterns of sugar molecules seemed to be conserved; therefore, the
refinement was successfully done by simply orienting the entire sugar chain to
the right position. A clearly fitted electron density for a sugar chain appeared,
and it showed an interaction with α-bungarotoxin. The position of the sugar
chain was different from the one in α211. It was stretched out horizontally,
and it held α-bungarotoxin from the bottom. Even though the orientation of a
sugar chain was different between two structures, α-bungarotoxin clearly
interacted with the sugar chain. It also reconfirmed the importance of the
sugar chain in α1 subunit, and it showed how snake toxins have naturally
evolved to target the α1 receptor.
Lastly, three mutations successfully increased the binding affinity of
α-bungarotoxin. The crystal structure revealed that the side chain of F191
stuck into a groove of α-bungarotoxin; therefore, it enhanced a binding.
However, two different positions were observed for the residue though it had
only one position in α211 ECD/α-bungarotoxin complex structure. A gel shift
assay showed that the single mutation, S191F was enough to shift a band. In
α211/α-bungarotoxin complex structure, W189 indirectly supported the
binding of α-bungarotoxin through interacting with a sugar chain. However, it
was directly interacting with a loop of α-bungarotoxin in the α9 complex
structure. The loop was located closer to α9 subunit in distance compared to
α211. A sugar chain went through the space between the loop of α-
bungarotoxin and α211 loop C in the structure of α1/α- bungarotoxin.
Although further studies and analysis are necessary to determine the
functional or importance of the sugar chain, it is clear that the sugar chain
196

indeed interacts with α9 subunit or α-bungarotoxin since it adopts a well
defined structure of all three molecules in the asymmetric unit. Our structure
with a sugar chain would provide more insights to understand nAChRs.
6.5 Materials and Methods
6.5.1 Construction of α9 expression vector
The cDNA of rat α9 was provided by Dr. Douglas Vetter, Department
of Neuroscience, Tufts University of Medicine. The insert was PCR amplified
and inserted in pPICZαA vector using EcoRI and XbaI sites for rat α9 and
XhoI and XbaI sites for rat Flag α9 and its mutants. All constructs contained
His-tag at the C-terminus, and Flag-tag was added at the N-terminus except
rat α9. Site-directed mutagenesis was performed to introduce mutations
following a protocol of QuikChange II Site-Directed Mutagenesis Kit (Agilent
Technologies, Inc.). Mutations were introduced to rat Flag α9 construct, and
those mutations were Val186Trp, Ser188Phe, and Gly190Ser for rat Flag α9
3Mut and Ser188Phe for rat Flag α9 single Mut to enhance α-bungarotoxin
binding.
6.5.2 α9 expression and purification
Linearized constructs by SacI were transformed into KM71H of P.
pastoris (Invitrogen) by electroporation and plated on YPDS plates which
contained 100 µg/mL Zeocin. Protocols of α9 expression and purification
were same as the ones of α211, and can be found in Chapter 2 materials and
methods (Section 2.5.2 and 2.5.3). One modification was that multiple runs (3
– 4 runs) of Superdex 75 10/300 (GE Healthcare) were necessary due to the
high expression of α9.
197

6.5.3 Gel shift assay
Equimolar ratio of α9 (rat Flag α9 3Mut/rat Flag α9 single Mut) and α-
bungarotoxin were mixed (0.5 µg of α9 was used), and the mixture incubated
on ice for 1 hour. Mouse α211 was used as a positive control. Refer to Chapter
2 material and methods for detailed protocols (Section 2.5.7).
6.5.4 Crystallization of rat Flag α9 mutants with α-
bungarotoxin
After obtaining rat Flag α9 mutants proteins, the complex of rat Flag
α9 mutants and α-bungarotoxin were purified. These proteins were mixed at
1:2 molar ratio and incubated on ice for an hour before being subjected on
Superdex S75 10/300 with 150 mM NaCl and 20 mM HEPES, pH 7.5 buffer.
Two peaks were obtained: one is for the complex, and the other is for excess
amount of α-bungarotoxin. After each fractions was identified and confirmed
by OD
280
and SDS-PAGE gels, fractions of the rat Flag α9 mutants/α-
bungarotoxin complex were pooled and concentrated for crystallization. First,
crystallization was attempted with rat Flag α9 3Mut/α-bungarotoxin.
Crystallization conditions were screened using Crystal Screen, Crystal Screen
2 and Index kits (Hampton Research) by the hanging drop method at room
temperature. Each drop contained 0.5 μL of protein and 0.5 μL of reservoir
solution, and 0.5 mL of screening solution was added in each well. Screening
was performed with 3 mg/mL protein, but various protein concentrations (1
mg/mL, 3 mg/mL, and 6 mg/mL) and different temperatures (4 °C and room
temperature) were also tested. Small cluster crystals were observed in two
conditions. One condition was 0.1 M ammonium acetate, 0.1 M BIS-TRIS pH
5.5, and 17% w/v PEG 10K, and the other condition was 0.2 M lithium sulfate
monohydrate, 0.1 M BIS-TRIS pH 5.5, 25% w/v PEG 3350. The former
198

condition was tried to optimize first, but there was a limitation in diffraction.
Thus, the latter condition was also optimized. The optimization of rat Flag α9
single Mut/α-bungarotoxin was performed around the conditions which the
complex of rat Flag α9 3Mut was crystallized. Harvest solution for the first
condition was 0.05 M ammonium acetate, 0.15 M sodium chloride. 0.1 M
BIS-TRIS pH 5.5, and 25% w/v PEG 10K (added 25% glycerol into cryo
solution). For the second solution, harvest solution was 0.1 M lithium sulfate
monohydrate, 0.15 M sodium chloride, 0.1 M BIS-TRIS pH 5.5 and 25% w/v
PEG 3350 (added 20% glycerol into cryo solution). Dehydration of crystals
(only the first crystallization condition) were attempted to improve crystal
packing using two different solutions: one with PEG 10K based and the other
with glycerol based. Concentration of PEG 10K was increased by 5% up to
40% (total three solutions), and glycerol concentration was also increased in
the similar manner. Crystals were crashed in 35% glycerol solution by
osmotic pressure, but they seemed to be fine in high concentration of PEG
10K solution. Crystals were incubated in each solution for different time
length (0 hour, overnight, and 72 hours).
6.5.5 Endo Hf digestion
Endo Hf was purchased from Invitrogen, and a protocol was optimized
based on the original protocol provided by the manufacture. A time course
experiment was performed in a native condition with a small scale (40 µg of
rat Flag α9 3Mut). Rat Flag α9 3Mut with (1:2 molar ratio) and without α-
bungarotoxin was incubated on ice for 2 hours, and then 1 µL Endo Hf per 4
µg of protein was added together with 10x G5 Reaction Buffer supplied
together with Endo Hf by the manufacture. Samples were incubated at room
temperature, and four different time point samples (0, 24 hours, 48 hours, and
72 hours) were taken. SDS-PAGE samples were made by mixing with 5x SDS
199

loading dye with heating and stored at -20 °C until SDS-PAGE gels were run
to prevent further Enfo Hf activity. The amount of rat Flag α9 3Mut was
scaled up to 1.5 mg for large scale Endo Hf digestion to perform
crystallization. The same ratio of Enfo Hf was used, and sample was
incubated at room temperature. Each time point sample (0, 24 hours, 48 hours,
and 72 hours) was taken as with the small scale experiment. Only one round
of purification was enough to obtain crystals of the complex; therefore, further
optimization for incubation time was not performed. The protein (rat Flag α9
3Mut) used for crystallization was from the large scale experiment, and it was
incubated with Endo Hf for 72 hours.
6.5.6 Endo Hf treated rat Flag α9 3Mut/α-bungarotoxin
complex purification
After the digestion was confirmed with an SDS-PAGE gel, Ni-NTA
purification was performed to remove Endo Hf. Detailed information of Ni-
NTA purification can be found in Chapter 2, Section 2.5.3. In this purification,
500 µL of Ni-NTA beads were sufficient. The elution was run over a size
exclusion column (Superdex 75 10/300 GL, GE Healthcare) with 150 mM
NaCl and 20 mM HEPES pH 7.5 buffer. A complex formation was confirmed
with an SDS-PAGE gel, and fractions of the complex were concentrated dwon
for crystallization.
6.5.7 Crystallization of Endo Hf treated rat Flag α9 3Mut/α-
bungarotoxin complex
Crystallization conditions were screened around the condition of rat
Flag α9 3Mut/α-bungarotoxin (0.1 M ammonium acetate, 0.1 M BIS-TRIS pH
5.5, and 17% w/v PEG 10K) using 6 mg/mL protein by the hanging drop
method at room temperature. Each drop contained 0.5 μL of protein and 0.5
200

μL of reservoir solution, and 0.5 mL of screening solution was added in each
well. Tube-like crystals were obtained, and those crystals were harvested by
transferring to different concentrations of harvest/cryo mixture solution to
protect crystals from osmotic shock (100% harvest solution, 1:3 harvest/cryo
solution, 1:1 harvest/cryo solution, 3:1 harvest/cryo solution, 100% cryo
solution). Harvest solution was 0.05 M ammonium acetate, 0.15 M sodium
chloride, 0.1 M BIS-TRIS pH 5.5, and 25% w/v PEG 10K, and 25% glycerol
was additionally added to cryo solution. Crystals were incubated in each
solution for 5 - 10 minutes before transferred to a next solution.
6.5.8 Data collection
Data were collected at the Advanced Photon Source (APS) beamline
23-ID-B at Argone National Laboratory using a 20 x 20 beam (λ = 1.0332 Å,
12.000 keV) with the attenuation factor of 2.0 and a Pilatus 3 6M detector
(423.6 x 434.6 mm
2
sensitive area, 172 x 172 µm
2
pixel size) (DECTRIS Ltd.,
Baden, Switzerland). The detector distance was 350.0 mm. The oscillation
range and the exposure time per frame were 0.5° and 0.5 sec, respectively.
Data were processed and scaled using HKL2000 package. The phase was
determined by molecular replacement using Phaser MR
162, 163
the coordinates
of α9 and α-bungarotoxin complex (PDB: 4UY2). Refmac5
162, 163
was used to
refine the structures, and further structural model buildings were also carried
out in O
166
. It was a twinning crystal; therefore, amplitude based twin
refinement was performed. After several rounds of refinement, some residues
were replaced to the correct residues (human to rat and three mutations). A
sugar chain was also inserted using the one of α211 complex as a template.
Further refinements were done with Refmac 5.

201

Chapter 7

Expression and Purification of
nAChR α7/AChBP Chimera

7.1 Abstract
The α7 nAChRs are expressed predominantly in the brain and play
important roles in cognition, learning and memory. They are deeply related to
neurological diseases such as Alzheimer’s disease and schizophrenia. Some
drugs which work as agonists to the receptors showed positive therapeutic
effects, but only a few drugs are approved to treat symptoms. One drug used
to treat Alzheimer’s disease is cholinesterase inhibitor, which works by
blocking the activity of cholinesterase which normally breaks down the ligand,
acetylcholine, of α7 nAChRs
246
. There are several drugs which directly target
α7 nAChRs. Structure-guided drug design would help to develop more
effective drugs for such diseases. In this chapter, purification of nAChR α7-
AChBP chimera is discussed. Optimization of purification is the first critical
step of a structural study. It would enable us to understand the receptor better
and develop effective drugs in the future.
202

7.2 Introduction
nAChR α7 subunit is one of the neuronal nAChRs, and it plays an
important role in mediating neurotransmission in the central nervous system.
Nine subunits (α2 – α7 and β2 – β4) are known to exist in the mammalian
brain
247-249
, and the α7 subtype nAChR is predominant as well as the α4β2
subtype receptors
17, 20, 23, 250-252
. Interestingly, the α7 nAChR has been reported
to be expressed in not only the brain
253, 254
but also immune cells
255-259
and
neurons
260
. It is also known for high calcium permeability, which would
trigger a signaling cascade and release neurotransmitters. α7 nAChR is
believed to be closely related to some neurodegenerative diseases (e.g.
Alzheimer’s disease and Parkinson’s disease), neurodevelopmental disorders
(e.g. autism and schizophrenia), and chronic pain and is, therefore, a good
target for therapeutic drug developments
261
.
Several drugs shown to target α7 nAChRs and α4β2 nAChRs have
been discovered, and they show some positive effects in animal models
262-265
.
GTS-21 was tested on healthy volunteers in clinical trials, and it showed
improvements in memory and attention
266
. EVP-6124 is a novel drug, which
is currently in Phase III clinical trials for Alzheimer’s disease. This compound
is known to be a partial agonist of α7 nAChRs
267-270
. It was also shown to be
α7 selective and does not have influence on α4β2 nAChRs
271
.
Crystal structures would enhance drug design and development. It
would be possible to improve functionality by making modifications on drugs
which failed in clinical trials based on crystal structures. Some structures are
available to study nAChRs including a cryo-EM structure of the Torpedo
AChR
70
and X-ray crystal structures of AChBP
38, 60
, ECD of muscle-type
nAChR α1 subunit
75
, prokaryotic orthologs
76-78
, and ECD of nAChR α9
subunit
86
. Those structures provide insights into detailed structure of nAChRs;
203

however, α7/AChBP chimera structure which was solved by our lab would be
the closest model for the α7 subunit considering the maintained functionality
of the receptor as a pentamer. It would help to guide structure-based drug
design for diseases related to α7 subunit.
In a previous study, a nAChR α7-AChBP chimera was successfully
expressed and purified as a pentamer
85
. The engineered protein was necessary
because native α7 nAChR (ECD) could not be produced due to solubility and
folding issues. Since AChBP was known to be expressed with high yield in a
soluble form, some AChBP residues were introduced to α7 nAChR. Some
hydrophobic residues were replaced to increase solubility by maintaining the
important residues and loops of α7 nAChR. The chimera protein would keep
the pentamer formation, and also showed similar ligand bindings as the native
α7 nAChR. The chimera construct shares 64% sequence identity and 71%
sequence similarity with native human α7 subunit (Figure 7.1).
It is necessary to establish a robust purification protocol of α7/AChBP
chimera protein to learn how α7 nAChRs and therapeutic compounds interact.
Knowledge which was obtained in the purification of α211 was applied, and a
new construct designed for an improved purification system. The chimera
proteins are purified with large quantity as pentamers. It is important to
produce the protein routinely not only to determine structures leading to drug
discovery but also to perform other biochemical studies such as mRNA
display as discussed in Chapter 5. α7 subunit-directed ligands would be used
to deliver agents across the blood-brain barrier as with immunotherapy. In
addition, the ligands would contribute to the understanding of function and
localization of the receptors. Newly designed α7/AChBP shows differences
from the previously purified protein. Structural studies will be carried out with
other lab members in future, but the first aim was achieved by successfully
purifying the α7/AChBP chimera protein.
204

Figure 7.1: Sequence information of α7/AChBP chimera.
The figure is taken from Li, S. et.al. Nature Neuroscience (2011)
85
. (A) Sequence
alignment of the α7/AChBP chimera and human α7. Orange indicates conserved
residues and yellow indicates homologous residues. (B) Structural image of
α7/AChBP. Residues of AChBP are shown in yellow, and the remaining residues in
blue are human α7.

A
B
205

7.3 Results
7.3.1 α7/AChBP chimera expression and purification
The expression and purification of α7/AChBP chimera protein was
attempted using P. pastoris as with other nAChR subunits, α211 and α9.
(Refer to Chapter 2 and Chapter 6, respectively). The construct design of
α7/AChBP was slightly different from the one created by Dr. Sine’s group
85
.
A Flag-tag was introduced at the C-terminus in their construct, but it was
moved to the N-terminus and a His-tag was added to the C-terminus in this
construct, basically following the same backbone as the α211 construct.
Another change was the removal of the first two amino acids, Glu1 and Phe2,
at the N-terminus, and the final six amino acids, Gly204 to Lys209, at the C-
terminus. Those amino acid deletions were decided based on the previous
crystal structure of the α7/AChBP chimera, in which these residues show no
density
85
. The purification protocol of α7/AChBP was also modified. Instead
of an anti-Flag M2 affinity purification, Ni-NTA purification was performed
in addition to the purification protocols for α211 and α9. However, there was
a major difference between α7/AChBP and other subunits. α7/AChBP was a
pentamer protein whereas α211 and α9 were monomers. Based on studies
carried out by both Dr. Sine’s group and ours, pentamerization of α7/AChBP
was pH sensitive and tended to dissociate and became a monomer at higher
pH (> pH 7.0). It was ideal to perform experiments at pH 6.0, but the lower
pH could cause a problem in Ni-NTA purification. The optimal binding range
of proteins to Ni-NTA beads is pH 7.2 – 7.8 based on the QIAGEN protocol,
but some protein could be still bound at pH 6.0. Crystallization of α7/AChBP
was previously tried at both pH 6.0 and pH 6.5, and crystals were successfully
produced at both conditions. Thus, pH 6.5 buffer was used for Ni-NTA
purification.
206

After Ni-NTA purification, the elute was subjected to Superdex 200
10/300. Both monomer and pentamer peaks were observed on the FPLC
chromatogram though the monomer peak was much smaller than the pentamer
peak (Figure 7.2, A, Peak 3). Each FPLC fraction was determined by an SDS-
PAGE gel, and multiple bands were detected on the gel (Figure 7.2, B). The
expression level of α7/AChBP was high, and ~1 mg of purified protein was
obtained from 1 L of the starting culture.

Figure 7.2: α7/AChBP purification.
(A) Chromatogram of α7/AChBP purification with Superdex 200 10/300. The peak 1
indicates some aggregated proteins with contaminants. The second peak indicates the
pentamer form of α7/AChBP, and the peak 3 is the monomer form. (B) 15% SDS-
PAGE gel of α7/AChBP purification (16 μL of sample + 4 μL of 5x loading dye
except Ni beads: 2 μL of beads + 6 μL of water + 2 μL of 5x loading dye and
Injection: 1 μL of sample + 7 μL of water + 2 μL of 5x loading dye). There was
decent expression shown in the media sample. Even though samples were fairly
clean, multiple bands indicate the existence of several species. Fractions 23 – 27 were
pooled and concentrated for future experiments.

A
B
1
2
3
207

7.3.2 Comparisons of α7/AChBP chimera proteins
Since multiple bands in the newly purified α7/AChBP protein were
detected on an SDS-PAGE gel, further analysis was performed by comparing
with a protein which was expressed from the construct designed by Dr. Sine’s
group (labeled as original). Their protein was successfully crystallized in our
lab before
85, 272
, so it served as a good reference for comparison. The sizes of
the proteins was slightly different, but both samples showed multiple bands
(Figure 7.3, A). Six total bands were observed in the original sample. There
seemed to be three major bands, but each band was actually a doublet,
resulting in six bands. On the other hand, the new sample had only three bands.
One hypothesis for the origin of these multiple bands was a variation
of glycosylation on α7/AChBP. There are two glycosylation sites on the
protein at N64 and N106; therefore, each band would be showing a different
level of glycosylation (top band: fully glycosylated, middle band: only
glycosylated on one site, and bottom band: no glycosylation). Both samples
showed the highest intensity on the top band (fully glycosylated) and the least
intensity on the bottom band (no glycosylation). Most of the protein seemed to
be fully glycosylated. There is no clear explanation for the doublet bands of
the original sample, but it could have resulted from different type of sugar
chains since glycosylation mechanisms are quite complicated. It is also
difficult to explain why different levels of glycosylation were obtained for
α7/AChBP. The α9 protein also has two glycosylation sites, but a single band
was detected showing homogeneity of the protein sample. It is still possible to
obtain crystals using the new samples containing several species since crystals
were obtained with the original sample which has the same issue.
208

7.3.3 Endo Hf digestion on α7/AChBP chimera
In order to confirm that multiple bands resulted from different levels
of glycosylation, Endo Hf digestion was performed. If the hypothesis was
correct, the digested sample would show a single band on an SDS-PAGE gel
after Endo Hf treatment. The experiment was performed in denatured
condition, and it was confirmed that multiple bands were caused by
glycosylation. A clear single band was obtained in both original and new
sample though the size of two samples was slightly different (Figure 7.3, B).

Figure 7.3: Comparison of α7/AChBP purified with different constructs.
(A) 15% SDS-PAGE gel for comparison of two α7/AChBP proteins. One was
expressed from the construct designed by Dr. Sine’s group (original), and the other
was from the construct designed by our lab (new). Different amounts of protein were
loaded to aid in distinguishing multiple bands. (B) 15% SDS-PAGE gel of
α7/AChBP protein treated with Endo Hf to remove glycans. Both α7/AChBP proteins
were treated with Endo Hf under a denaturing condition. The gel shows a clear
removal of sugar chains from both proteins. This gel proves that the multiple bands
observed in A resulted from different level of glycosylation.
A B
209

7.4 Discussion
The α7/AChBP construct was newly designed, and it was expressed
using P. pastoris with a yield of ~ 1 mg per liter of starting culture. A
purification protocol was developed to be suitable for the new construct. The
original construct designed by Dr. Sine’s group did not have a His-tag, but the
tag was introduced for Ni-NTA purification since it is easier and more cost
effective compared to anti-Flag bead purification. The difference could be
more prominent since α7/AChBP protein was secreted into media, that is,
proteins had to be purified from a large volume of liquid. One concern upon
using Ni-NTA beads was buffer pH since it is pH sensitive. It was also known
that the pentamerization of α7/AChBP is affected by high pH; pH 6 would be
an ideal pH for the protein. On the other hand, the optimal pH range of Ni-
NTA beads is pH 7.2 – pH 7.8. The purification protocol was successfully
optimized at pH 6.5, and pentamer formation of α7/AChBP was retained with
that pH.
Purified α7/AChBP showed multiple bands on SDS-PAGE gel. It was
confirmed to result from different levels of glycosylation. Three different
species existed in the sample: no glycosylated protein, glycosylated at one site,
and fully glycosylated protein. While the cause is unclear, the previously
purified α7/AChBP, which had a slightly different construct design, also
showed heterogeneity. The purity of proteins is generally important in
crystallization, but the protein was successfully crystallized in the previous
study. Therefore, it is possible that α7/AChBP expressed by the newly
designed construct could also be crystallized though crystal conditions would
need to be re-screened. Indeed, preliminary crystals have been obtained with
the newly purified α7/AChBP protein (Kevin Wu, unpublished results).
Removal of glycans, especially on the top of the receptor, may help
210

crystallization as was shown with the crystallization of nAChR α9 subunit by
our group (Refer to Chapter 6, Section 6.3.8) as well as another group
86
. Two
glycans are located on the top of the receptor, and they might be interfering in
crystal packing. Among many nAChR subunits, locations of glycosylation
sites are conserved, but α7 has different glycosylation sites. One major
difference is that there is no glycosylation site under loop C which was
discovered to be important in the study of α1/α-bungarotoxin complex
structure
30
. In the previous structures of α7/AChBP chimera, sugar chains
were not observed. That is, those sugar chains did not interact with α7/AChBP
or α-bungarotoixn as seen in the structure of α1 complex. They could be freely
moving; therefore, electron densities could not be detected. The functions of
sugar chains are still not clear, but it would be interesting if structures are
solved with them.
Overall, the success of the α7/AChBP purification would expand
studies of nAChR. The purification of α7/AChBP in large quantity is the first
step of crystallization which leads to detailed structural studies. Now, the
protein expression system is well-developed for crystallization; therefore, the
next step is to solve a structure which could be utilized for drug developments.
If co-crystallization of α7/AChBP with therapeutic compounds is achieved, it
is possible to obtain more detailed information of actual interactions. It would
lead to a further discovery of therapeutic drugs. Lastly, it was proven that our
purification protocol, which was developed with α211, could apply to other
subunits of nAChRs. A limited number of nAChR subunits have been solved,
and therefore, this method could be used for protein purification of other
subunits.
211

7.5 Materials and Methods
7.5.1 Construction of α7/AChBP chimera
The α7/AChBP construct was designed based on the previous
construct designed by Dr. Steven Sine
85
as well as the α211 construct (Refer
to Chapter 2, Section 2.5.1). Briefly, the α7/AChBP sequence was adapted
from Dr. Sine’s construct with few modifications; removal of Glu1 and Phe2
at the N-terminus and removal of six amino acids, Gly204 to Lys209 at the C-
terminus. The position and design of tags were applied from the α211
construct. Flag-tag was introduced at the C-terminus, and His-tag was added
at the N-terminus. The synthesized DNA was ordered from GenScript. The
gene was codon optimized for yeast, and was cloned into a pPICZαA vector
using EcoRI and XbaI sites.
7.5.2 α7/AChBP chimera expression
The α7/AChBP construct was linearized by digestion with Sac I
restriction enzyme and transformed into KM71H of P. pastoris (Invitrogen)
by electroporation. The detailed expression was same as α211 (Refer to
Chapter 2, Section 2.5.2).
7.5.3 α7/AChBP chimera purification
Ni-NTA agarose beads (QIAGEN) were incubated with the
supernatant at 4 °C overnight with end-over-end rotation. Since
pentamerization of α7/AChBP was pH sensitive, a lower pH buffer needed to
be used. The protein was eluted with elution buffer (50 mM NaH
2
PO
4
; pH 6.5,
0.3 M NaCl, 10% glycerol, and 500 mM imidazole) after washing with
washing buffer containing a lower amount of imidazole (20 mM) and 0.1%
212

Triton X-100 to remove loosely bound proteins. Eluted fractions were
determined by SDS-PAGE gels and concentrated down to further purify by a
size exclusion column (Superdex 200 10/300 GL. GE Healthcare). Fractions
of the peak containing pentamer α7/AChBP were concentrated down for
future experiments.
7.5.4 Endo Hf digestion
Endo Hf was purchased from Invitrogen, and the manufacture’s
protocol was followed. Briefly, 1 µL of Endo Hf was mixed with 2 µg of
α7/AChBP, which were expressed with two different constructs: one was
designed by Dr. Sine’s group, and the other was designed by our group. Prior
to the Endo Hf incubation, α7/AChBP proteins were denatured by adding 10x
Glycoprotein Denaturing Buffer, which was supplied with Endo Hf by the
company, followed by heating the reaction at 100 °C for 10 minutes. Then,
10x G5 Reaction buffer supplied by the company was added, and the mixture
was incubated at 100 °C for 1 hour.

213

Chapter 8

Conclusion

Many studies have shown that the nicotine acetylcholine receptors are
closely related to some diseases and substance addiction
5
. These receptors are
found in CNS and PNS as well as in non-neuronal cells. They belong to the
cys-loop ligand-gated ion channel superfamily, and they transmit electronic
signals when neurotransmitter, ACh, binds to them. One of the diseases linked
to nAChR is MG, a neuromuscular autoimmune disease, which is caused by
autoantibodies attacking the α1 subunit of nAChR. Some studies have shown
that these autoantibodies bind to a specific region of the α1 subunit, called
MIR, but very little is known about how they actually bind to it. Our studies
made a huge step to further understand MG and could potentially lead to new
treatments of the disease. The first X-ray crystal structures of the complex of
autoantibodies binding to the α1 subunit were solved at 2.6 Å resolution. The
interface of binding sites was finally revealed and it allows us to study the
antibody-receptor interaction at an atomic level. Two well-known MG mAbs,
mAb35 and mAb210 were used in this study. mAb35 is a prototypical MG
autoantibody and it is known to bind in a conformational dependent manner
131,
135, 148
, whereas mAb210 can bind to a denatured nAChRs. Crystal structures
214

of complexes α1 bound by those antibodies would provide detailed insights to
understand the disease.
To study an interface of nAChR and an autoantibody, a large quantity
of proteins was necessary. In a previous study, the extracellular domain of
nAChR α1 (α211) was obtained for a structural study and necessary mutations
and truncation were determined to stabilize the protein
144
. For high yield
protein purification, however, further optimization was needed to obtain the
protein in our lab. This thesis study showed that a critical modification from
the original protocol was a temperature change during an induction period.
Finally, both mouse and human α211 were successfully expressed and
purified from P. pastoris for crystallization. These purified proteins were then
mixed with Fab35/Fab210 and α-bungarotoxin to form ternary complexes. In
order to prepare Fab35 and Fab210, mAb35 and mAb210 were expressed and
purified from hybridoma cells. The mAbs were then treated with papain to
obtain only Fab portions. First, the complex structure of α1 bound by Fab35
was crystallized for structural study. The findings in the structure were
absolutely stunning. Fab35 indeed targeted the MIR motif of α1, and it
showed that a proper conformation is required to form the interaction network
at the MIR interface. Two critical residues, N68 and D71 played important
roles in binding and these observations are consistent with previous
biochemical studies
134-136
. There are two interfacial water molecules
mediating interactions of α1 and Fab35 at the MIR site. In addition to the
interactions at the MIR, other binding sites were detected which contribute to
the binding of Fab35. The N-terminal α-helix is one of the binding sites,
interacting with both CDR-H2 and CDR-H3. The CDR-H3 shows an unique
feature; R102 which is located the tip of the loop inserts into a groove of α1.
The loop is significantly different from the previously solved Fab structures,
Fab192 and Fab198
151, 152
. The CDR-H3 loops of those Fabs are shorter than
215

one of Fab35; therefore, they would have different interaction patterns than
Fab35.
Next, the crystal structure of Fab210 ternary complex was solved at
2.6 Å resolution. Differences between Fab35 and Fab210 would provide us
more knowledge to understand MG autoantibodies. Both of them are known
to bind the MIR loop of α1, but Fab210 is capable in binding to denatured
nAChRs. The structure of Fab210 enabled us to compare the binding interface
with the one from the Fab35 ternary structure. In fact, Fab210 also showed
extensive interactions at the interface of MIR site as well as with the N-
terminal α-helix of α1. The CDR-H3 loop was similarly inserted into the
surface pocket of nAChR α1. Then, why can Fab210 bind to denatured
nAChR α 1? Our answer is that the binding interactions of Fab210 to nAChR
α1 are much more extensive than those between Fab35 and the receptor. Many
of the interactions for Fab210 are highly inter-connected and focused on
clustered residues on nAChR α1. Moreover, the Fab210 CDRs contain
multiple aromatic residues that pack tightly with each other as well as with
residues on the receptors. Such aromatic packing interactions are similar to
those found at the interface between nAChR α1 and α-bungarotoxin, which
has an extremely high binding affinity
30
. As a conclusion, mAb210 can form
more extensive bindings with nAChR α1 compared to mAb35. If the
conformation of the MIR is retained at some degree, it would be enough for
mAb210 to recognize it, whereas mAb35 interactions are not strong enough to
form bindings. Our Fab35 ternary structure revealed the detailed interactions
at the interface for the first time, and it would be a major breakthrough in this
field. Another major discovery was also achieved by the X-tal structure of
Fab210 which has different behavior than Fab35 (mAB35). There are so many
MG mAbs that are available, and some of them such as Fab192 have been
216

studied extensively showing unique features. It would be useful to obtain
more structural information of various MG mAbs to understand MG further.
Our next interest was to develop a new detection method of MG
autoantibodies which would be simpler compared to the current method, RIA.
A major difficulty of receptor studies is obtaining receptor proteins. Actual
human muscle is often used to extract receptors for RIA, though an alternative
muscle source has been studied, such as TE671 cell line
181-187
. The problem
was overcome by recombinantly expressing α211 proteins. Compared to RIA
which requires radioactive iodine, our system would be inexpensive and thus
could be used as a routine test to monitor the disease status of patients.
Sensitivity of a detection method is necessary to diagnosis MG at the early
stages. Although our method has not been optimized for sensitivity, like the
RIA method (0.5 nM cut off for negative samples), it was achieved to detect a
wide range of MG antibody concentration (1 nM – 200 nM) with pseudo rat
MG samples. Further experiments and optimizations are necessary with
human MG samples including standardization, and it would be possible to
increase a sensitivity of our method to a level comparable with RIA method.
A usage of recombinantly expressed and purified α1 ECD is a big advantage.
It could have a constant supply of the receptor source with a low cost, and also
it would be easier to control an experimental set up by applying modifications
such as His-tag. Even though additional experiments are required, it is
promising to develop a more efficient detection method for MG in near future.
Another focus of our research was to develop a binder protein which
targets a specific nAChR subunit. It would be a useful tool for mechanism
study and drug development if it succeeds. A binder protein which specifically
targets nAChR α1 (mouse α211) was discovered using a technique called
mRNA display
91
. In the assay, the 10FnIII scaffold was used and two loops
resembling to antibodies were diverged to create a library for in vitro selection.
217

Several binder proteins showing high binding affinity were further tested for
expression in E. coli. The binding was also confirmed with a gel shift assay
and a binding specificity was also determined using nAChR α9 subunit.
Among several candidate binders selected by mRNA display, Clone 4.2
showed a high protein expression with the highest stability as well as tighter
binding to mouse α211. The protein still presented a limited stability, and it
was critical to add 10% glycerol to buffer for stabilization. Although proteins
were purified with fairly high purity, diffractable crystals could not be
obtained. Preliminary crystals were grown as small clusters and they were not
improved with optimizations. Mutations were introduced to further stabilize
Clone 4.2 binder protein and it indeed showed improvement, though it did not
provide any better crystals. Even though further attempts are necessary to
obtain a crystal structure of the complex of Clone 4.2 and α1, the mRNA
display study showed how powerful the technique is to select a binder protein
with high binding affinity. Clone 4.2 and its mutant, Clone 4.2 5b,
demonstrated that they could not only compete with α-bungarotoxin but also
they could bind to α1 tighter. More importantly, the natural α-bungarotoxin
cross-reacts with α1, α7, and α9, but mRNA display selected proteins show a
much higher selectivity. The X-ray crystal structure of complex α211/binder
protein could enable us to develop therapeutic and laboratory research tools.
This demonstrates the successful selection of binder proteins with mRNA
display and continued work is necessary for structural studies. The
information of the binding interface would be useful to modulate binders and
to further study nAChRs. The nAChRs are expressed throughout our bodies
and they have high homology. Therefore, it would be a significant advance if
subunit specific binders are developed in the future.
Most of my thesis projects were focused on the nAChR α1 subunit, but
the nAChR α9 was also worked on as another major project. The α9 subunit is
218

one of latest subunits identified in our system among nAChRs. It was first
found in hair cells of the inner ear, but recently it has been shown that it could
be related to pain, cancers, and skin diseases. The α9 is another unique
receptor that can form a homopentamer like the α7 nAChR and also can bind
to α-bungarotoxin. The X-ray structure of human α9 ECD was previously
solved, but sugar chains were removed for a crystallization purpose. The aim
of our study on α9 was a determination of a sugar chain since some studies
suggested it to be important for the receptor functions
30, 31, 74, 75
. Despite
crystal twinning, a sugar chain was detected in our structure, and it showed
differences from the one of α1. The sugar chain of α1 was stretched out
vertically and went through between the α1 receptor and α-bungarotoxin. On
the other hand, the sugar chain of α9 was pushed down under α-bungarotoxin.
It could be because the glycosylation site is shifted to upward in α9. It is still
possible that the sugar chain stretches up without α-bungarotoxin binding, but
there could be a reason for the differences in glycosylation. In addition, the
importance of some amino acids, especially F191 was demonstrated in this
study. Even though α9 is known to bind α-bungarotoxin, the binding is weaker
compared to α1. Thus, mutations were introduced to enhance α-bungarotoxin
binding, and it showed that F191 played an important role in the binding.
These observations show us that α-bungarotixn has evolved so that it can
maximize the binding to the α1 receptor.
Finally, the study of nAChRs was expanded to α7 nAChRs. These
receptors are of high interest to scientists because they are closely related to
neurological diseases such as Alzheimer’s disease and schizophrenia. There is
no cure for these diseases and limited treatments are available in slowing
down disease progression. Further understanding of the receptors is necessary
for treatment development, as well as drug discoveries. A structural study of
protein is often a powerful technique to design drugs. Tremendous
219

information can be obtained from structures to improve drugs. Even though it
is a useful tool for drug discovery, there are some difficulties. A main
challenge is obtaining a stable and soluble protein at a large quantity.
Receptors are especially difficult to recombinantely purify since they are
membrane proteins containing hydrophobic transmembrane regions. In
addition, nAChRs are composed of five subunits in order to form a functional
receptor. Dr. Sine’s group has accomplished a construct design which is
soluble and also forms a pentamer structure by making a chimera with
Lymnaea stagnalis AChBP
85
. Purified proteins were successfully crystallized
to provide us with insights of the receptors. Even though it was the chimera
protein, it has a high sequence homology (64% sequence identity and 71%
sequence similarity with native human α7 subunit). In our study, further
modifications were applied to optimize protein purification. This newly
designed construct produced a large amount of proteins using a protocol that
was modified and optimized. The resulting proteins were not homogenous and
it was concluded to result from different levels of glycosylation. Purity is
important for crystallization but it would be possible to obtain crystals with
those proteins since the original construct (designed by Dr. Sine’s group) also
contained several species. There are some developing drugs which target α7
nAChRs, such as GTS-21 and EVP-6124. Our next aim is to co-crystallize the
α7/AChBP chimera protein with those drugs. Even if those drugs would not
clear all check points of clinical trials, our structure could be a guide in further
develpoment for the drugs to be eventually marketable.
Lastly, I would like to propose a future project combining all
information obtained from our studies, as well as previous studies performed
by various scientists. This proposal would lead to a possible treatment of MG.
First of all, several unique observations were made in previous studies in
terms of the binding of MG mAbs to peptides of α1. Some studies have shown
220

that low affinity but specific binding of some MG mAbs to synthetic peptides
of Torpedo α subunit, α66 – 76
119, 128-131, 134, 135
. In other studies, some mAbs
did not bind to the peptides or denatured nAChRs
120, 135
. Furthermore, it was
discussed that lower detection of mAb bindings in solution compared to the
solid phase
135
. This phenomenon could be explained by an importance of
conformation of the receptors for antibody bindings. Our structures of α1
bound to MG mAbs revealed that a specific conformation is indeed necessary
for antibody binding. The solid phase method would create a more rigid
structure of the peptides by immobilizing to wells. In most experiments,
synthetic peptides were cross-liked with glutaraldehyde to polylysine-coated
wells and some peptides could have been immobilized with a conformation
that resembled to actual nAChRs (Figure 8.1). However, the ratio of peptides
with the right conformation on wells would not be high. In fact, a study
mentioned that very small percentage of peptides were actually active forms
for mAb bindings, even though the solid phase method showed a better
binding than in solution
135
. In summary, a conformation of the MIR is
important and synthetic peptides could mimic the binding site of the MIR by
being anchored on plates (e.g. ELISA).
By considering all information and applying a technique of mRNA
display, a new approach can be taken from a different direction. In
collaboration with Dr. Richards’ group, a process has already been started to
develop a binder protein that would compete with mAb35 to bind α1 by using
10FnIII scaffold. There is a concern for targeting α1subunit since it could
block the receptor function by binding of the selected protein. Since the
interface between α1 and MG mAbs was revealed at the atomic level and the
extensive binding with MIR loop was confirmed, it would be more efficient to
target mAb35 which makes close contacts with the MIR of α1. Although
mAb35 also forms interactions with other sites of α1, the binding of mAb35 to
221

α1 could be inhibited or dramatically reduced by blocking the MIR binding
site. Dr. Richards’ group also developed a technique to select cyclic peptides
which specifically binds to a target protein
273-275
. This approach would be
suitable for our purpose with several reasons. First, previous studies indicate
that anchoring peptides on wells increased mAb bindings and it could be due
to the correct formation of peptides. The cyclic peptides would give a higher
chance for peptides to form the conformation resembled to the MIR. Second,
the length of cyclic peptide matches well with the MIR which is composed of
11 amino acids. For cyclic peptides selection in mRNA display, the length is
usually 14 amino acids and 11 of them would be randomized for specific
binding. Third, it does not require purification of the binder protein using E.
coli. Difficulties of purifying the binder protein have been encountered with
Clone 4.2 experiments (Chapter 5). Therefore, a mass production of the stable
binder molecule would not be a problem compared to 10FnIII scaffold. Lastly,
there is no concern in blocking nAChRs since the cyclic peptides target MG
mAbs, not the receptor as mentioned earlier. This approach would work better
than the current method. The original sequence of the MIR loop would bind
well, but it could achieve even tighter binding with the selection using mRNA
display.
Our crystal structures of the α1/Fab 35 and α1/Fab210 complex
provided not only details insights of the interface of binding site but also a
new approach to develop therapeutic molecules. The proposal was focused on
the interaction at the MIR sites; however, more possible approaches could be
proposed based on our structures that reveal detailed interactions at multiple
binding sites. In addition, all MG antibodies would not bind in a similar
manner. Some would bind to α1 depending on the MIR, and others would
bind to α1 depending on the N-terminal α-helix, which were previously
discussed
132
. Therefore, it would be necessary to develop several binder cyclic
222

peptides which could bind to different type of MG mAbs, or completely new
approaches needs to be taken.

Figure 8.1: Hypothesized schematic view of the immobilized synthetic peptide
which corresponds to the MIR loop on a poly-D-Lys coated well with
glutaraldehyde.
Our crystal structures reveal the importance of the conformation of the MIR loop. In
solid phase experiments, peptides could be cross-liked to a well mimicking the
original conformation; therefore MG mAbs could bind them. The interaction of the
MIR loop with Fab35 provides an idea that cyclic peptides would bind better even
though liner peptides could not. With a mRNA approach, tighter binding of cyclic
peptides could be selected, which would be used as a therapeutic molecule for MG
treatment.

223

Bibliography

1. Kalamida, D. et al. Muscle and neuronal nicotinic acetylcholine receptors.
Structure, function and pathogenicity. FEBS J 274, 3799-845 (2007).
2. Lindstrom, J. Nicotinic Acetylcholine Receptors. Encyclopedia of Life
Sciences (2002).
3. Langley, J.N. On the contraction of muscle, chiefly in relation to the
presence of "receptive" substances: Part I. J Physiol 36, 347-84 (1907).
4. Dale, H.H. THE ACTION OF CERTAIN ESTERS AND ETHERS OF
CHOLINE, AND THEIR RELATION TO MUSCARINE. J Pharmacol Exp
Ther 6, 147-190 (1914).
5. Jackson, M.B. Ligand-gated channel: postsynaptic receptors and drug targets.
Adv Neurol 79, 511-24 (1999).
6. Arcangeli, A., Pillozzi, S. & Becchetti, A. Targeting ion channels in
leukemias: a new challenge for treatment. Curr Med Chem 19, 683-96
(2013).
7. Cesario, A., Russo, P., Nastrucci, C. & Granone, P. Is alpha7-nAChR a
possible target for lung cancer and malignant pleural mesothelioma
treatment? Curr Drug Targets 13, 688-94 (2012).
8. Schuller, H.M. Regulatory role of the alpha7nAChR in cancer. Curr Drug
Targets 13, 680-7 (2012).
9. Lee, C.H. et al. Overexpression and activation of the alpha9-nicotinic
receptor during tumorigenesis in human breast epithelial cells. J Natl Cancer
Inst 102, 1322-35 (2010).
10. Wu, C.H., Lee, C.H. & Ho, Y.S. Nicotinic acetylcholine receptor-based
blockade: applications of molecular targets for cancer therapy. Clin Cancer
Res 17, 3533-41 (2011).
11. Corringer, P.J., Le Novere, N. & Changeux, J.P. Nicotinic receptors at the
amino acid level. Annu Rev Pharmacol Toxicol 40, 431-58 (2000).
12. Lindstrom, J. Nicotinic Acetylcholine Receptors. Encyclopedia of Life
Sciences (2010).
224

13. Zouridakis, M., Zisimopoulou, P., Poulas, K. & Tzartos, S.J. Recent
advances in understanding the structure of nicotinic acetylcholine receptors.
IUBMB Life 61, 407-23 (2009).
14. Gahring, L.C. & Rogers, S.W. Neuronal nicotinic acetylcholine receptor
expression and function on nonneuronal cells. AAPS J 7, E885-94 (2005).
15. Chernyavsky, A.I., Arredondo, J., Vetter, D.E. & Grando, S.A. Central role
of alpha9 acetylcholine receptor in coordinating keratinocyte adhesion and
motility at the initiation of epithelialization. Exp Cell Res 313, 3542-55
(2007).
16. Dani, J.A. Overview of nicotinic receptors and their roles in the central
nervous system. Biol Psychiatry 49, 166-74 (2001).
17. Jones, S., Sudweeks, S. & Yakel, J.L. Nicotinic receptors in the brain:
correlating physiology with function. Trends Neurosci 22, 555-61 (1999).
18. Le Novere, N., Grutter, T. & Changeux, J.P. Models of the extracellular
domain of the nicotinic receptors and of agonist- and Ca2+-binding sites.
Proc Natl Acad Sci U S A 99, 3210-5 (2002).
19. McGehee, D.S. & Role, L.W. Physiological diversity of nicotinic
acetylcholine receptors expressed by vertebrate neurons. Annu Rev Physiol
57, 521-46 (1995).
20. Khiroug, S.S. et al. Rat nicotinic ACh receptor alpha7 and beta2 subunits co-
assemble to form functional heteromeric nicotinic receptor channels. J
Physiol 540, 425-34 (2002).
21. Liu, Q. et al. A novel nicotinic acetylcholine receptor subtype in basal
forebrain cholinergic neurons with high sensitivity to amyloid peptides. J
Neurosci 29, 918-29 (2009).
22. Murray, T.A. et al. alpha7beta2 nicotinic acetylcholine receptors assemble,
function, and are activated primarily via their alpha7-alpha7 interfaces. Mol
Pharmacol 81, 175-88 (2012).
23. Sudweeks, S.N. & Yakel, J.L. Functional and molecular characterization of
neuronal nicotinic ACh receptors in rat CA1 hippocampal neurons. J Physiol
527 Pt 3, 515-28 (2000).
24. Gotti, C. & Clementi, F. Neuronal nicotinic receptors: from structure to
pathology. Prog Neurobiol 74, 363-96 (2004).
25. Lindstrom, J. Nicotinic acetylcholine receptors in health and disease. Mol
Neurobiol 15, 193-222 (1997).
26. Miledi, R., Molinoff, P. & Potter, L.T. Isolation of the cholinergic receptor
protein of Torpedo electric tissue. Nature 229, 554-7 (1971).
27. Kasai, M., Changeux, J.P. & Monnerie, L. In vitro interaction of 1-anilino 8
naphthalene sulfonate with excitable membranes isolated from the electric
organ of Electrophorus electricus. Biochem Biophys Res Commun 36, 420-7
(1969).
28. Noda, M. et al. Primary structure of alpha-subunit precursor of Torpedo
californica acetylcholine receptor deduced from cDNA sequence. Nature
299, 793-7 (1982).
29. Karlin, A. Emerging structure of the nicotinic acetylcholine receptors. Nat
Rev Neurosci 3, 102-14 (2002).
225

30. Dellisanti, C.D., Yao, Y., Stroud, J.C., Wang, Z.Z. & Chen, L. Crystal
structure of the extracellular domain of nAChR alpha1 bound to alpha-
bungarotoxin at 1.94 A resolution. Nat Neurosci 10, 953-62 (2007).
31. Miller, P.S. & Aricescu, A.R. Crystal structure of a human GABAA receptor.
Nature 512, 270-5 (2014).
32. Pedersen, S.E. & Cohen, J.B. d-Tubocurarine binding sites are located at
alpha-gamma and alpha-delta subunit interfaces of the nicotinic acetylcholine
receptor. Proc Natl Acad Sci U S A 87, 2785-9 (1990).
33. Blount, P. & Merlie, J.P. Molecular basis of the two nonequivalent ligand
binding sites of the muscle nicotinic acetylcholine receptor. Neuron 3, 349-
57 (1989).
34. Neubig, R.R. & Cohen, J.B. Equilibrium binding of [3H]tubocurarine and
[3H]acetylcholine by Torpedo postsynaptic membranes: stoichiometry and
ligand interactions. Biochemistry 18, 5464-75 (1979).
35. Sine, S.M. Molecular dissection of subunit interfaces in the acetylcholine
receptor: identification of residues that determine curare selectivity. Proc
Natl Acad Sci U S A 90, 9436-40 (1993).
36. Karlin, A. & Akabas, M.H. Toward a structural basis for the function of
nicotinic acetylcholine receptors and their cousins. Neuron 15, 1231-44
(1995).
37. Unwin, N. Acetylcholine receptor channel imaged in the open state. Nature
373, 37-43 (1995).
38. Brejc, K. et al. Crystal structure of an ACh-binding protein reveals the
ligand-binding domain of nicotinic receptors. Nature 411, 269-76 (2001).
39. Le Novere, N., Zoli, M. & Changeux, J.P. Neuronal nicotinic receptor alpha
6 subunit mRNA is selectively concentrated in catecholaminergic nuclei of
the rat brain. Eur J Neurosci 8, 2428-39 (1996).
40. Gaimarri, A. et al. Regulation of neuronal nicotinic receptor traffic and
expression. Brain Res Rev 55, 134-43 (2007).
41. Boulter, J. et al. Alpha 3, alpha 5, and beta 4: three members of the rat
neuronal nicotinic acetylcholine receptor-related gene family form a gene
cluster. J Biol Chem 265, 4472-82 (1990).
42. Lindstrom, J.M. Acetylcholine receptors and myasthenia. Muscle Nerve 23,
453-77 (2000).
43. Galzi, J.L. et al. Identification of a novel amino acid alpha-tyrosine 93 within
the cholinergic ligands-binding sites of the acetylcholine receptor by
photoaffinity labeling. Additional evidence for a three-loop model of the
cholinergic ligands-binding sites. J Biol Chem 265, 10430-7 (1990).
44. Dennis, M. et al. Amino acids of the Torpedo marmorata acetylcholine
receptor alpha subunit labeled by a photoaffinity ligand for the acetylcholine
binding site. Biochemistry 27, 2346-57 (1988).
45. Kao, P.N. et al. Identification of the alpha subunit half-cystine specifically
labeled by an affinity reagent for the acetylcholine receptor binding site. J
Biol Chem 259, 11662-5 (1984).

226

46. Sine, S.M., Quiram, P., Papanikolaou, F., Kreienkamp, H.J. & Taylor, P.
Conserved tyrosines in the alpha subunit of the nicotinic acetylcholine
receptor stabilize quaternary ammonium groups of agonists and curariform
antagonists. J Biol Chem 269, 8808-16 (1994).
47. Chiara, D.C. & Cohen, J.B. Identification of amino acids contributing to high
and low affinity d-tubocurarine sites in the Torpedo nicotinic acetylcholine
receptor. J Biol Chem 272, 32940-50 (1997).
48. Chiara, D.C., Xie, Y. & Cohen, J.B. Structure of the agonist-binding sites of
the Torpedo nicotinic acetylcholine receptor: affinity-labeling and mutational
analyses identify gamma Tyr-111/delta Arg-113 as antagonist affinity
determinants. Biochemistry 38, 6689-98 (1999).
49. Sine, S.M. Identification of equivalent residues in the gamma, delta, and
epsilon subunits of the nicotinic receptor that contribute to alpha-
bungarotoxin binding. J Biol Chem 272, 23521-7 (1997).
50. Czajkowski, C. & Karlin, A. Structure of the nicotinic receptor
acetylcholine-binding site. Identification of acidic residues in the delta
subunit within 0.9 nm of the 5 alpha subunit-binding. J Biol Chem 270,
3160-4 (1995).
51. Taly, A., Corringer, P.J., Guedin, D., Lestage, P. & Changeux, J.P. Nicotinic
receptors: allosteric transitions and therapeutic targets in the nervous system.
Nat Rev Drug Discov 8, 733-50 (2009).
52. Bourne, Y., Talley, T.T., Hansen, S.B., Taylor, P. & Marchot, P. Crystal
structure of a Cbtx-AChBP complex reveals essential interactions between
snake alpha-neurotoxins and nicotinic receptors. EMBO J 24, 1512-22
(2005).
53. Celie, P.H. et al. Crystal structure of nicotinic acetylcholine receptor
homolog AChBP in complex with an alpha-conotoxin PnIA variant. Nat
Struct Mol Biol 12, 582-8 (2005).
54. Celie, P.H. et al. Crystal structure of acetylcholine-binding protein from
Bulinus truncatus reveals the conserved structural scaffold and sites of
variation in nicotinic acetylcholine receptors. J Biol Chem 280, 26457-66
(2005).
55. Celie, P.H. et al. Nicotine and carbamylcholine binding to nicotinic
acetylcholine receptors as studied in AChBP crystal structures. Neuron 41,
907-14 (2004).
56. Dutertre, S. et al. AChBP-targeted alpha-conotoxin correlates distinct
binding orientations with nAChR subtype selectivity. EMBO J 26, 3858-67
(2007).
57. Hansen, S.B. et al. Structures of Aplysia AChBP complexes with nicotinic
agonists and antagonists reveal distinctive binding interfaces and
conformations. EMBO J 24, 3635-46 (2005).
58. Ulens, C. et al. Structural determinants of selective alpha-conotoxin binding
to a nicotinic acetylcholine receptor homolog AChBP. Proc Natl Acad Sci U
S A 103, 3615-20 (2006).
227

59. Hibbs, R.E. et al. Structural determinants for interaction of partial agonists
with acetylcholine binding protein and neuronal alpha7 nicotinic
acetylcholine receptor. EMBO J 28, 3040-51 (2009).
60. Smit, A.B. et al. A glia-derived acetylcholine-binding protein that modulates
synaptic transmission. Nature 411, 261-8 (2001).
61. Fruchart-Gaillard, C. et al. Experimentally based model of a complex
between a snake toxin and the alpha 7 nicotinic receptor. Proc Natl Acad Sci
U S A 99, 3216-21 (2002).
62. Molles, B.E. et al. Residues in the epsilon subunit of the nicotinic
acetylcholine receptor interact to confer selectivity of waglerin-1 for the
alpha-epsilon subunit interface site. Biochemistry 41, 7895-906 (2002).
63. Schapira, M., Abagyan, R. & Totrov, M. Structural model of nicotinic
acetylcholine receptor isotypes bound to acetylcholine and nicotine. BMC
Struct Biol 2, 1 (2002).
64. Sine, S.M., Wang, H.L. & Bren, N. Lysine scanning mutagenesis delineates
structural model of the nicotinic receptor ligand binding domain. J Biol
Chem 277, 29210-23 (2002).
65. Sixma, T.K. & Smit, A.B. Acetylcholine binding protein (AChBP): a
secreted glial protein that provides a high-resolution model for the
extracellular domain of pentameric ligand-gated ion channels. Annu Rev
Biophys Biomol Struct 32, 311-34 (2003).
66. Anand, R. et al. Reporter epitopes: a novel approach to examine
transmembrane topology of integral membrane proteins applied to the alpha
1 subunit of the nicotinic acetylcholine receptor. Biochemistry 32, 9975-84
(1993).
67. Smit, A.B. et al. Acetylcholine-binding proteins: functional and structural
homologs of nicotinic acetylcholine receptors. J Mol Neurosci 30, 9-10
(2006).
68. Bouzat, C. et al. Coupling of agonist binding to channel gating in an ACh-
binding protein linked to an ion channel. Nature 430, 896-900 (2004).
69. Miyazawa, A., Fujiyoshi, Y. & Unwin, N. Structure and gating mechanism
of the acetylcholine receptor pore. Nature 423, 949-55 (2003).
70. Unwin, N. Refined structure of the nicotinic acetylcholine receptor at 4A
resolution. J Mol Biol 346, 967-89 (2005).
71. Changeux, J.P. et al. Brain nicotinic receptors: structure and regulation, role
in learning and reinforcement. Brain Res Brain Res Rev 26, 198-216 (1998).
72. Unwin, N. Nicotinic acetylcholine receptor at 9 A resolution. J Mol Biol 229,
1101-24 (1993).
73. Miyazawa, A., Fujiyoshi, Y., Stowell, M. & Unwin, N. Nicotinic
acetylcholine receptor at 4.6 A resolution: transverse tunnels in the channel
wall. J Mol Biol 288, 765-86 (1999).
74. Chen, L. In pursuit of the high-resolution structure of nicotinic acetylcholine
receptors. J Physiol 588, 557-64 (2009).
75. Dellisanti, C.D., Yao, Y., Stroud, J.C., Wang, Z.Z. & Chen, L. Structural
determinants for alpha-neurotoxin sensitivity in muscle nAChR and their
implications for the gating mechanism. Channels (Austin) 1, 234-7 (2007).
228

76. Bocquet, N. et al. X-ray structure of a pentameric ligand-gated ion channel in
an apparently open conformation. Nature 457, 111-4 (2009).
77. Hilf, R.J. & Dutzler, R. X-ray structure of a prokaryotic pentameric ligand-
gated ion channel. Nature 452, 375-9 (2008).
78. Hilf, R.J. & Dutzler, R. Structure of a potentially open state of a proton-
activated pentameric ligand-gated ion channel. Nature 457, 115-8 (2009).
79. Althoff, T., Hibbs, R.E., Banerjee, S. & Gouaux, E. X-ray structures of
GluCl in apo states reveal a gating mechanism of Cys-loop receptors. Nature
512, 333-7 (2014).
80. Hassaine, G. et al. X-ray structure of the mouse serotonin 5-HT3 receptor.
Nature 512, 276-81 (2014).
81. Boulter, J. et al. Isolation of a cDNA clone coding for a possible neural
nicotinic acetylcholine receptor alpha-subunit. Nature 319, 368-74 (1986).
82. Jensen, A.A., Frolund, B., Liljefors, T. & Krogsgaard-Larsen, P. Neuronal
nicotinic acetylcholine receptors: structural revelations, target identifications,
and therapeutic inspirations. J Med Chem 48, 4705-45 (2005).
83. Gotti, C., Zoli, M. & Clementi, F. Brain nicotinic acetylcholine receptors:
native subtypes and their relevance. Trends Pharmacol Sci 27, 482-91
(2006).
84. Schoepfer, R., Conroy, W.G., Whiting, P., Gore, M. & Lindstrom, J. Brain
alpha-bungarotoxin binding protein cDNAs and MAbs reveal subtypes of
this branch of the ligand-gated ion channel gene superfamily. Neuron 5, 35-
48 (1990).
85. Li, S.X. et al. Ligand-binding domain of an alpha7-nicotinic receptor
chimera and its complex with agonist. Nat Neurosci 14, 1253-9 (2011).
86. Zouridakis, M. et al. Crystal structures of free and antagonist-bound states of
human alpha9 nicotinic receptor extracellular domain. Nat Struct Mol Biol
21, 976-80 (2014).
87. Kuryatov, A., Onksen, J. & Lindstrom, J. Roles of accessory subunits in
alpha4beta2(*) nicotinic receptors. Mol Pharmacol 74, 132-43 (2008).
88. Liao, H.I. et al. mRNA display design of fibronectin-based intrabodies that
detect and inhibit severe acute respiratory syndrome coronavirus
nucleocapsid protein. J Biol Chem 284, 17512-20 (2009).
89. Tzartos, S., Hochschwender, S., Vasquez, P. & Lindstrom, J. Passive transfer
of experimental autoimmune myasthenia gravis by monoclonal antibodies to
the main immunogenic region of the acetylcholine receptor. J Neuroimmunol
15, 185-94 (1987).
90. Wang, F. et al. Chronic nicotine treatment up-regulates human alpha3 beta2
but not alpha3 beta4 acetylcholine receptors stably transfected in human
embryonic kidney cells. J Biol Chem 273, 28721-32 (1998).
91. Roberts, R.W. & Szostak, J.W. RNA-peptide fusions for the in vitro selection
of peptides and proteins. Proc Natl Acad Sci U S A 94, 12297-302 (1997).
92. Meriggioli, M.N. & Sanders, D.B. Autoimmune myasthenia gravis: emerging
clinical and biological heterogeneity. Lancet Neurol 8, 475-90 (2009).
93. Glickstein, M. Book Review. New England Journal of Medicine 348, 180-
181 (2003).
229

94. Patrick, J. & Lindstrom, J. Autoimmune response to acetylcholine receptor.
Science 180, 871-2 (1973).
95. Drachman, D.B., Adams, R.N., Josifek, L.F. & Self, S.G. Functional
activities of autoantibodies to acetylcholine receptors and the clinical severity
of myasthenia gravis. N Engl J Med 307, 769-75 (1982).
96. Fulpius, B.W., Miskin, R. & Reich, E. Antibodies from myasthenic patients
that compete with cholinergic agents for binding to nicotinic receptors. Proc
Natl Acad Sci U S A 77, 4326-30 (1980).
97. Gomez, C.M. & Richman, D.P. Anti-acetylcholine receptor antibodies
directed against the alpha-bungarotoxin binding site induce a unique form of
experimental myasthenia. Proc Natl Acad Sci U S A 80, 4089-93 (1983).
98. Souroujon, M.C., Mochly-Rosen, D., Gordon, A.S. & Fuchs, S. Interaction
of monoclonal antibodies to Torpedo acetylcholine receptor with the receptor
of skeletal muscle. Muscle Nerve 6, 303-11 (1983).
99. Appel, S.H., Anwyl, R., McAdams, M.W. & Elias, S. Accelerated
degradation of acetylcholine receptor from cultured rat myotubes with
myasthenia gravis sera and globulins. Proc Natl Acad Sci U S A 74, 2130-4
(1977).
100. Heinemann, S., Bevan, S., Kullberg, R., Lindstrom, J. & Rice, J. Modulation
of acetylcholine receptor by antibody against the receptor. Proc Natl Acad
Sci U S A 74, 3090-4 (1977).
101. Kao, I. & Drachman, D.B. Myasthenic immunoglobulin accelerates
acetylcholine receptor degradation. Science 196, 527-9 (1977).
102. Reiness, C.G., Weinberg, C.B. & Hall, Z.W. Antibody to acetylcholine
receptor increases degradation of junctional and extrajunctional receptors in
adult muscle. Nature 274, 68-70 (1978).
103. Conti-Fine, B.M., Milani, M. & Kaminski, H.J. Myasthenia gravis: past,
present, and future. J Clin Invest 116, 2843-54 (2006).
104. Carr, A.S., Cardwell, C.R., McCarron, P.O. & McConville, J. A systematic
review of population based epidemiological studies in Myasthenia Gravis.
BMC Neurol 10, 46 (2010).
105. Grob, D., Brunner, N., Namba, T. & Pagala, M. Lifetime course of
myasthenia gravis. Muscle Nerve 37, 141-9 (2008).
106. Zhang, X. et al. Clinical and serological study of myasthenia gravis in HuBei
Province, China. J Neurol Neurosurg Psychiatry 78, 386-90 (2007).
107. Berrih-Aknin, S., Frenkian-Cuvelier, M. & Eymard, B. Diagnostic and
clinical classification of autoimmune myasthenia gravis. J Autoimmun, 1-6
(2014).
108. Jayam Trouth, A., Dabi, A., Solieman, N., Kurukumbi, M. & Kalyanam, J.
Myasthenia gravis: a review. Autoimmune Dis 2012, 874680 (2012).
109. Sanders, D.B. Myasthenia Gravis. Encyclopedia of Life Sciences (2001).
110. Kusner, L.L. et al. Guidelines for pre-clinical assessment of the acetylcholine
receptor-specific passive transfer myasthenia gravis model-
Recommendations for methods and experimental designs. Exp Neurol 270, 3-
10 (2015).
230

111. Losen, M. et al. Standardization of the experimental autoimmune myasthenia
gravis (EAMG) model by immunization of rats with Torpedo californica
acetylcholine receptors - Recommendations for methods and experimental
designs. Exp Neurol 270, 18-28 (2015).
112. Luo, J., Kuryatov, A. & Lindstrom, J.M. Specific immunotherapy of
experimental myasthenia gravis by a novel mechanism. Ann Neurol 67, 441-
51 (2010).
113. Luo, J. & Lindstrom, J. Antigen-specific immunotherapeutic vaccine for
experimental autoimmune myasthenia gravis. J Immunol 193, 5044-55
(2014).
114. Guptill, J.T., Sanders, D.B. & Evoli, A. Anti-MuSK antibody myasthenia
gravis: clinical findings and response to treatment in two large cohorts.
Muscle Nerve 44, 36-40 (2011).
115. Hoch, W. et al. Auto-antibodies to the receptor tyrosine kinase MuSK in
patients with myasthenia gravis without acetylcholine receptor antibodies.
Nat Med 7, 365-8 (2001).
116. Lavrnic, D. et al. The features of myasthenia gravis with autoantibodies to
MuSK. J Neurol Neurosurg Psychiatry 76, 1099-102 (2005).
117. Higuchi, O., Hamuro, J., Motomura, M. & Yamanashi, Y. Autoantibodies to
low-density lipoprotein receptor-related protein 4 in myasthenia gravis. Ann
Neurol 69, 418-22 (2011).
118. Pevzner, A. et al. Anti-LRP4 autoantibodies in AChR- and MuSK-antibody-
negative myasthenia gravis. J Neurol 259, 427-35 (2011).
119. Tzartos, S.J. & Lindstrom, J.M. Monoclonal antibodies used to probe
acetylcholine receptor structure: localization of the main immunogenic
region and detection of similarities between subunits. Proc Natl Acad Sci U S
A 77, 755-9 (1980).
120. Tzartos, S.J., Rand, D.E., Einarson, B.L. & Lindstrom, J.M. Mapping of
surface structures of electrophorus acetylcholine receptor using monoclonal
antibodies. J Biol Chem 256, 8635-45 (1981).
121. Sargent, P.B. et al. Structure and transmembrane nature of the acetylcholine
receptor in amphibian skeletal muscle as revealed by cross-reacting
monoclonal antibodies. J Cell Biol 98, 609-18 (1984).
122. Dau, P.C. et al. Plasmapheresis and immunosuppressive drug therapy in
myasthenia gravis. N Engl J Med 297, 1134-40 (1977).
123. Newsom-Davis, J., Pinching, A.J., Vincent, A. & Wilson, S.G. Function of
circulating antibody to acetylcholine receptor in myasthenia gravis:
investigation by plasma exchange. Neurology 28, 266-72 (1978).
124. Haven, T.R. et al. An algorithm for acetylcholine receptor antibody testing in
patients with suspected myasthenia gravis. Clin Chem 56, 1028-9 (2010).
125. Masuda, T. et al. Antibodies against the main immunogenic region of the
acetylcholine receptor correlate with disease severity in myasthenia gravis. J
Neurol Neurosurg Psychiatry 83, 935-40 (2012).
126. Heldal, A.T., Eide, G.E., Romi, F., Owe, J.F. & Gilhus, N.E. Repeated
acetylcholine receptor antibody-concentrations and association to clinical
myasthenia gravis development. PLoS One 9, e114060 (2014).
231

127. Ratnam, M. et al. Location of antigenic determinants on primary sequences
of subunits of nicotinic acetylcholine receptor by peptide mapping.
Biochemistry 25, 2621-32 (1986).
128. Barkas, T. et al. Mapping the main immunogenic region and toxin-binding
site of the nicotinic acetylcholine receptor. Science 235, 77-80 (1987).
129. Barkas, T. et al. Monoclonal antibodies to the main immunogenic region of
the nicotinic acetylcholine receptor bind to residues 61-76 of the alpha
subunit. J Biol Chem 263, 5916-20 (1988).
130. Tzartos, S.J., Kokla, A., Walgrave, S.L. & Conti-Tronconi, B.M.
Localization of the main immunogenic region of human muscle acetylcholine
receptor to residues 67-76 of the alpha subunit. Proc Natl Acad Sci U S A 85,
2899-903 (1988).
131. Tzartos, S.J. et al. Main immunogenic region of Torpedo electroplax and
human muscle acetylcholine receptor: localization and microheterogeneity
revealed by the use of synthetic peptides. J Neurochem 54, 51-61 (1990).
132. Luo, J. et al. Main immunogenic region structure promotes binding of
conformation-dependent myasthenia gravis autoantibodies, nicotinic
acetylcholine receptor conformation maturation, and agonist sensitivity. J
Neurosci 29, 13898-908 (2009).
133. Morell, S.W. et al. Structural characterization of the main immunogenic
region of the Torpedo acetylcholine receptor. Mol Immunol 58, 116-31
(2013).
134. Bellone, M., Tang, F., Milius, R. & Conti-Tronconi, B.M. The main
immunogenic region of the nicotinic acetylcholine receptor. Identification of
amino acid residues interacting with different antibodies. J Immunol 143,
3568-79 (1989).
135. Das, M.K. & Lindstrom, J. The main immunogenic region of the nicotinic
acetylcholine receptor: interaction of monoclonal antibodies with synthetic
peptides. Biochem Biophys Res Commun 165, 865-71 (1989).
136. Saedi, M.S., Anand, R., Conroy, W.G. & Lindstrom, J. Determination of
amino acids critical to the main immunogenic region of intact acetylcholine
receptors by in vitro mutagenesis. FEBS Lett 267, 55-9 (1990).
137. Papadouli, I., Sakarellos, C. & Tzartos, S.J. High-resolution epitope mapping
and fine antigenic characterization of the main immunogenic region of the
acetylcholine receptor. Improving the binding activity of synthetic analogues
of the region. Eur J Biochem 211, 227-34 (1993).
138. Kubalek, E., Ralston, S., Lindstrom, J. & Unwin, N. Location of subunits
within the acetylcholine receptor by electron image analysis of tubular
crystals from Torpedo marmorata. J Cell Biol 105, 9-18 (1987).
139. Beroukhim, R. & Unwin, N. Three-dimensional location of the main
immunogenic region of the acetylcholine receptor. Neuron 15, 323-31
(1995).
140. Tzartos, S.J. et al. Anatomy of the antigenic structure of a large membrane
autoantigen, the muscle-type nicotinic acetylcholine receptor. Immunol Rev
163, 89-120 (1998).
232

141. Tzartos, S.J., Seybold, M.E. & Lindstrom, J.M. Specificities of antibodies to
acetylcholine receptors in sera from myasthenia gravis patients measured by
monoclonal antibodies. Proc Natl Acad Sci U S A 79, 188-92 (1982).
142. Wang, Z.Z., Hardy, S.F. & Hall, Z.W. Assembly of the nicotinic
acetylcholine receptor. The first transmembrane domains of truncated alpha
and delta subunits are required for heterodimer formation in vivo. J Biol
Chem 271, 27575-84 (1996).
143. Wang, Z.Z., Hardy, S.F. & Hall, Z.W. Membrane tethering enables an
extracellular domain of the acetylcholine receptor alpha subunit to form a
heterodimeric ligand-binding site. J Cell Biol 135, 809-17 (1996).
144. Yao, Y. et al. Yeast expression and NMR analysis of the extracellular
domain of muscle nicotinic acetylcholine receptor alpha subunit. J Biol Chem
277, 12613-21 (2002).
145. Conti-Tronconi, B., Tzartos, S. & Lindstrom, J. Monoclonal antibodies as
probes of acetylcholine receptor structure. 2. Binding to native receptor.
Biochemistry 20, 2181-91 (1981).
146. Gullick, W.J., Tzartos, S. & Lindstrom, J. Monoclonal antibodies as probes
of acetylcholine receptor structure. 1. Peptide mapping. Biochemistry 20,
2173-80 (1981).
147. Tzartos, S., Langeberg, L., Hochschwender, S. & Lindstrom, J.
Demonstration of a main immunogenic region on acetylcholine receptors
from human muscle using monoclonal antibodies to human receptor. FEBS
Lett 158, 116-8 (1983).
148. Kordossi, A.A. & Tzartos, S.J. Monoclonal antibodies against the main
immunogenic region of the acetylcholine receptor. Mapping on the intact
molecule. J Neuroimmunol 23, 35-40 (1989).
149. Tamamizu, S., Butler, D.H., Lasalde, J.A. & McNamee, M.G. Effects of
antibody binding on structural transitions of the nicotinic acetylcholine
receptor. Biochemistry 35, 11773-81 (1996).
150. Gomez, C.M. et al. Monoclonal hybridoma anti-acetylcholine receptor
antibodies: antibody specificity and effect of passive transfer. Ann N Y Acad
Sci 377, 97-109 (1981).
151. Kontou, M. et al. The crystal structure of the Fab fragment of a rat
monoclonal antibody against the main immunogenic region of the human
muscle acetylcholine receptor. Eur J Biochem 267, 2389-97 (2000).
152. Poulas, K. et al. Crystal structure of Fab198, an efficient protector of the
acetylcholine receptor against myasthenogenic antibodies. Eur J Biochem
268, 3685-93 (2001).
153. Lindstrom, J. Immunobiology of myasthenia gravis, experimental
autoimmune myasthenia gravis, and Lambert-Eaton syndrome. Annu Rev
Immunol 3, 109-31 (1985).
154. Lindstrom, J.M., Seybold, M.E., Lennon, V.A., Whittingham, S. & Duane,
D.D. Antibody to acetylcholine receptor in myasthenia gravis. Prevalence,
clinical correlates, and diagnostic value. Neurology 26, 1054-9 (1976).
233

155. Patrick, J. & Stallcup, W.B. Immunological distinction between
acetylcholine receptor and the alpha-bungarotoxin-binding component on
sympathetic neurons. Proc Natl Acad Sci U S A 74, 4689-92 (1977).
156. Tzartos, S.J., Sophianos, D. & Efthimiadis, A. Role of the main
immunogenic region of acetylcholine receptor in myasthenia gravis. An Fab
monoclonal antibody protects against antigenic modulation by human sera. J
Immunol 134, 2343-9 (1985).
157. Richman, D.P. et al. Monoclonal anti-acetylcholine receptor antibodies can
cause experimental myasthenia. Nature 286, 738-9 (1980).
158. Rousseaux, J., Rousseaux-Prevost, R. & Bazin, H. Optimal conditions for the
preparation of Fab and F(ab')2 fragments from monoclonal IgG of different
rat IgG subclasses. J Immunol Methods 64, 141-6 (1983).
159. Asai, D.J. Methods in Cell Biology (ed. Asai, D.J.) (Academic Press, Inc.,
1993).
160. Otwinowski, Z. & Minor, W. Processing of X-ray diffraction data collected
in oscillation mode. Methods Enzymol 276, 307-326 (1997).
161. Winn, M.D. et al. Overview of the CCP4 suite and current developments.
Acta Crystallogr D Biol Crystallogr 67, 235-42 (2011).
162. McCoy, A.J. Solving structures of protein complexes by molecular
replacement with Phaser. Acta Crystallogr D Biol Crystallogr 63, 32-41
(2007).
163. McCoy, A.J. et al. Phaser crystallographic software. J Appl Crystallogr 40,
658-674 (2007).
164. Murshudov, G.N. et al. REFMAC5 for the refinement of macromolecular
crystal structures. Acta Crystallogr D Biol Crystallogr 67, 355-67 (2011).
165. Murshudov, G.N., Vagin, A.A. & Dodson, E.J. Refinement of
macromolecular structures by the maximum-likelihood method. Acta
Crystallogr D Biol Crystallogr 53, 240-55 (1997).
166. Jones, T.A., Zou, J.Y., Cowan, S.W. & Kjeldgaard, M. Improved methods
for building protein models in electron density maps and the location of
errors in these models. Acta Crystallogr A 47 ( Pt 2), 110-9 (1991).
167. Emsley, P., Lohkamp, B., Scott, W.G. & Cowtan, K. Features and
development of Coot. Acta Crystallogr D Biol Crystallogr 66, 486-501.
168. Kleinjung, J. et al. The third-dimensional structure of the complex between
an Fv antibody fragment and an analogue of the main immunogenic region of
the acetylcholine receptor: a combined two-dimensional NMR, homology,
and molecular modeling approach. Biopolymers 53, 113-28 (2000).
169. Fleury, D., Wharton, S.A., Skehel, J.J., Knossow, M. & Bizebard, T. Antigen
distortion allows influenza virus to escape neutralization. Nat Struct Biol 5,
119-23 (1998).
170. Muller, Y.A. et al. VEGF and the Fab fragment of a humanized neutralizing
antibody: crystal structure of the complex at 2.4 A resolution and mutational
analysis of the interface. Structure 6, 1153-67 (1998).

234

171. Fostieri, E., Tzartos, S.J., Berrih-Aknin, S., Beeson, D. & Mamalaki, A.
Isolation of potent human Fab fragments against a novel highly
immunogenic region on human muscle acetylcholine receptor which protect
the receptor from myasthenic autoantibodies. Eur J Immunol 35, 632-43
(2005).
172. Aurangzeb, S., Tariq, M., Irshad, M., Badshah, M. & Khan, R.S.
Relationship between anti-acetylcholine receptor antibody titres and severity
of myasthenia gravis. J Pak Med Assoc 59, 289-92 (2009).
173. Drachman, D.B. Myasthenia gravis. N Engl J Med 330, 1797-810 (1994).
174. Thanvi, B.R. & Lo, T.C. Update on myasthenia gravis. Postgrad Med J 80,
690-700 (2004).
175. Seybold, M.E. & Lindstrom, J.M. Patterns of acetylcholine receptor antibody
fluctuation in myasthenia gravis. Ann N Y Acad Sci 377, 292-306 (1981).
176. Dwyer, D.S., Bradley, R.J., Urquhart, C.K. & Kearney, J.F. An enzyme-
linked immunoabsorbent assay for measuring antibodies against muscle
acetylcholine receptor. J Immunol Methods 57, 111-9 (1983).
177. Hinman, C.L., Kellogg, C.M., Ernstoff, R.M., Rauch, H.C. & Hudson, R.A.
Immunoenzymometric assay and radioimmunoassay measure different
populations of antibody against human acetylcholine receptor. Clin Chem 31,
835-40 (1985).
178. Kawanami, S., Tsuji, R. & Oda, K. Enzyme-linked immunosorbent assay for
antibody against the nicotinic acetylcholine receptor in human myasthenia
gravis. Ann Neurol 15, 195-200 (1984).
179. Bindu, P.S., Nirmala, M., Patil, S.A. & Taly, A.B. Myasthenia gravis and
acetylcholine receptor antibodies: a clinico immunological correlative study
on South Indian patients. Ann Indian Acad Neurol 11, 242-4 (2008).
180. Jailkhani, B.L., Asthana, D., Jaffery, N.F., Kumar, R. & Ahuja, G.K. A
simplified ELISA for anti-receptor antibodies in myasthenia gravis. J
Immunol Methods 86, 115-8 (1986).
181. Franciotta, D. et al. TE671 cell-based ELISA for anti-acetylcholine receptor
antibody determination in myasthenia gravis. Clin Chem 45, 400-5 (1999).
182. Martino, G., Twaddle, G., Brambilla, E. & Grimaldi, L.M. Detection of anti-
acetylcholine receptor antibody by an ELISA using human receptor from a
rhabdomyosarcoma cell line. Acta Neurol Scand 89, 18-22 (1994).
183. Syapin, P.J., Salvaterra, P.M. & Engelhardt, J.K. Neuronal-like features of
TE671 cells: presence of a functional nicotinic cholinergic receptor. Brain
Res 231, 365-77 (1982).
184. Kennel, P.F. et al. Myasthenia gravis: comparative autoantibody assays using
human muscle, TE671, and glucocorticoid-treated TE671 cells as sources of
antigen. Clin Immunol Immunopathol 74, 293-6 (1995).
185. Somnier, F.E. Anti-acetylcholine receptor (AChR) antibodies measurement
in myasthenia gravis: the use of cell line TE671 as a source of AChR antigen.
J Neuroimmunol 51, 63-8 (1994).
186. Voltz, R. et al. Myasthenia gravis: measurement of anti-AChR
autoantibodies using cell line TE671. Neurology 41, 1836-8 (1991).
235

187. Walker, R., Vincent, A. & Newsom-Davis, J. Immunological and
pharmacological heterogeneity of alpha-bungarotoxin binding sites extracted
from TE671 cells. J Neuroimmunol 19, 149-57 (1988).
188. Dey, S. et al. Recombinant LipL32 antigen-based single serum dilution
ELISA for detection of canine leptospirosis. Vet Microbiol 103, 99-106
(2004).
189. Harlow, L. et al. Identification of citrullinated hsp90 isoforms as novel
autoantigens in rheumatoid arthritis-associated interstitial lung disease.
Arthritis Rheum 65, 869-79 (2013).
190. Kanigicherla, D. et al. Anti-PLA2R antibodies measured by ELISA predict
long-term outcome in a prevalent population of patients with idiopathic
membranous nephropathy. Kidney Int 83, 940-8 (2013).
191. Loizou, S. et al. Measurement of anti-cardiolipin antibodies by an enzyme-
linked immunosorbent assay (ELISA): standardization and quantitation of
results. Clin Exp Immunol 62, 738-45 (1985).
192. Nassau, E., Parsons, E.R. & Johnson, G.D. The detection of antibodies to
Mycobacterium tuberculosis by microplate enzyme-linked immunosorbent
assay (ELISA). Tubercle 57, 67-70 (1976).
193. Binz, H.K. & Pluckthun, A. Engineered proteins as specific binding reagents.
Curr Opin Biotechnol 16, 459-69 (2005).
194. Gebauer, M. & Skerra, A. Engineered protein scaffolds as next-generation
antibody therapeutics. Curr Opin Chem Biol 13, 245-55 (2009).
195. Huston, J.S. et al. Protein engineering of single-chain Fv analogs and fusion
proteins. Methods Enzymol 203, 46-88 (1991).
196. Holt, L.J., Herring, C., Jespers, L.S., Woolven, B.P. & Tomlinson, I.M.
Domain antibodies: proteins for therapy. Trends Biotechnol 21, 484-90
(2003).
197. Pereira, B., Benedict, C.R., Le, A., Shapiro, S.S. & Thiagarajan, P.
Cardiolipin binding a light chain from lupus-prone mice. Biochemistry 37,
1430-7 (1998).
198. Ward, E.S., Gussow, D., Griffiths, A.D., Jones, P.T. & Winter, G. Binding
activities of a repertoire of single immunoglobulin variable domains secreted
from Escherichia coli. Nature 341, 544-6 (1989).
199. Muyldermans, S. Nanobodies: natural single-domain antibodies. Annu Rev
Biochem 82, 775-97 (2013).
200. Vincke, C. et al. General strategy to humanize a camelid single-domain
antibody and identification of a universal humanized nanobody scaffold. J
Biol Chem 284, 3273-84 (2009).
201. Liu, R., Barrick, J.E., Szostak, J.W. & Roberts, R.W. Optimized synthesis of
RNA-protein fusions for in vitro protein selection. Methods Enzymol 318,
268-93 (2000).
202. Takahashi, T.T., Austin, R.J. & Roberts, R.W. mRNA display: ligand
discovery, interaction analysis and beyond. Trends Biochem Sci 28, 159-65
(2003).

236

203. Nemoto, N., Miyamoto-Sato, E., Husimi, Y. & Yanagawa, H. In vitro virus:
bonding of mRNA bearing puromycin at the 3'-terminal end to the C-
terminal end of its encoded protein on the ribosome in vitro. FEBS Lett 414,
405-8 (1997).
204. Koide, A., Bailey, C.W., Huang, X. & Koide, S. The fibronectin type III
domain as a scaffold for novel binding proteins. J Mol Biol 284, 1141-51
(1998).
205. Olson, C.A. et al. Rapid mRNA-display selection of an IL-6 inhibitor using
continuous-flow magnetic separation. Angew Chem Int Ed Engl 50, 8295-8
(2011).
206. Olson, C.A., Liao, H.I., Sun, R. & Roberts, R.W. mRNA display selection of
a high-affinity, modification-specific phospho-IkappaBalpha-binding
fibronectin. ACS Chem Biol 3, 480-5 (2008).
207. Waldo, G.S., Standish, B.M., Berendzen, J. & Terwilliger, T.C. Rapid
protein-folding assay using green fluorescent protein. Nat Biotechnol 17,
691-5 (1999).
208. Dickinson, C.D. et al. Crystal structure of the tenth type III cell adhesion
module of human fibronectin. J Mol Biol 236, 1079-92 (1994).
209. Leahy, D.J., Hendrickson, W.A., Aukhil, I. & Erickson, H.P. Structure of a
fibronectin type III domain from tenascin phased by MAD analysis of the
selenomethionyl protein. Science 258, 987-91 (1992).
210. Main, A.L., Harvey, T.S., Baron, M., Boyd, J. & Campbell, I.D. The three-
dimensional structure of the tenth type III module of fibronectin: an insight
into RGD-mediated interactions. Cell 71, 671-8 (1992).
211. Gross, G.G. et al. Recombinant probes for visualizing endogenous synaptic
proteins in living neurons. Neuron 78, 971-85 (2013).
212. Olson, C.A. et al. Single-round, multiplexed antibody mimetic design
through mRNA display. Angew Chem Int Ed Engl 51, 12449-53 (2012).
213. Olson, C.A. & Roberts, R.W. Design, expression, and stability of a diverse
protein library based on the human fibronectin type III domain. Protein Sci
16, 476-84 (2007).
214. Lipovsek, D. Adnectins: engineered target-binding protein therapeutics.
Protein Eng Des Sel 24, 3-9 (2010).
215. Bloom, L. & Calabro, V. FN3: a new protein scaffold reaches the clinic.
Drug Discov Today 14, 949-55 (2009).
216. Koide, A., Abbatiello, S., Rothgery, L. & Koide, S. Probing protein
conformational changes in living cells by using designer binding proteins:
application to the estrogen receptor. Proc Natl Acad Sci U S A 99, 1253-8
(2002).
217. Hackel, B.J., Ackerman, M.E., Howland, S.W. & Wittrup, K.D. Stability and
CDR composition biases enrich binder functionality landscapes. J Mol Biol
401, 84-96 (2010).
218. Hackel, B.J., Kapila, A. & Wittrup, K.D. Picomolar affinity fibronectin
domains engineered utilizing loop length diversity, recursive mutagenesis,
and loop shuffling. J Mol Biol 381, 1238-52 (2008).
237

219. Hackel, B.J. & Wittrup, K.D. The full amino acid repertoire is superior to
serine/tyrosine for selection of high affinity immunoglobulin G binders from
the fibronectin scaffold. Protein Eng Des Sel 23, 211-9 (2010).
220. Parker, M.H. et al. Antibody mimics based on human fibronectin type three
domain engineered for thermostability and high-affinity binding to vascular
endothelial growth factor receptor two. Protein Eng Des Sel 18, 435-44
(2005).
221. Raymond-Delpech, V., Matsuda, K., Sattelle, B.M., Rauh, J.J. & Sattelle,
D.B. Ion channels: molecular targets of neuroactive insecticides. Invert
Neurosci 5, 119-33 (2005).
222. Kuryatov, A., Gerzanich, V., Nelson, M., Olale, F. & Lindstrom, J. Mutation
causing autosomal dominant nocturnal frontal lobe epilepsy alters Ca2+
permeability, conductance, and gating of human alpha4beta2 nicotinic
acetylcholine receptors. J Neurosci 17, 9035-47 (1997).
223. Steinlein, O.K. et al. A missense mutation in the neuronal nicotinic
acetylcholine receptor alpha 4 subunit is associated with autosomal dominant
nocturnal frontal lobe epilepsy. Nat Genet 11, 201-3 (1995).
224. Elgoyhen, A.B., Johnson, D.S., Boulter, J., Vetter, D.E. & Heinemann, S.
Alpha 9: an acetylcholine receptor with novel pharmacological properties
expressed in rat cochlear hair cells. Cell 79, 705-15 (1994).
225. Rothlin, C.V., Katz, E., Verbitsky, M. & Elgoyhen, A.B. The alpha9
nicotinic acetylcholine receptor shares pharmacological properties with type
A gamma-aminobutyric acid, glycine, and type 3 serotonin receptors. Mol
Pharmacol 55, 248-54 (1999).
226. Elgoyhen, A.B. et al. alpha10: a determinant of nicotinic cholinergic receptor
function in mammalian vestibular and cochlear mechanosensory hair cells.
Proc Natl Acad Sci U S A 98, 3501-6 (2001).
227. Indurthi, D.C. et al. Presence of multiple binding sites on alpha9alpha10
nAChR receptors alludes to stoichiometric-dependent action of the alpha-
conotoxin, Vc1.1. Biochem Pharmacol 89, 131-40 (2014).
228. Lustig, L.R., Peng, H., Hiel, H., Yamamoto, T. & Fuchs, P.A. Molecular
cloning and mapping of the human nicotinic acetylcholine receptor alpha10
(CHRNA10). Genomics 73, 272-83 (2001).
229. Plazas, P.V., Katz, E., Gomez-Casati, M.E., Bouzat, C. & Elgoyhen, A.B.
Stoichiometry of the alpha9alpha10 nicotinic cholinergic receptor. J
Neurosci 25, 10905-12 (2005).
230. Sgard, F. et al. A novel human nicotinic receptor subunit, alpha10, that
confers functionality to the alpha9-subunit. Mol Pharmacol 61, 150-9 (2002).
231. Glowatzki, E. & Fuchs, P.A. Cholinergic synaptic inhibition of inner hair
cells in the neonatal mammalian cochlea. Science 288, 2366-8 (2000).
232. Nguyen, V.T., Ndoye, A. & Grando, S.A. Novel human alpha9 acetylcholine
receptor regulating keratinocyte adhesion is targeted by Pemphigus vulgaris
autoimmunity. Am J Pathol 157, 1377-91 (2000).
233. Luebke, A.E., Maroni, P.D., Guth, S.M. & Lysakowski, A. Alpha-9 nicotinic
acetylcholine receptor immunoreactivity in the rodent vestibular labyrinth. J
Comp Neurol 492, 323-33 (2005).
238

234. Elgoyhen, A.B., Katz, E. & Fuchs, P.A. The nicotinic receptor of cochlear
hair cells: a possible pharmacotherapeutic target? Biochem Pharmacol 78,
712-9 (2009).
235. McIntosh, J.M., Absalom, N., Chebib, M., Elgoyhen, A.B. & Vincler, M.
Alpha9 nicotinic acetylcholine receptors and the treatment of pain. Biochem
Pharmacol 78, 693-702 (2009).
236. Ellison, M. et al. Alpha-RgIA, a novel conotoxin that blocks the
alpha9alpha10 nAChR: structure and identification of key receptor-binding
residues. J Mol Biol 377, 1216-27 (2008).
237. Ellison, M. et al. Alpha-conotoxins ImI and ImII target distinct regions of the
human alpha7 nicotinic acetylcholine receptor and distinguish human
nicotinic receptor subtypes. Biochemistry 43, 16019-26 (2004).
238. Kwon, H.M. et al. Enhanced coronary vasa vasorum neovascularization in
experimental hypercholesterolemia. J Clin Invest 101, 1551-6 (1998).
239. Livett, B.G. et al. Therapeutic applications of conotoxins that target the
neuronal nicotinic acetylcholine receptor. Toxicon 48, 810-29 (2006).
240. McIntosh, J.M. et al. A novel alpha-conotoxin, PeIA, cloned from Conus
pergrandis, discriminates between rat alpha9alpha10 and alpha7 nicotinic
cholinergic receptors. J Biol Chem 280, 30107-12 (2005).
241. Peng, C. et al. Discovery of a novel class of conotoxin from Conus litteratus,
lt14a, with a unique cysteine pattern. Peptides 27, 2174-81 (2006).
242. Sun, D. et al. Structure-function relationship of conotoxin lt14a, a potential
analgesic with low cytotoxicity. Peptides 32, 300-5 (2010).
243. Verbitsky, M., Rothlin, C.V., Katz, E. & Elgoyhen, A.B. Mixed nicotinic-
muscarinic properties of the alpha9 nicotinic cholinergic receptor.
Neuropharmacology 39, 2515-24 (2000).
244. Gehle, V.M., Walcott, E.C., Nishizaki, T. & Sumikawa, K. N-glycosylation
at the conserved sites ensures the expression of properly folded functional
ACh receptors. Brain Res Mol Brain Res 45, 219-29 (1997).
245. Lodish H, B.A., Zipursky SL, et al. in Molecular Cell Biology (New York:
W. H. Freeman, 2000).
246. Deardorff, J.D., Feen, E. & Grossberg, T.G. The use of cholonesterase
inhibitors acorss all stages of alzheimer's disease. Drugs aging (2015).
247. Dani, J.A. & Bertrand, D. Nicotinic acetylcholine receptors and nicotinic
cholinergic mechanisms of the central nervous system. Annu Rev Pharmacol
Toxicol 47, 699-729 (2007).
248. Hurst, R., Rollema, H. & Bertrand, D. Nicotinic acetylcholine receptors:
from basic science to therapeutics. Pharmacol Ther 137, 22-54 (2012).
249. Yakel, J.L. Cholinergic receptors: functional role of nicotinic ACh receptors
in brain circuits and disease. Pflugers Arch 465, 441-50 (2013).
250. Colombo, S.F., Mazzo, F., Pistillo, F. & Gotti, C. Biogenesis, trafficking and
up-regulation of nicotinic ACh receptors. Biochem Pharmacol 86, 1063-73
(2013).
251. Jones, S. & Yakel, J.L. Functional nicotinic ACh receptors on interneurones
in the rat hippocampus. J Physiol 504 ( Pt 3), 603-10 (1997).
239

252. Quick, M.W., Ceballos, R.M., Kasten, M., McIntosh, J.M. & Lester, R.A.
Alpha3beta4 subunit-containing nicotinic receptors dominate function in rat
medial habenula neurons. Neuropharmacology 38, 769-83 (1999).
253. Leiser, S.C., Bowlby, M.R., Comery, T.A. & Dunlop, J. A cog in cognition:
how the alpha 7 nicotinic acetylcholine receptor is geared towards improving
cognitive deficits. Pharmacol Ther 122, 302-11 (2009).
254. Levin, E.D. alpha7-Nicotinic receptors and cognition. Curr Drug Targets 13,
602-6 (2012).
255. Sharma, G. & Vijayaraghavan, S. Nicotinic cholinergic signaling in
hippocampal astrocytes involves calcium-induced calcium release from
intracellular stores. Proc Natl Acad Sci U S A 98, 4148-53 (2001).
256. Shen, J.X. & Yakel, J.L. Functional alpha7 nicotinic ACh receptors on
astrocytes in rat hippocampal CA1 slices. J Mol Neurosci 48, 14-21 (2012).
257. Shytle, R.D. et al. Cholinergic modulation of microglial activation by alpha 7
nicotinic receptors. J Neurochem 89, 337-43 (2004).
258. Wang, H. et al. Nicotinic acetylcholine receptor alpha7 subunit is an essential
regulator of inflammation. Nature 421, 384-8 (2003).
259. Wessler, I. & Kirkpatrick, C.J. Acetylcholine beyond neurons: the non-
neuronal cholinergic system in humans. Br J Pharmacol 154, 1558-71
(2008).
260. Si, M.L. & Lee, T.J. Alpha7-nicotinic acetylcholine receptors on cerebral
perivascular sympathetic nerves mediate choline-induced nitrergic
neurogenic vasodilation. Circ Res 91, 62-9 (2002).
261. Dineley, K.T., Pandya, A.A. & Yakel, J.L. Nicotinic ACh receptors as
therapeutic targets in CNS disorders. Trends Pharmacol Sci 36, 96-108
(2015).
262. Buccafusco, J.J. & Terry, A.V., Jr. A reversible model of the cognitive
impairment associated with schizophrenia in monkeys: potential therapeutic
effects of two nicotinic acetylcholine receptor agonists. Biochem Pharmacol
78, 852-62 (2009).
263. Levin, E.D., Bettegowda, C., Blosser, J. & Gordon, J. AR-R17779, and
alpha7 nicotinic agonist, improves learning and memory in rats. Behav
Pharmacol 10, 675-80 (1999).
264. Pichat, P. et al. SSR180711, a novel selective alpha7 nicotinic receptor
partial agonist: (II) efficacy in experimental models predictive of activity
against cognitive symptoms of schizophrenia. Neuropsychopharmacology
32, 17-34 (2007).
265. Van Kampen, M. et al. AR-R 17779 improves social recognition in rats by
activation of nicotinic alpha7 receptors. Psychopharmacology (Berl) 172,
375-83 (2004).
266. Kitagawa, H. et al. Safety, pharmacokinetics, and effects on cognitive
function of multiple doses of GTS-21 in healthy, male volunteers.
Neuropsychopharmacology 28, 542-51 (2003).

240

267. Deardorff, W.J., Shobassy, A. & Grossberg, G.T. Safety and clinical effects
of EVP-6124 in subjects with Alzheimer's disease currently or previously
receiving an acetylcholinesterase inhibitor medication. Expert Rev Neurother
15, 7-17 (2015).
268. Hashimoto, K. Targeting of alpha7 nicotinic acetylcholine receptors in the
treatment of schizophrenia and the use of auditory sensory gating as a
translational biomarker. Curr Pharm Des (2015).
269. Prickaerts, J. et al. EVP-6124, a novel and selective alpha7 nicotinic
acetylcholine receptor partial agonist, improves memory performance by
potentiating the acetylcholine response of alpha7 nicotinic acetylcholine
receptors. Neuropharmacology 62, 1099-110 (2011).
270. van Goethem, N.P., Prickaerts, J., Welty, D., Flood, D.G. & Koenig, G.
Continuous infusion of the alpha7 nicotinic acetylcholine receptor agonist
EVP-6124 produces no signs of tolerance at memory-enhancing doses in
rats: a pharmacokinetic and behavioral study. Behav Pharmacol 26, 403-6
(2015).
271. Prickaerts, J. et al. EVP-6124, a novel and selective alpha7 nicotinic
acetylcholine receptor partial agonist, improves memory performance by
potentiating the acetylcholine response of alpha7 nicotinic acetylcholine
receptors. Neuropharmacology 62, 1099-110 (2012).
272. Li, Y., Karlin, A., Loike, J.D. & Silverstein, S.C. A critical concentration of
neutrophils is required for effective bacterial killing in suspension. Proc Natl
Acad Sci U S A 99, 8289-94 (2002).
273. Howell, S.M. et al. Serum stable natural peptides designed by mRNA
display. Sci Rep 4, 6008 (2014).
274. Millward, S.W., Fiacco, S., Austin, R.J. & Roberts, R.W. Design of cyclic
peptides that bind protein surfaces with antibody-like affinity. ACS Chem
Biol 2, 625-34 (2007).
275. Millward, S.W., Takahashi, T.T. & Roberts, R.W. A general route for post-
translational cyclization of mRNA display libraries. J Am Chem Soc 127,
14142-3 (2005).

241

Appendix A

Figures for Structural Studies

242

Figure A.1: Superimposed view of the human Fab35 ternary structure and the
mouse Fab35 ternary structure.
The structure of human α211/Fab35/α-bungarotoxin (cyan: human α211, green: α-
bungarotoxin, magenta: Fab35 light chain, yellow: Fab35 heavy chain) is
superimposed with the structure of mouse α211/Fab35/α-bungarotoxin (brown) using
the C-alpha backbone of mouse α1. Fab35 portion is shifted to left, and the gap is
larger at the constant regions of Fab (C
H
and C
L
). However, variable regions (V
H
and
V
L
) shifted slightly, and interaction patterns at the binding interface are conserved.
243

Figure A.2: Ramachandran plot of the human Fab35 ternary complex
structure.
The final structure is determined using Procheck in CCP4 program. Most of side
chains are fixed based on the result. Although there are some residues outside of the
allowed range, they should not affect accuracy of our structure.

244

Figure A.3: Ramachandran plot of the human Fab210 ternary complex
structure.
The final structure is determined using Procheck in CCP4 program. Several residues
are outside of the allowed range, but they would not cause a major issue.

245

Figure A.4: Ramachandran plot of Endo Hf treated α9/α-bungarotoxin complex
structure.
The final structure is determined using Procheck in CCP4 program. Although there
are some residues outside of the allowed range, the accuracy of the structure should
not be affected.

246

Appendix B

Sequence Alignments

247

Comparison of Interface Interaction (Fab35 vs. Fab210)

1 10 20 30
Lym_AChBP ------------------------------LDRADILYNIRQTSRPDVIP-TQRDRPVAV
Aply_AChBP -------------------------------QANLMRLKSDLFNRSPMYPGPTKDDPLTV
Torpedo -----------------------------SEHETRLVANLLENYNKVIRPVEHHTHFVDI
a1_Human -----------------------------SEHETRLVAKLFKDYSSVVRPVEDHRQVVEV
a2_Human EEAKRPPPRAPGDPLSSPSPTALPQGGSHTETEDRLFKHLFRGYNRWARPVPNTSDVVIV
a3_Human --------------------LSLLPVARASEAEHRLFERLFEDYNEIIRPVANVSDPVII
a4_Human ------------------------HVETRAHAEERLLKKLFSGYNKWSRPVANISDVVLV
a5_Human --------RCGLAGAAGGAQRGLSEPSSIAKHEDSLLKDLFQDYERWVRPVEHLNDKIKI
a6_Human ------------------------KGCVGCATEERLFHKLFSHYNQFIRPVENVSDPVTV
a7_Human -----------------------------GEFQRKLYKELVKNYNPLERPVANDSQPLTV
a9_Human ---------------------------ADGKYAQKLFNDLFEDYSNALRPVEDTDKVLNV
a10_Human ---------------------------AEGRLALKLFRDLFANYTSALRPVADTDQTLNV
b3_Human ----------------------------IAENEDALLRHLFQGYQKWVRPVLHSNDTIKV
a3_Chicken -----------------------------SEPEHRLYAALFKNYNQFVRPVKNASDPVII
: * : :

40 50 60 70 80 90
Lym_AChBP SVSLKFINILEVNEITNEVDVVFWQQTTWSDRTLAWNSSHS--PDQVSVPISSLWVPDLA
Aply_AChBP TLGFTLQDIVKADSSTNEVDLVYYEQQRWKLNSLMWDPNEYGNITDFRTSAADIWTPDIT
Torpedo TVGLQLIQLISVDEVNQIVETNVRLRQQWIDVRLRWNPADYGGIKKIRLPSDDVWLPDLV
a1_Human TVGLQLIQLINVDEVNQIVTTNVRLKQQWVDYNLKWNPDDYGGVKKIHIPSEKIWRPDLV
a2_Human RFGLSIAQLIDVDEKNQMMTTNVWLKQEWSDYKLRWNPTDFGNITSLRVPSEMIWIPDIV
a3_Human HFEVSMSQLVKVDEVNQIMETNLWLKQIWNDYKLKWNPSDYGGAEFMRVPAQKIWKPDIV
a4_Human RFGLSIAQLIDVDEKNQMMTTNVWVKQEWHDYKLRWDPADYENVTSIRIPSELIWRPDIV
a5_Human KFGLAISQLVDVDEKNQLMTTNVWLKQEWIDVKLRWNPDDYGGIKVIRVPSDSVWTPDIV
a6_Human HFEVAITQLANVDEVNQIMETNLWLRHIWNDYKLRWDPMEYDGIETLRVPADKIWKPDIV
a7_Human YFSLSLLQIMDVDEKNQVLTTNIWLQMSWTDHYLQWNVSEYPGVKTVRFPDGQIWKPDIL
a9_Human TLQITLSQIKDMDERNQILTAYLWIRQIWHDAYLTWDRDQYDGLDSIRIPSDLVWRPDIV
a10_Human TLEVTLSQIIDMDERNQVLTLYLWIRQEWTDAYLRWDPNAYGGLDAIRIPSSLVWRPDIV
b3_Human YFGLKISQLVDVDEKNQLMTTNVWLKQEWTDHKLRWNPDDYGGIHSIKVPSESLWLPDIV
a3_Chicken QFEVSMSQLVKVDEVNQIMETNLWLKHIWNDYKLRWNPVDYGGAEFIRVPSGQIWKPDIV
. . : :: . :. .: : : * * *: . . :* **:

100 110 120 130 140 150
Lym_AChBP AYN-AISKPEVLTPQLARVVSDGEVLYMPSIRQRFSCDVSGVDTESG-ATCRIKIGSWTH
Aply_AChBP AYS-STRPVQVLSPQIAVVTHDGSVMFIPAQRLSFMCDPTGVDSEEG-ATCAVKFGSWVY
Torpedo LYNNADGDFAIVHMTKLLLDYTGKIMWTPPAIFKSYCEIIVTHFPFDQQNCTMKLGIWTY
a1_Human LYNNADGDFAIVKFTKVLLQYTGHITWTPPAIFKSYCEIIVTHFPFDEQNCSMKLGTWTY
a2_Human LYNNADGEFAVTHMTKAHLFSTGTVHWVPPAIYKSSCSIDVTFFPFDQQNCKMKFGSWTY
a3_Human LYNNAVGDFQVDDKTKALLKYTGEVTWIPPAIFKSSCKIDVTYFPFDYQNCTMKFGSWSY
a4_Human LYNNADGDFAVTHLTKAHLFHDGRVQWTPPAIYKSSCSIDVTFFPFDQQNCTMKFGSWTY
a5_Human LFDNADGRFEG-TSTKTVIRYNGTVTWTPPANYKSSCTIDVTFFPFDLQNCSMKFGSWTY
a6_Human LYNNAVGDFQVEGKTKALLKYNGMITWTPPAIFKSSCPMDITFFPFDHQNCSLKFGSWTY
a7_Human LYNSADERFDATFHTNVLVNSSGHCQYLPPGIFKSSCYIDVRWFPFDVQHCKLKFGSWSY
a9_Human LYNKADDESSEPVNTNVVLRYDGLITWDAPAITKSSCVVDVTYFPFDNQQCNLTFGSWTY
a10_Human LYNKADAQPPGSASTNVVLRHDGAVRWDAPAITRSSCRVDVAAFPFDAQHCGLTFGSWTH
b3_Human LFENADGRFEGSLMTKVIVKSNGTVVWTPPASYKSSCTMDVTFFPFDRQNCSMKFGSWTY
a3_Chicken LYNNAVGDFQVDDKTKALLKYTGDVTWIPPAIFKSSCKIDVTYFPFDYQNCTMKFGSWSY
:. : : * : .. * . * :.:* * :

Yellow: mAb35 Specific Blue: mAb210 Specific Red: Both

248

Figure B.1: Sequence alignments of human α subunits and other related
subunits to determine cross-reactivity as well as critical residues for different
mAb bindings (PISA results based).
Sequences of some subunits are aligned for comparison. Subunits which were
reported to have cross-reactivity with mAb35 and mAb210 are colored in black, and
the rest are colored in gray. α1 subunit is colored in red, since it is the original target
for those mAbs. Amino acids are highlighted with different colors depending on
interacting mAbs. Those amino acids are chosen based on the results of PISA
program which was run using the ternary complex structures of Fab35 or Fab210.
Amino acids of α1 which have possibility to form salt bridges or hydrogen bonds
with Fabs (mAbs) were first highlighted, and if other subunits carry the same
residues, those residues are also highlighted with the same color with α1 residues.
Amino acids which have interactions with mAb35 specific are highlighted in yellow,
and mAb210 specific is highlighted in blue. If they interact with both mAbs, they are
highlighted in red. This alignment gives ideas what residues would be causing the
cross-reactivity between subunits. For example, all subunits reported for cross-
reactivity (human α3, human α5, human β3, and chicken α3) have N64, D71, and
Y72 which would have important roles to interact with both mAb35 and mAb210.

249

Comparison of Interface Interaction (Fab35 vs. Fab210), 4.5Å cut-off

1 10 20 30
Lym_AChBP ------------------------------LDRADILYNIRQTSRPDVIP-TQRDRPVAV
Aply_AChBP -------------------------------QANLMRLKSDLFNRSPMYPGPTKDDPLTV
Torpedo -----------------------------SEHETRLVANLLENYNKVIRPVEHHTHFVDI
a1_Human -----------------------------SEHETRLVAKLFKDYSSVVRPVEDHRQVVEV
a2_Human EEAKRPPPRAPGDPLSSPSPTALPQGGSHTETEDRLFKHLFRGYNRWARPVPNTSDVVIV
a3_Human --------------------LSLLPVARASEAEHRLFERLFEDYNEIIRPVANVSDPVII
a4_Human ------------------------HVETRAHAEERLLKKLFSGYNKWSRPVANISDVVLV
a5_Human --------RCGLAGAAGGAQRGLSEPSSIAKHEDSLLKDLFQDYERWVRPVEHLNDKIKI
a6_Human ------------------------KGCVGCATEERLFHKLFSHYNQFIRPVENVSDPVTV
a7_Human -----------------------------GEFQRKLYKELVKNYNPLERPVANDSQPLTV
a9_Human ---------------------------ADGKYAQKLFNDLFEDYSNALRPVEDTDKVLNV
a10_Human ---------------------------AEGRLALKLFRDLFANYTSALRPVADTDQTLNV
b3_Human ----------------------------IAENEDALLRHLFQGYQKWVRPVLHSNDTIKV
a3_Chicken -----------------------------SEPEHRLYAALFKNYNQFVRPVKNASDPVII
: * : :

40 50 60 70 80 90
Lym_AChBP SVSLKFINILEVNEITNEVDVVFWQQTTWSDRTLAWNSSHS--PDQVSVPISSLWVPDLA
Aply_AChBP TLGFTLQDIVKADSSTNEVDLVYYEQQRWKLNSLMWDPNEYGNITDFRTSAADIWTPDIT
Torpedo TVGLQLIQLISVDEVNQIVETNVRLRQQWIDVRLRWNPADYGGIKKIRLPSDDVWLPDLV
a1_Human TVGLQLIQLINVDEVNQIVTTNVRLKQQWVDYNLKWNPDDYGGVKKIHIPSEKIWRPDLV
a2_Human RFGLSIAQLIDVDEKNQMMTTNVWLKQEWSDYKLRWNPTDFGNITSLRVPSEMIWIPDIV
a3_Human HFEVSMSQLVKVDEVNQIMETNLWLKQIWNDYKLKWNPSDYGGAEFMRVPAQKIWKPDIV
a4_Human RFGLSIAQLIDVDEKNQMMTTNVWVKQEWHDYKLRWDPADYENVTSIRIPSELIWRPDIV
a5_Human KFGLAISQLVDVDEKNQLMTTNVWLKQEWIDVKLRWNPDDYGGIKVIRVPSDSVWTPDIV
a6_Human HFEVAITQLANVDEVNQIMETNLWLRHIWNDYKLRWDPMEYDGIETLRVPADKIWKPDIV
a7_Human YFSLSLLQIMDVDEKNQVLTTNIWLQMSWTDHYLQWNVSEYPGVKTVRFPDGQIWKPDIL
a9_Human TLQITLSQIKDMDERNQILTAYLWIRQIWHDAYLTWDRDQYDGLDSIRIPSDLVWRPDIV
a10_Human TLEVTLSQIIDMDERNQVLTLYLWIRQEWTDAYLRWDPNAYGGLDAIRIPSSLVWRPDIV
b3_Human YFGLKISQLVDVDEKNQLMTTNVWLKQEWTDHKLRWNPDDYGGIHSIKVPSESLWLPDIV
a3_Chicken QFEVSMSQLVKVDEVNQIMETNLWLKHIWNDYKLRWNPVDYGGAEFIRVPSGQIWKPDIV
. . : :: . :. .: : : * * *: . . :* **:

100 110 120 130 140 150
Lym_AChBP AYN-AISKPEVLTPQLARVVSDGEVLYMPSIRQRFSCDVSGVDTESG-ATCRIKIGSWTH
Aply_AChBP AYS-STRPVQVLSPQIAVVTHDGSVMFIPAQRLSFMCDPTGVDSEEG-ATCAVKFGSWVY
Torpedo LYNNADGDFAIVHMTKLLLDYTGKIMWTPPAIFKSYCEIIVTHFPFDQQNCTMKLGIWTY
a1_Human LYNNADGDFAIVKFTKVLLQYTGHITWTPPAIFKSYCEIIVTHFPFDEQNCSMKLGTWTY
a2_Human LYNNADGEFAVTHMTKAHLFSTGTVHWVPPAIYKSSCSIDVTFFPFDQQNCKMKFGSWTY
a3_Human LYNNAVGDFQVDDKTKALLKYTGEVTWIPPAIFKSSCKIDVTYFPFDYQNCTMKFGSWSY
a4_Human LYNNADGDFAVTHLTKAHLFHDGRVQWTPPAIYKSSCSIDVTFFPFDQQNCTMKFGSWTY
a5_Human LFDNADGRFEG-TSTKTVIRYNGTVTWTPPANYKSSCTIDVTFFPFDLQNCSMKFGSWTY
a6_Human LYNNAVGDFQVEGKTKALLKYNGMITWTPPAIFKSSCPMDITFFPFDHQNCSLKFGSWTY
a7_Human LYNSADERFDATFHTNVLVNSSGHCQYLPPGIFKSSCYIDVRWFPFDVQHCKLKFGSWSY
a9_Human LYNKADDESSEPVNTNVVLRYDGLITWDAPAITKSSCVVDVTYFPFDNQQCNLTFGSWTY
a10_Human LYNKADAQPPGSASTNVVLRHDGAVRWDAPAITRSSCRVDVAAFPFDAQHCGLTFGSWTH
b3_Human LFENADGRFEGSLMTKVIVKSNGTVVWTPPASYKSSCTMDVTFFPFDRQNCSMKFGSWTY
a3_Chicken LYNNAVGDFQVDDKTKALLKYTGDVTWIPPAIFKSSCKIDVTYFPFDYQNCTMKFGSWSY
:. : : * : .. * . * :.:* * :

Yellow: mAb35 Specific Blue: mAb210 Specific Red: Both

250

Figure B.2: Sequence alignments of human α subunits and other related
subunits to determine cross-reactivity as well as critical residues for different
mAb bindings (CCP4 contact results based).
The main description follows Figure B.2. A different program, CCP4 contact is used
to determine amino acids which are in close contacts. The cut-off distance is set to
4.5 Å. All color cording is the same with Figure B.1. Figure B.1 showed the most
critical residues, and this gives wider range of contacts, which would have more
information to determine the differences of mAb bindings.

251

Sequence Alignments of Different Fabs (Light Chain)

Fab35_Light DIVITQSPSLLSASVGDRVTLTCKGSQNIDNYLAWYQQKLGEAPKLLIYKTNSLQTGIPS
Fab210_Light NIQLTQSPSLLSASVGDRVTLSCKGSQNINNYLAWYQQKLGEAPKLLIYKTNSLQTGIPS
:* :*****************:*******:******************************

Fab35_Light RFSGSGSGTDYTLTISSLHSEDLATYYCYQYINGYTFGTGTKLELKRADAAPTVSIFPPS
Fab210_Light RFSGSGSGTDYTLTISSLHSEDLATYYCYQYDNGYTFGTGTKLELKRADAAPTVSIFPPS
******************************* ****************************

Fab35_Light TEQLATGGASVVCLMNNFYPRDISVKWKIDGTERRDGVLDSVTDQDSKDSTYSMSSTLSL
Fab210_Light TEQLATGGASVVCLMNNFYPRDISVKWKIDGSEQQNGVLNSWTDQDSKDSTYSMSSTLTL
*******************************:*:::***:* ****************:*

Fab35_Light TKADYESHNLYTCEVVHKTSSSPVVKSFNRNEC
Fab210_Light TKDEYERHNSYTCEATHKTSTSPIVKSFNRNEC
** :** ** ****..****:**:*********

Fab35_Light DIVITQSPSLLSASVGDRVTLTCKGSQNIDNYLAWYQQKLGEAPKLLIYKTNSLQTGIPS
Fab210_Light NIQLTQSPSLLSASVGDRVTLSCKGSQNINNYLAWYQQKLGEAPKLLIYKTNSLQTGIPS
Fab198_Light DIKLTQSPSLLSASVGDRVTLSCKGSQNINNYLAWYQQKLGEAPKLLIYNTNSLQTGIPS
Fab192_Light DIQMTQSPPSLSASLGDKVTITCQASQDINKYIAWYQQKPGKAPRQLIRYTSILVLGTPS
:* :****. ****:**:**::*:.**:*::*:****** *:**: ** *. * * **

Fab35_Light RFSGSGSGTDYTLTISSLHSEDLATYYCYQYINGYTFGTGTKLELKRADAAPTVSIFPPS
Fab210_Light RFSGSGSGTDYTLTISSLHSEDLATYYCYQYDNGYTFGTGTKLELKRADAAPTVSIFPPS
Fab198_Light RFSGSGSGTDYTLTISSLQPEDVATYFCYQYNNGYTFGAGTKLELKR--TAPTVSIFPPS
Fab192_Light RFSGSGSGRDFSFSISNVASEDIASYYCLQYGNLYTFGAGTKLEIKRADAAPTVSIFPPS
******** *::::**.: .**:*:*:* ** * ****:*****:** :**********

Fab35_Light TEQLATGGASVVCLMNNFYPRDISVKWKIDGTERRDGVLDSVTDQDSKDSTYSMSSTLSL
Fab210_Light TEQLATGGASVVCLMNNFYPRDISVKWKIDGSEQQNGVLNSWTDQDSKDSTYSMSSTLTL
Fab198_Light TEQLATGGASVVCLMNNFYPRDISVKWKIDGTERRDGVLDSVTDQDSKDSTYSMSSTLSL
Fab192_Light TEQLATGGASVVCLMNNFYPRDISVKWKIDGTERRDGVLDSVTDQDSKDSTYSMSSTLSL
*******************************:*:::***:* ****************:*

Fab35_Light TKADYESHNLYTCEVVHKTSSSPVVKSFNRNEC
Fab210_Light TKDEYERHNSYTCEATHKTSTSPIVKSFNRNEC
Fab198_Light TKADYESHNLYTCEVVHKTSSSPVVKSFNRNEC
Fab192_Light TKADYESHNLYTCEVVHKTSSSPVVKSFNRNEC
** :** ** ****..****:**:*********

Yellow: indicate 3CDRs.
Blue: Possible false sequencing results. Used Fab35 residues at the position.
Pink: It might need to be changed to Fab35 residue.

Figure B.3: Sequence alignments of Fabs (Light Chain).
Sequence alignments of Fab35 and Fab210 (Top). Sequence alignment of Fab35,
Fab210, Fab198, and Fab192 (Bottom). Yellow highlight indicates CDRs. Some
residues of Fab210 did not agree with the electron density (highlighted in blue).
Based on the density, those residues would be same as Fab35 residues, and those
restudies are assigned in the final structure. The residue highlighted in pink was not
changed to Fab35 residue in the final structure, and the correct residue would be D
not E.

252

Sequence Alignments of Different Fabs (Heavy Chain)

Fab35_Heavy EVQLQESGPGLVQPSETLSLTCTVSGFSLTSYSVSWLRQPSGKGPEWMGRMWDDGGTVYN
Fab210_Heavy QVQLKESGPGLVQPSETLSLTCTVSGFSLTSYSVSWVRQPSGKGPEWMGRMWNDGYTAYN
:***:*******************************:***************:** *.**

Fab35_Heavy SGLKSRLSISRDTSKNQVFLKMNSLQTDDTGTYYCTRDERIR--AINWFAYWGQGTLVTV
Fab210_Heavy SALKSRLSINRDTSKNQVFLKMNSLQTDDTGTYYCTRDQRSSRMGYWYSDFWGPGTMVTV
*.*******.****************************:* . : :** **:***

Fab35_Heavy SSAETTAPSVYPLAPGTALKSNSMVTLGCLVKGYFPEPVTVTWNSGALSSGVHTFPAVLQ
Fab210_Heavy SSAETTAPSVYPLAPGTALKSNSMVTLGCLVKGYFPEPVTVTWNSGALSSGVHTFPAVLQ
************************************************************

Fab35_Heavy SGLYTLTSSVTVPSSTWPSQTVTCNVAHPGQQHQRWTRKLCPET
Fab210_Heavy SGLYTLTSSVTVPSSTWPSQTVTCNVAHPASSTKVDKKIVPRNC
*****************************... : .: : :

Fab35_Heavy EVQLQESGPGLVQPSETLSLTCTVSGFSLTSYSVSWLRQPSGKGPEWMGRMWDDGGTVYN
Fab210_Heavy QVQLKESGPGLVQPSETLSLTCTVSGFSLTSYSVSWVRQPSGKGPEWMGRMWNDGYTAYN
Fab198_Heavy QVQLLESGPGLVRPSETLSLTCTVSGFSLTSFSVSWVRHPSGKGPEWMGRMWYDGYTAYN
Fab192_Heavy EVKLLESGPGLVQPSQTLSLTCTVSGFPLTTNGVSWVRQPPGKGLEWIAAISSGGSPYYN
:*:* *******:**:***********.**: .***:*:*.*** **:. : .* . **

Fab35_Heavy SGLKSRLSISRDTSKNQVFLKMNSLQTDDTGTYYCTRDERIRAIN---WFAYWGQGTLVT
Fab210_Heavy SALKSRLSINRDTSKNQVFLKMNSLQTDDTGTYYCTRDQRS-SRMGYWYSDFWGPGTMVT
Fab198_Heavy SALKSRLSISRDTSKNQVFLKMNSLQTDDTGTYYCTRDLYGGYPLGFWYFDFWGPGTMVT
Fab192_Heavy SALKSRLSINRDTSKSQVFLKMNSLQTEDTAIYFCTRED------GWNYFDYWGPGTMVT
*.*******.*****.***********:**. *:***: : :** **:**

Fab35_Heavy VSSAETTAPSVYPLAPGTALKSNSMVTLGCLVKGYFPEPVTVTWNSGALSSGVHTFPAVL
Fab210_Heavy VSSAETTAPSVYPLAPGTALKSNSMVTLGCLVKGYFPEPVTVTWNSGALSSGVHTFPAVL
Fab198_Heavy VS-------SVFPLAPGSAAQTNSMVTLGCLVKGYFPEPVTVTWNSGALSSGVHTFPAVL
Fab192_Heavy VSSAQTTAPSVYPLAPGCGDTTSSTVTLGCLVKGYFPEPVTVTWNSGALSSDVHTFPAVL
** **:***** . :.* **************************.********

Fab35_Heavy QSGLYTLTSSVTVPSSTWPSQTVTCNVAHPGQQHQRWTRKLCPET
Fab210_Heavy QSGLYTLTSSVTVPSSTWPSQTVTCNVAHPASSTKVDKKIVPRNC
Fab198_Heavy QSGLYTLTSSVTVPSSTWSSQAVTCNVAHPASSTKVDKKIVPRDC
Fab192_Heavy QSGLYTLTSSVTS--STWPSQTVTCNVAHPASSTKVDKKLERR--
************ ***.**:********... : .:

Yellow: indicate 3CDRs

Figure B.4: Sequence alignments of Fabs (Heavy Chain).
Sequence alignments of Fab35 and Fab210 (Top). Sequence alignment of Fab35,
Fab210, Fab198, and Fab192 (Bottom). Yellow highlight indicates CDRs. Major
differences can be seen in the CDR3 region.

Abstract (if available)

Abstract Nicotinic acetylcholine receptors (nAChRs) play a key role in neuronal communication by sending electric signals upon binding of neurotransmitters, acetylcholines. nAChRs relate closely to various neuronal diseases including Alzheimer’s disease and schizophrenia. Understanding the mechanism of nAChRs would lead to better understanding of those diseases and eventually to new treatments. This dissertation covers six different projects, four of which are related to a nAChRs α1 subunit. The other two projects involve nAChR subunits, α7 and α9. For many of our studies, the X-ray crystallography technique is utilized for structural determination, which provides useful information to develop drugs or deeper understandings of molecular mechanisms. ❧ The main objective of this dissertation is to understand how a disease called myasthenia gravis (MG) is triggered at molecular level. It is an autoimmune neuromuscular disease, and it is known to be caused by autoantibodies targeting nAChRs. Tremendous researches have been done, but it is still not known how the autoantibodies actually bind the receptors. The first part of discussion is the structural determination of nAChR α1 subunit bound by one of MG mAbs named mAb35, which is a prototypical MG mAb. The first X-ray crystal structure of the complex was determined at 2.6 Å, and detailed interactions of the interface were revealed. The study was then extended to the determination of another complex structure of nAChR α1/mAb210. Various types of MG mAbs have been generated, and they have shown different characteristics. Thus, it is important to learn the differences of these mAbs, specifically mAb35 and mAb210, to understand MG better. Furthermore, this knowledge could be applied for drug developments in the future. The crystal structures of these two distinct complexes provide, for the first time, molecular details for understanding the disease mechanisms and for structure-guided drug design of MG. As the third aim, a new detection method of MG antibody was developed by utilizing all materials and knowledge obtained in the first section of MG study. The current diagnosis of MG has several limitations due to the usage of radiolabeled materials. Even though additional experiments are required for further improvement, our newly developed assay system is more time efficient and cost effective and could contribute to a MG diagnosis in the feature. The last aim related to MG is to discover proteins which can specifically target the nAChR α1 subunit at the interface between the receptor and the autoantibody, and also in near the future, those which target the binding site of MG antibodies. This project is a collaborative work with Dr. Rorberts’ group at USC and is in progress. It has already been achieved to develop a binder protein which targets the specific subunit. Structural information would enable us to design a variety of binder proteins (e.g. weaker binders or stronger binders), which could be used as therapeutic molecules. Then, the topic is moved onto other nAChR subunits. The 5th project is the structural determination of a nAChR α9 subunit. The focus of this project is a sugar chain on the α9, which is highly conserved on the cys-loop of nAChR α subunit and might have important roles in the gating mechanism of nAChRs. The X-ray crystal structure of nAChR α9 bound by α-bungarotoxin was solved at 3.0 Å. Our structure provides more insights into the possible role of the sugar chain. Lastly, this dissertation ends with purification of another nAChR subunit, α7. Due to the difficulty in expressing of the wild type α7 subunit, a chimera of α7/AChBP was produced and successfully purified using purification techniques and knowledge from the nAChR α1 study. The structure determination of the α7 bound with therapeutic compounds would lead for further drug development. ❧ In summary, our structural studies on nAChRs related to MG revealed the detailed information of the binding interface between the receptor and autoantibodies, resulting in further understanding of the disease. In addition, the new diagnostic method has been developed as well as possible therapeutic protein molecules for MG. The X-ray crystal structure of another nAChR subunit, α9 was successfully solved with a clear picture of a sugar chain, which might play an important role in functions of nAChRs. Finally, a robust purification system for the nAChR subunit α7 was developed and optimized for additional structural studies in order to develop therapeutic drugs.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

mRNA display as a method to generate stable fibronectin ligands for the alpha 1 subunit of the nicotinic acetylcholine receptor

PDF

Biochemical regulation of hunger by the melanocortin 4 receptor

PDF

New understanding of RIC-3, a regulator of alpha7 nicotinic acetylcholine receptor expression

PDF

Structural studies of the human leukotriene B4 receptor 1

PDF

Exploring the structure of G protein-coupled receptors

PDF

Structure and function of the orphan G protein-coupled receptor 6

PDF

Structure and regulation of lymphoid tyrosine phosphatase (LYP) in autoimmune response

PDF

X-ray structural studies on DNA-dependent protein kinase catalytic subunit:DNA co-crystals

PDF

Developing ultrasensitive ethanol receptors (USERS) as novel tools for alcohol research: optimizing Loop 2 mutations in α1 GlyRs

PDF

Structural and biochemical studies of two DNA transaction enzymes

PDF

mRNA oxidation and its relation to p53 amyloid formation and disease

PDF

Structural and biochemical analyses on substrate specificity and HIV-1 Vif mediated inhibition of human APOBEC3 cytidine deaminases

PDF

Ultra-sensitive ethanol receptors as novel tools for alcohol and brain research: optimizing loop 2 mutations in α2 glycine receptors, γ2 and α1 γ-aminobutyric acid type A receptors

PDF

Development of stable cell lines expressing α1β2γ2 and α5β3γ2 GABAA receptors

PDF

Functional analyses of androgen receptor structure pertaining to prostate cancer

PDF

A structure based study of the HIV restriction factor APOBEC3G

PDF

Using novel small molecule modulators as a tool to elucidate the role of the Myocyte Enhancer Factor 2 (MEF2) family of transcription factors in leukemia

PDF

Data-driven approaches to studying protein-DNA interactions from a structural point of view

PDF

Site-specific effects of ubiquitin and ubiquitin-like modifier proteins on α-synuclein aggregation

PDF

Structure and function of archaeal McM helicase

Asset Metadata

Creator Noridomi, Kaori (author)

Core Title Structural studies of nicotinic acetylcholine receptors and their regulatory complexes

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Chemistry

Publication Date 08/10/2017

Defense Date 06/22/2015

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag crystal structure,crystallography,myasthenia gravis,nicotinic acetylcholine receptors,OAI-PMH Harvest

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Chen, Lin (committee chair), Arnold, Donald B. (committee member), Cherezov, Vadim (committee member), Roberts, Richard W. (committee member)

Creator Email kaori0821@hotmail.com,noridomi@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-632077

Unique identifier UC11307030

Identifier etd-NoridomiKa-3820.pdf (filename),usctheses-c3-632077 (legacy record id)

Legacy Identifier etd-NoridomiKa-3820.pdf

Dmrecord 632077

Document Type Dissertation

Format application/pdf (imt)

Rights Noridomi, Kaori

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA