STRUCTURAL INSIGHT IN TO THE MECHANISM OF
WAVELENGTH TUNING IN A RHODOPSIN MIMIC
AND
A SINGLE MUTATION RESULTED AN EXTENSIVE
3D DOMAIN SWAPPED DIMERIZATION IN HCRBPII
By
Zahra Nossoni

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Chemistry - Doctor of Philosophy
2014

ABSTRACT
STRUCTURAL INSIGHT OF THE MECHANISM OF WAVELENGTH TUNING A
RHODOPSIN MIMIC
AND
A SINGLE MUTATION RESULTED AN EXTENSIVE
3D DOMAIN SWAPPED DIMERIZATION IN HCRBPII
By
Zahra Nossoni
In the human eye the three types of the rhodopsin (blue, green, red) are
responsible for color vision. All of these color pigments bound to a single
chromophore: 11-cis-retinal. It is not still clear that how the interaction of this
chromophore with these different opsins leads to different wavelength spectrum.
Since working with rhodopsins (membrane proteins) are challenging therefore in
this study human Cellular retinol binding protein II (hCRBPII) small, cytosolic and
soluble protein has been selected to be used as rhodopsin mimic. The first step
was generating of the active lysine in the binding pocket that can bond to alltrans-retinal and forms the Schiff base. Systematic mutations on this protein
followed by x-ray crystallography lead us conversion of hCRBPII to mimic of
rhodopsin that covers the full visual spectrum. Crystal structure of several holo
mutants bound to retinal at high resolution illustrate that changes in the ordered
water networks (making water net work, removing eater net work or changing in
dipole moment) in the binding pocket leads to tremendous changes of the
absorbance of protein bound to retinal. The mutants that do not effect on the
ordered water networks do not make the big differences in wavelength. The
crystal structures also showed that in this system despite rhodopsin the counter

anion is not necessary and the positive charge of the protonated Schiff base can
be stabilized with the other interactions.
Based on these data we proposed the mechanism of the mechanism of
wavelength regulation in one rhodopsin mimic. We could also correlate the pKa of
protonated Schiff base with the water networks and wavelength. We have
observed that the electrostatic interactions between the amino acids and also
polar groups (water) in the binding pocket have an undeniable role on spectral
tuning.
By introducing one specific mutation on hCRBPII, this protein forms a very
stable dimer and domain swapped protein. This unique protein folding
mechanism so far has been observed in only around 40 different proteins and is
believed to be a mechanism of evolution of the dimers and oligomers. The high
resolution crystal structure of the dimer and domain swapped hCRBPII shows
that we could trapped or produce a very stable partially folded protein and with
this data we were able to explain the mechanism of protein folding In hCRBPII.

This Dissertation is lovingly dedicated to grandfather, Agha joon

iv

ACKNOWLEDGEMENTS

"By the power of truth, I while living have conquered the universe." Dr. John Faust

PhD time is like spring weather: sometimes stormy, sometimes rainy, sometimes
sunny and beautiful. It is certainly not stable, but still beautiful. It passed much faster
than I expected. I really don’t know at this step if I am happy to finish it or sad to leave
the lab, the office, the bench and everything I used to have for more than 5 years and
move on. While my heart aches, I am glad that I was able to finish one of the toughest
stages of my life. I was not alone on this journey and honestly would not have finished
this without the help of those close to me.

I am glad and thankful to have great

supporters that were with me shoulder to shoulder. I would like to express my deepest
appreciation to my advisor Prof. James Geiger. Jim taught me how to think as a
scientist, how to think independently, how to face and solve the problems scientifically.
Not only was Jim my academic advisor, he was also my role model. I learned from Jim
how to be a brilliant “Thinker” yet still be down to earth. The transitions I’ve made from
the person I was to who I am now I owe in great part to him. I am so glad and honored
to have him as my advisor as he will be part of my life forever. Thank you Jim, for
everything.
I would like to thank my second reader and committee member Prof. Babak
Borhan. He was the first person who believed in me. Our weekly bio meetings with
Babak were not always smooth and easy, instead often challenging. As such I learned

v

more in each and every meeting. Babak really changed me from a very tender and
fragile person to a scientific fighter. Thank you Babak, I really appreciate it. I really do.
I would like to appreciate Dr. Kevin Walker for being on my committee and
reading my dissertation. Dr Walker always gave expressed positive energy when I was
walking in the hallway in the second floor.
In addition, thank you to Professor Bill Henry for also reading my dissertation.

I

received some great advice from him when I struggled with problems in cloning.
I want to thank Dr. Wenjing Wang for being the best lab partner that anyone
would wish to work with. Wenjing is not only was a great co-worker, but also has a
great personality making her a joy to work with.
I also would like to thank Dr. Chrysoula Vasileiou for her help. I knew Chrysoula
before I joined graduate school in the chemistry department. My brother was working
with her and always talked highly about her.

Honestly, I thought he must be

exaggerating about how smart and helpful she was until I witnessed it personally. Any
description I might provide isn’t likely to do her justice. She has such presence in the
organic group and personally I cannot imagine 5th floor without her. Thank you,
Chrysoula.
I consider myself to be very fortunate to be in an awesome group with great
group members. I would like to thank all current and former members in Geiger’s group
for being friendly and helpful to me. Specifically, thank you Remie Fawaz, Dr. Camille
Watson, Meisam Nosrati, Zahra Assar, Dr. Xiafei Jia, as well as Rebecca Ober, Lindsey
Gilbert and Janice Chiou. I also would like to thank the bio girls in Pro. Borhan’s group
specifically Tetyana Berbasova and Ipek Yapici.

vi

I would like to thank LS-CAT staff at the Argonne National Laboratory for all of
their help during data collection.
I would also like to express thanks to my dear friends who supported me on this
path, particularly my very best Friend Tate Schaar. When I was frustrated and unhappy
during the hard days of graduate school, he was my greatest support. Thank you Tate.
Finally, I would like to thank my dear family. My mom, dad, sister (Dr. Goli
Nossoni), brother (Dr. Farid Nossoni) and baby brother Hessam. Without them, their
love and support, it would not be possible for me to finish this stage.
I owe my success to all of the wonderful people who support, love, and help me.
I would also like to express my thanks to Spongebob, Patrick and Piglet. During
the darkest of times you kept my mind positive and were a constant reminder that some
things in this universe don’t have to always follow obvious paths of logic or sense.

Thank you all.

“All of the pages of this book are full but the story is still unfinished…” Saadi Shirazi,
Iranian Poet of the 1200

vii

TABLE OF CONTENTS

LIST OF TABLES…………..............................................................................................xi
LIST OF FIGURES……………………………………………………………….……………xiii
KEY TO SYMBOLS AND ABBREVIATIONS……………………..………………………..xxi
CHAPTER I. Introduction of Color Vision………………………………………………..1
I-1
Introduction…………………………………...………….…………………...…….…..1
I-1-1 Conformational Changes of 11-cis-retinal Upon Light Absorption……..….8
I-1-2 The Phototransduction Cascade………………………………….……..…..13
I-1-3 Color Vision with a Single Chromophore ……………….………………..…14
I-1-4 Modeling of the Cone Rhodopsin Based on the Crystal
Structure of Bovine Rhodopsin….….………………….…………..…..…….17
1-2
The Proposed Theories For the Mechanism of the Wavelength
Regulation……………..………………………………………………………….……22
I-2-1 Wavelength Regulation Due to the Conformation of the
Chromophore……………………………………………………………….….24
I-2-2 Computational Studies on the Wavelength Regulation……..…….……....27
I-3
Using Model Studies to Understand the Mechanism of Wavelength
Regulation and Rhodopsin System………………….…………………………....…29
I-3-1 Modeling the Color Pigments Based on the Structure of Bovine
Rhodopsin.…………………………………………………………………..…29
I-3-2 Characteristic of a Potential Protein To Be Used As a Rhodopsin
Mimic………………………………………………………………………..….35
I-3-3 The First Generation of Rhodopsin Mimic……………………......……..….35
I-3-4 The Second Generation Rhodopsin Mimic (Human Cellular
Retinol Binding Protein II)…………..….…..........................................…..39
REFERENCES……………………………….…………………………………….....……….42
CHAPTER II. The Crystal Structure of Holo-Wild-Type human Cellular
Retinol Binding protein II (hCRBPII) bound with Retinol and Retinal………….…..50
II-1
Structure Determination from X-ray Diffraction……………………………….…….50
II-2
The Crystal Structure of wt hCRBPII Bound with Retinol and Retinal….…....….54
II-2-1 Introduction………………………………………………………………….….54
II-2-2 Crystal Structure of Holo wt human Cellular Retinol Binding Protein II
(hCRBPII) Bound with Retinol and Retinal………………….………………61
II-3
Experimental……….……………………………………………….………………….68
II-3-1 Material and Method………………………...…………….…………….…….68
II-3-2 Crystallization and Structure Determination………………..……….….…..70
II-3-3 Data Collection, Refinement and Solution……………….………………...71
REFERENCES………………...………………………………………….………….……..…73

viii

CAHPTER III . The Structural Insight of hCRBPII as Rhodopsin Mimic…..……….76
III-1 The Design of a Rhodopsin Mimic Based on a
Heterogonous Protein System…………………………………….……………..…..76
III-1-1 The Crystal Structure of Q108K:K40L (KL) Bound to Retinal…………….78
III-1-2 The Crystal Structure of Q108K:K40L:T51V (KLV) Bound to Retinal.......82
III-1-3 The Crystal Structure of Q108K:K40L:T53C (KLC) Bound to Retinal……84
III-1-4 The Crystal Structure of Q108K:K40L:T51V:T53C(KLVC) Bound to
Retinal………………………………………………………………………..…88
III-1-5 The Crystal Structure of Q108K:K40L:T51V:T53S
(KLVS) Bound to Retinal……………………………………………………….90
III-2 The Effect of Introduction of Tryptophan in the Middle of Polyene
Chain and Rigidifying the Chromophore…….……………………………………..94
III-2-1 The Crystal Structure of Q108K:K40L:T51V:Y19W:R58W
(KLWW) Bound to Retinal……….…………....…..…………………………..96
III-3
Effect of Mutation at Arg58 Position……...…...……………………………………99
III-3-1 The Crystal Structure of Q108K:K40L:T51V:R58F (KLVF)
Bound to Retinal………………………………..…………………………...100
III-3-2 The Crystal Structure of Q108K:K40L:T51V:R58Y:Y19W
(KLVYW) Bound to Retinal…………………………………..……..…….…105
III-3-3 The Crystal Structure of Q108K:K40L:T51V:T53C:R58L
(KLVCL) Bound to Retinal………………………………………….……….110
III-3-4 The Crystal Structure of Q108K:K40L:T53V: R58W bound to
retinal……………………………………………………………………….....112
III-3-5 The Crystal Structure of Apo-Q108K:K40L:T51V:T53C:R58W
:T29L………………………………………………………………………..…119
III-3-6 Conclusion of Mutation on Arg58……………………...……………….......122
III-4 Toward the Most Red Shifted Mutant with Cis-Iminium
Conformation…………………………………………………………….….………..123
III-4-1 The Crystal Structure of
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:A33W
Mutant Bound to Retinal……………….…………...………………....…..125
III-5 Trans-iminium and Details Studies on Gln4…………………………….…….…..128
III-5-1 The Crystal Structure of
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4H
Mutant Bound to Retinal………….………….………….…….……………130
III-5-2 The Crystal Structure of Q108K:K40L:T51V:T53C:R58F
:T29L:Y19W:Q4H Mutant Bound to Retinal……….………………..…….133
III-5-3 The Crystal Structure of Q108K:K40L:T51V:T53C:R58W
:T29L:Y19W:Q4A and Q108K:K40L:T51V:T53C:R58W:
T29L:Y19W:Q4R Mutant Bound to Retinal……………………….…..……135
III-6 The Water Network Between Gln38 and Gln128…………..…………………..…147
III-7 The Effect of the Overall Electrostatic Potential and the Water Networks and
Interaction with Chromophore………………………...……….……………………150
III-8 Effect of Conformation of the Ionone Ring on the Wavelength
Regulation……………………………………………………………………...……..152
III-8-1 The Crystal Structure of Q108K:K40L:T51V:R58Y:Y19W:T53L

ix

Bound to Retinal…………..…….…….…………...…………………..…….153
III-8-2 The Crystal Structure of Q108K:K40L:T51V:R58W:Y19W:T53L:L77T
Bound to Retinal………………...………………...…………………………156
III-9 Conclusion………………………………………….…………………………………157
REFERENCES …………………………………..………………………….…….…………159
CHAPTER IV. Single Mutation Resulted an Extensive 3D Domain
Swapped Dimerization in hCRBP II……………………………………………………...161
IV-1 Introduction……………………………………………………………..…..….……..161
IV-2 Results and Discussion…………………………………………….….……….…..164
IV-2-1 Dimerization by applying a single site
mutationY60W……………………………………………….…………...…..169
IV-2-2 The Crystal Structure of Monomer Q108K:K40L:Y60W (KLY60W)........170
IV-2-3 Structure of KLY60W Domain Swapped Dimer………………..…..…….172
IV-2-4 Structure of the Y60W dimer……………………………………………..…175
IV-2-5 The Nature of Domain Swap Dimerization in HCRBPII……………........180
IV-3 Experimental Procedure………………………………………………………….….182
IV-3-1 Crystallization of hCRBPII Mutants………………………………...………183
IV-3-2 Data collection and Data Refinement…………………..……………...…..184
REFERENCES ………………………….…………………………………………………...186
CHAPTER V. Mutation, Over-expression, Purification, Crystallization, X-ray
Diffraction, Structural solution and Refinements of hCRBPII Mutants…………..189
V-1 Material and Method………………………………………………………..……….189
V-2 Crystallization and Data Collection……………………………..……..…………..194

x

LIST OF TABLES
Table I-1

Sequence homology between human rhodopsins…………………...…….18

Table I-2

Characterization of different ring-locked retinal analogues……………….26

Table I-3

CRABPII mutants with all-trans-retinal and all-trans-C15……………..…..38

Table II-1

X-ray crystallography data and refinement statistics of wt hCRBPII.…….72

Table III-1

The generation of high pKa mutants………..……………………….…..…..78

Table III-2

Comparison of mutations at the Thr53 position…………………….………85

Table III-3

The effect of removal of the polar residues around the chromophore...…90

Table III-4

Combination of Y19W mutation with the other mutations………...……….95

Table III-5

The effect of mutation of R58 on the wavelength…………..………....….100

Table III-6

The effect of R58F mutation in the presence of the other mutations…...103

Table III-7

The enhancement of red shifting based on R58F mutation………...…...104

Table III-8

The enhancement of the red shifting based on R58W mutation….....….104

Table III-9

The enhancement of the red shifting based on R58Y mutation…………109

Table III-10 Mutation of Thr51 and Thr53 to more hydrophobic residue......…………113
Table III-11 Mutation of Arg58 to charged and nonpolar amino acid…………………122
Table III-12 The protein shift caused by A33W…….……………………….……..…....125
Table III-13 Introduction of A33W in different hCRBPII mutant………...……………..128
Table III-14 The effect of Q4 mutation...………...……………………………………….130
Table III-15 Mutation on Gln38 and Gn128 in hCRBPII…….………………………….148
Table IV-1

The monomer psi and phi angles for 10 amino acids……………..…..….179

Table IV-2

The Phi and Psi angles in KLY60W and Y60W dimer structures…….....179

Table IV-3

X-ray crystallography data and refinement statistics……………..…..…..185

xi

Table V-1

The PCR protocol for wt hCRBPII mutants………………………………189

Table V-2

X-ray crystallography data and refinement statistics……….………...….197

xii

LIST OF FIGURES
Figure I-1

The electromagnetic spectrum……………………………….….............……2

Figure I-2

Structure of mammalian eye……………….………………….………….…...3

Figure I-3

The schematic structure of rod and cone photoreceptor cells………..……4

Figure I-4

Crystal structure of bovine rhodopsin (PDB entry: 1F88)…….………...…..5

Figure I-5

The Rhodopsin spans the membrane disk of rod and cone cells………….7

Figure I-6

The cycle of retinal conformational change after light absorption
of bovine rhodopsin………………………………………………………..……8

Figure I-7

The crystal structure overlay of bathorhodopsin (blue)
(PDB entry 2G87) and ground state rhodopsin (green)
(PDB entry 1U19)……………………………………………………………….9

Figure I-8

The overlaid crystal structure of ground state bovine rhodopsin (dark
blue, PDB entry: 1F88) with Meta II (cyan, PDB entry: 3PQR)……....….11

Figure 1-9

Crystal structure of bovine rhodopsin (PDB entry: 1F88)………..….…....12

Figure I-10 The schematic diagram of Phototransduction cascade…………..…..…...13
Figure I-11 The different absorption of retinal as free aldehyde methanol bound as
SB to n-butylamine and protonated form of Schiff base.....…………….…15
Figure I-12

The crystal structure of bovine rhodopsin with highlighted positions,
Which were mutated by Sakmar’s group…………………………………..17

Figure I-13 The crystal structure of bovine rhodopsin PDB entry (1F88)
with the highlighted amino acid residues that were mutated by
Sakmar’s group………………………………………………………..….….21
Figure 1-14 Positioning of point charge or dipoles along the polyene chain of the
chromophore may cause spectral tuning of the chromophore………..….23
Figure I-15 The torsion angle around C6-C7 single bond of retinol-PSB……………..26
Figure I-16 The computational analysis of the gas phase visual chromophore…..….28
Figure I-17 The bond length of the polyene chain of chromophore.
a. Retinal. b. Cyanine dye………………………………...……………...….32

xiii

Figure I-18 Electrostatic potential calculation around the chromophore. Electrostatic
potential calculation of blue, rod, green, and red opsin ……….........…....34
Figure I-19 The Crystal structure of hCRABPII bound with retinoic acid
highlighted interactions of the chromophore carboxylic acid
and CRABPII residues (PDB entry 2FR3)…………………..…………..…36
Figure I-20 The detail of hydrogen bonds between CRABPII and all-transretinoic acid. b. Illustration of the Bürgi-Dunitz trajectory………...............36
Figure I-21

The crystal structure overlaid of R132K:R111L:L121E KLE, PDB entry
2G7B) bound with retinal, (blue) and structure of KLE-R59W
(PDB entry 3F8A) bound with retinal, (blue) and structure of KLE-R59W
(PDB entry 3F8A) bound to C15, with R59W position is highlighted. The
overlaid structure illustrates that C15 is fully embedded within the protein
binding pocket, while retinal (blue) is exposed b. Chemical structures of
all-trans-retinal. c. all-trans C15 analogue…………………...…………......39

Figure I-22

The superimposed structure of hCRBPII bound to all-trans-retinal
(blue) and CRABPII bound to all-trans-retinoic acid (green)
b. The space model of hCRBPII (the exposed part of the chromophore
is blue). c. The space model of hCRBPII (the exposed part
of the chromophore is red)………………………………………………..….41

Figure II-1

The Crystal structure of overlay wt hCRBPII with all-trans-retinal
(blue) and all-trans-retinol (green)………………...................................…56

Figure II-2

The crystal structure overlay of hCRBPII (salmon red),
solution structure of human CRBPI (magenta), 1KGL,
crystal structure of hCRBPIII (red) 1GGL, crystal structure of
hCRBPIV (cyan) 1LPJ………………………………………………...……....57

FigureII-3

The crystal structure CRBP with retinal and interaction with
Phe16 and Leu77…………………………………………………..…….……58

Figure II-4

The crystal structure overlaid of hCRBPII (blue) and CRABPII (red).
The chromophore is highlighted in these two structures...……..….…..…59

Figure II-5

The multiple sequence alignment between human
CRBPs, rat CRBPI and II and zebrafish CRBP………………...…………..61

Figure II-6

The detail of the hydrogen bonds in CRBPII bound to all-transretinal……………………………………………………………………...….…63

Figure II-7

The overlaid structure of CRBPII with all-trans-retinal (green)
and solution structure of CRBPI (blue, PDB entry 1KGL).………….……64

xiv

Figure II-8

The water network and the binding site in hCRBPII with
all-trans-retinal…………………………………………………………………65

Figure II-9

The crystal structure of wt hCRBPII with all-transretinal…………………………………………………………….……………..65

Figure II-10 The crystal structure of single mutant T51V with all-transretinol………………………………………………………………..….…...….67
Figure III-1 The crystal structure of hCRBPII bound to all-trans-retinal with the
highlighted amino acid residues that are important for
generation of the platform of all further mutations…..……………………..77
Figure III-2 The structure overlay of Q108K:K40L mutant with wt hCRBPII with
all-trans-retinal……………………………………………...………….………79
Figure III-3 The π-cation interaction and water mediate stabilize the
positive charge of the iminium…………………..……………………...……80
Figure III-4 The water network inside hCRBPII close to the chromophore………..….81
Figure III-5 The crystal structure of Q108K:K40L (KL) shows a water molecule
forming a hydrogen bond with residues Thr51 and Thr53………..……….81
Figure III-6 The crystal structure overlaid of Q108K:K40L mutant and
Q108K:K40L:T51V mutant………………………………………………...….83
Figure III-7 The overlaid structure of Q108K:K40L andQ108K:K40L:T51V
bound to all-trans-retinal………………………….………………………….83
Figure III-8 Overlay of the structures of three mutants: Q108K:K40L,
Q108K:K40L:T51V and Q108K:K40L:T53C ……………………...………..87
Figure III-9 The crystal structure overlay of four different mutant: Q108K:K40L
,Q108K:K40L :T51, Q108K:K40L:T53C, and
Q108K:K40L:T51V:T53C…………....………………………………….....…88
Figure III-10 The crystal structure of KLVC with the highlighted mutation
positions………………………………………………………………………...89
Figure III-1 The crystal structure of KLVS mutant at 1.4Å resolution.
b. The superimposed KLVS and KLVC mutants.
As it is shown in this figure Ser53 makes a strong H-bond to
Gln38, which causes changing the conformation in the chromophore…..92
Figure III-12 The superimposed structure of KLVC and KLVS….…..……..………...…93

xv

Figure III-13 The crystal structure of Q108K:K40L mutant with highlighted Y16W…...94
Figure III-14 The crystal structure overlay of KLVF and KLVWW ……………...……….97
Figure III-15 The superimposed crystal structure of Q108K:K40L:T51V:R58W:
Y19W and Q108K:K40L:T51V:R58F with the highlighted
Gln38 and 128 with the chromophore distance ………………...…………98
Figure III-16 The space filling model based on the crystal structure of wt hCRBPII…...99
Figure III-17 The space filling model of Q108K:K40L:T51V and the
chromophore is shown in pink . b. The space filling model of
Q108K:K40L:T51V:R58F. c. The superimposed structure of
Q108K:K40L:T51V(green) and Q108K:K40L:T51V:R58F (purple)…….102
Figure III-18 The crystal structure overlay of KL:T51V:R58F, KL:T51V:Y19W:
R58Y and KL:T51V:Y19W:R58W . The conformation of the
ionone ring of chromophore is 6-s-cis. b. The overlaid structure of
all-trans-retinal in Q108K:K40L (blue), Q108K:K40L:T53C (red) and
Q108K:K40L:T51V:Y19W:R58Y (green).……….……………….………..105
Figure III-19 The crystal structure of Q108K:K40L:T51V:Y19W:R58Y mutant
with all-trans-retinal and the ionone ring of chromophore is 6-s- cis……107
Figure III-20 The crystal structure of wt hCRBPII with all-trans-retinol
b. The superimposed crystal structure of KLVR58WY19W
(pink) with KLVR58YY19W (blue)………………………………………...107
Figure III-21 The crystal structure of penta mutant Q108K: K40L: R58L: T51V:
T53C………………………………………………………………………..…111
FigureIII-22 The superimposed structure of Q108K:K40L:T51V:T53C:R58W:
T29L:Y19W and Q108K:K40L:T51V:T53S:R58W:T29L:Y19W.…….….115
Figure III-23 The crystal structure of Q108K:K40L:T51V:T53S:R58W
:T29L:Y19W mutant…………………………...……………………………..116
Figure III-24 The crystal structure of chain A of The binding pocket of
Q108K:K40L:T51V:T53S:R58W:T29L:Y19W mutant.............................118
Figure III-25 The superimposed crystal structure of
Q108K:K40L:T51V:T53S:R58W:T29L:Y19W (light pink) and
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W (green).…………...…...…119
Figure III-26 The crystal structure of apoQ108K:K40L:T51V:T53C:R58W:T29L with the highlighted mutated

xvi

residues)............................................................................................…..120
Figure III-27 The superimposed structure of apoQ108K:K40L:T51V:T53C:R58W:T29L (green) and holoQ108K:K40L:T51V:T53C:R58W:T29L:Y19W (cyan)………………...…..121
Figure III-28 The superimposed structure of KL (508 nm, green)
KL:T51V:R58W:Y19W (577 nm, blue) KL:T51V:R58W:Y19W
(565 nm, red)…………………………………………………………….…..123
Figure III-29 a. The superimposed structure of and
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W (cyan) and
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:A33W (green). b. The spacefilling model of the crystal structure of and
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:A33W with the highlighted
residues. c. The superimposed structure of
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W (blue) and
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:A33W (cyan) …………...127
Figure III-30 The crystal structure of Q108K:K40L:T51V:T53C:R58W:T29L:
Y19W:Q4H with the highlighted residues………………………………....132
Figure III-31 The superimposed crystal structure of
K40L:T51V:T53C:R58W:T29L:Y19W:Q4H (green) and
K40L:T51V:T53C:R58W:T29L:Y19W (cyan)……………………………..133
Figure III-32 The superimposed crystal structure of
KL:T51V:T53C:R58W:T29L:Y19W:Q4H (purple)
and KL:T51V:T53C:R58F:T29L:Y19W (pink)………………..……………135
Figure III-33 The most red shifted mutants (such as
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R) also
have the lowest pKa…………………………………………………………137
Figure III-34 The crystal structure of Q108K:K40L:T51V:T53C:R58W:T29L:
Y19W:Q4R. The side chain of Arg4 is buried in the N-term of
the protein…………………………………………………………………….138
Figure III-35 The π-cation interaction of Trp109 with cis-iminium. b. Trans-iminium...139
Figure III-36 The superimposed structure of
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R (green) and
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W (cyan)…...………………..140

xvii

Figure III-37 a. The space filling model of the crystal structure of
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R mutant the
chromophore is red. b. The sphere structure of
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W mutant.……………...…....141
Figure III-38 The details of the water network between Gln38
and Gln128 in Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4H
mutant……………………………………………………………………......142
Figure III-39 The crystal structure of Q108K:K40L:T51V:T53C:R58W:T29L:Y19W
(green) and model of Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4E
(blue)……………...…………………………………………..………………143
Figure III-40 The crystal structure of Q108K:K40L:T51V:T53C:Y19W:R58W:
T29L:Q4A with highlighted mutated amino acid……………..…….……..144
Figure III-41 a. The sphere model of the most red shifted mutant
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R.
b. The sphere model based on the crystal structure of
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W…………………………….146
Figure III-42 The details of the water network in the binding pocket of
Q108K:K40K:T51V:R48W:Q38F mutant. The water network
is originally between Gln38 and Gln128………………………….…….….149
Figure III-43 Full spectrum of hCRBPII mutants. a. UV-vis spectra of
different hCRBPII mutants bound with all-trans-retinal.
b. Protein solution of different hCRBPII mutants incubated
with all-trans-retinal………………………………………………………….150
Figure III-44 Electrostatic projections (kbTec-1) using the crystal structure of
a. KL (508 nm). b. KL:T51V:T53C:Y19W:R58W:T29L:Q4R
b. (622 nm) hCBRP II mutants (for each calculation both side
of the chromophore, all-trans-retinal, are shown………………….......…152
Figure III-45 a. The crystal structure of Q108K:K40L:T51V:R58Y:Y19W.
b. The superimpose structure of Q108K:K40L:T51V:R58Y:Y19W
(cyan) and Q108K:K40L:T51V:R58Y:Y19W:T53L (blue)……................155
Figure III-46 The crystal structure of Q108K:K40L:T51V:R58W:Y19W:T53L:L77T
with the highlighted residues………………………………………….……157
Figure IV-1 The crystal structure of wt hCRBPII bound with all-trans-retinal
with the highlighted Y60 position in purple. b. The relative position
of Tyr60 to all-trans-retinal.…………………………………………..……..165
Figure IV-2 a. Chromatogram of KLY60W mutant at 280 nm,

xviii

attributed to absorptions of tryptophan residues from the protein. b.
Chromatogram of KLY60W mutant at 280 nm after refolding with
urea.………...…………………………………………………………..….....166
Figure IV-3 a. UV-vis spectral overlay of 40 mM salt elution (red) and 150
mM salt elution (blue) of Q108K:K40L:Y60W incubated with
all-trans-retinal. b.UV-vis comparison of native 40 mM elution
and 40 mM elution of refolded 150mM elution.….…...…………………..166
Figure IV-4 The melting point experiment on KLY60W monomer and dimer.….……167
Figure IV-5 Y60 position in hCRBPII…………….……………………………………….168
Figure IV-6 a. The overlaid crystal structure of Q108K:K40L mutant (light
purple), the chromophore is highlighted in hot pink with
Q108K:K40L:Y60W monomer (salmon) with highlighted
chromophore in green. b. The highlighted chromophore in both
structures..……...……....………………………………………………….…172
Figure IV-7 The crystal structure of KLY60W dimer.……………….…………………..173
Figure IV-8 a. The two chains of the Q108K:K40L:Y60W dimer overlaid
(blue, chain B, green chain A showing a minor structural difference. b.
The surface structure of the dimer, showing the extensive binding
cavity.…………………………………………………………………...……..173
Figure IV-9 a. The crystals structure of the domain swapped Y60W dimer (chain
A). b. The 2Fo-Fc electron density map around the residues between
50-60 (contoured at 1σ) illustrates the continuous electron density that
crosses the dimer interface.………………...……………...……………….174
Figure IV-10 The overlaid structure of KLY60W and Y60W dimer………..............….175
Figure IV.11 The two chains of the Y60W dimer overlaid (red, chain A,
yellow chain B showing a major structural difference in these two
chains.……………..……………………….……….….…………….…..….176
Figure IV-12 The overlaid of the four chains of the Q108K:K40L:Y60W dimer
(green, chain A and blue, chain B) with Y60W dimer
(yellow, chain B and red chain A).…….…………………………………..178
Figure IV-13 The proposed mechanism for formation of monomer
and dimer in Y60……………..……………………………….…………..…182
Figure V-1

The Ramachandran plot of Q108K:K40L mutant…………...……………202

xix

Figure V-2

The Ramachandran plot of Q108K:K40L:T53C mutant………………….203

Figure V-3

The Ramachandran plot of Q108K:K40L:T51V:T53S mutant…………..204

Figure V-4

The Ramachandran plot of Q108K:K40L:T51V:R58F mutant…………..205

Figure V-5

The Ramachandran plot of Q108K:K40:T51V:T53S:R58W:T29L:Y19W
Mutant……………………………………………………………………...…206

Figure V-6

The Ramachandran plot of Q108K:K40L:T51V:R58Y:Y19W mutant..…207

Figure V-7

The Ramachandran plot of Q108K:K40L:Y60W dimer…………………..208

Figure V-8

The Ramachandran plot of Q108K:K40L:Y60W monomer…………...…209

Figure V-9

The Ramachandran plot of
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4H mutant………….…..210

Figure V-10 The Ramachandran plot of wt hCRBPII bound to all-trans-retinal …..…211
Figure V-11 The Ramachandran plot of Y60W dimer………………………...….……..212
Figure V-12 The Ramachandral plot of wt hCRBPII bound to all-trans-retinol……….213

xx

KEY TO SYMBOLS AND ABBREVIATIONS
Ala, A

Alanine

Arg, R

Arginine

Asn, N

Asparagine

Asp, D

Aspartate

Cys, C

Cysteine

Gln, Q

Glutamine

Glu, E

Glutamic acid

Gly, G

Glycine

His, H

Histidine

Ile, I

Isoleucine

Leu, L

Leucine

Lys, K

Lysine

Met, M

Methionine

Phe, F

Phenylalanine

Pro, P

Proline

Ser, S

Serine

Thr, T

Threonine

Trp, W

Tryptophan

Tyr, Y

Tyrosine

Å

Angstrom

MW

molecular weight

xxi

µM

micromolar

µL

microliter

Amp

ampicillin

bR

bacteriorhodopsin

cGMP

cyclic guanosine monophosphate

Clm

Chloramphenicol

CRABPII

cellular retinoic acid binding protein II

CRBPII

cellular retinol binding protein II

Da

Dalton

DMSO

dimehtylsulfoxide

DNA

deoxyribonucleic acid

dNTP

deoxynucleotide triphosphates

dsDNA

double-stranded DNA

dTTP

deoxythymidine triphosphate

Equiv

equivalent

EDTA

ethylenediaminetetraacetic acid

E. Coli

Escherichia coli

ε

extinction coefficient

FABP

fatty acid binding protein

FPLC

Fast protein liquid chromatography

FTIR

Fourier transform infrared spectroscopy

GDP

guanosine diphosphate

xxii

GPCR

G-protein couple receptor

h

hour

iLBP

intracellular lipid-binding protein

IPTG

isopropylthio-β-galactoside

Kd

dissociation constant

L

Liter

LB

Luria Broth

λmax

maxima wavelength

Ni-NTA

Nickel-nitrilotriacetic acid

M

molar

PCR

polymerase chain reaction

PEG

polyethylene glycol

PSB

protonated Schiff base

R

rhodopsin

R-factor

reliability factor

RT

Room temperature

RMSD

root-mean-square deviation

SB

Schiff base

SDS

sodium dodecyl sulfate

SDS-PAGE

sodiumdodecyl sulfate polyacrylamide gel electrophoresis

TM

Transmembrane

xxiii

7TM

Seven transmembrane

UV

Ultraviolet light

Vis

Visible light

WT

wild type

xxiv

CHAPTER I
Introduction of Color Vision

I-1:Introduction

“Somewhere, something incredible is waiting to be known.” Carl Sagan
Colors are part of our life, we start our life with them and sometimes we forget to
ask this question of how we can see them. If we just think about the human color vision,
we know that we can distinguish different absorbances from 400-700 nm. But how can
we see the different colors?
But before we can answer this question we need to know about the properties and
character of light. Light is a range of electromagnetic radiation that to detectable by
human eye. The different color has different wavelength and as the result have different
energies. The energy of each wavelength can be defined by the Plank-Einstein equation
E=hν, where h is Plank’s constant and ν is the frequency. The frequency is related to
the wavelength by the following equation:
Speed of light (C)
Frequency =
Wavelength

1

The energy of the light can be absorbed by the photoreceptor of the eyes and
transformed in to an electrical signal to be processed by the brain. To better understand
the mechanism of light detection, let’s look at the structure of human eye.

THE ELECTROMAGNTIC SPECTRUM
K9?/E/%LGM$4&7$
89:;<$
*#!$

=;2.<>9?/$

@%A.9./:$

*#I,$

*#I($

B;C;DE/$

FEG.9?;<E/G$

HI893$

J9&&9$893$

*#I+$

*#I*#$

*#I*,$

*#I)$

-./01/%23$4567$

*#($

'"#$%&$

*#+$

)"#$%&$

*#*,$

*#*"$

""#$%&$

*#*"$

("#$%&$

*#*+$

*#,#$

!"#$%&$

Figure I-1: The electromagnetic spectrum

Light is absorbed by photoreceptor cells in the retina, which is the most inner layer of
the eye and plays the most important role in vision and color detection. There are two
different types of cells are in the retina, cone cells and rod cells and as it is shown in
Figure I-2 these cells are located in the outer layer of retina. The name of rod and cone
cells comes from the shape of these photoreceptor cells.
2

Figure I-2: Structure of mammalian eye1

Rod cells are responsible for the dim and black-and-white vision under dim light.
These cells are responsible to see the black-and-white images in the dark. These cells
are very sensitive to light but do not mediate color vision.2,3 The other types of
photoreceptors are cone cells are responsible for color vision and carry these color
pigments (blue, green and red rhodopsin). The three cone cells are blue, green and red,
the blue rhodopsin is sensitive to short wavelength (425 nm) while green rhodopsin is
sensitive to medium wavelength (533 nm) and red rhodopsin is sensitive to long
wavelength (560 nm).2,4 in the other hand the cone cells are for the color vision and
daylight and high resolution vision. These cells are mostly concentrated on the retina,
also called “yellow spot”, which is located right across from the pupil of the eye and is
300-700 µm in diameter.

4

In human eye there are approximately 120 millions rod cells
3

while the cone cells are only 6 to 7 millions. These cells are located at the most outer
layer of the retina. The cone and rode cells are neural cells that can detect the light and
convert it to an electrical signal which process in the visual section of the brain. The rod
cells are responsible for the dim vision and they are very sensitive the light. The optical
cells are divided into two parts: the inner segment (IS) and the outer segment (OS). In
both type of cells the outer segment contains the light absorption molecular structure5.
As it is shown in Figure I-3 the outer layer contains a highly dense pack of tiny disks,
these disks contains the light receptor proteins, rhodopsins5.

membrane disk
Outer sigment
mitochondria

nucleous
Inner sigment

synaptic end

Figure I-3: The schematic structure of rod and cone photoreceptor cells

The inner segment (IS) contains the nucleus and other cell parts such as golgi
body and mitochondria which are necessary for the cell functioning. Beside these parts
4

the inner segment is also comprised of a synapse terminus, which can send the
electrical signals to the neurons in the inner layer of retina. The signals are transmitted
through an elaborated array of synapses in the retina toward the visual section of the
brain.6,7,8
As it was mentioned above both outer segment layer of cone and rode cells
contains the photoactive protein, rhodopsin. In the 1930’s it was discovered that
rhodopsin consists of two parts: a protein part, which is called opsin and a chromophore
bound to the protein, which is identified,] as 11-cis-retinal. 9,10
Opsin is a seven membrane helical protein from the family of the G-protein couple
receptors;11,12 this protein can bind covalently to 11-cis-retinal via Schiff base to an
active lysine residue. The Schiff base can be protonated and deprotonated (FigureI-4).

a.

b.
Lys296

c.

NH +
Lys 296

Figure I-4: Crystal structure of bovine rhodopsin (PDB entry: 1F88).

5

Rhodopsin belongs to the superfamily of protein, guanine nucleotide-binding
protein G protein coupled receptors (GPCR).12,13 These proteins can be activated by an
external stimuli, such as light, chloride concentration, hormones, odors and etc.
Activation of the G-protein initiates an enzymatic cascade reaction, which produces the
neuro signals. All of the GPCRs are composed of the seven membrane helices (T7M),
and they connect to each other by six loops, three of them intracellular and three
extracellular. Over 90% of the protein is in the disk membrane with a very high density
of 25,000 rhodopsin/µm.14,2 The N-terminus of the protein is located at the extracellular
surface and the C-terminus at the intracellular (cytoplasmic). Rhodopsin is a unique
member of the GPCR family.

Despite the other family members has no covalent

interaction with the ligand, rhodopsin (both cone and rode) bind covalently to the ligand
11-cis-retinal. The only crystal structure of visual rhodopsin was determined in 2000 by
Palczewski at

2.8 Å resolution (PDB ID: 1F88).15 This structure reveals that 11-cis-

retial has a covalent bound to Lys296 as a cis-Schiff base (SB). Glu113 is the counter
anion and makes hydrogen bond to the SB, stabilizing the protonated state of the
pigment. This feature increases the pKa of PSB to higher than 16. The chromophore is
oriented mostly parallel to the surface of the membrane. This orientation allows the
chromophore to absorb the maximum amount of light and increase the light
sensitivity.18,19,20 The chromophore is deeply buried in the protein and the binding site is
not accessible to the cytosolic or extracellular surface. The chromophore has interaction
with the residues of the neighboring helices and therefore it is locked in the position.
The crystal structure shows no gaps between the helices that would let the retinal exit

6

the binding pocket. However like the other ligands retinal must move out of the binding
pocket for regenerating since all-trans-retinal must isomerize to 11-cis-retinal.16

rhodpsin

!
disks

C-term

cytoplasmic surface

N-term

Figure I-5: The Rhodopsin spans the membrane disk of rod and cone cells.

7

I-1-1: Conformational Changes of 11-cis-retinal Upon Light Absorption

O
Opsin

N
H+

Lys296

min

11-cis-retinal, PSB, 408 nm
Meta-II
Deprotonated Schiff Base
Active State
380 nm

hv
200 fs

NH +
Lys296
11

12

6ms
H+

highly distorted
all-trans-retinal
555 nm

Meta-I
480 nm

ns

NH +

Lumi
497 nm

Lys296
distorted all-trans-retinal
Batho
543 nm

Figure I-6: The cycle of retinal conformational change after light absorption of
bovine rhodopsin

8

Within a few picosecond of light absorption the chromophore goes through a
series of conformational changes, which are described in Figure I-7 and the
chromophore’s structure changes from 11-cis-retinal to all-trans-retinal.17 This
conformational change of the chromophore triggers conformational changes on the
protein. This retinal cis-trans isomerization is usually monitored at low temperature
using various spectroscopic techniques such as UV-Vis.
When rhodopsin (R) absorbs the light, it converts to meta II, the active state (R*).
Formation of meta II causes conformational changes in the T7M helical structure of the
opsin protein and initial the biochemical cascade.21,

22

This biochemical cascade is

illustrated in Figure I-6. Upon light activation, 11-cis-retinyledene isomerizes to the alltrans-isomer.

a.

b.

Figure I-7: The crystal structure overlay of bathorhodopsin (blue) (PDB entry 2G87)
and ground state rhodopsin (green) (PDB entry 1U19)

9

This isomerization from cis to trans happens in picoseconds. The first stable
intermediate form, which is called bathorhodopsin, has λmax at 543 nm and the crystal
structure of bathorhodopsin was determined by the exposure of bovine rhodopsin
crystal to 488 nm light at low temperature of 98K (PDB entry 2G87). 23
The overlay structure of the bathorhodopsin and ground state rhodopsin indicates that
the overall structure of the backbone of the protein does not change at this step and
only the chromophore isomerizes from 11-cis to distorted all-trans. Since in the structure
of bathorhodopsin the ligand isomerizes to a highly twisted form, therefore the energy of
the ground state increase and the energy gap between the ground and excited states
decreases. As the result a dramatic bathochromic shift is observed from 498 nm
(ground state rhodopsin) to 555 nm (bathorhodopsin). After this step rhodopsin forms
two unstable intermediate Lumi and Meta I. Meta I further convert to meta II (the active
state). In this step the proton transfers from the PSB to the counter anion. The crystal
structure of this photo intermediate was determined in 2011 by two different groups
(PDB entry 3PQR).21,22 The crystal structure of this intermediate is available with the
short Gα-helices from transducine. The crystal structure of this intermediate illustrates
significant conformational changes in compared to ground state rhodopsin.

10

a.

b.

E113

TM3
5.29 Å

3.88 Å

K296

Figure I-8: The overlaid crystal structure of ground state bovine rhodopsin (dark
blue, PDB entry: 1F88) with Meta II (cyan, PDB entry: 3PQR).

In Meta II, the transmembrane helix 3 moves from their position to open up the
binding site for Gα peptide of transducine. As it is illustrated in Figure I-8.a due this
significant structural changes in Meta II, the chromophore (all-trans-retinal) is bound in a
very different way and this conformational changes can explain the drop of the PSB-pKa
of this intermediate. In Meta II the PSB deprotonated and the proto transfer to the
counter anion Gli113. As it is shown in Figure I-8.b the distance of the counter anion,
Glu113, from PSB of Meta II rhodopsin increases to 5.3 Å. Therefore the hydrogen bond
between the PSB and the counter anion disturb and as the result the pKa decreases.
Among all of rhodopsin intermediates, Meta II has the longest lifetime as compared to
the other rhodopsin intermediates.

The conformational changing of retinal causes

changes of the transmembrane helices and leads to activation of the phototransduction

11

cascade.25 At the end of the visual cycle the iminum bond of Meta II hydrolyzes and
Meta II breaks down to all-trans-retinal and opsin and the visual cycle repeats.24,25

a.

K296

b.

11-cis-retinal
3.6 Å
3.3 Å

E113

Figure 1-9: Crystal structure of bovine rhodopsin (PDB entry: 1F88)

12

Light
R

R* Active State
Ion channel closed
Light

CGMP concentration drop

CGMP

CGMP

Phosphotransferase (PDE)

CGMP

CGMP

Figure I-10: The schematic diagram of Phototransduction cascade

I-1-2: The Phototransduction Cascade

The biological cascade of phototransduction and activation of G-proteins is
illustrated in Figure I-10. The rhodopsin active state (R*) binds to GDP-bound
transducin (Gt), resulting of this part is changing GDP changing to GTP.26,27 Gt has
three subunits: α, β, γ. The GTP and subunit α (Gtα-GTP) dissociate from the other two
subunits (γβ). Gtα-GTP binds cGMP PDE (phosphodiesterase) and remove the
inhibitory subunit (γ). The activated PDE converts cGMP (3’, 5’ cyclic monophosphate)
13

to GMP.24 The reduction of the level of cGMP in the cell makes cGMP gate-ion channel
close and as the result is hyperpolarization of the membrane and generation the electric
signal from the basis of vision.25,28,29 The active form of rhodopsin (R*) separates from
the inactive part of Gt and recycles to active more transducins. Since hydrolysis of R* is
a slow process therefore R* has enough time to activate 102 PDEs. Each PDE is able to
convert 103 cGMPs to GMPs, therefore each activated rhodopsin is capable to convert
105 cGMPs to GMPs.72. It starts with activation of rhodopsin by light. The α subunit of Gt
separates and activates the phosphotransferase which coverts CGMP to GMP and the
ion channel closes. This conversion polarized membrane leads in to neural signal.
Since the hydrolysis of the imine bond between all-trans-retinal and the active rhodopsin
(R*) is very slow R* has enough time to activate more than 102 of PDEs, which each of
them can covert ~ 103 cGMPs to GMPs. Therefore one R* has the ability to covert 105
cGMPs to GMPs. At the end of the cycle all-trans-retinal dissociates from the opsin for
regeneration. The first step involves a dehydrogenase that reduces the aldehyde to alltrans-retinol and moves it back to the retinal pigment epithelium (EP) for regeneration to
the original11-cis-retinal. The opsin can bind to a new regenerated 11-cis-retinal and
initiate the new cascade.

I-1-3: Color Vision with a Single Chromophore

All the different color pigments (cone rhodopsins) bind with a single
chromophore, 11-cis-retinal. However they all have different absorbance. Free all-transretinal and 11-cis-retinal solution on the organic solvent like ethanol absorbs at 380 nm.
14

After formation of Schiff base (SB) with n-buthylamine (to model rhodopsin) the pigment
absorbs at 365 nm. However protonated SB (PSB) absorbs at 440 nm (Figure I-11).

30

By changing the concentration and also the solvent this absorbance can shift to 500
nm.34

O

11-cis-retinal
λmax = 380 nm

N

11-cis-retinal-PSB
λmax = 365 nm

+H
N

11-cis-retinal-SB
λmax = 440 nm

Figure I-11: The different absorption of retinal as free aldehyde in methanol, bound
as SB to n-butylamine and protonated form of Schiff base.

In 2005 the absorbance of all-trans-retinal with protonated n-buthylamine Schiff
base as the model of rhodopsin complex were was measured in the vacuum and it was
found the λmax at 610 nm.31,32 Since in this system there is no counter anion present that
can stabilize the positive charge of the iminium, therefore the positive charge can be
delocalized to the largest extent and this can lead to the moret bathochromic shift.
Although formation of the PSB is enough to explain the absorption of the blue and rode
pigments but it is not enough to explain the red shifted absorption for the green and red
rhodopsin. 35
The protein-chromophore interactions must play a critical role for the spectral
tuning of the chromophore.

36,37,38

Therefore the single chromophore in the different

rhodopsins can absorb from 400 to 600 nm. The difference between the λmax of the
15

pigment from the PSB-butylamine model is called “ Opsin Shift”33. This term also can
define the effect of the opsin binding site on the chromophore. Although the Opsin shifts
have been known for the long time,39 still the mechanism of opsin shift is not clear.
Different theories were suggested to explain the opsin shifts. One of the theories
suggests that interaction of the residues inside the binding pocket of the opsin with the
chromophore causes red shift as compare to the retinal in solution.40 Unfortunately
there is not any crystal structure of any of the color pigments to study on the interaction
of the residues inside the binding pocket with the chromophore. As it was mentioned
above the first and only available structure of the rhodopsins is bovine rhodopsin. This
structure was determined at 2.8Å by Palczewski, however the higher resolution
structures of this protein were also determined later. The higher resolution crystal
structure at 2.2Å of rhodopsin illustrates that the β-ionone ring of the chromophore
adopts the 6-s-cis conformation.15 This structural data also indicates the isomerization
around the C11-C12 double bond upon absorption of photons.41 As of today there is no
crystal structure of any cone rhodopsin available therefore the only information about
the structure of these color pigment can be obtained through indirect methods such as
NMR labeling, site mutation and advanced theoretical calculations.42,43,44

16

I-1-4: Modeling of the Cone Rhodopsin Based on the Crystal Structure of Bovine
Rhodopsin

As was mentioned above so far the only available crystal structure of a visual

A164

A269
F261

K296

Figure I-12: The crystal structure of bovine rhodopsin with highlighted positions,
which were mutated by Sakmar’s group.

pigment is the structure bovine rhodopsin.
This crystal structure was used as the platform for rational mutation and modeling of the
three cone rhodopsins. The crystal structure of bovine rhodopsin illustrates that the
17

binding pocket of this protein is made of the hydrophobic residues. Even though no
crystal structure of color pigment is available, the theoretical models of the three cone
rhodopsins illustrate the same fact. Nathan and his co-workers sequenced the gene of
red, green and blue rhodopsin in 1986.45 Red and green pigments are 96% identical.
They are only different in 15 amino acid residues in the whole protein and they have
almost identical in the binding pockets, while blue rhodopsin is very different from the
other two. Both green and red rhodopsin bind to the chromophore via Lys312 (which is
identical to Lys296 in bovine rhodopsin) and counter anion is Glu129 (analogous to
Glu113 in bovine rhodopsin). The second glutamic acid (Glu102) is available in the
green and red rhodopsin; this amino acid is absent in blue rhodopsin.

Table I-1: Sequence homology between human rhodopsins

Percentage
Rod

Blue

Red

Green

Rod

100

75

73

73

Blue

41

100

79

79

Red

37

43

100

99

Green

38

44

96

100

In 1991 Neitz and his coworkers show that three amino acids that are different in
the binding pocket of green and red rhodopsin are probably responsible for the 30 nm
absorbance differences between green (562 nm) and red rhodopsin (530 nm)
18

46

. The

crystal structure of bovine rhodopsin illustrates that these three amino acid residues are
located at the ionone ring region and mutation of these hydrophobic amino acid
residues to polar amino acids leads to more red shifted absorption (Figure I-12). In the
blue pigment Tyr262 is approximately located in the ionone region while at the same
position there is a Tryptophan in both green and red rhodopsin. Mutation of Tyrosine to
Tryptophan (Y262W) leads to 10 nm red shift in the absorption spectra.

47

This study

illustrates the important role of Trp262 in red shifting of the wavelength. However
without the crystal structure it is not clear the mechanism and the effect of each amino
acid. On modulating the spectra in the protein. Oprian and his group mutated all of the
15 amino acid residues, which are different in green and red rhodopsin.48 They
observed that seven of the mutations are responsible for the wavelength for spectral
tuning, whiles Neitz proposed three of these positions previously. Sakmar’s group made
the mutation on the three amino acids residues, which were proposed by Neitz in bovine
rhodopsin. The three positions are Ala269, Phe261 and Ala164. Sakmar’s group did
mutations on nine different positions in the bovin rhodopsin and they could modulate the
wavelength from 500 nm to 438 nm

(62 nm) more blue shift with overall nine

mutations.49 These mutations are A299C, A292S, G90S, A117G, M86L, A299C, A124T,
W265Y, and E122L.51 The general conclusion from these studies was that introduction
the negative charge at the PSB region and removal of a negative charge from the
ionone region can lead to more blue shifted pigment. However, since the crystal
structures of these mutants are not available therefore this conclusion could not be
verified. In 1991 Jacobs and his coworkers demonstrated that mutation of three nonpolar amino acids in green rhodopsin with polar amino acid (hydroxyl bearing) could
19

cause the majority of the absorption difference between blue and green rhodopsins.46
These mutations are A108S (~4nm), F277Y (~10 nm) and A285T (~ 16 nm). Based on
the theoretical model of green rhodopsin these residues are located around the ionone
of the chromophore. In 1992 Sakmar’s group made the same mutations at the
equivalent positions of rod rhodopsin. The tetra mutant A:164S:F261Y:A269T did not
bind to the chromophore. However the double mutant F261Y:A269T displays 20 nm
more red shifted, while the single mutant A164S showed a slight red shift (~2nm) as
compared to the wild type structure .Therefore it seems that only two amino acids, F261
and A269, are responsible for the differences between green and red pigment in human
eye. These data show that spectral tuning of the chromophore by the different mutation
is not exactly additive and can enhance protein-chromophore interaction.46 Different
studies have been done on the counter anion region as well. Oprian and his group
mutated all of the glutamic acid and aspartic acid residues in the rhodopsin and found
that Glu113 is the counter anion since the E113Q mutation leads to significantly lower
pka of the PSB of the bovine rhodopsin.50 After this study Sakmar’s group mutated
Glu113 and they showed that that by shortening the length of the counteranion by
introducing an aspartic acid instead or by removing Glu113 completely and introducing
the hydrophobic amino acid like alanin up 30 nm more red shift is obtained.52

20

!

Glu112

Ala117

Met86

Gly90

Ala299
Ala124

Lys296
Ala292
Trp265

Figure I-13: The crystal structure of bovine rhodopsin PDB entry (1F88) with the
highlighted amino acid residues that were mutated by Sakmar’s group.

The result of this study suggest that removing the negative charge close to the
PSB region leads to more delocalization of the positive charge along the chromophore
and therefore red shift. The second glutamic acid in both green and red rhodopsin is
believed to be responsible for the additional blue shifting in these two rhodopsins.
In both green and red rhodopsin, chloride-binding sites have been observed and it was
believed that this anion is responsible for the point charge in modulation of the
absorbance in the chromophore.53 The other two positively charge amino acids (His197
and Lys200) that are conserved residues in the long-wavelength pigment are believed
to forming the chloride binding site. These residues are not present in the shortwavelength pigment. Although all of these studies have resulted in significant
improvement of our knowledge about the mechanism of wavelength regulation and
21

color vision still no crystal structure of any of these mutants rhodopsins are available. To
fully understand this mechanism and study on the interactions between the amino acid
residues and the bond chromophore the crystal structure of the WT opsin and their
mutants are necessary.

1-2: The Proposed Theories For the Mechanism of the Wavelength Regulation

Over more than 50 years several different mechanisms have been in an attempt
to explain for the mechanism of the wavelength regulation by the different color
pigments. These theories are mostly focused on the delocalization of the π-electrons of
the chromophore.

54

Based on the general knowledge on conjugated systems, more

delocalized π electrons lead to more red shifting.
The role of the counteranion and its interaction with the PSB on the mechanism of the
modulation the absorption maxima was studied. In 1993 Sheves and his coworkers
used a model compound to mimic the PSB in rhodopsin and they illustrated that if the
counteranion is moved away from the PSB, a significant red shift absorption maximum
of the rhodopsin is obtained. They also showed that changing the angle of approach for
the counter anion does not have a significant effect on the wavelength, however it could
perturb the pKa of the PSB.55
Another possible explanation of the opsin shift is based on the point charge
model. Based on this model placement of another negative charge beside the
counteranion along the polyene chain of the PSB-chromophore can change the
delocalization (or conjugation) of the positive charge of the PSB along the polyene.56,57
22

It is obvious that the presence of the counteranion is necessary for stabilization of the
retinal-PSB in order to achieve a high pKa for the rhodopsin. By removing the
counteranion in rhodopsin the PSB becomes very unstable. Therefore removing of the
counteranion cannot explain the bathochromic shift from green rhodopsin (530 nm) to
red rhodopsin (560 nm).
Nakanishi and Honig proposed the point charge is located in a different place in
bacteriorhodopsin and bovine rhodopsin.56,57 Based on this suggestion in the
bacteriorhodopsin the point charge is located close to the ionone ring, while in bovine
rhodopsin is closer to in the middle of the polyene chain of the chromophore. Therefore
the bacteriorhodopsin (560 mm) is more red shifted than bovine rhodopsin (500 nm).

_

_
_

N
H+
Figure 1-14: Positioning of point charge or dipoles along the polyene chain of
the chromophore may cause spectral tuning of the chromophore.

This theory has lost popularity since mutagenesis studies did not have a large
effect on the wavelength expect for the counteranion.58 Besides that, the crystal
23

structures of bovine and bacteriorhodopsin also illustrate that the only negative charged
amino acid in the binding pocket is the counter anion.15 Since no permanent negatively
charged residue was found in the binding pocket of rhodopsin, therefore the
polarizability of the residues could play an important role in the opsin shift. In the other
words, positioning of the polarizable groups in the binding pocket and along the
chromophore can favor or disfavor the delocalization of the charge depending on the
orientation and therefore cause the shift in the wavelength.39 In 1976 Mathies and Stryer
showed that at the excited state of the protonated Schiff base the negative charge
transfer from the ionone ring to the PSB upon electronic excitation. In the other word at
excited state the positive charge of the PSB is transferred and delocalized to the ionone
ring.54 Therefore the permanent dipoles of the amino acid residues, which have
interaction with retinal could generate inducible dipoles upon electronic excitation and
the maximum and optimal interaction with the ground states. This theory is supported by
experiments with polarizable solvents. The data shows that the polarizable solvents like
benzene or dichloromethane yield more red shifted species are compared to methanol
and ethyl ether. Other studies also suggest that the presence of well-ordered water
molecules in the binding pocket can increase the pKa of the PSB.

I-2-1: Wavelength Regulation Due to the Conformation of the Chromophore

The conformation of the chromophore has been suggested to have an important
role in the wavelength regulation. Based on this theory the steric factor in the
chromophore may lead to twisting around the single chromophore’s and therefore
24

change the level of the delocalization of the π-electrons or p-orbital overlap. Changing
the delocalization of the π-electrons of the polyene backbone leads to changes in the
degree of cationic conjugation along the polyene chain. In other word changing the level
of conjugation may change the absorption of the chromophore. Although retinal would
rather adopt a more planar conformation to maximize the overlap of the p-orbital, 11-cisretinal is not a planer molecule. The ionone ring has a 6-s-cis or 6-s-trans conformation
since steric interactions between C5-Me and C8-H as well as the C10-H and C13-Me
result in twisting of the C12-C13 single bond.60 Since there is no crystal structure of any
of the color pigments available so far, there is no solid data to prove this theory.59 One
of the major conformational changes is believed to happen is twisting around the single
bond C6-C7. The crystal structure of the bovine rhodopsin illustrates that the
conformation of the chromophore is 6-s-cis. This twisting may happen because the C5methyll and C8-H have steric interaction and therefore the ionone ring rotates around
the C6-C7 single bond. The result of this twisting is less conjugation than 6-s-trans. As
mentioned above the conformation of the chromophore in bovine rhodopsin is 6-s-cis
while in microbial rhodopsin is 6-s-trans. This theory was proposed to explain why the
microbial rhodopsin is more red shifted than bovine rhodopsin.32
Other than that the planarity of the chromophore is believed to be important for the
wavelength tuning. When the ionone ring of the chromophore locked in the planar
position the most red shifted pigment is to be obtained.

25

C6

C7

O

1: n-0
2: n=1
3: n=2
4: n=3

N

Figure I-15: The torsion angle around C6-C7 single bond of retinol-PSB

Table I-2: Characterization of different ring-locked retinal analogues.
Chromophore

Absorbance (nm)

11-cis-retinal
PSBaldehyde

bound to rhodopsin

free

500

444

380

1

506

539

422

2

496

546

416

3

457

503

386

4

440

483

374

As it is shown in table I-2 retinal analogs 1 and 2 are the most planar ones as
compared to the 3 and 4 and therefore they are the most red shifted aldehyde in the
solution. Garavelli and his group showed that 6-s-trans-conformer of all-trans-retinalPSB in vacuum absorbs at 610 nm while 6-s-cis conformer absorbs at 530 nm. This
26

dramatic different in the absorbance is due to the highly twisting along C6-C7 single
bond.32 Olivucci and Garavelli computed that in gas phase 11-cis-retinal-PSB absorbs
at 545 nm and the ionone ring of the chromophore is 68° twisted from the flat plane.61 In
their studies they also showed that in 6-s-cis conformer if the rotation angle (θ) between
the ionone ring and the plane of the polyene chains is 90° and the conjugation of πelectrons of the polyene completely breaks from π-electron of the ionone ring, the
chromophore absorbs at 535 nm.

I-2-2: Computational Studies on the Wavelength Regulation

By the rapid development in computational studies the computational simulation
methods used widely in the modern studies. One of the calculations have been done by
Olivucci and Garavelli on the vertical excitation energy of the S0S1 vertical excitation
of 11-cis-retinal in gas phase and they predicated the absorption with different torsion
angles between the ionone ring and the polyene chain of chromophore.

62

They

calculated that when the torsion angle between the β-ionone ring and the polyene chain
is zero and the double bond of the ring is on the same plane as the backbone double
bonds the maximum conjugation is obtained and the most red shifted absorbance is
resulted, while when this angle is changed to 90° and the double bond is completely out
of the plane the most blue shifted pigment is obtained (Figure I-14).

27

θ = 0 λmax = 673 nm
θ = 68 λmax = 545nm
θ = 90 λmax = 535 nm

N
H+
Figure I-16: The computational analysis of the gas phase visual chromophore.

Based on this data they proposed that in red rhodopsin the chromophore is more
planar and has the most conjugated π-system, while blue rhodopsin is expected to have
the most twisting around the C6-C7 single bond and therefore ionone ring double bond
is completely out of conjugation.
Another calculation that has been done was studies on the presence of the counter
anion in the structure of the rhodopsin. As it was mentioned above, the crystal structure
of rhodopsin illustrate that the positive charge of the PSB is stabilized by the presence
of a negatively charged residue (Glu113). Buss and his coworkers showed that by
introduction of the counter anion the absorption of the retinal-PSB in vacuum from 610
nm decreases to 486 nm.63 The different amino acids around the chromophore lead to
slightly bathochromic shifts due to their charge and dipolar, while the conformational
changes in the structure of the chromophore do not show any significant role in tuning
the wavelength. They suggest that tuning of wavelength is dictated by the counter anion
and the rest of the amino acids inside the binding pocket do not play any significant role.
28

However mutagenesis of the counter anion leads to just 30 nm more red shift and does
not cover the whole spectrum while the pKa drops significantly by this mutation.

I-3: Using Model Studies to Understand the Mechanism of Wavelength Regulation
and Rhodopsin System

As it was explained above different theories were suggested to explain the
mechanism of the wavelength regulation. However still no single conclusion has been
obtained from these theories because unfortunately so far no crystal structure of any of
the color pigments are available. Lack of direct information from the crystal structures
made the scientists to study on the wavelength regulation with the indirect methods. Not
only site mutagenesis but also biophysical methods were used to study on the
rhodopsin system. Raman spectroscopy was used to study the conformational changes
of retinal and the interaction of the chromophore with the residues in the binding pocket.

I-3-1: Modeling the Color Pigments Based on the Structure of Bovine Rhodopsin

The X-ray crystal structure of bovine rhodopsin was a revolution in the study on
the wavelength regulation. This crystal structure provides the structural information that
later on was used in the other theories which tries to explain the tuning the spectrum in
human eye. Since rhodopsin is a membrane protein, therefore expression, purification
and crystallization of this protein are very challenging. Therefore more flexible strategies
such as computational studies and also using a small-reengineered protein as the
29

mimic of rhodopsin were used. Using a small protein, which is reengineered as the
binding pocket of the rhodopsin, can provide us with a better platform to find out the
mechanism of wavelength regulation.
Due to difficulties and challenges of working with rhodopsin, over the past few decades
many different computational and modeling have been done to understand the
mechanism of spectrum tuning which takes place in each of the color pigments.
Although the crystal structures of the cone rhodopsins are the ultimate answer
regarding the mechanism of wavelength regulation in human eye. However, due to lack
of the real crystal structure we believe that a protein mimic is would be the most realistic
system to study the color pigments. Professor Babak Borhan in Department of
chemistry at Michigan State University and his group initiated the idea to use a soluble
and relatively small protein and reengineer the binding pocket to make it a mimic of the
binding pocket of the visual rhodopsin. Such a realistic model might provide more
information and better platform to study the spectral tuning.
The electronic profile of retinal-PSB shows that at the ground state the positive charge
of the iminium is distributed on the different carbons of the chromophore. Generally the
carbon, which is directly attached to the nitrogen of the iminium, carries the highest
concentration of the positive charge and by increasing the distance from the iminium the
amount of the positive charge decreases. However the profile of the charge distribution
in the excited state of all-trans-retinylidene is very different form the ground state. In the
excited state the charged is transferred to the end of the chromophore and carbon 5 to
9 carried the most amount of the positive charge.65

30

Another electric parameter in retinal-PSB is the alternating single/double bond
character. Buss shows in his calculation on the bond distances of bond form of all-transretinal in PSB and SB form in solution and in rhodopsin. The data clearly shows that in
the most blue shifted form has the most obvious distances in the single and double
bonds. By protonation of the Schiff base and red shifting from 360 nm to 440 nm these
alternations of the single/double bond distances become less significant. The result of
this effect is charge delocalization on the chromophore through resonance structure
(Figure I-15). The data indicate that the more stable and smaller alternation differences
between the bond length of single and double bonds lead to more delocalization of the
positive charge of the PSB and more red shifting will obtain. The cyanine dyes can be
used as an example. This fact can explain why a cyanine dye is more red shifted than
retinal, although they both have the same number of double bonds. Since in the
cyanine-PSB the positive charge is fully delocalized along the whole conjugated system.
Based on these calculations in order to achieve the red shift, the positive charge needs
to be delocalized on the polyene and the resonance structure needs to be stable
through favorable electrostatic interactions with the residues in the binding pocket. The
interactions of chromophore-protein in rhodopsin were studied by different groups.
Matthias, Ullmann and coworkers calculated the electrostatic potential on the surface of
the chromophore and the protein environment in sensory rhodopsin II (500 nm),
bacteriorhodopsin (560 nm) and halorhodopsin n (570 nm). For this calculation the high
resolution crystal structure is necessary, since the conformation of each residue
specially the ones that have interaction with the chromophore is important.65

31

1.48

Bond Length ( A)

1.38

1.28

C5 C6

C7

C8 C9 C10 C11 C12 C13 C14 C15 N16

Carbon Atom
b.
a.
N
H

Figure I-17: The bond length of the polyene chain of chromophore.64
a. Retinal. b. Cyanine dye

Dr. Lee in Professor Borhan’s lab performed the same calculation on the theoretical
models of the three different color pigments.66 The results of this calculation are
summarized in Figure 1-17. The chromophore is divided into three segments; the
32

qualitative average score for each segment represents the overall electrostatic
potentials that lead to the wavelength regulation of each opsin.66 (calculations of blue,
green, and red are based on available homology models) projected on the Van der
Waals surface of retinal (the electrostatic potential of the retinylidene chromophore was
set to zero as to only illustrate the electrostatic contribution of the protein).
As it is illustrated in this figure in blue rhodopsin the negative electrostatic
potential is mostly concentrated on the SB region while in the red rhodopsin the
negative charge is more transferred on the ionone ring area. Based on this calculation,
since blue rhodopsin has more negative electrostatic potential around the iminium part
therefore the positive charge of the PSB is more stabilized on that region. On the other
hand in red rhodopsin the positive charge is more delocalized. This electrostatic
calculation is used as the platform of the series of mutants that Dr. Wenjing Wang in
Prof. Borhan’s group generated on the protein mimic. In this study human Cellular
Retinol Binding Protein II (hCRBPII) has been used as the platform of the rhodopsin
mimic.

33

-40
0
40
Blue 420 nm

Rod 500 nm

Green 530 nm

Red 530 nm

Qualitative Average Scoring
I

II

III

Rh

Blue/Rh

Blue

Red

Green

Rh

Less Polar

Blue/Green

Red

Red/Green

Non Polar

NH

I

II

III

Polar

Figure I-18: Electrostatic potential calculation. Electrostatic potential calculation of
blue, rod, green, and red opsin.

34

I-3-2: Characteristic of a Potential Protein To Be Used As a Rhodopsin Mimic

The protein that is used for this purpose must have several characteristics. (1) It
must be very soluble and should not aggregate during expression and purification and
therefore easy to handle and study. Additionally it should be expressed in an E.coli
expression system. (2) Since during this study several site-mutations need to be done
on the protein, therefore this protein should have a robust structure and tolerance to the
mutations. (3) Also since during this study the crystal structures of different mutants are
necessary, the mimic protein must be relatively easy to crystallize and easy to
determine its structure. (4) The candidate protein as rhodopsin mimic also must have an
embedded binding pocket. This feature of the protein makes the changes in the spectral
tuning just by applying the different mutation not the effect of the solvents. The
reengineering of the rhodopsin mimic is based on the electrostatic potential calculation,
which was performed in Professor Borhan’s group.

I-3-3: The First Generation of Rhodopsin Mimic

Human Cellular Retinoic Acid Binding Protein (hCRABPII) from the family of lipid
binding proteins (iLBP) with 137 amino acid residue was initially used as the template
for the rhodopsin mimic.67 Since 11-cis-retinal is very light sensitive compound therefore
in this study all-trans-retinal has been used, since it is easier to handle. As it was
mentioned above the chromophore is covalently bond to the protein via an active lysine.
For formation the first generation of rhodopsin mimics R132 position was mutated to a
35

lysine. In order to have a successful nucleophilic attack to the carbonyl group of the
retinal, the Bürgi-Dunitz trajectory is important and the amine group of lysine must
attack to the aldehyde plane, with an optimal angle of 107°.68
The crystal structure of this protein is shown in Figure 1-18.
!

R132
Y134
2.98Å

2.34Å
2.91Å
2.34Å

R111

Figure I-19: The Crystal structure of wt CRABPII bound with retinoic acid with
highlighted interactions of the chromophore carboxylic acid and CRABPII residues (PDB
entry 2FR3)
a.

b.

107

!

Figure I-20 a. The detail of hydrogen bonds between CRABPII and all-transretinoic acid. b. Illustration of the Bürgi-Dunitz trajectory.

36

R111L mutation is introduced to remove the observed water molecules, since
this water molecule locks the aldehyde in an unfavored position. However no PSB
formation have been observed with R132K:R111L mutation at the physiological pH=7.3
due to the low pKa of the PSB of the retinal (<6.5). As it was mentioned above the pKa
of the bovine rhodopsin is higher than 16 and the crystal structure of this protein pointed
to that Glu113 as a counter anion that has an important role to increasing the pKa.
Inspection of the crystal structure of CRABPII protein shows that Leu121 is a favored
position for the introduction of counter anion (L121E mutation). Introducing this mutation
could increase the pKa of the PSB to higher than biological pKa. The final pKa of
R132K:R111L:L121E was 8.7.69,70
However mutagenesis studies and switching the polarity of the amino acid residue
(R59) which is located at the “mouth” of the protein binding pocket do not have any
effect on the wavelength tuning. The crystallographic data of rhodopsin showed that the
chromophore is deeply buried in the binding pocket, In CRABPII the ionone ring of the
chromophore sticks out from the binding pocket and is exposed to the bulk solution and
the effect of the polarity of the amino acids (specially in zone I) is buffered by the media
(table I-3). Since water in the bulk solution has high polarity, it can make retinal inert to
the changes of the polarity of the protein. To investigate weather exposure of the
chromophore to the solution neutralized the effect of the amino acid residues, two
different approaches have been used:
1) Using a shorter chromophore analogue which has the same structure but shorter,
so it can be embedded in the binding pocket of the CRABPII protein.

37

2) Using another protein mimic, which has deeper binding pocket and reengineer it
the same as CRABPII.
The first approach has been carried by Dr. Lee in Professor Borhan’s Lab and the
crystallographic data were collected by Dr. Jia in Professor Geiger’s lab.69,70

Table I-3: CRABPII mutants with all-trans-retinal and all-trans-C15
CRBAP II mutants

λmax with retinal

λmax with C15

R132K-R111L-L121E-R59

449

n.d.a

R132K-R111L-L121E-R59E

450

424

R132K-R111L-L121E-R59Q

444

413

R132K-R111L-L121E-R59L

443

391

R132K-R111L-L121E-R59W

442

404

a

due to the low pKa of PSB and therefore no PSB peak has been observed.

The comparison of the mutagenesis studies on both all-trans-retinal and all-transC15 is listed in Table I-3. As it is shown in the table by changing the polarity of the R59
position no spectral tuning has been observed when the protein is bound to retinal,
however with the C15 analogue mutation of the same position leads to modulation of
wavelength. This data proves that for tuning the wavelength the embedding binding
pocket is essential and the chromophore need to be deep in the binding pocket and not
exposed to the bulk solution. Based on these data the second generation of rhodopsin
mimics with deeper binding pockets were reengineered.
38

a.
b.
all-trans-retinal
c.

C15-analogue

Figure I-21: a. The crystal structure overlaid of R132K:R111L:L121E (KLE, PDB
entry 2G7B) bound with retinal, (blue) and structure of KLE-R59W (PDB entry
3F8A) bound to C15, with R59W position is highlighted. The overlaid structure
illustrates that C15 is fully embedded within the protein binding pocket, while
retinal (blue) is exposed b. Chemical structures of all-trans-retinal. c. all-trans C15
analogue.

I-3-4: The Second Generation Rhodopsin Mimic (Human Cellular Retinol Binding
Protein II)

Since the data from C15 analogue illustrate the fact the embedded and covered
binding pocket is an essential factor for a protein, which is used as mimic of rhodopsin,
we switched to another protein from the same family of proteins, Cellular retinol binding
protein (hCRBPII).71 This is also relatively small and very soluble protein with 133 amino
acid residues, which is responsible to carry all-trans-retinol inside the cell. The crystal
structure of this protein shows that the chromophore is not exposed to the bulk solution.
Since these two proteins (hCRBPII and CRABPII) have different functions in body
39

therefore the binding site of them are different. HCRBPII is binding to all-trans-retinal
and protecting that against oxidation and other environmental damages, therefore the
binding pocket is very preserved, while CRABPII is responsible for carrying the all-transretinoic acid and therefore that the biding pocket is more open and accessible. The
sequence identity between hCRBPII and CRABPII is only 35%, however they share the
same structure of the iLBPs, which is 10 β-stranded sheets with two short α-helixes that
cover the binding pocket while the ligand binds deeply in the binding pocket. The crystal
structure and function of CRBPII is explained in the next chapter.
The comparison of the CRABPII and hCRBPII shows that the ligand (all-trans-retinal) is
almost 5 Å deeper in the binding pocket. Figure I-20, shows the superimposed structure
of CRABPII and hCRBPII. As it is clear in this figure (b, c) the ligand is more covered in
CRBPII and not accessible to the bulk solution.

40

a.

b.

c.

Figure I-22: a. The superimposed crystals structure of hCRBPII bound to alltrans-retinal (blue) and CRABPII bound to all-trans-retinoic acid (red). b. The
space model of hCRBPII (the exposed part of the chromophore is blue). c. The
space model of hCRBPII (the exposed part of the chromophore is red).

41

REFERENCES

42

REFERENCES

1.

Shichida, Y.; Matsuyama, T., Evolution of opsins and phototransduction. Phil.
Trans. R. Soc. B, 2009, 364 (1531), 2881-2895.

2.

Shichi, H. Biochemistry of Vision, Academic Press, Inc., London, UK, 1983.

3.

Frank, R. N., Cavanagh, H. D. & Kenyon, K. R. Light-Stimulated
Photophosphorylation of Bovine Visual Pigments by AdenosineTransphosphate. Journal of Biochemical Chemistry, 1973, 248, 596-609.

4.

Padgham, C. A& Saunders, J. E. The Perception of Light and Colour, G. Bell
& Sons Ltd, London, UK, 1975.

5.

Molday, R.S. Photoreceptor membrane protein, Phototransduction and retinal
degenerative disease – The Friedenwald lecture. Investigative Ophthalmology
& Visual Science, 1998, 39, 2493-2513.

6.

Streyler, L. The Molecules of Visual Excitation. Scientific American, 1987,
257, 42-50.

7.

Ripp, H.& Weale, R. A. The Visual Photoreceptors, The Eye (Davson, H., Ed.)
2.6 vols, Academic Press, Inc., 1969, New York, NY.

8.

The Schanpf, J. Rhodopsin-Structure
Biochemistry, 1992, 4923-4931.

9.

Wald, G. Cartenoids and the visual cycle. Journal of General physiology,
1935, 19, 351-371.

10.

Wald, G. Pigments of the retinal I. The bull frog. Journal of General
Physiology, 1936, 19, 781-795.

11.

Stryer, L.; Hurley, J. B.; Fung, B. K. K., Transducin - An amplifier
protein in Vision. Trends Biochem. Sci. 1981, 6 (9), 245-247

12.

Gilman, A. G., G-Proteins - Transducers of receptor-generated
signals. Annu. Rev. Biochem. 1987, 56, 615-649.

13.

Hofmann, K. P.; Scheerer, P.; Hildebrand, P. W.; Choe, H. W.;
Park, J. H.; Heck, M.; Ernst, O. P., A G protein-coupled receptor at work:
43

and

Function

and

Genetics.

rhodopsin model. Trends Biochem. Sci. 2009, 34 (11), 540-552
14.

Pugh, E. N. & Lamb, T. D. Amplification and Kinetics of the Activation
Step in Phototransduction, Biochimica, Et, Biophysica Acta, 1993, 1141,
111-149

15.

Palczewski, K.; Kumasaka, T.; Hori, T.; Behnke, C. A.; Motoshima,
H.; Fox, B. A.; Le Trong, I.; Teller, D. C.; Okada, T.; Stenkamp, R. E.;
Yamamoto, M.; Miyano, M., Crystal structure of rhodopsin: a g protein-coupled
receptor. Science 2000, 289 (5480), 739-745.

16.

(A) Doukas, A. G.; Junnarkar, M. R.; Alfano, R. R.; Callender, R. H.;
Kakitani, T.; Honig, B., Fluorescence quantum yield of visual pigments evidence for subpicosecond isomerization rates. P. Natl. Acad. Sci-Biol.1984, 81
(15), 4790-4794; (B) Buchert, J.; Stefancic, V.; Doukas, A.G.; Alfano, R. R.;
Callender, R. H.; Pande, J.; Akita, H.; Baloghnair, V.; Nakanishi, K., Picosecond
absorption and fluorescence kinetic-studies of bovine rhodopsin with a fixed 11ene. Biophys. J. 1983, 41 (2), A338.

17.

a. Winter, N. S.; Bratt, J. M.; Banaszak, L. J., Crystal-structures of
holo and apo-cellular retinol-binding protein-II. J. Mol. Biol. 1993, 230 (4),
1247-1259; (b) Tarter, M.; Capaldi, S.; Carrizo, M. E.; Ambrosi, E.; Perduca,
Monaco, H. L., Crystal structure of human cellular retinol-binding protein II to
1.2 angstrom resolution. Proteins 2008, 70 (4), 1626-1630.

M.;

18.

Sakmar, T, P. Structure of rhodopsin and seven-helical receptors: the
same and not the same. Current Opinion in Cell Biology, 2002, 14, 189195.

19.

Cahbre, M. Breton, J., Michelvillaz, M., Saibil, H., Linear Dichroism
Studies in the Visible, UV and IR on Oriented Rod Suspension, Method
In Enzymology,1982, 81, 605-616.

20.

Nathane, J. Determinants of Visual Pigment Absorbance-Role of Charged
Amino-Acids In the Putative Transmembrane Segment, Biochemistry,
1990, 29, 937-942.

21.

Choe, H.-W.; Kim, Y. J.; Park, J. H.; Morizumi, T.; Pai, E. F.;
Krauss, N.; Hofmann, K. P.; Scheerer, P.; Ernst, O. P., Crystal structure of
metarhodopsin II. Nature, 2011, 471 (7340), 651-U137.

22.

Standfuss, J.; Edwards, P. C.; D'antona, A.; Fransen, M.; Xie, G.;
Oprian, D. D.; Schertler, G. F. X., The structural basis of agonist-induced
activation in constitutively active rhodopsin. Nature 2011, 471 (7340), 656- 660.
44

23.

Nakamichi, H.; Okada, T., Crystallographic analysis of primary
visual photochemistry. Angew. Chem. Int. Ed. Engl. 2006, 45 (26), 42704273.

24.

Leibovic, K. N. Vertebrate Photoreceptors. In Science of Vision, pp. 16-52.
Springer-Verlag, 1990, New York, NY.

25.

Stenkamp, R. E., Teller, D. C. & Palczewski, K. Rhodopsin: A structural
for primer for G-protein couple receptors, Archive Der Pharmaze, 2005,
338, 209-216.

26.

(A) Vuong, T. M.; Chabre, M.; Stryer, L., Millisecond activation of
transducin in the cyclic-nucleotide cascade of vision. Nature 1984, 311
(5987),659-661; (B) Guy, P. M.; Koland, J. G.; Cerione, R. A., Rhodopsinstimulated activation deactivation cycle of transducin - kinetics of the intrinsic
fluorescence response of the alpha-subunit. Biochemistry 1990, 29 (30), 69546964.

27.

Gilman, A. G., G-Proteins - Transducers of receptor-generated signals.
Annu. Rev. Biochem. 1987, 56, 615-649.

28.

Filipek, S., Stenkamp, R.E., Teller, D. C. & Palcezewki, K. G protein couple receptor rhodopsin: A prospectus. Annual Review of Physiology
2003, 65, 851-879.

29.

Stryer, L. Molecular-Basis of Visual Excitation, Cold Spring Harbor,
Symphosia on Quantitative Biology, 1988, 53, 283-294.

30.

Blatz, P. E.; Baumgart.N; Balasubr.V; Balasubr.P; Stedman, E.,
Wavelength regulation in visual pigment chromophore - large induced
bathocromic shifts in retinol and related polyenes. Photochem. Photobiol.
1971,14 (4), 531-549.

31.

Andersen, L. H.; Nielsen, I. B.; Kristensen, M. B.; El Ghazaly, M. O.
A.; Haacke, S.; Nielsen, M. B.; Petersen, M. A., Absorption of schiff-base
retinal chromophores in vacuo. J. Am. Chem. Soc. 2005, 127 (35),
12347-12350.

32.

33.

Rajput, J.; Rahbek, D. B.; Andersen, L. H.; Hirshfeld, A.; Sheves,
M.; Altoe, P.; Orlandi, G.; Garavelli, M., Probing and modeling the
absorption
of retinal protein chromophores in vacuo. Angew. Chem. Int. Ed. Engl. 2010, 49
(10), 1790-1793.
Motto, M. G.; Sheves, M.; Tsujimoto, K.; Baloghnair, V.; Nakanishi,
K., Opsin shifts in bovine rhodopsin and bacteriorhodopsin - comparison of
45

2external point-charge models. J. Am. Chem. Soc. 1980, 102 (27), 7947-7949.
34.

Erickson, J. O. & Blatz, P. E. N-Retinylidene-1-Amino-2-Propanol-a
Schiff Base Analog for Rhodopsin, Vision Research, 1968, 8, 1397-1357.

35.

Kochendoerfer, G. G., Lin, S. W., Sakmar, T, P. & Mathies, R. A. How
color visual pigments are tuned. Trends in Biochemical Science, 1999,
24, 300-305.

36.

Wald, G., Molecular Basis of Visual Excitation, Science, 1968, 162, 230239.

37.

Wald, G., Molecular Basis of Visual Excitation, Nature, 1968, 219, 800807.

38.

Neitz, M.; Neitz, J.; Jacobs, G. H., Spectral tuning of pigments underlying
red-green color-vision. Science, 1991, 252 (5008), 971-974.

39.

Irving, C. S.; Byers, G. W.; Leermake.Pa, Spectroscopic model for
visual pigments - influence of microenvironmental polarizability Biochemistry,
1970, 9 (4), 858-864.

40.

Albeck, A., Livnah, N., Gottlieb, H., & Sheves, M. C-13 NMR-Studies
of Model Compounds for Bacteriorhodopsin-Factors Affecting the Retinal
Chromophore Chemical-Shifts and Absorption Maximum. Journal of the
American Chemical Society, 1992, 114, 2400-2411.

41.

Okada, T., Sugihara, M., Bondar, A. N., Elstner, M., Entel, P., & Buss, V.
The retinal conformation and its environment in rhodopsin in light of a new
2.2 angstrom crystal structure. Journal of Molecular Biology, 2004, 342,
571-583.

42.

Sakmar, T. P. Rhodopsin: A prototypical G protein coupled receptor.
Progress in Nucleic Acid Research and Molecular Biology, 1998, Vol 59
1-34.

43.

Schertler, G. F. X & Hargrave, P. A, Projection Structure of Frog Rhodopsin
in 2 Crystal Forms, Proceeding of National Academy of Sciences of the
States of America, 1995, 92, 11578-11582.

44.

Unger, V. M., Hargrave, P. A., Baldwin, J. M. & Schertler, G. F. X,
Arrangement of rhodopsin transmembrane alpha-helices. Nature, 1997,
389, 203-206.

45.

Nathans, J.; Thomas, D.; Hogness, D. S., Molecular-genetics of
46

human color-vision - the genes encoding blue, green, and red pigments. Science,
1986, 232 (4747), 193-202.
46.

Neitz, M.; Neitz, J.; Jacobs, G. H., Spectral tuning of pigments
underlying red-green color-vision. Science,1991, 252 (5008), 971-974.

47.

Fasick, J. I., Lee, N. and Oprian, D. D. Spectral tuning in the human blue
cone pigment. Biochemistry, 1999, 38, 11593-11596.

48.

Asenjo, A. B.; Rim, J.; Oprian, D. D., Molecular determinants of
human red/green color discrimination. Neuron, 1994, 12 (5), 1131-1138.

49.

Lin, S. W.; Kochendoerfer, G. G.; Carroll, H. S.; Wang, D.; Mathies,
R. A.; Sakmar, T. P., Mechanisms of spectral tuning in blue cone visual
pigments- visible and raman spectroscopy of blue-shifted rhodopsin mutants. J.
Biol. Chem. 1998, 273 (38), 24583-24591.

50.

Zhukovsky, E. A.; Oprian, D. D., Effect of carboxylic-acid sidechains
on the absorption maximum of visual pigments. Science 1989, 246
(4932), 928-930.

51.

Nakayama, T. A.; Zhang, W. M.; Cowan, A.; Kung, M., Mutagenesis
studies of human red opsin: trp-281 is essential for proper folding and protein
retinal interactions. Biochemistry, 1998, 37 (50), 17487-17494.

52.

Sakmar, T. P.; Franke, R. R.; Khorana, H. G., The role of the
retinylidene schiff-base counterion in rhodopsin in determining wavelength
absorbency and schiff-base pKa. P. Natl. Acad. Sci. Usa, 1991, 88 (8), 30793083.

53.

Wang, Z,Y., Asenjo, A, B,& Oprian, D. D. Identification of the Cl- binding
site in the human red and green color vision pigments, Biochemistry, 1993,
32, 2125-2130.

54.

Mathies, R.; Stryer, L., Retinal has a highly dipolar vertically excited
singlet-state - implications for vision. P. Natl. Acad. Sci. Usa, 1976, 73 (7),
2169- 2173.

55.

Livnah, N.; Sheves, M., Model compounds can mimic spectroscopic
properties of bovine rhodopsin. J. Am. Chem. Soc. 1993, 115 (1), 351-353.

56.

Motto, M. G.; Sheves, M.; Tsujimoto, K.; Baloghnair, V.; Nakanishi,
K., Opsin shifts in bovine rhodopsin and bacteriorhodopsin - comparison of 2
external point-charge models. J. Am. Chem. Soc. 1980, 102 (27), 7947-7949.
47

57.

Honig, B.; Dinur, U.; Nakanishi, K.; Baloghnair, V.; Gawinowicz, M.
A.; Arnaboldi, M.; Motto, M. G., External point-charge model for wavelength
regulation in visual pigments. J. Am. Chem. Soc. 1979, 101 (23), 7084-7086.

58.

Beppu, Y.; Kakitani, T., Theoretical-study of color control
mechanism in retinal proteins .1. role of the tryptophan residue, tyrosine residue
and water molecule. Photochem. Photobiol. 1994, 59 (6), 660-669.
Wada, A.; Tsutsumi, M.; Inatomi, Y.; Imai, H.; Shichida, Y.; Ito, M.,
Retinoids and related compounds. part 26. synthesis of (11 z)-8,18- propanoand methano-retinals and conformational study of the rhodopsin chromophore.
J.Chem. Soc. Perk. T. 1 2001, (19), 2430-2439.

59.

60.

Tan, Q., Lou, J. H. Borhan, B., Karnaukhova, E., Berova, N. & Nakanishi,
K. Absolute sense of twist of rhe C12-C13 bond of the retinal chromophore
In bovine rhodopsin based on excitation CD spectra of dihydroretinal
Analogues, Angewandte Chemis-International Edition in English, 1997,
36, 2089-2093.

61. Cembran, A.; Gonzalez-Luque, R.; Altoe, P.; Merchan, M.;
Bernardi, F.; Olivucci, M.; Garavelli, M., Structure, spectroscopy, and spectral
tuning of the gas-phase retinal chromophore: The beta-ionone "handle" and alkyl
group effect. J. Phys. Chem. A , 2005, 109 (29), 6597-6605.
62.

Sakurai, M.; Sakata, K.; Saito, S.; Nakajima, S.; Inoue, Y., Decisive
role of electronic polarization of the protein environment in determining the
absorption maximum of halorhodopsin. J. Am. Chem. Soc. 2003, 125 (10),
3108-3112.

63. Sekharan, S.; Sugihara, M.; Buss, V., Origin of spectral tuning in
rhodopsin - it is not the binding pocket. Angew. Chem. Int. Ed. Engl. 2007, 46 (12), 269-271.
64.

Sekharan, S.; Weingart, O.; Buss, V., Ground and excited states of
retinal schiff base chromophores by multiconfigurational perturbation
theory.Biophys. J. 2006, 91 (1), L7-L9.

65.

Kloppmann, E.; Becker, T.; Ullmann, G. M., Electrostatic potential
at the retinal of three archaeal rhodopsins: implications for their different
absorption spectra. Proteins, 2005, 61 (4), 953-965.

66.

Lee, K. S. S. Ph.D Thesis, Michigan State University, East Lansing, MI,
2011.

67. (a) Vaezeslami, S.; Mathes, E.; Vasilelou, C.; Borhan, B.; Geiger, J.
H., The structure of apo-wild-type cellular retinoic acid binding protein II at 1.4Å
48

and its relationship to ligand binding and nuclear translocation. J. Mol. Biol.2006,
363 (3), 687-701; (b) Kleywegt, G. J.; Bergfors, T.; Senn, H.; Lemotte, P.; Gsell, B.;
Shudo, K.; Jones, T. A., Crystal-structures of cellular retinoic acidbinding protein-i
and protein-ii in complex with all-trans-retinoic acid and a synthetic retinoid.
Structure 1994, 2 (12), 1241-1258.
68. Burgi, H. B.; Dunitz, J. D., From crystal statics to chemicaldynamics.
Accounts Chem. Res. 1983, 16 (5), 153-161.
69.

Vasileiou, C.; Vaezeslami, S.; Crist, R. M.; Rabago-Smith, M.;
Geiger, J. H.; Borhan, B., Protein design: Reengineering Cellular Retinoic Acid
Binding Protein II into a rhodopsin protein mimic. J. Am. Chem. Soc. 2007, 129
(19), 6140-6148.

70. (a) Crist, R. M.; Vasileiou, C.; Rabago-Smith, M.; Geiger, J. H.;
Borhan, B., Engineering a rhodopsin protein mimic. J. Am. Chem. Soc. 2006,
128(14), 4522-4523; (b) Vasileiou, C.; Wang, W. J.; Jia, X. F.; Lee, K. S. S.;
Watson, C. T.; Geiger, J. H.; Borhan, B., Elucidating the exact role of engineered
CRABPII residues for the formation of a retinal protonated Schiff
base. Proteins, 2009, 77 (4), 812-822.
71. (a) Winter, N. S.; Bratt, J. M.; Banaszak, L. J., Crystal-structures of
holo and apo-cellular retinol-binding protein-II. J. Mol. Biol. 1993, 230 (4), 12471259; (b) Tarter, M.; Capaldi, S.; Carrizo, M. E.; Ambrosi, E.; Perduca, M.;
Monaco, H. L., Crystal structure of human cellular retinol-binding protein II to 1.2
angstrom resolution. Proteins,` 2008, 70 (4), 1626 1630.
72. Leibovic, K. N. Vertebrate Rhotoreceptors in Science of Vision, pp. 16-52,
1990. Spring-Verlad, New York, NY.

49

CHAPTER II
The Crystal Structure of Holo-Wild-Type human Cellular Retinol Binding protein II
(hCRBPII) bound with Retinol and Retinal

II-1: Structure Determination from X-ray Diffraction
One single crystal is made of huge number of molecules in the highly ordered
and repetitive pattern. When a crystal is exposed to X-ray beam the scattered waves
from the imaginary plane of the crystals will only take place when the path difference
between the rays are equal to the integral number of wavelength. This phenomena is
known as Bragg’s law:

2d Sinθ= nλ

(2.1)

When the d is the distance between imaginary reflected planes and θ is the angle
between the incident and reflected beam (λ Is the wavelength of the incident beam).
The process of the reflection is incident and reflection of the rays than make an angle θ
with the fixed crystal plane. Since the electrons are the scatters in therefore each atom
has the different effectiveness as the scatters. The X-ray diffraction is widely used for
biological macromolecules such as proteins. The X-ray diffraction patterns reflect all the
spatial distribution of the electrons of all the atoms in the crystal. The 3-D picture of the
electron density of a macromolecule can be derived from the X-ray diffraction patterns
with using the Fourier transformation. The scattering intensity of the individual atoms
and the phase of the wave from each scattering source is known as the individual
structure factor (ƒj) and the structure factor of each plane (hkl) is the summation of the
50

individual structure factors for individual atoms, ƒj, times a phase factor for each atom
α(hkl). The structure factor can be represented by this equation:

F(hkl)=Σ ƒ je2 i (hxj+kyj+lzj)= F(hkl)ei
π

α

(hkl)

(2.2)

In this equation F(hkl) is wave function with frequency while α(hkl) is phase
angles of the reflections. The frequency of F(hkl) is identical to the X-ray source and the
amplitude is proportional to the square root of the intensity of the reflection, which can
be measured from the X-ray diffraction data. In data collection the x-ray detector can
only record the intensities and not the phase and therefore phase is lost and only the
intensity of the diffracted beam is recorded. The intensity of each spot can is given by:

I(hkl) = [ F(hkl)]2

(2.3)

The Fourier transform takes the structure factor, which is function of the electron density
ρ(r) and invert the functional dependence and therefore the electron density is
expressed as the function of the structure factor. The atomic coordination can be
extracted from it.
ρ(r) = ∫ diffraction space F(s)e -2 ir.s dνs
π

(2.4)

dνs is a small unit of volume in the diffraction space. The integration can be replaced by
summation since F(s) is not continues and is non-zero only at the reciprocal lattice
points. Therefore:
51

ρ(xyx) = 1/V ∑ ∑ ∑ F(hkl)e

2πi (hx+ky+lz)

(2.5)

h k l

= 1/V ∑ ∑ ∑ F(hkl)e

iα(hkl)

h k l

e2 i (hx+ky+lz)
π

In this equation ρ(xyx) is electron density at any point (x,y,z) and F (hkl) is the amplitude
and proportional to the square root of the measured intensity of each reflection. As it
was mentioned above the problem is missing the phase. There are few methods to
solve the phase problem. One of them is known, as Molecular replacement is a known
structure is used to find the phase of the unknown structure.
Molecular Replacement: One method to solve the phase problem is molecular
replacement. In this method one know structure is used to calculate the electron density
map of a biological entity that X-ray diffraction data has been collected. The two
structures need to share at least 25% sequence Identity, therefore the structure can be
used as the initial search model. We could easily used molecular replacement to
determine the initial phase of each of the structures of CRBP II mutant complexes with
all-trans-retinol. For finding the correct orientation of the protein structure the two
following steps are used by molecular replacement:
Rotation Function: In rotation function, the Patterson map of the unknown
structure is compared to the Patterson map of the initial searching model and in the
different orientations.

52

The Patterson function is calculated form the amplitude using the formula:
P (u,v,w) = Σ |Fhkl|2 e -2 (hu+kv+lw)
π

(2.6)

The Patterson function is essentially is Fourier transform of the intensities. Evaluation of
the rotation function is only necessary over the asymmetric unit in rotation space. The
rotation function is usually calculated as the function of Eulerian angles, α, β and γ.
Translation Function: The translation function places the oriented unknown
biological molecule in the defined asymmetric unit to the best fit. Crowther and Blow
defined the standard Patterson map based on the translation function:

T (t) = ∫ P obs (u) . P calc (u,t ) du

(2.7)

In translation function the initial model moves and the new Patterson map is calculated.
The calculated Patterson map compares to the unknown structure.
Therefore this is basically defined the similarity of observed and calculated Patterson
map of the observed and calculated Pattersons.
Structure Refinement: After the phase problem has been solved and the solution
if found. The atomic location of each atom was defined in the electron density, the high
speed computer is need to be used for adjust some parameters to minimized and
reduced some parameter such as Σ (Fobs – Fcal)2. Each amino acid residue of the protein
is fitted to the density until the bias introduces by the initial model is reduces. A correct
structure should have a reasonable Rfactor with absolutely no major unexplained
discrepancies. The satisfactory Rfactor value depends on the resolution of the structure.
53

Usually at the higher resolution the lower Rfactor can be gained. Before starting the
refinement 10% of the data randomly is separated and the refinement is performed on
the 90% of the data. Based on the separated data another Rfactor, which is called Rfree, is
calculated. As the Rfactor by continuing the refinement should decrease the Rfree should
concurrently decrease. Usually the Rfactor is lower than R-free however there is a
limitation between their difference and this value should not be higher than 6-7 %. The
value of Rfactor which is also called Rwork are the agreement between the values of the
observed structure factor and calculated by the following equations:
∑ || Fobs | - | Fcal||
Rwork =

(2.8)

∑ || Fobs |

For all the structures with the resolution range between 1.2-1.8 Å the Rfactor and Rwork
bellow 20% is acceptable but bellow 18% is more desirable.

II-2: The Crystal Structure of wt hCRBPII Bound with Retinol and Retinal
II-2-1: Introduction

Vitamin A or retinol is necessary for vision, cell growth, reproduction, metabolism,
function and function. Because of low solubility in aqueous solution and instability of this
compound, therefore retinol needs a carrier and soluble protein for transporting.
Different proteins in different tissues were identified and characterized to play important
role in transport, storage and metabolism of vitamin A. They are Retinol Binding
54

Proteins (RBP), Cellular Retinoic Acid Binding Proteins (CRABP), Cellular Retinol
Binding Protein (CRBP), Cellular Retinal Binding Protein (CRALBP), Interphotoreceptor
Retinol Binding Protein (IRBP).23 In all of these proteins the hydrophobic ligand and
protein have non covalent interactions.24
CRBPs are relatively small cytosolic protein with 133 amino acid that play
important role in the metabolism of vitamin A in the cells. Four isoforms of human
Cellular Retinol Binding Protein (CRBP) have been identified: I, II, III and IV.

1,2

They

belong to the superfamily of intracellular Lipid Binding Proteins (iLBP). This family also
includes another two group members, which are besides Cellular Retinoic Acid Binding
Proteins (CRABP), and Fatty Acids Binding Proteins (FABP).2 However in CRBPs
subfamily only CRBPI and II are in the process of intracellular retinol. Despite the fact
that these proteins have low sequence identity, surprisingly the structures of them are
highly homologues.

55

!

b
a
1

c
d

j
2
i

h

g
e
f

Figure II-1: The Crystal structure of overlay wt hCRBPII with all-trans-retinal
(blue) and all-trans-retinol (green).

They share 10 β-sheet domains (a-j) and two short α-helices (1, 2) (Figure II-2). These
two iso-forms of CRBP share 56% sequence identity3. CRBPI and II bind both -transretinol and 13-cis-retinol, while neither bind 9-cis-retinol4. These two proteins are
localized in different organs. CRBPI is found predominantly in the liver and all kidney
while CRBPII is more prevalent in the small intestine5,6,7. It was reported that CRBPII
regulates retinoid metabolism in intestine and facilitation of reduction of retinal to retinol
and also subsequent of esterification of retinal to retinyl ester.25

56

The different binding constant of the two CRBPs beside the different distribution of them
in human body make them specific for the different propose.

Figure II-2: The crystal structure overlay of hCRBPII (salmon red), solution
structure of human CRBPI (magenta), 1KGL, crystal structure of hCRBPIII (red)
1GGL, crystal structure of hCRBPIV (cyan) 1LPJ

The first crystal structures of rat apo and all-trans-retinol bound CRBPII were
published in 19928. Comparison of the two structures showed that no significant
differences were observed upon ligand binding. (The crystal structures of human
CRBPII and zebra fish CRBP bound with all-trans-retinol at high resolution were
published in 2007 and 2002 respectively.9,10 In all three structures the main interactions
with the ligand is with Gln108 and Lys40. These two residues is responsible for
hydrogen bond in CRBPI and II while in CRBPIII and IV, His108 makes this interaction.
57

CRBPIII and CRBPIV have weaker binding constant with all-trans-retinol compare to the
mammalian CRBPs (I and II).2

3.70Å
3.88Å

Phe16
Lue77

!
Figure II-3: The crystal structure CRBP with retinal and interaction with Phe16
and Leu77

By comparison of the binding constant between CRBPI, II with III and IV, it can
be concluded since Gln108 is able to make two strong hydrogen bonds with retinol
Therefore the dissociation constant of CRBPI and II for retinol is higher than CRBPIII
and IV because the later proteins both have Histidine instead of Gln at the 108 position.
Both of these residues (Gln108 and Lys40) make hydrogen bonds with the hydroxyl
group of the ligand. All CRBP family members consist of 10 anti parallel β-strands,
which together form a beta barrel. The hydrophobic ligand is deeply buried inside the
binding pocket formed by this barrel. Two short α helixes cover the binding pocket like a
lid, isolating the ligand from bulk solution.

58

Figure II-4: The crystal structure overlaid of hCRBPII (blue) and CRABPII (red).
The chromophore is highlighted in these two structures.

In addition to these two helices, the binding pocket opening is further covered by
the side-chain of Arg58, which is located on the loop connecting two β-strands (C and
D) and is held in position by a hydrogen bond between its guanadinium group and the
carbonyl group of Ser77. The ligand is stabilized within this binding pocket, allowing it to
preserve its yellow color at room temperature and in light for more than three days,
while retinol solution in ethanol lost ¼ of the activity when exposed to the air.26 CRBP II
binds both all-trans-retinol and all-trans-retinal with similar affinity (dissociation constant
(Kd) of CRBPII with all-trans-retinal is 23±10 nM, while with retinol it is 10 nM)11,12 We
have determined the first crystal structure of wt hCRBPII with retinal at 1.6Å.

59

The complex crystallizes in the P1 space group with four molecules per
asymmetric unit. Electron density corresponding to the bound ligand was very clear in
three of the four chains (A, B and C). All of the crystallographic data is listed in table II1. The position of retinal is almost superimposable with retinol. In both ligands the βionone ring adopts the 6-s-trans conformation. Two hydrophobic residues, Phe16 and
Leu77, surround the ionone ring of the ligand in hCRBPII. This position of these
residues lock the chromophore in its position restricting rotation of the ionone ring about
the C6-C7 single bond, locking it in the 6-s-trans conformation. The distances between
the C5 methyl group of retinal and Leu77 and Phe16 are 3.70Å and 3.83Å respectively
(Figure II-3). It is clear that rotation of the ionone ring would lead to steric clashes
between the gem-dimethyl group and these two residues. Sequence alignment of
hCRBP II with the other human CRBPs, rat CRBPI13 and II, and zebrafish CRBP10
shows that the ionone ring is surrounded by bulky hydrophobic residues in all (Figure II5). Despite CRBPs the crystal structure of the other family member of iLBPs In contrast
CRABPII has a more open binding pocket, allowing its ligand (all-trans-retinoic acid) to
adopt a 6-s-cis conformation14 which is 0.6 Kcal/mol more stable than the trans
conformation15.

60

!
CRBPIV
CRBPI
CRBPII
CRBPIII

!

Zebrafish
rat CRBPI
rat CRBPII

CRBPIV
CRBPI
CRBPII
CRBPIII
Zebrafish
rat CRBPI

!!

rat CRBPII

Figure II-5: The multiple sequence alignment between human CRBPs, rat CRBPI
and II and zebrafish CRBP

II-2-2: Crystal Structure of Holo wt human Cellular Retinol Binding Protein II
(hCRBPII) Bound with Retinol and Retinal

The overall structures of hCRBPII bound to all-trans-retinal and all-trans-retinol
no significant differences (Figure II-4). However inspection of the details of binding of
all-trans-retinol and all-trans-retinal show they bind differently in the binding pocket. The
most important interaction in both structures is between the ligand’s oxygen and the
Gln108 side chain inside the binding pocket. The side chain of Gln108 is fixed in
position by a small network of water molecules and residues, beginning with Thr1,
continuing to Gln4, and concluding with interactions to Gln108 (Figure II-6). This water
network is observed only in CRBPII, not in zebra fish CRBP, rat CRBPII and human
61

CRBPI, III10 and IV2. This water molecule makes hydrogen bonds with the main chain
carbonyls of Thr51 and Asp91. This water buttresses the orientation of Gln4 in human
CRBPII (Figure II-8). Since water makes hydrogen bonds with two carbonyl groups of
the main chain, only a lone pair on the oxygen of water is left to make the hydrogen
bond to the amide NH2 of Gln4, therefore defining its orientation. Therefore it is the
carbonyl group of Gln4 that makes a hydrogen bond, necessarily with the amide of
Gln108, defining its orientation as well. This water molecule is only seen in the all-transretinal-bound structure and is not observed in any of the CRBP members bound to
retinol. As shown in Figure II-6 the carbonyl group of retinal does not make a hydrogen
bond with Gln108. It is obvious that carbonyl group of retinal do not make any hydrogen
bond with carbonyl of Gln108 and therefore only NH2 group of the side chain is available
to make a hydrogen bond, however the distance between NH2 of Gln108 and the
carbonyl group of retinal is about 3.6Å, which is longer than a moderate hydrogen bond.
However, a water mediated interaction between Gln108 and the retinal carbonyl is
formed. The hydrogen bond between this water and the retinal carbonyl is 2.49Å, which
is an ideal distance for a strong low-barrier hydrogen bond. This water makes an
additional hydrogen bond with the ε-amino group of Lys40, which is fixed in space by
two hydrogen bonds. An overlay of the crystal structure of retinal-bound hCRBPII and
the solution structure of hCRBPI (PDB entry 1KGL)16 illustrate that in hCRBPII the
ligand is more than 1Å deeper in the binding pocket. The same translation of the ligand
is observed in rat CRBPI (PDB entry 1CRB)13 and CRBPII (PDB entry 1OPB).17 The
identity of the amino acid at position 51 appears to be the deciding factor for ligand
position. The sequence alignment of CRBPI and II of rat and shows that position 51 in
62

the binding pocket of CRBPI (rat and human) is Isoleucine while this position is
Threonine in CRBPII. The larger steric bulk of Isoleucine prevents deeper penetration of
the ligand in the binding pocket. CRBPI has weaker binding affinity with retinal.
Inspection of the structures of human and rat CRBPI (PDB entry 1CRB)13 shows that
retinal can not bind in CRBPI in the same way as hCRBPII, possibly explaining the
weaker binding affinity of CRBPI.
The side chain of Isoleucine also disrupts the water network seen in CRBPII. The
CRBPII T51I mutant was produced and its binding affinity (Kd) for retinal measured. The
dissociation constant of this mutant for all-trans-retinal was two fold weaker (45±10 nM)
relative to wt CRBPII, but was still much smaller than all-trans-retinal binding to CRBPI,

all-trans-retinal

2.43Å
Thr51

3.89Å

2.92Å
2.52Å
2.27Å

2.92Å

Gln108

Gln38

Lys40

Figure II-6: The details of the hydrogen bonds in hCRBPII bound to alltrans-retinal.
indicating that position 51 is not solely responsible for the differences in affinity for
retinal in the two proteins. Unfortunately attempts at crystallization of CRBPII T51I
bound all-trans-retinal was not successful.
63

1.2Å
Gln108
2.98Å

2.74Å

all-trans-retinol

Thr51
Ilu51

Figure II-7: The overlaid structure of CRBPII with all-trans-retinal (green)
and solution structure of CRBPI (blue, PDB entry: 1KGL)16

The crystal structure of hCRBPII (PDB entry 2RCT)9 at 1.2Å shows that the
hydroxyl group of retinol adopts two different conformations (PDB ID: 2RCT), one of
which having a torsion angle about the C13-C14 bond of 113 degree which is far from
the 180 degree angle expected for all-trans-retinol. Surprisingly, and consistent with the
previous observation, the C13 of retinol does not indicate the expected planar sp2
configuration but instead appears to be tetrahedral, consistent with sp3 hybridization.
The two different conformations for the hydroxyl group were also observed in Zebrafish
CRBP (PDB entry 1KQW)10 as well. In our high-resolution crystal structure of all-transretinol-bound CRBPII, we observed a single conformation, which is equivalent to the
more highly occupied position found in the previous structure of the complex (Figure II9).

64

!
Asp91

all-trans-retinal

2.52Å

2.43Å
2.49Å

2.79Å

Gln4
2.93Å

2.76Å

Gln108
2.74Å
Thr1

Lys40

Figure II-8: The water network and the binding site in hCRBPII with all-trans-retinal

2.77Å
all-trans-retinal

Gln108

3.12Å
Lys40

Figure II-9: The crystal structure of wt hCRBPII with all-trans-retinal

It is clear in our structure that the double bond between C13-C14 is has also
been reduced as both centers indicate sp3 hybridization and the position of C15 of
retinol also shows that the torsion angle about the C13-C14 bond is about 50 degrees,
65

not the 180 degrees expected for all-trans-retinol. As the figure illustrates the water
molecule that hydrogen bonds with the carbonyl, and bridges to Gln108 and Lys40 in
our all-trans-retinal structure has been replaced by the hydroxyl group of retinol, which
now makes direct interaction with Gln108 and Lys40. In the crystal structure wt hCRBPII
with retinal this dihedral angel is only 3.95°, which is consistent with the expected sp2
hybridization. It is obvious that the C13-C14 double bond in retinol has been reduced to
a single bond. At first we suspected an impurity in the retinol as a result of retinol
degradation over time. Therefore we produced new crystals with fresh retinol. The HNMR of retinol was taken just prior to crystallization and confirms the purity of this
compound. Also the absorbance of retinol solution in ethanol was measured to be 325
nm as previously reported.22 The crystals were grown and flash frozen within three
days. The structure was redetermined, this time in a P1 space group with eightmolecules per asymmetric unit at 1.7Å resolution. We observed the same retinol
conformation as previously described in all eight molecules in the asymmetric unit. Wt
hCRBP or not, T51V mutant was generated. The crystal structure of this mutant in P1
space group with two molecules per asymmetric unit at 1.9Å resolution was determined.
Further the structure of the CRBPII T51V mutant was also determined and the retinol
electron density in this structure also shows the same puzzling distortion (Figure II-10).
As mentioned above this distorted structure for retinol was observed by at least two
different research groups for both zebra fish and human10,17. Unfortunately no crystal
structure of CRBPI in complex with all-trans-retinol is available, though the solution
structure of this complex shows a single conformation for retinol consistent with all-trans
retinol.
66

A possible solution to this puzzle came during data collection of these complexes
at the synchrotron. It was observed that the yellow color of the crystal changed to dark
orange after the first few frames of data collection and this color is maintained to the
end of data collection. It is therefore possible that some rearrangement occurred during
the synchrotron data collection leading to the ligand structure observed. All of the
structures that display the severely distorted retinol conformation were collected at
synchrotron sources.
In contrast, the first reported crystal structure of rat CRPPII bound with all-trans-retinol
(at 1.9Å resolution) shows a single conformation for retinol consistent with all trans13.
Rigaku RU-200. The most likely explanation for the anomalous retinol structure is that
!
T51V

2.73Å

Lys40

2.71Å

Gln108

Figure II-10: The crystal structure of single mutant T51V with all-trans-retinol

The diffraction data for this structure was collected using CuKα X-rays at home source.
This data indicates that retinol by was degraded during synchrotron data collection
67

Based on the rat CRBPII crystal structure we can conclude that real structure of
hCRBPII is similar to the rat CRBPII and hydroxyl group makes two hydrogen bonds
with both the carbonyl and amino groups of Gln108.

II-3: Experimental
II-3-1: Material and Method

The hCRBPII gene was purchased from ATCC Company and cloned in pET17b
vector, with using the NdeI and XhoI restriction enzyme. 10 µl of the resulting plasmid
were transformed in 100 µL of XL-1 Blue super (Novagen®) competent cells. The cells
were incubated for 30 min in the ice and heat shocked at 42 °C for 45 seconds, then 1
mL of Luria-Bertani broth (LB) was added and the cells were incubated at 37°C for an
hour. The result mixture were spread on the LB agar plate with (ampicillin 100 µg/mL
and incubated at 37°C for 16 hour. A single colony was picked from the plate and
inoculated in 5 mL of LB media contains 100 µg/mL ampicillin. The cell culture was
grown over night (12-16 hours) at 37°C, then the media were centrifuged at 15000 rpm
for 1 min. DNA extraction and isolation from the cell pellet was done by using the
standard procedure of Promega Wizard. Plus SV Miniprep (A1330) DNA purification kit.
The target gene was transformed into BL21(DE3) pLysS (Invitrogen™) E.coli competent
cells for protein expression. A single colony was picked and inoculated in 5 mL of LB,
containing 100 µg/mL ampicillin antibiotics at 37°C for 12-16 hours. The media
transferred to 50 mL of LB with the same antibiotic at 37°C overnight. This media
68

transferred into 1L of LB media with ampicillin (100 mg/L) and incubated at 37°C until
OD600 reached 0.5-1.0. The protein expression was induced with 1 mM isopropyl β-D-1
thiogalactopyranoside (IPTG) (IPTG, purchased from Gold Biotechnology) overnight at
22°C. The cells were harvested by centrifugation at 5000 rpm for 20 min. The harvest
cells were resuspended again with Tris buffer (10 mM Tris, 10 mM NaCl pH = 8.0,
50mL). The suspended cells were lysed by sonication and the lysed cells were
centrifuged at 4°C (14,000 rpm, 20 min).
The supernatant were loaded on Fast Q anion exchange resin, which was
equilibrated with (10 mM Tris, 10 mM NaCl pH = 8.0). The after the binding, the resin
was washed three times with 50 mL of (10 mM Tris, 10 mM NaCl pH = 8.0). The protein
was eluted with 40 mL of (10 mM Tris, 100 mM NaCl pH = 8.0).
The eluted protein was desalted by using Centriprep® centrifugal filter at 2500 rpm. The
protein was applied on BioLogic DuoFlow system loaded with 15Q anion exchange
resin for the second step of purification. The purity of the protein was determined with
SDS-PAGE.
Kd Determination via Fluorescence Titration: All-trans-retinol and all-trans-retinal
were purchased from Sigma. The dissociation constant Kd was determined by
fluorescence titration with previously reported method18

69

II-3-2: Crystallization and Structure Determination

The pure protein was concentrated down to 5-10 mg/mL. The complexes of
protein with all-trans-retinal and all-trans-retinol were prepared by adding 3-4 equiv of
retinal and retinol solution (30 mM retinal and retinol in ethanol). The final concentration
of ethanol in protein solution should be lower than 10% V/V. The mixture of protein and
ligands were incubated at room temperature in Black LiteSafe™ Microcentrifuge Tubes
for an hour. The crystal of the complex of protein and ligands were prepared at room
temperature by hanging drop vapor diffusion method in Limbro plates and plates were
wrapped in aluminum foil to prevent light-initiated degradation of the retinal and retinol.
The best crystals were grown by using Hampton Research Screens contains (30-35%
PEG 4000 (Sigma-Aldrich), 0.1 M sodium acetate (Columbus Chemical Industry) pH
4.6-4.8 and 0.1 M ammonium acetate (J. T. Baker) with 1 µL of hCRBPII-ligand
complex. The crystals appeared after 3 days and reached to the maximum size in one
week. For preventing degradation of ligand at room temprature the crystals were flash
frozen by liquid nitrogen and using cryoprotectant solution (30% PEG 4000, 0.1 M
sodium acetate pH 4.6-4.8, 0.1 M ammonium acetate, 10-20% glycerol) after three
days.

70

II-3-3: Data Collection, Refinement and Solution

All of the diffraction data were collected at beamline 21-ID-D, LS-CAT (Argonne
National Laboratory, Advanced Photon Source, Chicago, IL) using a MAR300 detector
and 1.00Å wavelength radiation at 100K. The diffraction data were indexed using
HKL200019 software package. The initial phase were solved by MOLREP using human
CRBPII Protein data bank, accession code 2RCQ as model in the CCP4 program
package and REFMAC5 in the CCP4 suite produced the initial electron density
maps20,21.

The entire rebuilding, placement of ordered water molecules was done

manually using COOT. The chromophore was created using jligand and manually fitted
in the electron density at the end of the refinement. All refinements were carried out
using REFMAC5 program.

71

Table II-1: X-ray crystallography data and refinement statistics of wt hCRBPII
Holo-WT-CRBPIIretinol
Space group

P21

a(Å)

34.65

b(Å)

75.14

c (Å)

90.00

β(° )
δ(° )
Molecule per
asymmetric unit

100.78
90.00

Total reflection

306503
87302

Rmerge (%)

36.35

29.37

54.94

35.96

128.16

63.82

92.19

90.73

92.74

94.19

103.21

112.81
2

2

Unique reflection
Average I/σ
Resolution (Å) (last
shell)

T51V-retinol

54.65

α(° )

Completeness (%)

Holo-WT-CRBPIIretinol
P1

20.19-2.04
50.00-1.19(1.211.19)
a
9.1(43.4)

Wavelength (Å)

0.97872

Rwork/Rfree (%)

21.53-18.50

36507

89048

18179

35.44
54.89
68.71
107.75
97.66
103.08
4

a

99.8(84.0)

P1

8
181745

92.0(93.1)

a

Holo-WT-CRBPIIretinal

33.40(1,98)

a

241247
a

86.3(95.9)

34.24-6.48

20.00-1.83(1.831.80)
a
5.5(46.7)

50.00-1.90(1.931.91)
3.7(27.4)

1.0781

0.97872

35.82/27.47
RMSD from ideal value

29.57/21.77

76879
a

91.5-(83.9)

51.85-4.13
50.00-1.50(1.531.50)
a

3.2(32.1)
0.97856

25.65/20.40

Bond length (Å)

Bond angle (°)
Number of water
molecule

0.0297

0.0230

0.0205

0.0240

2.3815

2.0600

1.8208

2.200

530

790

150

320

72

REFERENCES

73

REFERENCES

1. Folli, C., Calderone, V., Ottonello, S., Bolchi, A., Zanotti, G., Stoppini, M., and
Berni, R. 2001, Proc. Natl. Acad. Sci. U. S. A. 98, 3710–3715
2. Folli, C., Calderone, V., Ottonello, S., Bolchi, A., Zanotti, G., Stoppini, M.,
and Berni, R. 2002, J.Biol.Chem. 277, 41970-41977
3. Li, E., Demmer, L. A., Sweetzer, D. A., Ong, D. E., and Gordon,
J. I. 1986, Proc. Natl. Acad. Sci. U. S. A. 83,5779-5783
4. MacDonald, P.N., Ong, D.E. 1987, J.Biol.Chem, 262, 10550-10556
5. Crow, J. A., and Ong, D. E. 1985, Proc. Natl. Acad. Sci. U. S. A.
82, 4707-4711
6. Ong, D. E. 1984, J. Biol. Chem. 259, 1476-1482
7. Chytil, F., and Ong, D. E. 1984,The Retinoids (Sporn, M. B., Roberts, A. B., and
Goodman, D. S., eds) Vol. 2, pp. 89-123, Academic Press, Orlando, FL
8. Inagami S, Ong DE.,1992, J Nutr 122(3):450-6
9. Tarter, M., Capaldi, S., Carrizo, M.E., Ambrosi, E., Perduca, M., Monaco, H.L.
2007, Proteins 70: 1626-1630
10. Calderone V, Folli C, Marchesani A, Berni R, Zanotti G . 2002, J. Mol.
Biol.321, 527-535
11. Li, E., Qian, S. J., Winter, N. S., d’Avignon, A., Levin, M. S. & Gordon, J.I.
1991 J. Biol. Chem. 266, 3622–3629
12. Malpeli, G., Stoppini, M., Zapponi, M. C., Folli, C. & Berni, R. 1995, Eur.
J. Biochem. 229, 486–493.
13. Cowan, S. W., Newcomer, M. E. & Jones, T. A. 1993. J. Mol. Biol. 230,
1225-1246
14. Kleywegt, G.J., Bergfors, T.,Senn, H., Le Motte, P., Gsell,B., Shudo, K,
Jones, T.A. 1994 Structure 2: 1241-1258
15. Rajput, J.; Rahbek, D. B.; Andersen, L. H.; Hirshfeld, A.; Sheves,
74

M.; Altoe, P.; Orlandi, G.; Garavelli, M., Probing and modeling the absorption
retinal protein chromophores in vacuo. Angew. Chem. Int. Ed. Engl.
2010, 4910, 1790-1793
16. Franzoni, L., Lucke, C., Perez, C., Cavazzini, D., Rademacher,
M., Ludwig, C., Spisni, A., Rossi, G.L., Ruterjans, H. 2002.
J.Biol.Chem. 277: 21983-21997
17. Winter, N.S., Bratt, J.M., Banaszak, L.J. 1993 J.Mol.Biol. 230, 1247-1259
18. Vasileiou, C., PhD. Dissertation, Michigan State University, Eas Lansing,
2005
19. Z. Otwinowski, W. Minor, Macromolecular Crystallography, Pt A 276, 307
1997
20. S. Bailey, Acta Crystallographica Section D-Biological Crystallography 50,
760,1994
21. A. Vagin, A. Teplyakov, J. Appl. Crystallogr. 30, 1022,1997
22. Dawson, R.M.C. et al., eds. Data for Biochem. Res. 3rd edition, Clarendon
Press, Oxford, 136,1987
23. David E. Ong and David L. Page, J. Lipid Res. 1987. 28: 739- 745.
24. Veerkamp, J. H., Peeters, R. A., and Maatman, R. G. H. J. ,1991, Biochim.
Biophys. Acta 1081, 1–24
25. Vogel.s, Mendelsohn. C, Mertz.J, Piantedosi R, Waldburgeri C,
Gottesmani. M and Blaner.W, Journal of Biological Chemistry, 2001, 276,
1353-1360
26. Sigma-Aldrich Production & Quality Control

75

CHAPTER III
The Structural Insight of hCRBPII as Rhodopsin Mimic
III-1: The Design of a Rhodopsin Mimic Based on a Heterogonous Protein System

As was mentioned in Chapter I, in this study human Cellular Retinol Binding
Protein II (hCRBPII) was selected for use as a rhodopsin mimic to study wavelength
regulation. Since in rhodopsin the chromophore, which is 11-cis-retinal, has a covalent
bond to the protein via an active lysine in the binding pocket and makes a protonated
Schiff base therefore for making the mimic of this protein generation the active lysine is
necessary. Inspection of the crystal structure of wt hCRBPII shows that Gln108 makes
a hydrogen bond with the hydroxyl group of all-trans-retinal. This residue seemed to be
in the best position for mutation to make this active lysine so that it would have a
favorable Bürgi-Dunitz trajectory to attack the aldehyde. Therefore Dr. Wenjing Wang in
Professor Borhan’s group generated the first mutant (Q108K). However the UV-vis
spectrum of the Q108K mutant incubated with all-trans-retinal shows a very small peak
at ~506 nm and the major peak was observed at 356 nm. This data shows that the
formed Schiff base (SB) was not protonated in this mutant at a physiological pH.
Inspection of the binding pocket of wt hCRBPII by Dr. Wenjing Wang illustrated that
positively charged amino acid Lys40 is close to the PSB and may decrease the pKa of
the Q108K mutant, perturb the pKa of the retinal-PSB and prevent formation of the PSB
due to charge repulsion. To improve the pKa of the PSB two different strategies were
tried: 1- Introduction of a counter anion or a negatively charged residue close to the SB
to stabilized the positive charge. This strategy was successfully applied on the first
76

generation of rhodopsin mimics (CRABPII). 2- Removing the positive charge near the
PSB by mutation of Lys40 to a hydrophobic residue like leucine. The two mutants were
generated by Dr. Wang. Both of these mutants have higher pKa than the physiological
pKa (Table III-1). Q108K:K40L not only has a reasonable pKa but also is more red
shifted compared to Q108K:T51D and therefore the double mutant Q108K:K40L was
chosen as the platform for all further mutations.

all-trans-retinal

3.613.61
Å Å

Gln108
2.87 Å

Å
2.872.51
Å
2.72 Å
2.51 Å

Thr51

Lys40
Gln108
Lys40

Figure III-1: The crystal structure of wt hCRBPII bound to all-trans-retinal with
the highlighted amino acid residues that are important for generation of the
platform of all further mutations.

77

Table III-1: The generation of high pKa mutants
Mutants

λmax (nm)

pKa

Q108K

506

5.2

Q108K:K40L

508

7.9

Q108K:T51D

474

9.2

III-1-1: The Crystal Structure of Q108K:K40L (KL) Bound to Retinal
The crystal structure of KL was solved by using the Molecular Replacement (MR)
method and the structure wild type hCRBPII was used as the initial model. The structure
was refined at 1.7Å in P1 space group with two molecules per asymmetric unit and with
crystallographic Rfactor = 21.22% and Rfree = 26.72%. The density of the ligand was clear
in one chain. The overlaid structure of this mutant with wt hCRBPII shows that the plane
of the polyene chain of the chromophore rotated in hCRBPII mutants relative to the wt
hCRBPII retinal and retinol-bound structures. However the crystal structures show that
the ionone ring of retinal is located in almost the same place, meaning that the polyene
is rotated relative to the wt hCRBPII retinol structure (Figure III-2).

78

all-trans-retinal

4.93Å

Q108

Figure III-2: The overlay of Q108K:K40L mutant with wt hCRBPII with all-transretinal.

The crystal structure of this mutant demonstrates that in this system the positive
charge of the PSB appears to be stabilized in two different ways. First an ordered water
molecule interacts with both Gln4 and the iminium of the PSB (Figure III-3). This water
molecule was observed in all of the mutant structures that contained Gln4. Mutation of
this residue disturbs this water network and the iminium adopts trans conformation and
dropping the pKa of retinal-PSB. The second interaction that stabilizes the pKa is the πcation interaction between the PSB and the side chain of Trp109, which is almost 3.8Å
away from the protonated Schiff base. This interaction has been observed in all of the
hCRBPII mutants. Like the structure of wt hCRBPII, only a few well-ordered water
molecules were observed in the protein binding pocket.

79

Ilu110
Thr1

Gln4

all-trans-retinal
Q108K

4.13Å

3.22 Å 2.97 Å

3.06 Å
2.75 Å

3.29 Å
3.28 Å

Asp91
Trp106

Figure III-3: The π-cation interaction and water mediate stabilize the
positive charge of the iminium.

One of them is located between Gln38 and Gln128, which is 3.6Å away from the
polyene of the chromophore (Figure III-4). The other one is between the hydroxyl group
of Thr51 and Thr53. The binding pocket of bovine rhodopsin is made of hydrophobic
residues while in hCRBPII, the binding pocket is relatively non polar.
The next step was to make systematic mutations of the polar residues inside the
binding pocket to the non-polar amino acids. Inspection inside the binding pocket of the
structure of both wt hCRBPII and KL showed that Thr51 is 5.4Å away from the Schiff
base (Figure III-4), the closest polar residue to the PSB. One water molecule was
observed between Thr51 and Thr53. This water molecule is very well ordered and
observed in the structure of wt hCRBPII as well (PDB entry 2RCT). This water molecule
is near the middle chromophore backbone (4.17Å), 6.1Å away from C13 of retinal and
almost 7Å away from the PSB.
80

3.11Å
3.14Å

Q108K

3.64Å

Gln38

Gln128

Figure III-4: The water network inside the binding pocket of hCRBPII,
close to the chromophore.
.

Q108K

4.17Å

5.85Å

6.19Å

W1
3.18Å

3.09Å
Thr53

Thr51

K40L

Figure III-5: The crystal structure of Q108K:K40L (KL) shows a water
molecule forming a hydrogen bond with residues Thr51 and Thr53.

81

III-1-2: The Crystal Structure of Q108K:K40L:T51V (KLV) Bound to Retinal

The crystal structure of wt hCRBPII and Q108K:K40L shows the binding pocket
of hCRBPII consists primarily of polar amino acids. Theoretically, removing the more
polar groups from around the PSB-retinal can help the delocalization of the positive
charge of the chromophore and lead to more red shifted absorbance. The closet polar
residue to the PSB is Thr51. As shown in Figure III-4, the hydroxyl group of this residue
is about 6Å away from the PSB; however the hydroxyl group does not have direct
interaction with the PSB. The crystal structure shows that the hydroxyl group of Thr51 is
pointing in the direction of the ionone ring of the chromophore and makes water
mediated interaction with the Thr53 hydroxyl group. Mutation of Thr51 to a non polar
amino acid valine (T51V mutation) was performed by Dr. Wenjing Wang. This mutation
leads to a significant bathochromic shift (25 nm) moving the λmax from 508 nm to 533
nm. The crystal structure of this mutant was determined in the same condition as
Q108K:K40L mutant and refined at 1.7Å with crystallographic Rfac = 21.85% and Rfree =
24.64%. All of the structural data is listed in table IV-1. The space group is P1 and the
unit cell is virtually identical to the Q108K:K40L structure. The overlaid structures of
Q108K:K40L and Q108K:K40L:T51V (Figure III-6) show that the overall structures are
quite similar. However the water molecule between Thr51 and Thr53 has disappeared
in both chains due to the T51V mutation.

82

Thr53
2.95Å

2.87Å

T51V
T51

Q108K

all-trans-retinal

Figure III-6: The crystal structure overlaid of Q108K:K40L mutant (green)
and Q108K:K40L:T51V mutant (dark blue). The water mediate between
Thr51 and Thr53 belongs to Q108K:K40L mutant.

Figure III-7: The overlaid structure of Q108K:K40 (green)
andQ108K:K40L:T51V (dark blue) bound to all-trans-retinal.

83

No other major differences between these two structures were observed. The pKa of
Q108K:K40L:T51V mutant was measured to be 8.3 which is 0.4 unit higher than
Q108K:K40L and therefore the data shows that the PSB hydrophilicity of Thr51 did not
seem to play a role in stabilization of the pKa .
Removing this polar residue by T51V mutation has two effects: 1: Removing the
polarity of the amino acid residue; 2: removing a water molecule in the binding pocket
that is 7.09Å away from the PSB. The combination of these two effects leads to 25 nm
red shifting compared to Q108K:K40L. The spectroscopic data and the crystal structure
show that less polarity in the vicinity of the PSB of the chromophore is needed for more
red shifting absorbance.

III-1-3: The Crystal Structure of Q108K:K40L:T53C (KLC) Bound to Retinal

Theoretical calculations suggested that introducing a polarizable amino acid may
lead to more red shifting.1,2,3 Since the positive charge of the protonated Schiff base is
transferred from the iminium region to the ionone ring during the electronic excitation, a
polarizable environment could lower the energy level of the excited state and decrease
the energy gap between the ground and excited states, leading to more red shifted
spectra. The crystal structure of hCRBPII shows several polar residues in the binding
pocket of hCRBPII. Thr53, situated near the middle of the polyene chain, was chosen to
test this hypothesis. Tryptophan, other aromatic residues and cysteine are all relatively
polarizable. However since the binding pocket of hCRBP is narrow, Dr. Wenjing Wang
84

generated the T53C mutation combined with Q108K:K40L to give Q108K:K40L:T53C
(KLC).

Table III-2: Comparison of mutations at the Thr53 position
Mutants

λmax (nm)

pKa

Q108K: K40L:T53

508

7.9

Q108K:K40L:T53V

503

7.3

Q108K:K40L:T53C

513

8.3

Crystals of Q108K:K40L:T53C (KLC) are quite similar to the previous two mutants
(almost identical P1 space group with two molecules per asymmetric unit). The structure
was refined at 1.4Å (Rfac = 21.20% and Rfree = 28.20%). Again the overall structure,
backbone, etc. of this mutant shows little deviation from the Q108K:K40L and
Q108K:K40L:T51V structures (Figure III-7). However the conformation of the bound
chromophore is considerably altered. The plane of the polyene chain is rotated almost
90° relative to KL that has 6-s-trans conformation, resulting in a 6-s-cis conformation.
Since the position of the b-ionone ring did not change the conformational change is due
to the rotation of the polyene chain. It is not clear the T53 mutation caused this
conformational change. Though this is one of the mutations that proved a platform for
the mutations leading to wavelength tuning, this distorted structure was not observed in
any of the other T53C containing mutants. The conformation of the iminium is cis which
is the same as Q108K:K40L and Q108K:K40L:T51V, and the water mediated interaction
85

between the PSB and Gln4 is present in both chains in this structure. In this structure
the water molecule between Thr51 and Thr53 is still observed in both chains. The side
chain of Cys53, which is pointing in the direction of the PSB, makes a hydrogen bond to
the aforementioned water molecule. The absorbance of Q108K:K40L:T53C is 513 nm,
only 5 nm more red shifted compared to Q108K:K40L (508 nm). It seems that the
polarity of the amino acid in the middle of polyene does not have a significant effect on
the wavelength tuning, compared to the effect of the water molecule located near the
middle of the polyene chain, which seems to have a far more important role in tuning
the wavelength. Since water is the most polar group is in the binding pocket, having a
larger dipole moment than virtually any amino acid side chain, removing or at least
changing the polarity of the water which is located almost in the middle of the polyene
chain and 4.17Å away from the back bone of the chromophore can have a very
significant effect on the spectral tuning. Since in Q108K:K40L:T53C the position of the
water did not change, only a 5 nm red shift was observed, while in the T51V mutation,
which resulted in complete removal of the water molecule, a 25 nm bathochromic shift
was observed.

86

T53C
W1

T51V

all-trans-retinal

Figure III-8: Overlay of the structures of three mutants: Q108K:K40L (blue),
Q108K:K40L:T51V (green) and Q108K:K40L:T53C (red). As it is shown in
this figure the plane of the backbone of the chromophore rotates almost 90°.
The water molecule between Thr51 and Thr53 is available.

87

III-1-4: The Crystal Structure of Q108K:K40L:T51V:T53C(KLVC) Bound to Retinal

Combination of the Q108K:K40L mutations with the T51V and T53C mutations
resulted in tetra mutant Q108K:K40L:T51V:T53C that absorbs at 539 nm. The λmax of
this mutant is the combination of the 25 nm from T51V and 5 nm from T53C by adding
to Q108K:K40L (508 mm). In other words the absorbance of this mutant is only
controlled by these mutations. This mutant also crystallized in the P1 space group with
two molecules per asymmetric unit. The data was refined at 1.6Å with crystallographic
Rfac = 21.31% and Rfree = 26.46 %. This structure is superimposed on the other three
previous structures (RMS = 0.153Å for 133 amino acid residues) (Figure III-8). The
similarity of this combination with Q108K:K40K indicates that different mutations do not

Figure III-9: The crystal structure overlay of four different mutants in P1 space
group: Q108K:K40L (green), Q108K:K40L :T51V (blue), Q108K:K40L:T53C
(orange) and Q108K:K40L:T51V:T53C (red).

88

change the integrity of the structure. The conformation of the chromophore in
Q108K:K40L:T51V:T53C is located in the same place as in the other mutant structures
with only minor conformational change relative to KL and KLV. The side chain of Cys53
adopts two different conformations, one of them pointed to the ionone ring and the other
one pointed to the Schiff base region. The Occupancy between these two
conformations is about 50%. The water molecule located between Thr51 and Thr53 is
completely removed in this mutant. The different conformation can be explained by
removing the water molecule, since there is no H-bond available to hold the Cys53 in a
specific location.

T53C

T51V

Q108K
all-trans-retinal
Figure III-10: The crystal structure of KLVC with the highlighted mutation
positions. Cys53 adopts different conformations.

Based on the crystal structure of these four mutants and comparison of their
absorbances in table III-3 we can conclude that two parameters have important roles for
tuning the wavelength to more red shifted absorbance: 1- Removing the polar residues
close to the PSB region. This effect can help the positive charge to be more delocalized
89

along the polyene chain. 2- Removing the polar groups in the binding pocket which
potentially can have an interaction with the chromophore. These polar groups can be
any polar or charged molecules in the binding pocket for example in hCRBPII water
molecules. The effect of removing or changing the polarity (decrease or increase) of
these groups has an important role in tuning the wavelength.

Table III-3: the effect of removal of the polar residues around the chromophore
Mutants

λmax(nm)

pKa

Q108K:K40L

508

7.9

Q108K:K40L:T51V

533

8.3

Q108K:K40L:T53C

513

8.4

Q108K:K40L:T51V:T53C

539

8.4

III-1-5: The Crystal Structure of Q108K:K40L:T51V:T53S (KLVS) Bound to Retinal

To find out how much the polarity of the amino acid residues in the polyene chain
effects the wavelength regulation, Dr. Wenjing Wang generated other mutants at the
Thr53 position. In one of these mutants serine was introduced instead of Thr53. Serine
is less polarizable than cysteine but is a more polar amino acid. The lmax of this mutant
is 534 nm which is 5 nm blue shifted compared to Q108K:K40L:T51V:T53C (539 nm)
with a pKa of 8.7. Unlike the other crystal structures of hCRBPII mutants the crystal
structure of Q108K:K40L:T51V:T53S (KLVS) was determined in the P2 space group
with two molecules per asymmetric unit. The data were refined to 1.4Å with
90

crystallographic Rfree = 22.77% and Rfac = 18.57%. The crystallographic data of this
mutant is summarized in table V-1. This space group was unique for this mutant. The
structure of the bound chromophore shows a major conformational difference when
compared to Q108K:K40L, Q108K:K40L:T51V and Q108K:K40L:T51V in the ionone
ring region with the ionone ring almost perpendicular to the polyene chain. In this
structure the polyene chain remains in essentially the same place (Figure III-10-b). In
this structure similar to Q108K:K40L:T51V

and Q108K:K40L:T51V:T53C, the water

molecule between Thr51 and Thr53 is removed completely. However despite the
Q108K:K40L:T51V:T53C structure, where Cys53 adopts two different conformations,
Ser53 has just one conformation, pointing toward the ionone ring area. The difference in
conformation in the chromophore is the result of the T53S mutation. Ser53 makes a
strong hydrogen bond with Gln38 and the result of this interaction moves the side chain
of Gln38 from its original position that has been observed in all other structures
including wt, and this movement leads to conformational changes in the structure of the
chromophore, with the chromophore adopting a 6-s-cis conformation. Otherwise, as
shown in Figure III-10-b, steric clash happens between the side chain of Gln38 and
methyl5 of retinal. The conformational difference in the structure the chromophore
indicates that this 6-s-cis conformation does not lead to a major hypsochromic shift.
Since in Q108K:K40L:T51V:T53C the chromophore adopts the 6-s-trans conformation
while Q108K:K40L:T51V:T53S the structure of the ionone ring of the chromophore is
cis and just 4 nm hypsochromic shift has been observed.

91

a.

b.
all-trans-retinal

Q108K
2.67Å
T51V

T53C
T53S

Gln38

Figure III-11: a. The crystal structure of KLVS mutant at 1.4Å resolution bound
to Retinal. b. The superimposed KLVS and KLVC mutants. As it is shown in
this figure Ser53 makes a strong H-bond to Gln38, which causes changing the
conformation in the chromophore.

92

In contrast to all other hCRBPII crystal structures, in the Q108K:K40L:T51V:T53S
crystal structure the water network between Gln38 and Gln128 is not observed. The
removing of this water molecule is a result of the T53S mutation. However the by the
dislocation of Gln38 positions the hydroxyl group of this residue locates almost in the
same position, as the water molecule was located. Removing of the water molecule is a
result of the hydrogen bond between the Ser53 side chain and Gln38. This hydrogen
bond is not observed. As shown in Figure III-12, the water molecule (W1) is 3.62Å away
from the chromophore, while the carbonyl group of Gln38 of Q108K:K40L:T51V:T53S is
3.66Å away form all-trans-retinal. At the end of this chapter it will be explained that
these water molecules are crucial both for wavelength regulation and also for controlling
the pKa of the PSB.

3.66Å

3.79Å

2.37Å

2.61Å

2.61Å
W1

Q108K
Gln128

Gln38

Figure III-12: The superimposed structure of KLVC (cyan) and KLVS
(green). W1 (water molecule) belongs to the KLVC structure.
93

Removing one polar group (water) and replacing it with another polar residue (the side
chain of Gln38) appears not to have a significant effect on the wavelength and pKa.
However new studies in our group (performed by Meisam Nosrati in Prof. James
Geiger’s group) illustrate clearly that the absence of a polar group leads to the most
hypsochromic shift in the wavelength.

III-2: The Effect of Introduction of Tryptophan in the Middle of Polyene Chain and
Rigidifying the Chromophore

The crystal structure of bovine rhodopsin indicates that several Tryptophan
residues are present in the chromophore binding pocket.4 Tryptophan is an electron rich
and polarizable group and has the ability to encourage the delocalization of the positive
charge of the PSB along the chromophore. Since introducing Cys53 leads to more red

Trp19
4.30Å

all-trans-retinal

Q108K

Figure III-13: The crystal structure of Q108K:K40L mutant with highlighted
Y16W position
94

shifted absorbance, therefore Tyr19 which is near the middle of the polyene chain,
about 5Å form the C9-methyl group and on the opposite side of the chromophore from
Thr53 (Figure III-12), was mutated to tryptophan and Y19W mutations were generated.
The combination of this mutation with Q108K:K40L (508 nm) leads to a modest 5 nm
red shift, similar to that seen for the T53C mutation. It seems that removing the more
polar group (tyrosine) and replacing it with an electron rich and polarizable amino acid
(tryptophan) can lead to a small red shifting.

Table III-4: Combination of Y19W mutation with the other mutations
Mutants

λmax(nm)

pKa

Q108K:K40L:Y19

508

7.9

Q108K:K40L:Y19W

513

8.9

Q108K:K40L:T51V:Y19W

537

9.3

Q108K:K40L:T51V:R58W:Y19

565

8.3

Q108K:K40L:T51V:R58W:Y19W

577

10.3

However adding this mutation to Q108K: K40L: R58W: T51V (565 nm) leads to 12 nm
more red shift. The effect of the R58W mutation will be explained in the next section.
Even The effect of the Y19W mutation is not significant in short series of mutations still
the red shifting can be related to two different parameters: 1.The planarity of the
chromophore, which leads to more delocalization of the positive charge of the PSB. The
crystal structure of all of the mutants with Y19W shows that retinal is more linear
compared to the ones without this mutation like Q108K:K40L. 2. Removing the tyrosine
95

from this position that contains hydroxyl group and replacing of that with the softer
polarizable group leads to better π-π interaction and more red shifted has been
achieved.

III-2-1: The Crystal Structure of Q108K:K40L:T51V:Y19W:R58W (KLWW) Bound to
Retinal

The

crystal

structure

of

Q108K:K40L:T51V:Y19W:R58W

(KLVWW)

was

determined in P1 space group with two molecules per asymmetric units at 1.5Å
resolution. The data were refined with Rfac = 23.16% and Rfree = 31.04%. The λmax of this
mutant is 577 nm. As shown in table III-4 the effect of the Y19W was 12 nm in the
presence of R58W. The overall structure of this mutant is the same as the other
mutants. However the details of the binding pocket of this mutant shows a translation in
the position of the chromophore. In this structure the “ionone ring part” of the
chromophore is pushed more to the opposite side and leads to a more planar
chromophore. Introduction of this mutation causes the chromophore to be pushed by
the bulky side chain of Tryptophan therefore the free motion of the chromophore is
reduced (Figure III-13). This effect rigidifies the position of the chromophore and may
make the polyene chain and the double bond of the ring more conjugated. In all of the
crystal structures that contain the Y19W mutation the translation in the location of the
chromophore was observed. Table III-3 shows the λmax absorbance in the different
mutation. Another important difference between the structure with and without the
Y19W mutation is the distance from Gln38. By applying the Y19W mutation the distance
96

between the chromophore and the side chain of Gln38 decreases and this residue
possibly can help the delocalization of the positive charge and therefore lead to greater
red shift (Figure III-14).

Tyr19
Y19W
Q108K
3.95Å
4.27Å

Trp58
K40L

T51V

R58W

Figure III-14: The crystal structure overlay of KLVF (purple, 561 nm at 1.4Å)
and KLVWW (dark blue, 577 m, at 1.5Å)

The details of the ordered water molecule in the binding pocket of this mutant illustrates
that in one chain of this mutant (chain B) an electron density (2.4Å) away from the
hydroxyl group of Thr53 is observed. This electron density could be a water molecule
that makes a strong hydrogen bond with the hydroxyl group of Thr53.

97

Trp19
Y19W

3.74Å
Gln128
3.15Å

Gln38

Figure
III-15:
The
superimposed
crystal
structure
of
Q108K:K40L:T51V:R58W:Y19W (dark blue) and Q108K:K40L:T51V:R58F (cyan)
with the highlighted Gln38 and 128 with the chromophore distance.

98

III-3: Effect of Mutation at R58 Position

The binding pocket of wt hCRBPII is covered by two short helices. In addition,
because the guanidium group of Arg58 makes a hydrogen bond with the main chain of
Ser55, the aliphatic side chain of Arg58 is stretched over the mouth of the binding
pocket. These features make the binding pocket quite covered (Figure III-15).

Figure III-16: The space-filling model based on the crystal structure of wt
hCRBPII. The chromophore is shown in blue.

Arg58 is located at the “entrance” of the ligand to the binding pocket and mutation of
this residue to aromatic amino acid residues leads to tremendous red shifting in the
absorbance. As shown in table III-5 introducing the rigid and aromatic amino acids (Trp,
Try and Phe) leads to red shifting absorbance, while mutation of this position to the
other amino acids (polar and non polar) causes a hypsochromic shift (table III-5). In this

99

section the crystal structure of the different mutants that have R58 mutations and the
effect of them in tuning of the wavelength will be described.

Table III-5: The effect of mutation of R58 on the wavelength
Mutants

λmax(nm)

pKa

Q108K:K40L:R58

508

8.3

Q108K:K40L:R58D

500

8.6

Q108K:K40L:R58A

499

8.4

Q108K:K40L:R58L

500

8.1

Q108K:K40L:R58Q

499

8.4

Q108K:K40L:R58E

500

8.1

Q108K:K40L:R58F

523

7.9

Q108K:K40L:R58Y

535

9.5

Q108K:K40L:R58W

527

7.9

III-3-1: The Crystal Structure of Q108K:K40L:T51V:R58F (KLVF) Bound to Retinal

The crystal structure of this mutant was determined in the P1 space group with
two molecules per asymmetric unit. The data was refined to 1.4Å resolution with Rfac =
21.48% and Rfree = 27.46%. Unfortunately the density of the chromophore was clear
only in one of the two chains (chain B) but the density of retinal in chain A is very clear.
The overall structure of this mutant is again the same as the others. Figure III-15-c
100

shows the superimposed structure of Q108K:K40L:T51V with Q108K:K40L:T51V:R58F.
As mentioned above the side chain of Arg58 almost covers the “ mouth” of the protein in
hCRBPII structures and retinal is buried inside the binding pocket (The structure of the
wt hCRBPII protein has been explained in chapter II). Applying the R58F mutation
makes the side chain of Phe57 flip from its original location (Figure III-15-c) and makes
the binding pocket of hCRBP more exposed. The effect of R58F enhances the effect of
other mutations. This effect is also observed in the R58W mutants but the enhancement
factor is much more uniform (factor of 2).5 As shown in Figure III-17, the position of
Tyr58, Phe58 and Trp58 is similar and therefore the effect is similar for Tryptophan and
phenyl alanine, while no enhancement has been observed for R58Y. However the
mutation of Arginine, which has a positive charge, to phenylalanine or any other
aromatic amino acid leads to more delocalization of the positive charge along the
polyene chain and the result of this mutation is more red shifted absorbance (Table III6). This effect has been also observed for the other mutations that have R58Y and
R58W as well.

101

a.

b.

Arg58

R58F
Phe57

Phe57

c.

Arg58
R58F
Phe57

Figure III-17: a. The space filling model of Q108K:K40L:T51V. The chromophore
is shown in pink. b. The space filling model of Q108K:K40L:T51V:R58F. c. The
superimposed
structure
of
Q108K:K40L:T51V
(green)
and
Q108K:K40L:T51V:R58F (purple).

However the enhancement of the bathochromic shifts is almost two times in the longer
series (combination with T51V:T53C:Y19W) with R58W mutations. This can be related
to the size of the Tryptophan vs. phenylalanine side chain (Table III-6, III-7).

102

Table III-6: The effect of R58F mutation in the presence of the other mutations
Mutants

λmax(nm)

pKa

Q108K:K40L:R58F

523

8.6

Q108K:K40L:T51V

533

8.3

Q108K:K40L:T51V:R58F

561

8.5

Q108K:K40L:Y19W

513

8.9

Q108K:K40L:Y19W:R58F

537

9.3

Q108K:K40L:T53C

513

7.3

Q108K: K40L: T53C:R58F

537

8.5

103

Table II-7: The enhancement of red shifting based on R58F mutation
Mutant

!max(nm) Protein

R58F

Shift
(nm)

Protein Enhancement
shift
(nm)
(nm)

Q108K:K40L

508

0

523

-----

-----

Q108K:K40L:T51V

533

25

561

38

13 (1.5x)

Q108K:K40L:T53C

513

5

537

14

9 (2.8x)

Q108K:K40L:Y19W

513

5

537

14

9 (2.8x)

Q108K:K40L:T51V:T53C

539

31

571

48

17(1.5x)

Table III-8: The enhancement of red shifting based on R58W mutation
Mutant

!ax(nm)

Protein

R58W

Shift
(nm)

Protein shift
(nm)

Enhancement
(nm)

Q108K:K40L

508

0

527

-----

-----

Q108K:K40L:T51V

533

25

570

43

18 (1.7x)

Q108K:K40L:Y19W

513

5

540

13

8 (2.6x)

Q108K:K40L:T53C

513

5

538

11

6 (2.2x)

Q108K:K40L:T51V:T53C

539

31

585

58

27 (1.9x)

Q108K:K40L:T51V:Y19W

537

29

577

11

21 (1.7x)

Q108K:K40L:T51V:T53C:Y19W

538

30

590

63

33 (2.1x)

The crystal structure overlaid of three mutants with R59F, R58W and R58Y is
shown in the figure III-16. It is clear the positions of these three side chains are
superimposable and therefore a similar enhancement was observed for R58Y and R58F
104

mutants. The electron rich and more rigid side chain of these three aromatic residues
lead to more delocalization of the positive charge.

Figure III-18: The crystal structure overlay of KL:T51V:R58F (orange, 561 nm),
KL:T51V:Y19W:R58Y (cyan, 577 nm and KL:T51V:Y19W:R58W (dark blue,
565 nm)

III-3-2: The Crystal Structure of Q108K:K40L:T51V:R58Y:Y19W (KLVYW) Bound to
Retinal

The bound form of this mutant with all-trans-retinal absorbs at 565 nm while the
parent mutant Q108K:K40L:T51V:Y19W, absorbs at 537 nm; thus the R58Y leads to a
28 nm bathochromic shift. This mutant also has the highest pKa of all of the generated
mutants (10.3). Though the overall conformation of the protein backbone did not
change, the conformation of the chromophore changes from 6-s-trans to 6-s-cis (Figure
III-17). In contrast to the previously described structures (Q108K:K40L:T53C) the
ionone ring itself rotates while the polyene is similarly oriented relative to the other
105

structures. In the other word in Q108K:K40L:T53C mutant the plane of the polyene
chain rotated about 90° while in Q108K:K40L:T51V:Y19W, ionone ring rotates about the
C6-C7 single bond (Figure III-18). As mentioned above the mutants with R58Y, R58F
and R58W have the higher pKa when compared to mutants that do not have an
aromatic group at position 58. Among all the R58 mutants, the ones with R58Y have the
highest pKa. It seems that the hydroxyl group of the side chain makes it even more
electronegative compare to the other aromatic residues, increasing the pKa. It is clear
from these studies that stabilization of the protonation state can be achieved by
manipulation of the polyene environment far from the site of protonation. Though not
definitive, comparison of the Q108K:K40L:T51V:Y19W:R58Y structure to others may
proved an explanation for the cis conformation of the ionone ring. The Phe16 side chain
is moved in this mutant (Figure III-18.b). A explained above in section III, the position of
two amino acid residues (Phe16 and Leu77) lock the ionone ring of the chromophore in
the

trans

conformation.

Q108K:K40L:T51V:Y19W:R58W,

Study
shows

on

that

the

another

crystal

chromophore

of

structure
has

trans

conformation same as the most of hCRBPII mutants. Since these two mutants are just
different in R58 mutation therefore the comparison has been done on these two
structures. In Q108K:K40L:T51V:Y19W:R58Y structure the side chain of Phe16 is
moved (~ 0.5 Å) therefore there is more space available for the rotation around the C6C7 single bond. The detail is shown Figure III-19.b.

106

b.

a.
T51V
R58Y

all-trans-retinal

Q108K

4.1Å

Y19W

Figure III-19: a. The crystal structure of Q108K:K40L:T51V:Y19W:R58Y mutant
with all-trans-retinal. The conformation of the ionone ring of chromophore is 6-scis. b. The overlaid structure of all-trans-retinal in Q108K:K40L (blue),
Q108K:K40L:T53C (red) and Q108K:K40L:T51V:Y19W:R58Y (green).

Leu77

b.

a.

Phe16
3.52Å
4.45Å

3.56Å

3.98Å

R58W
R58Y

`

Figure III-20: a. The crystal structure of wt hCRBPII with all-trans-retinol. b. The
superimposed crystal structure of KLVR58WY19W (pink) with KLVR58YY19W
(blue).

107

However the overall structure of the protein did not shows any different at all and
even Trp and Tyr side chains at 58 position are exactly in the same place. However it is
not clear that why the side chain of Phe16 translated from the original position but we
can conclude that in presence of more space, the C6-C7 single bond freely can rotates
and therefore chromophore adopts the structure which is more stable (6-s-cis) (Figure
III-18).

Even though the chromophore in Q108K:K40L:T51V:Y19W:R58Y mutant

adopts

with

6-s-cis

conformation

it

absorbs

at

565

nm

while

the

Q108K:K40L:T51V:Y19W:R58W mutant which the ionone ring of the chromophore
adopts the trans conformation 577 nm (12 nm) more red shifted. As it was predicated
applying R58W mutation leads to more bathochromic shift. It seems that changing from
6-s-trans to 6-s-cis does not lead to significant hypsochromic shift.
As mentioned above introducing of the aromatic residues (Tryptophan, Tyrosine
and phenyl alanin) at R58 position leads to the red shifted absorbance and at the same
time these mutations enhanced the effect of the other mutations (T51V, T53C, Y19W).
This enhancement effect in R58F mutation is around 2x (Table III-7) of the original red
shifting as well as R58W. On the other hand, introducing R58Y despite R58F and R58W
does not enhance the effect of other mutants and the red shifting effect is additional. It
seems that by mutation of Arg58 to tryptophan and phenyl alanine (both are not polar
residues), the hydrophobicity at the binding pocket of protein increase, in the other
word, creation of hydrophobic environment around the protein binding pocket can lead
to isolating of binding cavity from the bulk solution. Tyrosine does not have the same
effect since the hydroxyl group can make the protein “entrance” more polar. It seems
that by increasing the size of the aromatic residue at the “ month”, of the protein, the
108

binding pocket is covered more efficiently and the chromophore is more isolated form
the aqueous solution, decreasing the polarity of the binding pocket. Introduction of the
R58 mutation to the aromatic residues in the short series of mutants (Q108L:K40L) for
tryptophan and phenyl alanine shows a similar effect.

Table III-9: The enhancement of the red shifting based on R58Y mutation
Mutant

!max(nm) Protein

R58Y

Shift
(nm)

Protein
shift
(nm)

Enhancement
(nm)

Q108K:K40L

508

0

535

-----

-----

Q108K:K40L:T51V

533

25

563

28

3 (1.1x)

Q108K:K40L:T53C

513

5

540

5

0 (1.0x)

Q108K:K40L:T51V:Y19W

537

29

565

30

1 (1.03x)

Q108K:K40L:T51V:T53C

539

31

576

30

-1 (0.96x)

However the R58Y variant leads to more red shifted absorption in the short series (table
III-8), while in the longer series the tryptophan variant gives the larger bathochromic
shift, which might related to the size of the side chain. It seems that in the shorter series
the electron rich Tyr hydroxyl group plays an important role in regulation, while in the
more red shifted, longer series of mutants the better solvent sequestering resulting from
the larger Trp residue dominates.
Few water molecules have been observed in the binding pocket of hCRBPII and
these water networks have crucial effect on the modulation the wavelength in hCRBPII.
The dielectric constant of water is 78, while by changing the environment to more
hydrophobic area this number decreases dramatically and reaches to 2 to 10.
109

Inspecting the crystal structure of mutants with R58W mutation and comparison of them
with the structures that R58 position is not mutated illustrates that the well ordered
water molecules in the binding pocket did not change at all, however the dielectric
constant of them can be affected by the polarity of the binding pocket. (Covering the
binding pocket with bulkier hydrophobic groups like Tryptophan or/and creating the
more hydrophobic environment in the entrance of protein binding site can reduce the
effect of dielectric constant of water molecules and as the result enhance the effect of
the more non-polar mutations).
Mutation of R58 to other amino acids such as R58E, R58A or R58L results in
blue shifted spectra. It seems that mutation of Arg58 to any other shorter amino acids
open the binding pocket more and therefore the blue shifting will be resulted (Table III5).

III-3-3: The Crystal Structure of Q108K:K40L:T51V:T53C:R58L (KLVCL) Bound to
Retinal

The crystal structure of Q108K: K40L:T51V:T53C:R58L was determined in the P1
space group with two molecules per asymmetric units. The λmax of the bound form of
this

mutant

with

all-trans-retinal

absorbs

at

531

nm

while

the

λmax

of

Q108K:K40L:T51V:T53C is 8 nm more red shifted (539 nm). Inspection of the crystal
structure of this mutant reveals interesting information. Compared to the crystal
structure of Q108K:K40L:T51V:T53C the chromophore is twisted in this mutant.

110

However the crystal structure of this mutant provides structural details that
support the idea that the conformational of the chromophore does not result in a blue
shift in this mutant. Figure III-21 shows the structural details of the binding pocket of this
!

5.33Å

2.74Å
R58L

2.86Å

Q108K
T51V

T53C

Figure III-21 The crystal structure of penta mutant Q108K: K40L: R58L: T51V:
T53C with the highlighted residues.

mutant. Surprisingly and in contrast to the other structures with both T51V and T53C
mutations (crystal structure of Q108K:K40L:T51V:T53C mutant), in this structure the
water molecule between residues 51 and 53 is present. However in none of the other
crystal structures with both T51V and T53C mutations structures) is this water molecule
observed. The presence of this water molecule may explain the relatively blue shifted
absorbance of this mutant compared to Q108K:K40L:T51V:T53C and it has almost the
same lmax as Q108K:K40L:T51V (533nm versus 531 nm for KLVCR58L). It is not clear
why the water molecule is still present in this mutant since neither the valine nor
cysteine side chains are capable of, or in position to make a hydrogen bond with the
111

water and still the water molecule has the perfect hydrogen bond distance from them. It
seems that because this water is able to make only one hydrogen bond, it may rotating
and does not fix in the same position and therefore the dipole moment of this water is
not pointing in a single direction and it’s electrostatic thereby partially naturalized. This
data is the evidence that the dielectric constant of water is reduced in the hydrophobic
environment so this water, which is beside the valine, does not have the same dialectic
constant as is close to the threonine and by decreasing the dielectric constant of water
the polarity of them decrease as well.
It can be concluded that modulation of the chromophore spectrum by the protein
environment is not only effected by the polarity of the side chain, but also by the
ordered, and perhaps even disordered water molecules. Therefore as was shown above
the only mutations that show significant red shifted spectra, are those that either directly
influence ordered water molecules, or sequester the binding pocket from the external
aqueous solvent.

III-3-4: The Crystal Structure of Q108K:K40L:T53V: R58W Bound to Retinal

This mutant has been generated by Dr. Wenjing Wang in Professor Borhan group
to test the effect of different mutations in combination with the R58W mutation.
Q108K:K40L:T53V:R58W absorbs at 541 nm which is almost the same as
Q108K:K40L:T53C:R58W (540 nm). These two mutants are almost 20 nm blue shifted
relative to Q108K:K40L:T51V:R58W. The crystal structure of Q108K:K40L:T53C shows
that the water molecule between Thr51 and Thr53 is in exactly the same location,
112

between

T51

and

T53,

the

expectation

was

that

the

structure

of

Q108K:K40L:T53C:R58W would show the same ordered water molecule, since the
R58W mutation is on the other end of the binding pocket.

Table III-10: Mutation of Thr51 and Thr53 to more hydrophobic residue
Mutant

λmax(nm)

pKa

Q108K:K40L:T51V:R58W

565

8.4

Q108K:K40L:T53V:R58W

541

9.0

Q108K:K40L:T53C:R58W

540

8.3

As previously mentioned the T51V mutation results in the absence of this
ordered water molecule in all structures so far determined, and results in a large red
shift in each case. The crystal structure of Q108K:K40L:T53V:R58 was determined in
the P1 space group with two molecules per asymmetric unit. The crystallographic data
was refined to 1.53Å resolution with crystallographic Rfac = 24.63% and Rfree = 32.82 %.
The electron density of the ligand was very clear in one chain (chain B), but missing in
chain A. The structure of this mutant shows that the water molecule between Val53 and
Thr51 is, surprisingly, not present. Further the Thr51 side chain has rotated relative to
KL, and the hydroxyl group makes a water-mediated interaction with the PSB. In all the
other structures with Thr51 present, the hydroxyl group is pointing toward the ionone
ring and therefore cannot make a direct interaction with the PSB. This conformation
would appear to be correlated with the absence of the water molecule, which allows the
hydroxyl group of Thr51 to make an interaction with the PSB instead of the water
113

molecule lodged between residues 51 and 53 (Figure III-20). Presumably the increase
in hydrophobicity of the region results in rotation of the hydroxyl group to the more
hydrophilic environment in the vicinity of the iminium.
As shown in Table III-9, the pKa of this mutant is higher than Q108K:K40L:T51V:R58W
and Q108K:K40L:T53C:R58W. The increase in pKa can be the result of the extra water
mediated interaction with the PSB, which make the hydrogen bond to the PSB and
therefore PSB is more stable and the pKa increased. The presence of two water
molecules interacting with the PSB is reminiscent of the bacteriorhodopsin structure,
which also shows two water-mediated interactions with the PSB resulting in higher pKa.
The flipping of the Thr51 side chain and presence of two water molecules
interacting with the PSB causes almost 20 nm hypsochromic shift compared to
Q108K:K40L:T51V:R58W. It is interesting to note that the KLT53V mutant is blue
shifted 5 nm relative to KL to 508 nm. This hypsochromic shift may be caused by a
similar water mediated interaction between T51 and the iminium in this structure,
resulting in more localization of positive charge at the iminium.

Surprisingly

Q108K:K40L:T53V:R58W and Q108K:K40L:T53C:R58W have the same absorbance
(544 nm). Comparison of the absorbances of KLT53V with KLT53VR58W shows an
enhancement caused by the introduction of 58W that is much larger than that seen with
the other mutants in the series (Table 9 include the KLT53V and KLT53VR58W in
table).
Combination of all of the mutations so far discussed leads to the most red shifted
mutant

discussed

thus

far.

These
114

two

mutants

are

hepta

mutants

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W

(absorbs

at

591

nm)

and

Q108K:K40L:T51V:T53S:R58W:T29L:Y19W ( 600 nm).

Thr51
R58W

2.72Å

T53V

1.95Å
3.14Å

2.

96

Å

5.2Å

3.53Å

Gln4

Q108K

Figure III-22: The crystal structure of Q108K:K40L:T53V:R58W mutant with alltrans-retinal.

The crystal structures of these two mutants were determined in P1 space group
with two molecules per asymmetric unit. The crystal structure of the hepta mutant
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W was refined at 1.5Å resolution and the
electron density of the chromophore is clear on one chain. Figure III-23 shows the
overlaid structure of Q108K:K40L:T51V:T53C:R58W:T29L:Y19W (591 nm) and
Q108K:K40L:T51V:T53S:R58W:T29L:Y19W (600 nm)

115

Y19W
Q108K
T29L
T51V
K40L
T53C
R58W

FigureIII-23:
the
superimposed
structure
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W
(blue)
Q108K:K40L:T51V:T53S:R58W:T29L:Y19W (Orange).

of
and

As shown the two proteins show little structural difference. However in the
structure of Q108K:K40L:T51V:T53S:R58W:T29L:Y19W, Ser53 adopts only one
conformation

which

is

pointing

toward

the

ionone

ring,

while

Cys53

in

Q108K:K40L:T51V:T53S:R58W:T29L:Y19W adopts two conformations. The crystal
structure of these two mutants shows the chromophore to be relatively flat and locked in
position, with little difference in conformation in the two structures. However the
structural

details

of

the

water

network

inside

Q108K:K40L:T51V:T53S:R58W:T29L:Y19W

the

is

binding

pocket

different

of
from

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W in one chain (chain A) while in chain B is
the

same

as

the

other

structures (Figure
116

III-24).

In

the

structure

of

Q108K:K40L:T51V:T53S, Ser53 makes a direct hydrogen bond to Gln38, while in
Q108K:K40L:T51V:T53S:R58W:T29L:Y19W, the interaction between Gln38 and Ser53
is water mediated. As mentioned above and shown in the crystal structure of wt
hCRBPII, Q108K:K40L and Q108K:K40L:T53C, a water molecule bridges Thr51 and
Thr53. This water restricts the hydroxyl group of Thr53 to a single conformation.
However,

this

water

molecule

is

removed

in

Q108K:K40L:T51V:T53S

and

Q108K:K40L:T51V:T53S:R58W:T29L:Y19W. The hydroxyl group of Ser53 is then free
to make an alternate hydrogen bond to Gln38. On the other hand, the sulfur group of
Cys53 has two different conformations while the hydroxyl group of Thr53 only points
toward the PSB region. Other than this conformational difference between these two
hepta mutant structures, the water molecule between Gln38 and Gln128 is present in
chain A as well. Another structural difference between these two chains is the
conformation of Trp58. In one chain Trp58 has the same conformation as
Q108K:K40L:T51V:T53C:R58W:Y19W:T29L:Y19W and the other chain this Trp flips to
cover the binding pocket more. Figure III-25 shows the details of the structure inside
the binding pocket of the protein. It is not very clear that what exactly causes the 9 nm
more red shifted. Absorbance in Q108K:K40L:T51V:T53S:R58W:T29L:Y19W mutants,
since the two chains are only slightly different. The differences in ordered water
between these two structures is the most likely cause.

117

a.
R58W
Q108K

W1

Ser53

2.4Å
2.9Å

2.8Å
W2

2.7Å
Phe57

Ser56

4.9Å

Gln128

2.7Å
Gln38

b.

R58W
Q108K
W1
2.7Å
3.4Å
Phe57
Gln38

Gln128

Figure III-24: a. The crystal structure of chain A of The binding pocket of
Q108K:K40L:T51V:T53S:R58W:T29L:Y19W mutant. b. The details of the
binding of chain B of the same mutant. As it is shown in this figure the position
of Gln38 and Trp58 are different in two chains. Trp58 position in chain A is
same as Q108K:K40L:T51V:T53C:R58W:T29L:Y19W mutant.

118

R58W

Q108K

3.42Å
Phe57

2.67Å

2.82Å

2.8Å
2.41Å

Gln128

Gln4

Figure
III-25:
The
superimposed
crystal
Q108K:K40L:T51V:T53S:R58W:T29L:Y19W
(light
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W (green).

structure
pink)

of
and

III-3-5: The Crystal Structure of Apo-Q108K:K40L:T51V:T53C:R58W :T29L

The crystal structure of this hexa mutant has been determined in the P312 space
group and despite most of the bound hCRBPII mutants with retinal this structure has
one molecule per asymmetric unit. The unbound crystal structure is needed to compare
whether conformational changing happens when the chromophore bound to the protein
or not, since the goal is to study the effect of the different mutations on the wavelength
regulation, therefore at least one crystal structure of one apo protein was necessary.
The structural details of this protein illustrate some information, which are important for
understanding the effect of each mutation. Figure III-26 shows the highlighted mutated
residues of this protein.
119

As is shown by T51V and T53C mutations, the water molecule between them is
removed and the side chain of Cys53 has only one conformation, pointing toward the
“mouth” of the binding pocket and one water molecule has been observed in the binding
pocket between Gln4 and Lys40. This water molecule is the one, which is responsible
for stabilizing the pKa and cis-conformation of the iminium. It will be explained further
that removing this water molecule leads to a reduced pKa and also a change in the
conformation of the Schiff base from cis to trans. The overall structures of the apo hexa
and holo (Q108K:K40L:T51V:T53C:R58W:Y19W:T29L) do not change significantly.

T51V
T53C

R58W
K40L

T29L

4.60Å

3.00Å

Q108K

Figure

III-26:

The

crystal

structure

Gln4

of

apo-

Q108K:K40L:T51V:T53C:R58W:T29L with the highlighted mutated residues.

120

T51V

Gln4

T53C

R58W

3.36Å
3.45Å

T29L
Q108K

Figure
III-27:
The
superimposed
structure
of
Q108K:K40L:T51V:T53C:R58W:T29L
(green)
and
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W (cyan)

apoholo-

Figure III-27 shows the superimposed structure of these two with RMS = 0.963Å
for 133 amino acids residues. The main structural difference that has been observed is
the conformation of the loop between two beta-strands. Trp58 is situated in this loop.
When the protein is not bound to all-trans-retinal, the Trp58 side chain goes deeper
inside the binding pocket, while in the bound protein most of the binding pocket is
occupied by the ligand and therefore Trp58 does not have sufficient space inside the
binding Pocket and is flipped out of the binding pocket. Otherwise there would be a
clash between the chromophore and Trp58.

121

III-3-6: Conclusion of Mutation on Ar58

As mentioned above mutation of Arg58 to just to aromatic amino acids (Trp, Phe
and Tyr) cause red shifting. However only Trp and Phe can enhance the effect of the
other mutation in the binding pocket, while Tyr does not have this effect. The mutation
data shows that R58E mutation despite the theories leads to more blue shift (Table III10), while R58L does not have important effect. It seems that by mutation of Arg58 to
more hydrophobic amino acid, we create a “hydrophobic tunnel” in front of the entrance
of the protein and with mechanism the binding pocket isolates from the bulk solution.
R58E mutation leads to more blue shift, because Glutamic acid is charged and compare
to Arginine is shorter. This hydrophilic environment can increase the water exchange
rate in the binding pocket and more blue shift will be resulted. Inspecting the structure of
the mutation of R58F, R58Y and R58W show that Phe57 side chain flipped back to the
bulk solution while in none of the other structures this effect have been observed. In the
other word by flipping back the phenyl alanin in the bulk solution and creating this
hydrophobic tunnel the red shifting will be resulted.

Table III-11: Mutation of Arg58 to charged and nonpolar amino acid

Mutants

!max(nm) pKa

Q108K:K40L:T51V:T53C:R58

539

7.4

Q108K:K40L:T51V:T53C:R58L

531

7.1

Q108K:K40L:T51V:T53C:R58E

517

8.3
122

R58Y
R58W
Arg58
Phe57

Figure III-28: The superimposed structure of KL (508 nm, green)
KL:T51V:R58W:Y19W (577 nm, blue) KL:T51V:R58W:Y19W (565 nm,
red).

III-4: Toward the Most Red Shifted Mutant with Cis-Iminium Conformation
So far, based on the crystal structures and rational mutagenesis of hCRBPII, we
could show that different parameters seem to have crucial roles in the modulation of the
wavelength in the hCRBPII rhodopsin mimic. We can categorized these parameters as:
1-The polarity of the binding pocket is important and removing the polar groups
around the chromophore leads to more red shifting due to delocalization of the PSB
charge.
2-The mutations that have the most dramatic effect on the modulation of the
wavelength to more bathochromic absorbance are the ones that affect the water
networks in the binding pocket.
3- Isolation of binding pocket from the bulk solution by introduction of large
aromatic residue at the “entrance” of protein binding pocket.
123

4-the most bathochrmoic mutants are the ones where the chromophore is more
linear and more locked up in the specific positions, therefore it seems that making the
binding pocket more narrow and preventing the free rotation of all-trans-retinal is
essential.
Based on this information, Dr. Wenjing Wang in Professor Borhan’s group designed
another mutant, which was the most red shifted one in the series of cis-iminum
structures.
The crystal structure of Q108K:K40L:T51V:T53C:R58W:T29L:Y19W reveals that
the binding pocket of hCRBPII is not efficiently covered even by R58W mutation.
The amino acid position considered for covering the binding pocket more efficiently was
Ala33. However modeling the A33W mutation in silico shows that Trp33 has two
potential clashes: Trp58 or the ionone ring of the chromophore. However Tanya
Berbasova in Professor Babak Borhan’s group introduced the same mutation in
another, homologous rhodopsin mimic, CRABPII, and observed a significant red shift.
Based

on

these

data

Dr.

Wang

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W.

introduced
The

the

same
octa

mutation

to

mutant

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:A33W absorbs at 606 nm and leads to 15
nm more red shift, while introduction of this mutant in proteins with R58 caused a
hypsochromic shift.

124

Table III-12: The protein shift caused by A33W
Mutant

Protein shift
of A33W

!max(nm)

!max(nm)

with A33

with A33W

Q108K:K40L

508

498

-10

KL:T51V:T53C

539

533

-6

KL:T51V:Y19W

537

522

-15

KL:T51V:T53C:Y19W

538

522

-16

KL:T51V:T53C

539

533

-6

KL:R58F:Y19W

537

543

6

KL:T51V:T53C:R58W:T29L:Y19W

591

606

15

(nm)

To understand the mechanism of how the A33W mutation leads to red shifting, the
crystal structure of a mutant with the combination of A33W and R58W was necessary.

III-4-1: The Crystal Structure of Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:A33W
Mutant Bound to Retinal

The crystal structure of Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:A33W also
crystallizes in the P1 space group with two molecules per asymmetric unit. The
crystallographic data was refined to 1.4Å resolution with Rfac = 20.15% and Rfree =
24.15%. The electron density of the chromophore is not very well defined and is only
seen

in

one

of

the

two

chains.

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:A33W
125

The

overlaid

structure

of
and

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W

is

show

in

Figure

III-26.c.

The

superimposed structure of these two mutants shows the overall structure of protein did
not change and even introduction of a large residue like Trp did not effect the position
and conformation of the Trp33-containing helices. The RMSD of the two structures is
0.398 Å for 133 amino acids. However the crystal structure reveals that the
chromophore is translated slightly due to the Introduction of the A33W mutation. The
bulky side chain of Trp33 moved the chromophore more to the Y19W position and this
residue also shifted 0.6 Å (figure III-28.c). Figure III-28.b illustrates a space-filling
representation

of

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:A33W

with

the

chromophore shown in red. As predicated, A33W isolates the binding pocket more
efficiently and the binding pocket of the chromophore is more embedded. The crystal
structure shows the different conformation of the side chains of Trp58 and Arg58.

126

a.

b.b.
A33W

R58W

c.

Figure
III-29:
a.
The
superimposed
structure
of
and
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W
(cyan)
and
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:A33W (green). b. The space-filling
model
of
the
crystal
structure
of
and
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:A33W with the highlighted residues.
c. The superimposed structure of Q108K:K40L:T51V:T53C:R58W:T29L:Y19W
(blue) and Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:A33W (cyan)

127

Trp58 has flipped out of the binding site to avoid a clash with Trp33 and the density of
the side chain of Phe57 is weak, indicating significant disorder of this side chain. Since
Trp58 is flipped back the distance of the side chain of this residue increased form the
original distance (4.14Å) but the side chain of Trp33 is 3.28Å away from the ionone ring
of the chromophore. This close distance of an aromatic residue can have two effects:
helping the delocalization of the positive charge of the PSB and red shifting absorbance
and increasing the Pka of the PSB. As shown in the Table III-11 the pKa of the mutants
with A33W is around 0.5 units higher than the ones with no mutation of A33.

Table III-13: Introduction of A33W in different hCRBP mutant
Mutant

pKa

pKa

with A33

with A33W

Q108K:K40L (KL)

7.9

8.4

KL:T51V:T53C

8.4

7.7

KL:T51V:T53C:Y19W

8.4

8.9

KL:T51V:T53C:R58W:T29L

7.9

8.1

KL:R58F:Y19W

9.3

9.7

KL:T51V:T53C:R58W:T29L:Y19W

8.2

8.7

III-5: Trans-iminium and Details Studies on Gln4

As described above, a water mediated interaction between Gln4 and the iminium
molecule is not observed in any of the structures of wt CRBPII, nor is it seen in the apo
128

Q108K:K40L:T51V:T53C:R58W:Y19W:T29L mutant structure. This water-mediated
interaction requires the iminium double bond to be in the cis conformation, which is the
conformation seen in every structure in which Gln4 is present. In an effort to obtain even
more red-shifted absorbances in hCRBPII.
Dr. Wenjing Wang in Professor. Borhan’s group made various mutations at the Gln4
position. Table III-12 summarizes mutations at the Gln4 position in the context of the
with hepta mutant Q108K:K40L:T51V:T53C:R58W:T29L:Y19W. Clearly mutation of Q4
to a variety of amino acids all lead to a dramatic red shift and simultaneously a
significant drop in pKa. The lone exception is the Q4H mutation (Table III-12 entry 9)
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4H) which gives a small blue shift with
virtually no change in pKa. (relative to Q108K:K40L:T51V:T53C:R58W:T29L:Y19W).
The maximum red shift in this series of mutants are those that introduce a positively
charged amino acid at the Q4 position (Q4R (622 nm with the lowest pKa = 6.5) and
Q4K

(618 nm)). To understand the effect caused by Gln4 mutation, a number of

mutants were screened for crystallization. Unfortunately, most Q4 mutants failed to
crystallize. This is similar to what we have seen with other mutants in that mutants with
relatively low pKa’s are more difficult to crystallize, if they crystallize at all.

129

Table III-14: The effect of Q4 mutation
Mutants

a

!max(nm)

pKa

KLVCWLW:Q4a

591

8.2

KLVCWLW:Q4A

612

7.0

KLVCWLW:Q4L

614

7.9

KLVCWLW:Q4F

613

7.5

KLVCWLW:Q4W

613

7.7

KLVCWLW:Q4T

608

7.8

KLVCWLW:Q4K

618

7.2

KLVCWLW:Q4R

622

6.5

KLVCWLW:Q4H

585

7.9

KLVCWLW:Q4E

590

n.d

a

: KLVCWLW is Q108K:K40L:T51V:T53C:R58W:T29L:Y19W

Entry 10 was not stable and its pKa could not be determined.

III-5-1: The Crystal Structure of Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4H
Mutant Bound to Retinal

In

contrast

to

the

other

mutants

in

Table

III-12

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4H absorbs at 585 nm and is 6 nm blue
shifted compared to Q108K:K40L:T51V:T53C:R58W:T29L:Y19W (591 nm). The crystal
structure of this protein with all-trans-retinal crystallized in the P1 space group with two
molecules per asymmetric units and the data were refined to 1.3Å resolution with
crystallographic Rfac = 18.18 % and Rfree = 21.98%. The electron density of the
130

chromophore is very clear in both chains. The structure of the binding pocket with
highlighted mutated residues is shown in Figure III-27. The structure shows that
mutation of Gln4 to Histidine removed the water between Gln4 and the PSB and as a
result the iminium adopts the trans conformation. The trans-imine is usually a few
Kcal/mol more stable than the cis-imine, however the cis-imine in the hCRBPII mutants
can be stabilized by the water-mediated interaction between the iminium nitrogen and
Gln4.

Further potential stabilization may come from the p-cation interaction with

Trp106. Given the trans conformation seen in this structure, it appears that removal of
the water-mediated interaction is sufficient to convert the conformation of the PSB from
cis to trans. Indeed, in all structures that have Gln4 mutated, the PSB is found in the
trans conformation, consistent with the idea that this water-mediated interaction
is critical for stabilizing the cis conformation in the binding pocket.
However, mutation of Gln4 to other amino acids abolishes the water-mediated
interaction. In the structure of Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4H there
are also other differences in the water network near the PSB. Figure III-31 shows the
crystal

structures

of

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4H

and

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W overlaid. The two structures did not
change significantly (RMSD = 0.868Å for all 266 amino acid residues). As shown His4
makes a water mediated interaction with the main chain carbonyl of Thr1 via two water
molecules while in the structure of K40L:T51V:T53C:R58W:T29L:Y19W, Gln4 makes
the water mediated interaction with Thr1 using only one water molecule.

131

T53C

R58W

T51V
K40L

Q4H
5.78Å

30.81

Q108K

Y19W

Figure
III-30:
The
crystal
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4H
residues.

It

is

suspected

that

this

extra

structure
of
with the highlighted

water

molecule

in

K40L:T51V:T53C:R58W:T29L:Y19W:Q4H may result in the small blue shift and
moderate stabilization of the pKa. This mutant has the highest pKa of any of the Gln4
mutants (Table III-12). However this water molecule does not have any direct interaction
with the PSB, and is in fact 7Å away from the iminium nitrogen.
In this structure the chromophore adopts a more planer conformation than seen
in many of the other structures so far determined. For example the dihedral angle
between Met5-C5-C6-C7 in this structure is 36.81°and the angle between Met9-C9C13-Met13 is -3.68°, this is the smallest torsion angle observed in any of the structures
of Retinal-bound hCRBPII mutants.
132

!
Trp106
2.76 Å
2.55 Å
4.01Å

all-trans-retinal

3.04 Å

2.91 Å

2.94 Å

Gln4
Q4H

Figure
III-31:
The
superimposed
K40L:T51V:T53C:R58W:T29L:Y19W:Q4H
K40L:T51V:T53C:R58W:T29L:Y19W (cyan).

2.76 Å

Thr1

crystal
structure
of
(green)
and

III-5-2: The Crystal Structure of Q108K:K40L:T51V:T53C:R58F:T29L:Y19W:Q4H
Mutant Bound to Retinal

As explained above mutation of Arg58 to aromatic residues leads to a large red
shift, and enhances the red shift of other mutants over what would be expected form an
additive effect. The different mutants with Q4H and Arg58 mutation to different aromatic
residue were generated by Dr. Wenjing Wang. With these data we could see how much
the size of mutant at Arg58 position has effect in the longer series of mutations and
therefore Q108K:K40L:T51V:T53C:R58F:T29L:Y19W:Q4H has been generated. This
133

mutant

absorbs

at

547

nm

(11

nm

blue

shifted

relative

to

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4H) and has a pKa of 7.7. The mutant
was crystallized in the P1 space group with two molecules per asymmetric unit. All of
the crystallographic information of all of the mutants is listed in Table I in chapter IV.
The crystal structure of the two mutants with R58F and R58W do not change a major
different and the RMSD is 1.654Å for 266 amino acid residues (Figure III-29). The
superimposed structure of these two mutants illustrated that the only structural
difference is the position of the side chain of Phe58 and Trp58. The crystal structure
shows the Trp58 side chain not only cover the binding pocket more efficiently but is also
closer to the chromophore. This effect seems to be result of changing the conformation
of PSB from cis to trans and the chromophore is deeper in the binding pocket. This data
confirms that the nature and the distance of the other polar residue even far from the
PSB can stabilized the charge. The comparison of these two crystal structure shows
that while Cys53 is found in a single conformation, pointing toward the ionone ring in
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4H.

In

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W, Cyc53 adopts two conformations. This is
quite consistent with what has been seen in most other mutants, where the larger Trp58
leads to a more-red shifted spectrum than does Phe58.

134

R58W

R58F

Figure
III-32:
The
superimposed
KL:T51V:T53C:R58W:T29L:Y19W:Q4H
KL:T51V:T53C:R58F:T29L:Y19W (pink).

crystal
structure
(purple)

of
and

III-5-3: The Crystal Structure of Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4A
and Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R Mutant Bound to Retinal

Since the Q4H mutation leads to a blue shift while virtually all other mutations at
Gln4 resulted in red shifted spectra, it was necessary to confirm structurally that these
mutants also give the trans conformation. To this end structures were sought. Since
mutants with aromatic residues at position four had failed to crystallize, and the proteins
were found generally to be less stable, the crystallization of a Q4A mutant was
attempted, with the assumption being that less steric bulk may lead to a more stable
135

protein and better odds of crystallization. This was born out as well-diffracting crystals
were obtained of Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4A.

Further, since

mutation of Gln4 to a positively charged amino acid lead to even further red shifting,
culminating in the most red shifted mutant obtained to date (2013), a structure of Q4R
or

Q4K

mutant

was

also

sought.

To

this

end

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R was crystallized and its structure
determined to high resolution. Both mutants crystallized in the P1 space group, with 2
molecules

in

the

asymmetric

unit.

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R
crystallographic

Rfree

=

20.08

%

and

Rfac

The

crystallographic

was
=

refined

17.07

%.

at
The

data
1.5Å
structure

of
with
of

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4A bound to retinal was determined and
refined at 1.5 Å and Rfac = 18.70% and Rfree = 22.49%. All four Q4 mutant structures
are very similar. A trans iminium is observed in all four, and the conformation of the
chromophore is also very similar. Taken together, it appears that Gln4 results in the
iminium cis conformation while mutation of this residue to almost anything else leads to
the usually more stable trans iminium conformation. The trans conformation gives a
more red shifted spectrum and a lower PSB pKa (by about 1 pH unit), due to the loss of
stabilizing interactions with both Trp106 and the water mediated interaction with Gln4.
Figure III-33 shows the superimposed structure of these three mutants.

136

Y19W
Q4H
Q4R
Q4A
T29L
T39L
R58W
T53C
K40L

Figure
III-33:
The
most
red
shifted
mutants
(such
as
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R) also have the lowest pKa.

Inspection inside the binding pocket of the Q4R mutant shows that Arg4 with is
buried in the N-terminus of the protein and does not directly interact with the PSB, which
explains why the pKa is not severely depressed (Figure III-34). Arg58 makes hydrogen
bonds with the main chain carbonyls of Thr1, Lys109 and Asp91. Several ordered water
molecules, seen in virtually all other hCRBPII mutants are displaced by the position of
the Arg4 guanadinium. Thus the guanadinium is buried in a relatively hydrophilic region
of the binding cavity, explaining how the protein is capable of remaining relatively stable
upon burial of a large positively charged side chain. Even though the guanadinium of
Arg4 is relatively far from the chromophore (more than 6Å from the PSB), introduction of
the positive charge leads to a significant red shift of almost 10 nm relative to the Q4W
mutation (Table III-12) and a reduction in the PSB pKa (6.5 relative to between 7.2 and
7.9 in other Q4 mutants (Table III-12)).
137

Asp91
Q108K
2.93Å

2.70Å
2.87Å
Thr1

Q4R

Figure

III-34:

The

crystal

structure

of

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R. The side chain of Arg4 is
buried in the N-term of the protein.

Changing the conformation of the PSB from cis to trans also leads to a
translation of the chromophore of about 1Å. This translation allows the side chain of
Trp58 flip into the binding site, packing against the chromophore perfectly. This was
also observed in all of the crystal structures with the Q4 mutations (Figure III-33). Figure
III-33

illustrates

that

the

orientation

of

Trp58

in

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R covers the binding pocket of hCRBP
more efficiently. Further Crys53 adopts a single conformation, with its side chain
pointing toward the ionone ring of the chromophore, while in all of the other mutants
with both T51V and T53C mutations the side chain of Cys53 has two conformations.

138

a.
Trp106

3.9Å

4.3Å
5.2Å

5.5Å
Leu117

Leu40

b.

Ilu42

Trp106

4.8Å

4.5Å

3.8Å
4.9Å

Leu115

Leu40
Ilu40

Figure III-35: a. The π-cation interaction of Trp109 with cis-iminium. b. Transiminium.

139

The conformation of Cys53 may be correlated to the accessibility of the binding
pocket to the bulk media as the binding pocket in the trans mutants is even better
sequestered due to the conformation of Trp58 by closing the binding pocket the side
chain points to the ionone ring which probably carries more positive charge than when
the binding pocket is more open to the aqueous media, then Cys53 has two different
conformations. Another consistent difference seen in the Q4 mutants is in the water
network around Gln38 and Gln128. In the crystal structure of Q108K:K40L and most of
the other mutants with Gln4 one water molecule is found between Gln38 and Gln128.

4.46 Å

Gln4
Q4R

K40L

T53C

R58W

Thr1

T51V
2.99Å

2.87Å

W1
3.2Å
3.38Å

W2

2.97Å

T29L

Y19W

Figure
III-36:
The
superimposed
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W (cyan).

140

structure
(green)

of
and

However, in the structures of the Gln4 mutants, two water molecules are
observed (Figure III-35). One water makes hydrogen bond with two nitrogen’s of the
amide while the second one makes the two carbonyl groups. This is also observed in
the structure of wt hCRBPII.

a.
a.

b.

Figure III-37: a. The space filling model of the crystal structure of
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R mutant the chromophore is red.
b. The sphere structure of Q108K:K40L:T51V:T53C:R58W:T29L:Y19W mutant.

141

all-trans-retinal

3.37Å
3.23Å

2.88Å
2.72Å
2.77Å

Gln128

Gln38
Figure III-38: The details of the water network between Gln38 and Gln128 in
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4H mutant.

In the category of trans-iminum structures, the maximum red shifted mutant was
644 nm, while in the cis-iminums with Gln4; the most red shifted mutant absorbs at 606
nm. Therefore it seems that mutation of Gln4 was required to access the most red
shifted spectra. With the most red shifted of these the Q4R mutations, which introduces
a positive charge in the vicinity of the PSB.
Unfortunately

we

could

not

have

the

crystal

structure

of

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4E (590 nm) which is same as hepta
mutant Q108K:K40L:T51V:T53C:R58W:T29L:Y19W (591 nm). Modeling of this mutant
in Pymol suggests since the size of both glutamine and glutamic acid have the same
size therefore the glutamate side chain should be in the same place of glutamine and so
the water intermediate is most likely is in the same place and for this reason no red
shifting has been observed (Figure III-39).
142

3.36Å

3.46Å

Gln4
Q4E
Figure
III-39:
The
crystal
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W
(green)
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4E (blue).

structure
and
model

of
of

It seems that removing this water molecule is necessary for the most red shifting
absorbance and this strategy has been applied for the most red shifted mutant.

143

T29L

Y19W

Q108K
T51V

R58W

Q4A
K40L

T53C

Figure
III-40:
The
crystal
structure
of
Q108K:K40L:T51V:T53C:Y19W:R58W:T29L:Q4A with highlighted mutated
amino acids.

144

As previously discussed, adding the A33W mutation to the hepta mutant
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W (591 nm) the most

red shifted mutants

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:A33W with λmax 606 nm in the category of
cis-iminium was generated. The crystal structure of this mutant illustrates that A33W
can cover the binding pocket of hCRBPII and better isolate the pocket from bulk
solution. This isolation can prevent ions and water molecules to freely enter and exit the
binding pocket and produce a more uniform environment around the chromophore has
been

obtained.

In

the

category

of

the

trans-iminium

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R with 622 nm was the most red shifted
mutant. The combination of the most red shifted cis and trans-iminum mutant leads to
the most red shifted retinylidene–bound protein complex (644 nm) and the nona
mutants Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R:A33W was generated by Dr.
Wenjing

Wang.

Unfortunately

due

to

the

difficulty

of

crystallization

of

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R. The crystal structure of the most red
shifted mutant is not available. A model of the structure of this mutant (using Pymol) is
shown in Figure III-41.

145

b.

a.

Figure III-41: a. The sphere model of the most red shifted mutant
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R. b. The sphere model
based
on
the
crystal
structure
of
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W.

A

comparison

of

the

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R:A33W

model
and

the

of

structure

of

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R shows the binding pocket of the most
red shifted mutant to be completely covered with the chromophore deeply buried in the
binding pocket and to have almost no access to the bulk solution. The pKa of this
mutant

is

6.7

which

is

0.2

unit

higher

than

the

pKa

of

the

Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4R mutant (pKa 6.5). It seems that the
position of two aromatic residues (A33W and R58W) causes a small increase in the
pKa. The two important effects in the most red shifted mutants have important role:
1- Delocalization of the positive charge along the chromophore that can be
achieved by removing all the polar residues around it. We also need to consider
the water molecules in the binding pocket as some of the most polar groups in
the binding pocket.
146

2- Isolating the binding pocket from the bulk solution and preventing the water or
any charged ions to enter the binding pocket.

III-6: The Water Network Between Gln38 and Gln128

In the crystals structures of WT and mutant forms of hCRBPII, several ordered
water molecules are found inside the binding pocket. One water network is located
between the side chains ofGln38 and Gln128, which are almost 4Å away from the

polyene chain of retinal. It seems that these water molecules have a very crucial
role in the stabilization of pKa and also in wavelength regulation. In the structure of WT
and all of the Gln4 mutants (Q4H, Q4R and Q4R) two ordered water molecules are
suited between Gln38 and Gln128, while in the structures of other mutants a single
water is observed. We observed that mutation of the polar residues around the
chromophore to less polar or hydrophobic groups typically lead to bathochromic shift
(for example T51V, T53V, Q4W, Y19W R58W) but surprisingly, Mutations of Gln38 and
Gln128 to hydrophobic residues have a dramatic effect on both pKa and wavelength. As
is listed in table III-13 mutations of one or both to hydrophobic residues lead to a
decrease in the pKa and dramatic blue shifting. In fact mutation of both Gln38 and
Gln128 to non polar residues (entry 4 in table III-13), results in the largest hypsochromic
shift generated to date, neutralizing simultaneously the effect of all of the mutants that
give a red shift. Mutation of Gln38 to a charged or polar residue also leads to
hypsochromic shifts but the effect is not as dramatic as with non polar residues. Without
crystal structure it was suspected that by mutation of Gln38 or Gln128 the water
147

mediates as removed but fortunately the crystal structure of one of the hCRBPII mutant
with 5-demethyl-retinal has been solved. This chromophore is all-trans-retinal analogue
that has 6-s-cis conformation for ionone ring. This analogue was synthesized by Ipek
yapici in Professor Borhan group for study on the effect of cis and trans conformation on
the wavelength regulation. However the chromophore is not the same as retinal but as it
is shown in figure III-40 they are structurally are very similar and therefore we can study
on the water molecules in this structure. Figure III-39 shows the crystal structure of
Q108K:K40K:T51V:R48W:Q38F (The mutant has been generated by Ipek Yapici in
Professor Borhan’s group) bound to 5-demethyl all-trans-retinal in P1 space group has
been determined.

Table III-15: Mutation on Gln38 and Gn128 in hCRBPII
Mutants

λmax(nm)

pKa

Q108K:K40L:R58W:T51V:T53C:T29L

585

7.9

Q108K:K40L:R58W:T51V:T53C:T29L:Q38M

513

7.5

Q108K:K40L:R58W:T51V:T53C:T29L:Q128L

532

5.8

Q108K:K40L:R58W:T51V:T53C:T29L:Q128L:Q38M

504

7.2

Q108K:K40L:R58W:T51V:T53C:T29L:Y19W:Q4F

613

7.7

Q108K:K40L:R58W:T51V:T53C:T29L:Y19W:Q4W:Q38E

590

6.0

Q108K:K40L:R58W:T51V:T53C:T29L:Y19W:Q4F:Q38W

577

5.9

As is shown in this figure in both chains the water molecule has been observed
that makes a hydrogen bond to the Gln128. This data shows at least in this mutant the
148

water molecules are not removed and it is possible that in the other Gln38 mutants this
water is available, therefore the polarity of this water molecule seems to be important.
The environment can effect on the dielectric constant of water. The dielectric constant of
water drops inside the protein from 80 to 10-3.6 The polarity of these water molecules
when they make hydrogen bond to Glutamine is different when they are near
hydrophobic side chain of Luecine. Therefore the polarity of these water molecules is
very important, especially since it is located almost in the middle of the polyene chain,
disturbing the polarity of them can leads to less delocalization of the charge on along
the chromophore. More study on this water network is preformed by Meisam Nosrati in
Professor Geiger’s group.

Q108K
3.49Å

4.03Å
2.2Å

2.79Å

3.98Å
2.92Å

Q38F

Gln128

Figure III-42: The details of the water network in the binding pocket of
Q108K:K40K:T51V:R48W:Q38F mutant. The water network is originally
between Gln38 and Gln128.

149

III-7: The Effect of the Overall Electrostatic Potential and the Water Networks and
Interaction with Chromophore
The different rhodopsins are able regulate the wavelength from 420 nm to 560
nm with just one single chromophore, 11-cis-retinal. It was the first time that by
combination of mutations (maximum 9 different mutations) one protein could modulate
the full visual spectrum from 425 nm to 6 nm in one chromophore (all-trans-retinal)
Figure III-43. Combination of the rationalized mutations and X-ray crystallography
provide wide range of information to explain the mechanism of the wavelength

Normalized Absorbance

regulation in a rhodopsin mimic.

KD
KS
KL
KL-T53C
KL-R58F
KL-T51V
KLWC-Y19W
KLV-R58W
KLVC-R58W
KLVW-Q4W
KLVCWLWR
KLVCWLWR-A33W

0.8

0.6

0.4

0.2

450

525

600

380 425 474 482 508 533 570 577 590

675
613

622 644

retinal

Figure III-43: Full spectrum of hCRBPII mutants. a. UV-vis spectra of different hCRBPII
mutants bound with all-trans-retinal. b. Protein solution of different hCRBPII mutants
incubated with all-trans-retinal.
150

As was mentioned before the electrostatic calculation on the different rhodopsins
shows, that overall electrostatic potential of the protein around the chromophore is
important to modulate the wavelength. The different crystal structures of mutants helped
us to calculate the overall electrostatic potential of the protein around the chromophore.
For tuning the wavelength, we need to have a closed and isolated binding pocket, which
is exactly like rhodopsin structure. In hCRBPII as mimic of rhodopsin R58W mutation
and additionally A33W mutations cover the binding pocket and separate the
chromophore from the bulk solutions.
Therefore in these electrostatic calculations the dielectric constant of water was
assigned for the mutations with and without these two mutations.
The electrostatic calculation illustrates that for at the most blue shifted structure
the chromophore region has more positive charge, while the PSB has more negative
electrostatic potential. However this pattern is not reversed for the “super red shifted”
mutant. As shown in the Figure III-43-b for super red shifted spectrum we need to have
a uniform electrostatic potential around the chromophore. The calculation the overall
electrostatic potential of each mutants illustrate the different mutants with different
electrostatic potential have different absorbance. However looking closely to the details
of the each crystal structures shows that the mutants that have effects on the water
networks in the binding pocket have a dramatic influence on the wavelength while the
ones which do not have any interactions with the water networks are not able to have a
large effect on the wavelength. For example removing the water network near the PSB
leads to extra red shifting. Unfortunately the crystal structures do not show the direction
of the hydrogen atoms in water molecules and also the polarity of the water molecules,
151

because as it was shown above changing the polarity of the water molecules in the
binding pocket (Gln38 and Gln128 mutations) can lead to dramatic blue shifting.

a.

b.

Figure III-44: Electrostatic projections (kbTec-1) using the crystal structure of a. KL
(508 nm). b. KL:T51V:T53C:Y19W:R58W:T29L:Q4R (622 nm) hCBRP II mutants
(for each calculation both sides of the chromophore, all-trans-retinal, are shown).

III-8: Effect of Conformation of the Ionone Ring on the Wavelength Regulation

As it was explained the effect of conformation of the chromophore was
suggested as an important role in the different absorption maxima in PSB-bound to
protein. The more planer chromophore leads to more conjugation of the p-orbitals and
as the result, more red shifting is observed. One of the most common conformational
differentiations is believed to happen along the C6-C7 single bond. By twisting this

152

single bond the double bond of the ionone ring cannot have overlap with the rest of the
double bond of the backbone. Therefore hypsochromic shift will be observed.
Based on these theories Ipek Yapici in Prof. Babak Borhan group designed the series of
mutations. The idea was introducing the series of mutations in hCRBPII bound to alltrans-retinal that can change the conformation on the ionone ring from 6-s-trans to 6-scis and sturdy in the wavelength of these mutants. Since cis conformation is slightly
(0.6 Kcal/mol) is more stable than the trans one, therefore theoretically by providing
enough space around the ionone ring, the C6-C7 single bound can rotates and 6-s-cis
conformation will be formed. The different mutants have been generated and I solved
the structure of two different mutants. Dr. Camille Watson from Professor Geiger’s lab
also worked crystal structures of some of those mutants.

III-8-1: The crystal Structure of Q108K:K40L:T51V:R58Y:Y19W:T53L Bound to
Retinal

The crystals structure of Q108K:K40L:T51V:R58Y:Y19W mutant illustrated that
the chromophore adopts the 6-s-cis conformation. It is not clear that why introduction of
R58Y mutation leads to cis conformation, however as it was explained in the structure
of this mutant the position of side chain of Phe16 is moved and enough space was
provided for free rotation along C6-C7 bond. Based on these data Ipek Yapici designed
Q108K:K40L:T51V:R58Y:Y19W:T53L mutant. The crystals structure of this mutant was
determined in P1 space group with two molecules per asymmetric units. The data were
refined at 1.2Å resolution with crystallographic Rfree = 20.83 % and Rfac = 27.93%. Since
153

the ionone ring of Q108K:K40L:T51V:R58Y:Y19W has 6-s-cis conformation The idea of
this mutant was introduction a bulky group (T53L) that is able to makes crowded around
the gem dimethyl group, force it to rotate and lock the chromophore in the cis
conformation, however the crystal structure of this mutants clearly illustrated that the
ionone ring has trans conformation and the gem dimethyl group flipped Leu53 back. It
was not very surprising to find out that the protein could not able to dictate the
conformation to the chromophore and from the first time, it was not clear why
Q108K:K40L:T51V:R58Y:Y19W adopts the cis conformation. The superimposed
structure of Q108K:K40L:T51V:R58Y:Y19W and Q108K:K40L:T51V:R58Y:Y19W:T53L
mutants shows that the chromophore is exactly is the same position expect the ionone
ring (RMS is 0.394Å for 133 amino acid residue). The only difference in the structure of
these two mutants is the position of Phe16. The side chain of this residue is 0.4Å more
far from the chromophore. This distance might able to makes enough space for the
chromophore to rotate.

154

a.

Y19W

3.61Å

3.71Å

Phe16

Q108K

4.14Å
R58Y
T53L

Y19W

b.
Phe16

L77

Q108K
R58Y

T53L
Thr53

Figure III-45: a. The crystal structure of Q108K:K40L:T51V:R58Y:Y19W. b.
The superimpose structure of Q108K:K40L:T51V:R58Y:Y19W (cyan) and
Q108K:K40L:T51V:R58Y:Y19W:T53L (blue).

155

III-8-2: The Crystal Structure of Q108K:K40L:T51V:R58W:Y19W:T53L:L77T Bound
to Retinal

The crystal structure of Q108K:K40L:T51V:R58W:Y19W:T53L:L77T mutant with
all-trans-retinal was determined at 1.17Å in P1 space group with four molecules per
asymmetric unit. This mutant was generated by Ipek Yapici in Professor Borhan‘s
group. The data were refined with crystallographic data Rfree = 20.32% and Rfac =
24.15%. This mutant has been generated based on the idea of forcing chromophore to
adopt the 6-s-cis conformation. From one side mutation of Thr53 to Leucine which is
more bulky group and makes steric interaction with the gem-dimethyl group and the
other side mutation of Leu77 to threonine, which is less bulkier and makes enough
space for the chromophore to rotate. The crystal structure of this mutant shows the
chromophore adopts the same 6-s-trans conformation. Interestingly it was observed
Leu53 side chain flipped back to the beta-sheet to avoid the clash with the gemdimethyl of chromophore. Figure III-43 shows the crystal structure of this mutant and it
is clear the chromophore adopts the perfect trans conformation. So far no mutant has
been generated to force the chromophore to adopt another conformation. Dr. Camille
Watson in Professor Geiger’s group refined other structure for this project.

156

L77T

3.86Å
Q108K

R58W

T53L

Figure III-46: The crystal structure of
Q108K:K40L:T51V:R58W:Y19W:T53L:L77T with the highlighted residues.

III-9: Conclusion
In this study on the wavelength regulation in a cytosolic protein as a rhodopsin
mimic and based on more than two hundred mutations and more than twenty crystal
structure, we can conclude that, usually the mutations that have effect on the ordered
water molecules have usually the most significant effect on the wavelength spectrum in
both red and blue shifted. We could not find the relation between the conformation of
the chromophore and wavelength regulation. However in this study, the most red shifted
mutants also have the most planer chromophore, but there were not any correlation
between the torsion angles and wavelength. Other than that based on the crystal
structures of the mutants, it is concluded that the direct counter anion is not necessary
for this system and the PSB can be stabilized in different mechanism such as π-cation
interaction or hydrogen bond with a polar group. We were also able to show creating a
157

hydrophobic area around the entrance of binding pocket can leads to isolation of the
binding pocket from the bulk solution. In this case study water mediate was responsible
for stabilization of the PSB. The electrostatic calculation on based on the crystal
structure of hCRBPII mutants clearly shows that for super red shifted spectrum, evenly
distribution of charge is necessary.

158

REFERENCES

159

REFERENCES

1.

Baasov, T., Sheves, M., Alteration of pka of the bacteriorhodopsin
protonated schiff-base - a study with model compounds. Biochemistry 1986, 25
(18), 5249-5258.

2.

Rosenber.B; Krigas, T. M., Spectral shifts in retinal schiff base
complexes. Photochem. Photobiol. 1967, 6 (10), 769-&.

3.

Baasov, T.; Sheves, M., Model compounds for the study of
spectroscopic properties of visual pigments and bacteriorhodopsin. J. Am. Chem.
Soc. 1985, 107 (25), 7524-7533.

4.

Mogi, T.; Marti, T.; Khorana, H. G., Structure-function studies on
bacteriorhodopsin .9. substitutions of tryptophan residues affect protein-retinal
interactions in bacteriorhodopsin. J. Biol. Chem. 1989, 264 (24), 14197-14201.

5.

Wang W, Wang, Wenjing, Nossoni Z, Berbasova T, Watson C, Yapici I,
Lee, Kin S, Vasileiou C, Geiger J, Borhan B (2012) Tuning the Electronic
Absorption of Protein-Embedded All-trans-Retinal. Science 338 (6112):1340-1343.

6.

J. J. Dwyer et al. , Biophys. J. 79 , 1610, 2000.

160

CHAPTER IV
A Single Mutation Resulted an Extensive 3D Domain Swapped Dimerization in
HCRBPII

IV-1: Introduction

Human Cellular Retinol Binding Protein II (hCRBPII) is a stable, cytosolic,
monomeric protein responsible for the intracellular transport of retinol and retinal. It is a
member of the intracellular lipid binding protein (iLBP) family.1 CRBPII and other
members of this family have been shown to be structurally remarkably stable to
mutation.

Herein we report that a single tyrosine to tryptophan mutation causes a

unique and extensive domain swapped dimer. The mutation is not in the hinge loop
region but resides four amino acids away. The domain swapped region encompasses
almost half the protein, representing one of the largest domain swapped regions so far
discovered. The domain swapped dimer represents a novel structure with the mouths
of the two binding sites facing each other to produce a new cavity that stretches across
the dimer. Since the monomeric form of the mutant represents the thermodynamically
lower energy species, the domain swapped dimer represents a kinetically trapped
species that does not interconvert to monomer at neutral pH and room temperature.
Structural studies of two forms of the domain swapped dimer suggest a possible
mechanism for its formation. Three-dimensional (3D) domain swapping is a process in
which two or more monomer protein molecules exchange identical structural elements
to form dimers or higher-order oligomers. It was first observed in ribonuclease by the
161

Eisenberg’s group in 1995.2 To date over 40 different proteins are reported to show 3D
domain swapping.3 This mechanism, combined with gene duplication, may be
responsible for the evolution of a number of larger protein domains from smaller ones.
In a recent study 80% of protein domains were found to have potentially evolved in this
way. Domain swapping can also cause aggregation, amyloid and fibril formation, which
lead to protein misfunction.4,5 It is believed that protein misfolding and aggregation is the
cause of diseases such as Parkinson’s disease,6,7 Alzheimer disease,8,9 diabetes,11
Huntington’s disease,11 Gaucher’s disease and many others. However dimerization and
oligomerization of can also lead to some advantages such as increasing enzyme activity
due to the larger active and binding sites, larger surface area and therefore more
binding affinity, etc.12 The mechanism of domain swapping is not completely
understood, but it usually occurs during protein expression at high concentrations and
when the energy barrier between an “open” and “closed” monomer for any reason
decreases. This effect can be caused by pH, concentration, temperature, mutation and
ligand binding. Previously it was reported that domain swapping usually occurs at high
protein concentration.14,15 The first and most studied domain swapped protein was
Bovine pancreatic ribonuclease (Rnase A).1 The first two crystal structures of this
enzyme showed two distinct dimers depending on the concentration of mild acid. The
minor dimer forms by swapping the N-terminal a-helixes while the major one forms by
swapping the C-Terminal b-strands. So far more than 290 different structures of domain
swapped proteins are available in the protein data bank (PDB) and the number of these
proteins is still increasing.13 In most cases the exchanged region is located at the C- or
N-terminus of the protein. However in a few cases the exchanged part is in the middle
162

of the sequence. Domains swapped structures have been observed previously by
introducing point mutations in some proteins. For example a single site mutation G55A
in protein L from Peptostreptococcus Magnus leads to 3D domain swapping.16
Human cellular retinol binding protein II (hCRBPII) is a relatively small (133
residue), soluble, cytosolic, monomeric protein that plays an important role in
metabolism of vitamin A and is responsible for intracellular retinol transport. A member
of the intracellular lipid binding protein (iLBP) family, the structure consists of a tenstranded b-barrel, and two short a-helices that cover the binding pocket (Figure IV-1-a).
The ligand is buried deeply within the binding site.17 The structures of both hCRBPII and
other members of this family have been shown to be remarkably resilient to mutation.
For example more than 150 structurally stable mutants of hCRBPII have been
characterized in our lab in the course of our studies of protein-chromophore
interactions.18
Herein we report the first domain swapped structure of a mutant of human
cellular retinol binding protein II (hCRBPII). The extent of domain swapping is one of the
largest reported, involving almost 50% of the protein structure. This is, to our
knowledge, the first example of domain swapping caused by mutation outside the hinge
loop region. Further, we show that domain swapped dimerization is favored by high
expression levels in vivo and high concentrations in vitro. Further, melting studies show
that the domain swapped dimer is the less stable, kinetic product of refolding.

163

IV-2: Results and Discussion

HCRBPII was redesigned to be used as a rhodopsin mimic to study the
mechanism of wavelength tuning18. The first step in this process was to introduce a
lysine inside the binding pocket that would react with all-trans-retinal to produce an
Schiff base. Inspection of the crystal structure of the hCRBPII retinol complex shows
Gln108 to be hydrogen bonded to the hydroxyl group of the ligand and properly
positioned for nucleophilic attack when mutated to lysine, leading to the Q108K mutant.
Lys40, proximal to the Schiff base, was then mutated to leucine to remove a positive
charge proximal to the Schiff base and thus promote protonated Schiff base (PSB)
formation. The Q108K:K40L mutant resulted in stable PSB formation with retinal (ref).
The crystal structure of this mutant bound to all-trans-retinal at 1.7Å (PDB entry 4EXZ)
shows the PSB chromophore following a trajectory through the binding pocket similar to
that of Retinol in the wt hCRBPII retinol complex. The residues situated around the
chromophore were then selected for systematic mutation in an effort to understand the
mechanism of wavelength tuning. Tyr60 is located in beta-strand C near the loop that
connects beta-strands C and D (Figure IV-2). This residue is approximately 4Å from the
center of the polyene chain of all-trans-retinal (Figure VI-2-b).

164

It is a conserved residue in the CRBP family (found in hCRBPI, hCRBPII, hCRBPIII,
CRBPIV, zebrafish CRBP, rat CRBPI and II. In silico modeling suggested that mutation
of tyrosine to tryptophan, an electron rich and

b.b.

a.
C
B

2.53Å

D

2.53 Å

E
A
H
I

4.23 Å

F

4.23Å

G

J
2
1

Figure IV-1: a. The crystal structure of wt hCRBPII bound with all-transretinal with the highlighted Y60 position in purple. b. The relative position
of Tyr60 to all-trans-retinal.

polarizable residue, could result in a π-π interaction with the chromophore that could
potentially delocalize the positive charge of the PSB along the chromophore. This
modeling also suggested that the side chain of tryptophan would adopt a conformation
that would be parallel to the polyene of the chromophore. Therefore the mutant
Q108K:K40L:Y60W (KLY60W) was generated, expressed and purified as previously
described in chapter II.

165

a.

b.

Figure IV-2: a. Chromatogram of KLY60W mutant at 280 nm,
attributed to absorptions of tryptophan residues from the protein. b.
Chromatogram of KLY60W mutant at 280 nm after refolding with
urea.

a.
496 nm

b.
514 nm

native 80 mM elution
refolded 150 mM elution
native 150 mM elution

Absorption

Absorption

Wavelength (nm)

Wavelength (nm)

Figure IV-3: a. UV-vis spectral overlay of 40 mM salt elution (red) and 150
mM salt elution (blue) of Q108K:K40L:Y60W incubated with all-trans-retinal. b.
UV-vis comparison of native 40 mM elution and 40 mM elution of refolded 150
mM elution.

166

b.

a.

Figure IV-4: The melting point experiment in monomer and dimer of
KLY60W: a. monomer KLY60W b. KLY60W dimer.

The second purification step, Source-Q ion exchange chromatography,
surprisingly gave two protein fractions, one eluting at 80 mM NaCl, where hCRBPII
mutants typically elute, and a second eluting at 150 mM NaCl; both fractions contained
pure mutant CRBP II. The UV-Vis spectra of the all-trans-retinylidene PSB complex with
protein from each of the fractions gave distinct spectra.

The 80mM and 150 mM

fractions showed a λmax of 496nm and 514nm, respectively. (Figure IV-4A) Increasing
the expression temperature from 16°C to 30°C resulted in higher amounts of the 150
mM fraction protein while the amount of the 80 mM fraction was unchanged.

167

At low expression temperature the ratio of high to low salt fractions was 80 to 20 while
by increasing the temperature to 32°C this ratio reached 50% of each species.

2.53 Å

4.23 Å

Figure IV-5: Y60 position is located on starting of the β strand D.

Size-exclusion chromatography showed that while the 80 mM protein eluted as a
15 kD monomer, the 150 mM fraction eluted as a 30 kD dimer. No equilibrium between
the monomer and dimer forms was observed, even after several days at room
temperature. Melting curves of each fraction showed the 150 mM fraction to be less
stable than the 80 mM fraction, with the 150 mM protein denaturing at 65 °C, and the 80
mM protein stable to 95°C (Figure IV-5).
The fact that the two species displayed no evidence of interconversion at room
temperature indicates the barrier to interconversion to be relatively high. Therefore, to
show that the two protein species can indeed interconvert denaturation/renaturation
experiments were carried out. First the dimer species was denatured in 8 M urea at a
168

relatively low protein concentrations (0.03 mM) and rapidly renatured by dilution and
rapidly renatured by dilution. This results in renaturation occurring over a short period of
time and at a relatively low concentration of protein.

As shown by size exclusion

chromatography, this resulted in exclusive production of the monomeric species,
showing conclusively that the dimeric species can indeed be converted to the monomer.
On the other hand, renaturation by dialysis at 4 °C at and high protein concentration
(0.6 mM) gave a mixture of monomer and dimer, with the monomer amount about four
times that of the dimer. Together, this indicates that the monomeric species is still
thermodynamically favored over dimer, with the dimer species only forming at higher
concentrations, because oligomerization must be reasonably fast relative to monomer
folding. The UV-vis spectrum of the retinal PSB-bound refolded monomer was identical
to the originally isolated protein, strongly indicating the two proteins to be structurally
very similar if not identical (Figure IV-4-b).

IV-2-1:Dimerization by Applying a Single Site MutationY60W

Since both size exclusion chromatography and the crystal structure of the
Q108K:K0L mutant clearly show this species to be exclusively monomeric, with
dimerization only observed in the Q108K:K40L:Y60W (KLY60W) mutant, the natural
question was whether the single Y60W mutation was sufficient to produce a dimer. To
test this hypothesis Tyr60 was mutated to other hydrophobic and aromatic residues. To
this end Dr. Wenjing Wang and Zahra Assar generated Q108K:K40L:Y60F (KLY60F)
and single mutant Y60L. These two mutants were purified as previously described. In
169

both in both cases both proteins showed the dimer fraction and the ratio of dimer to
monomer in Y60L is about 80% of the protein concentration. However, the single Y60W
mutant exhibits behavior similar to that of KLY60W, eluting from the Source Q ion
exchange column in two distinct salt fractions (80 and 150 mM), with the monomeric
protein eluting at lower salt and the dimeric species eluting in higher salt, based on size
exclusion chromatography. Together these results strongly indicate that Y60W forms a
dimeric species similar to that of KLY60W.

IV-2-2: The Crystal Structure of Monomer Q108K:K40L:Y60W (KLY60W)

The crystal structure of the 80 mM species bound to all-trans-retinal as a PSB
confirmed that it was indeed monomeric hCRBPII, with a structure very similar to that of
wt hCRBPII and virtually all other mutant hCRBPII structures (Figure. 6A). The crystals
have the same space group and similar unit cell dimensions to those of the other CRBP
II mutants, with two molecules per asymmetric unit. Comparison of the hCRBPII
KLY60W-retinal complex to the hCRBPII Q108K:K40L-retinal complex shows that
though the overall fold of the two proteins is similar (with an RMSD 1.317Å) the
introduction of the Y60W mutation caused significant local structural changes. Trp60
adopts a conformation similar to that seen for Tyr60 in the Q108K:K40L mutant, but the
significantly larger size of the side chain has caused a significant displacement of the Cstrand (by 1.35Å) compared to either the Q108K:K40L mutant, or wt hCRBPII, serving
to slightly enlarge the binding cavity. This results in a rotation of the polyene portion of
the chromophore by almost 90° relative to that of the Q108K:K40L mutant, while the
170

ionone ring remains stationary. This results in a “6-s-cis’ conformation about the ionone
ring-polyene torsion (Figure IV-6-b). Since the 6-s-cis conformation is thought to be
more stable than the 6-s-trans conformation seen in Q108K:K40L hCRBPII. A possible
interpretation of our results is that the extra space provided by the displacement upward
of Trp60 relative to Tyr60 allows the chromophore to adopt this lower energy
conformation. 24

a.

b.

Figure IV-6. a. The overlaid crystal structure of Q108K:K40L mutant (light
purple), the chromophore is highlighted in hot pink with Q108K:K40L:Y60W
monomer (salmon) with highlighted chromophore in green. b. The highlighted
chromophore in both structures.
Mutation of Tyr60 to Trp removes the hydrogen bonds between Trp60 and
Glu72, which also has been observed in wt hCRBPII bound with retinol. It seems that
the loss of this interaction and insertion of the larger amino acid may lead to
displacement of the beta-C strand.

171

IV-2-3: Structure of KLY60W Domain Swapped Dimer

Crystals of the KLY60W dimer proved much harder to grow, appearing only after
several months. Crystals that produced reasonable diffraction were rare, and only one
incomplete data set was collected of this form (only 70% completion) due to decay in

Figure IV-7: The crystal structure of Q108K:K40L:Y60W.

the X-ray beam. The structure was nevertheless determined and refined to 2.5Å
resolution with two molecules per asymmetric unit. The electron density clearly shows
that a domain-swapped dimer has formed, with a crystallographic two-fold axis of
symmetry relating the two monomers in the dimer. Figure IV-7 shows electron density of
55-60 showing continuous density to the next molecule). Thus, the two monomers in the
asymmetric unit each give rise, through a crystallographic two-fold axis of symmetry, to
two independent dimers. An overlay of these two dimers (Figure IV-8-a) shows that
they are very similar to one another (RMS= 0.491Å).

172

The resulting domain swapped dimer represents a novel structural architecture.
The opening to the binding cavities are now face to face, creating a continuous internal
cavity that stretches from one side of the dimer to the other (Figure IV- 8-b). The
bottom of the structure forms a continuous beta sheet that also spans the length of the

a.

b.

Figure IV-8: a. The two chains of the Q108K:K40L:Y60W dimer overlaid
(blue, chain B, green chain A showing a minor structural difference. b. The
surface structure of the dimer, showing the extensive binding cavity.

dimeric structure, reminiscent of the continuous beta sheet that spans the two direct
repeats of the TATA-binding protein. A search of the various structural fold databases
(SCOP 1.75) reveals no other protein of similar architecture.
The major conformational changes that lead to domain swapping in this structure are
not confined to a single residue in this case, but involve large differences in main-chain
torsion angles of several of the residues between 55-61. Interestingly, Trp60 has flipped
from the interior of the protein to the exterior in both dimers, exposing the hydrophobic
residue to solvent. Together the conformational change in these residues give rise to
the domain swapped dimer as the remainder of the structure is virtually identical to the
monomeric KLY60W structure (Figure IV-10). The domain swapping is extensive in this
173

case, involving almost half of the molecule, from residue 1-56 which includes 3 beta
strands and two helices. This corresponds to an 83% extent of swapping (as defined in
the

domain

swapping

3D

knowledge

database

(http://caps.ncbs.res.in/3dswap/help.html). The structure of the KLY60W domain
swapped dimer provides a potential explanation for the bathochromic shift of the
absorption spectrum of the bound retinylidene PSB in the dimer versus monomer. It
was previously shown that sequestration of the retinal PSB chromophore from bulk
solvent invariably leads to red-shifted spectra.18 The face to face juxtaposition of the
mouths of the binding cavities in the domain swapped dimer result in a binding cavity
significantly isolated from solvent relative to the monomer, likely explaining the
significantly red-shifted spectrum of the dimer relative to monomer.

b.

a.

Figure IV-9: a. The crystals structure of the domain swapped Y60W dimer
(chain A). b. The 2Fo-Fc electron density map around the residues between
50-60 (contoured at 1σ) illustrates the continuous electron density that
crosses the dimer interface.

174

IV-2-4: Structure of the Y60W Dimer

To ascertain whether a single mutation could produce the dimer, the hCRBPII Y60W
mutant was generated, expressed and purified.

Figure IV-10: The overlaid structure of KLY60W monomer (green) and Y60W
dimer, chain A in red and chain B in blue.

This mutant behaved very similarly to KLY60W, eluting in two fractions from the
SOURCE Q ionic exchange column, one at 80 mM salt and a second at 150 mM salt.
As previously observed, the 80 mM fraction eluted consistent with a monomeric protein
by size exclusion chromatography, while the 150 mM fraction eluted exclusively as a
dimer. The dimeric protein crystallized relatively easily, this time in the P21212 space
group, with one dimer in the asymmetric unit (Figure IV-9-a). The electron density map
clearly showed continuous electron density across the two domains, confirming the
175

presence of a domain swapped dimer (Figure IV-9-b).

The high quality of the

crystallographic data (to 1.7Å) allowed a far more precise picture of the nature of the
domain swapping. The structure of the two domains of the dimer are each very similar
to monomeric KLY60W hCRBPII with RMS= 0.651Å (Figure IV-10).

Y60

N-terminus
C-terminus

Figure IV.11: The two chains of the Y60W dimer overlaid (red, chain A, yellow
chain B showing a major structural difference in these two chains.

Significant differences in the two structures are localized to the region spanning Ser55
to Trp60.

The most significant change occurs in the phi/psi angle of Thr56.

Surprisingly, comparison of the two dimer structures (KLY60W and Y60W) shows that
the relative orientation of the two domains is significantly different in the two structures.

176

As shown, when the C-terminus (from 56-133) of the two domains are overlaid, the Nterminus is displaced by as much as 10Å in the two structures. Moreover, overlay of the
four chains (1 from each of the two distinct dimers in the KLY60W structure and the two
from the Y60W dimer) 3 of the four appear to be quite similar while the fourth (Y60W
molecule A) displays a large displacement relative to the other three. Table 1 shows
the backbone torsion angles in and around the hinge region for the monomer structures.
As the data indicates the torsion range in the dimer structures (wt CRBP and KLY60W
monomer) is not a wide range, while It is immediately evident from Table IV-2 that there
are significant differences in the torsion angles that give rise to domain swapping in the
four chains.

Most notable is the fact that only one angle, the psi angle of Thr56,

deviates significantly from its monomer value in Y60W chain A. It is therefore possible
to produce the domain swapped conformation by rotation about a single torsion angle in
the hinge region. It would appear that the additional large torsion angle deviations in
Y60W chain B are required for mating of the two chains to form the dimer. On the other
hand, several torsion angles deviate significantly from their monomeric values in the two
KLY60W chains. Seven deviate substantially from the monomer values. It is important
to remember that a crystallographic two-fold axis relates two KLY60W chain A’s to
make a single dimer, while a second crystallographic two fold axis relates two KLY60W
chain B’s to make a second dimer. The two additional mutations Q108K and K40L, are
far from the hinge region and do not appear to contribute to the conformational
differences in the domain swapped dimers (Crystal structure of Q108K:K40L, PDB entry
4EXZ).18

177

Figure IV-12: The overlaid of the four chains of the Q108K:K40L:Y60W
dimer (green, chain A and blue, chain B) with Y60W dimer (yellow, chain B
and red chain A).

Therefore, it seems the additional torsional rotations are the result of the
crystallographically enforced two-fold axis. This leads to the conclusion that the four
distinct molecules seen in the two mutant crystal structures are the result of crystal
packing, allowed by the significant flexibility in the hinge region, with only the large
rotation in the psi angle of Thr56 a requirement of domain swapping. Besides the
difference in backbone torsion angles, there is also a significant difference in the
dispensation of the mutated Trp60 residue. While Trp60 is flipped out of the binding
pocket and exposed to solvent in both molecules of the KLY60W mutant structure, and
in chain B of the Y60W structure, it is rotated inward in chain A of Y60W. As previously
described, Tyr60 is completely buried inside the binding pocket in all known structures
of hCRBPII. It is important to remember here that while there is a two-fold axis of
178

symmetry that relates the two chains that make up a dimer in the KLY60W structure,
this is not the case in Y60W, because the two molecules that make up a dimer are quite
different in this structure.

Table IV-1: The monomer psi and phi angles for 10 amino acids
Residue

Monomer range
Phi

Monomer range
Psi

THR/51

-118.75

-128.77

144.78

150.60

LYS/52

-131.96

-142.20

119.02

127.14

THR/53

-106.75

-114.46

118.75

132.93

THR/54

-124.59

-130.11

146.34

152.92

SER/55

-141.03

-171.97

164.46

177.47

THR/56

-66.02

-78.04

-7.84

-28.65

PHE/57

-76.60

-96.89

-41.71

-100.34

ARG/58

-93.84

-165.78

116.95

159.20

ASN/59

-90.83

-120.79

145.26

161.72

TRP/60

-141.30

-151.03

113.78

153.91

Table IV-2: The phi and Psi angles in KLY60W and Y60W dimer structures
Residue

KLY0W-dimer chain A
phi
psi

KLY0W-dimer chain B
phi
psi

Y60W-dimer chain A
phi
psi

Y60W-dimer chain B
phi
psi

THR/51

-127.211

117.631

-115.367

147.721

-126.267

159.521

-121.246

143.204

LYS/52

-96.111

148.782

-124.911

131.791

-151.915

145.905

-130.250

121.593

THR/53

-138.837

115.942

-106.746

156.263

-123.728

139.693

-116.726

129.982

THR/54

-96.897

152.451

-149.731

115.200

-131.953

159.823

-121.777

135.471

SER/55

-124.717

130.491

-125.271

120.419

-145.887

136.661

-96.772

145.545

THR/56

-130.999

131.099

-150.457

-166.945

-65.386

146.007

-138.037

138.997

PHE/57

-145.711

153.215

-128.353

-160.388

-115.020

-74.831

-91.315

131.447

ARG/58

-148.754

-68.664

-128.200

-60.396

-130.341

152.976

-122.702

125.515

ASN/59

172.364

-105.767

76.781

-118.583

-124.810

136.948

-87.716

178.362

TRP/60

60.384

75.556

50.189

100.065

-136.578

141.385

-82.181

141.376

179

IV-2-5: The Nature of Domain Swap Dimerization in HCRBPII

Though the detailed mechanism of domain swapping in hCRBPII is not known,
the data so far obtained points to several important aspects of the process. First the
fact that the KLY60W monomer is significantly more stable than its domain swapped
dimer indicates that in this case domain swapping is a kinetic and not thermodynamic
product in the process. The fact that no dimer to monomer interconversion is ever
observed for either mutant indicates that the dimer represents a kinetic trap in the
monomer refolding process.

Both in vitro refolding experiments, where monomer

formation is strongly favored at low concentrations, but less favored at high
concentrations, and in vivo expression, where the monomer is strongly favored at lower
temperatures where the expression rate is slower and the concentration of protein in the
process of folding is expected to be lower, indicate that high concentrations of partially
folded intermediates is required for dimer formation. Indeed folding at higher
concentrations often favors the production of domain-swapped oligomers. Third is the
fact is mutation of Tyr60 to more hydrophobic residues is both necessary and sufficient
for domain swapped dimerization, though this residue is 4 amino acids away from the
hinge loop. Fourth, it is found in the non-native flipped out conformation in both domain
swapped dimers. Together this data indicates that folding of the protein happens in two
steps, with the first step being the independent folding of the N- and C-terminal regions
independently. The N-terminus forms 4 beta-stands (bA-bD) and two helices (a1,2) and
the C-terminus forms 8 beta-strand (bE-bJ). Formation of these two subdomains is
thought to be the fast step of protein folding and after this step the slow step which is
180

formation of the tertiary structure of the protein occurs. In this step the hinge region
plays the crucial role. Formation of the monomer has been slowed by the presence of
Trp60 and that the rate-limiting step in folding for these mutants may be the flipping of
the two sheets onto one another, allowing the extended sheet intermediate to dimerize
when the concentration is high. Static structures cannot explain the details of kinetic
transition states, but the presence of the flipped out conformation of Trp60 in the dimers
may provide a clue. It is possible that the transition from this flipped out conformation to
the flipped in conformation seen in the monomer may play a role in slowing the rate of
monomer folding. Mutation of Tyr60 to smaller aromatic residues (KLY60H and
KLY60F) does not result in domain swapping, indicating that the size of the residue is
important to the mechanism. Further, mutation of another residue in the same region to
Trp (R58W) does not lead to dimerization, indicating that the Tyr60 position may have
special significance. However, this in no way implies that the native protein undergoes a
similar pathway, as the presence of Trp60 may be responsible for the presence of the
folding intermediate required for domain swapping. By applying this mutation the energy
barrier between the open and closed states is increased and therefore the intermediate
form in the open state was trapped. Nevertheless the domain swapped dimerization of
Y60W CRBP II shows that a single mutation, far from the actual hinge region, can give
rise to an extensive domain swapping event, involving almost 50% of the protein.

181

C-terminus
N-terminus

high concentration

N-terminus

hinge loop
C-terminus

N-terminus

low concentration

dimer CRBP

N-terminus

C-terminus

partially folded protein
C-terminus

formation of the tertiaty structure

monomer CRBP

Figure IV-13: The proposed mechanism for formation of monomer and dimer in
Y60W hCRBPII.

IV-3: Experimental Procedure

The cloning, expression and purification of hCRBPII mutants was performed as
previously described. To increase the ratio of the dimer protein the expression
temperature was increased to 32°C overnight. Size exclusion chromatography was
performed using a Superdex 200 16/75 column (GE Health Sciences), packed in-house
in a buffer containing 10 mM Tris, 100 mM NaCl, pH 8.0.
182

PCR Primers for Y60W Mutant:
Forward: 5ʼ- CACATTCCGCAACTGGGATGTGGATTTCAC-3ʼ
Reverse: 5ʼ- GTGAAATCCACATCCCAGTTGCGGAATGTG-3’

IV-3-1: Crystallization of HCRBPII Mutants

Each mutant protein was concentrated to 10 mg/mL in a buffer containing 10 mM
Tris, 100 mM NaCl, pH 8.0. The complex of protein with all-trans-retinal was prepared
by adding 2 equivalents of retinal dissolved (30 mM retinal in ethanol) in ethanol and
incubated for one hour at room temperature. The ethanol concentration should not be
more than 10%. Crystals were grown using the hanging drop vapor diffusion method
using 1 mL of protein solution and 1 mL of crystallization solution in the drop and 1 mL
of crystallization solution in the reservoir. All crystallization trials were wrapped in
aluminum foil to limit light exposure to the chromophore. The best crystals grew in 5
days and reached maximum size in one week using a crystallization solution of 30%
PEG 4000, 0.1M CH3COONa.3H2O pH 4.6, 0.1 M CH3COONH4. The crystals were
soaked in a cryoprotectant solution (30% PEG 4000, 0.1M sodium acetate pH 4.6-4.8,
0.1 M ammonium acetate, 15-20% glycerol), flash frozen in loops in liquid nitrogen and
stored in a liquid nitrogen dewar prior to data collection at the synchrotron.

183

IV-3-2: Data Collection and Data Refinement

All of the diffraction data were collected at the Advanced Photon Source (APS)
(Argonne National Laboratory IL) LS-CAT, (sector 21-ID-D,F,G) using a MAR300
detector and 1.00 Å wavelength radiation at 100K. The initial diffraction data were
indexed, processed and scaled using the HKL2000 software package.19 The structures
were determined by molecular replacement using MOLREP in the CCP4 software
package and human CRBPII (PDB entry 2RCQ) as a search model.20,21 Though the
apparent space group of the KLY60W structure was P21212, due to almost perfect
merohedral twinning, the actual space group was P21. The initial electron density map
was produced by using REFMAC5 in the CCP4 package. All of the, rebuilding,
placement of water molecules and etc. were done using COOT.22 The chromophore
was built into the monomeric KLY60W mutant structure using the jligand program to
generate restraints.23 All of the refinements were done by REFMAC (restrained
refinement) until the final R-factor was between 18-20%.

184

Table IV-3: X-ray crystallography data and refinement statistics
Q108K:K40L:Y60
W (monomer)
P1

Y60W (dimer)
P21212

Q108K:K40L:Y60W
(dimer)
P1

a(Å)

30.08

62.62

36.62

b(Å)

36.05

109.77

61.07

c (Å)

63.92

36.32

64.30

α(° )

90.94

90.00

89.82

β(° )

91.99

90.00

90.13

δ(° )

113.93

Space group

90.00

90.18

Molecules per
Asymmetric Unit
Total reflection

2

2

2

124534

345311

214807

Unique Reflection

41751

55068

Completeness (%)
Average I/σ
Rmerge(%)
Resolution (Å) (Last Shell)
Rwork/ Rfree (%)

92.8(68.0)

a

50.30(3.58)

a

98.6(100.0)

a

66.13(7.71)

a

5.2(31.7)

4.1(29.3)
21.30-1.47(1.52-1.50)
19.86/25.06

a

50.00-1.70(1.73-1.70)
18.17/23.94

19239
62925
72.9(57.6)
29.48(1.38)
4.3(49.8)

40.00-2.50(2.54-2.50)
21.48/34.20

RMSD from Ideal Values

Bond Length (Å)

0.025

0.023

0.012

Bond Angle (°)
Average B factor
Number of water
Molecules

2.36
19.98
228

2.11
19.92
217

1.45
30.17
33

Values in the parenthesis refer to the last resolution shell

185

a

a
a

REFERENCES

186

REFERENCES

1.

Ong De (1984) a novel retinol-binding protein from rat - purification and partial
Ccharactrization. Journal of Biological Chemistry 259(3):1476-1482.

2.

Bennett MJ, Schlunegger MP, & Eisenberg D (1995) 3D domain swapping - a
mechanism for oligomer assembly. Protein Science 4(12):2455-2468.

3.

Liu L GA (2012) Protein and Nucleic Acid Folding: Domain Swapping in Proteins
(Oxford Academic Press).

4.

Dobson CM (2003) Protein folding and misfolding. Nature 426(6968):884-890.

5.

Selkoe DJ (2004) Folding proteins in fatal ways (vol 426, pg 900, 2003). Nature
428(6981):445-445.

6.

Shtilerman MD, Ding TT, & Lansbury PT (2002) Molecular crowding accelerates
fibrillization of alpha-synuclein: Could an increase in the cytoplasmic protein
concentration induce Parkinson's disease? Biochemistry 41(12):3855-3860.

7.

Fandrich M, Fletcher MA, & Dobson CM (2001) Amyloid fibrils from muscle
myoglobin - Even an ordinary globular protein can assume a rogue guise if
conditions are right. Nature 410(6825):165-166.

8.

Sunde M & Blake C (1997) The structure of amyloid fibrils by electron
microscopy and X-ray diffraction. Advances in Protein Chemistry, Vol 50: Protein
Misassembly, Advances in Protein Chemistry, ed Wetzel R), Vol 50, pp 123-159.

9.

Cleary JP, et al. (2005) Natural oligomers of the amyloid-protein specifically
disrupt cognitive function. Nature Neuroscience 8(1):79-84.

10.

Scherzinger E, et al. (1997) Huntingtin-encoded polyglutamine expansions form
amyloid-like protein aggregates in vitro and in vivo. Cell 90(3):549-558.

11.

Hayden MR, Tyagi SC, Kerklo MM, & Nicolls MR (2005) Type 2 diabetes mellitus
as a conformational disease. JOP : Journal of the pancreas 6(4):287-302.

12.

Marianayagam NJ, Sunde M, & Matthews JM (2004) The power of two: protein
dimerization in biology. Trends in Biochemical Sciences 29(11):618-625.

13.

http://caps.ncbs.res.in/3dswap/help.html.
187

14.

Minton AP (2000) Implications of macromolecular crowding for protein
assembly. Current Opinion in Structural Biology 10(1):34-39.

15.

Ellis RJ (2001) Macromolecular crowding: an important but neglected aspect of
the intracellular environment (vol 11, pg 114, 2001). Current Opinion in Structural
Biology 11(4):500-500.

16.

Single-site mutations induce 3D domain swapping in the B1 domain of protein L
from Peptostreptococcus magnus. Structure 9(11):1017-1027.

17.

Macdonald PN & Ong DE (1987) BINDING SPECIFICITIES OF CELLULAR
RETINOL-BINDING PROTEIN AND CELLULAR RETINOL-BINDING PROTEIN,
TYPE-II. Journal of Biological Chemistry 262(22):10550-10556.

18.

Wang W, et al. (2012) Tuning the Electronic Absorption of Protein-Embedded
All-trans-Retinal. Science 338(6112):1340-1343.

19.

Z. Otwinowski, W. Minor, Macromolecular Crystallography, Pt A 276, 307
(1997).

20.

S. Bailey, Acta Crystallographica Section D-Biological Crystallography 50, 760
(1997)

21.

A. Vagin, A. Teplyakov, J. Appl. Crystallogr. 30, 1022 (1997).

22.

G. N. Murshudov, A. Vagin, E. J. Dodson, J. Appl. Crystallogr. 30, 1022 (1997).

23.

G. N. Murshudov et al., Acta Crystallogr. D Biol. Crystallogr. 67, 355 (2011).

24.

Rajput, J.; Rahbek, D. B.; Andersen, L. H.; Hirshfeld, A.; Sheves,
M.; Altoe, P.; Orlandi, G.; Garavelli, M., Probing and modeling the absorption of
retinal protein chromophores in vacuo. Angew. Chem. Int. Ed. Engl. 2010, 49
(10), 1790-1793.

188

CHAPTER V
Mutation, Over-expression, Purification, Crystallization,
Structural solution and Refinements of CRBPII Mutants

X-ray

diffraction,

V-1: Material and Methods
Expects few mutations (T51I, Y60W, Q108H, T51V) the rest of mutations were
done by Dr. Wenjing Wang in Prof. Borhan group at department of chemistry in
Michigan State University. The all the mutations were constructed by Stratageneʼs
QuickChange Site-Directed Mutagenesis Protocol, which is a high performance
protocol that takes only one day for mutation. This method is widely used for
insert/delete and mutations. This method is performed by pfuTutrbo DNA polymerase
and temperature circle in PCR instrument. The hCRBPII was cloned into pET17b
expression vector by Dr. Wenjing Wang using NdeI and XhoI restriction enzymes.

Table V-1: The PCR protocol for wt hCRBPII mutants

DNA
template

Primer
forward

Primer
reverse

pfuTurbo

dNTP

DNA
polymerase
50 ng

20 pmol

20 pmol

1 unit

Bufer

dd water

Total
volume

Up to 50 µL

50 µL

(x10)

1µL

189

5µL

PCR Protocol for hCRBPII mutagenesis:
1 x

95°C

3 min

20 x

95 °C
(Tm-4)°C
72 °C

30 sec
1 min
5 min

1
1

72 °C
25 °C

10 min
5 min

x
x

Tm stands for the melting temperature of the primer for each mutation, which is
calculated using the following website:
http://www.basic.northwestern.edu/biotool/OligoCal.html

Wt HCRBPII Gene Sequence
ATGACGAGGGACCAGAATGGAACCTGGGAGATGGAGAGTAATGAAAACTTT
GAGGGCTACATGAAGGCCCTGGATATTGATTTTGCCACCCGCAAGATTGCA
GTACGTCTCACTCAGACGCTGGTTATTGATCAAGATGGTGATAACTTCAAGA
CAAAAACCACTAGCACATTCCGCAACTGGGATGTGGATTTCACTGTTGGAGT
AGAGTTTGACGAGTACACAAAGAGCCTGGATAACCGGCATGTTAAGGCACT
GGTCACCTGGGAAGGTGATGTCCTTGTGTGTGTGCAAAAGGGGGAGAAGA
GAACCGCGGCTGGAAGAAGTGGATTGAGGGGGACAAGCTGTACCTGGAGC
TGACCTGTGGTGACCAGGTGTGCCGTCAAGTGTTCAAAAAGAAGTGA

190

Wt HCRBPII Amino Acid Sequence
TRDQNGTWEMESNENFEGYMKALDIDFATRKIAVRLTQ
TLVIDQDGDNFKTKTTSTFRNYDVDFTVGVEFDEYTKS
LDNRHVKALVTWEGDVLVCVQKGEKENRGWKKWIEGD
KLYLELTCGDQVCRQVFKKK
The PCR product is treated with DpnI endonuclease. This enzyme is specific to
digest the unmutated and parental DNA and methylated DNA and the target is GATC
sequence. This enzyme digests and destroys the entire parental DNA and methylated
ones and leaves the mutated one. The process takes for 20 min at 37°C.
The Primes Used Were as Follows:
T51I
For: GGT GAT AAC TTC AAG ATA AAA ACC ACT AGC ACA
Q108H
For: AAC CGC GGC TGG AAG CAC TGG ATT GAG GGG GAC
Q108L
FOR: GGC TGG AAG TGG ATT GAG
F57S
For: ACA AGC ACA TCC CGC AAC TAT
Transformation of Mutated Plasmid in to Competent Cells: The modified protocol
by Sambrook et al has been used for all transformation. The sterile condition should be
maintained during the procedure. 100µL aliquot of competent cell (DH5α) was thawed
on ice. Depends on the quantity of PCR between 1 to 10µL (10-100 ng) DNA was
added to the cells and mixes gently (without pipette it up and down). The competent
cells should never be vortexed. The cells incubated in ice for 30 min and they were heat
191

shocked in the water bath at 42°C for 20 seconds. The sample immediately needs to
return back to the ice for 2 min. 0.5 ml of prewarmed (37°C) LB media was added and
incubated at 37°C for one hour. After one hour the solution spread on the LB/agar plate
containing appropriate antibiotic (Amp 100 µg/mL). The plates finally were incubated for
16-20 hours at 37°C.
Plasmid Purification: A single colony was grown up in 3 mL LB ( containing AmP)
and grows over night and then purified using the Qiagen® mini plasmid purification kit
with min prep protocol. The recovered DNA was eluted with 30 mL of EB buffer and
analyzed by nanodrop UV-vis for determination of the purity and concentration. The
DNA is sent for sequencing to The Research Technology Support Facility (RTSF) at
MSU using T7promoter.
Protein Over-Expression: Human recombinant hCRBPII was expressed and
purified by Wang et al described procedure. The expression plasmid pET-17b was
transformed into E.coli strain BL21( DE3)pLysS cells. The transformed cells were grown
at 37°C in LB/agar plates containing antibiotics (Amp100). A single colony was inoculated
in 50 ml of LB containing the same amount of antibiotics and grown overnight with
shaking at 37°C. This culture then transferred to 1L of LB with the same antibiotics.
When OD600 reached between 0.6 to 0.8. The sample was induced by 0.1 nM IPTG and
grows over night at room temperature (22°C). The cells were harvested next day (500
rpm, 20 min, 4°C) by centrifugation. The cells can be stored at -20°C for later
purification.

192

Protein Purification: All the steps of purification were done I cold room or using
ice bath. The frozen cells (obtained from 2L) were resuspendedn In 80 mL of (10mM
Tris-HCl, 10mM NaCl pH8.0). The resuspended cells were lysed by sonication (three 1
min with 1 min relaxation between each) and centrifuge for 20 min, 13000 rpm. The
supernatant was subjected to FastQ, which is a strong anion exchanger (Q-sepharose
Fast Flow Resin) column chromatography. (100 mL resin with capacity ~ 20 mg protein,
Amersham Biosciences). The column at first per-equilibrated with the 100 mL of 10mM
TRIS-HCl pH 8.0 buffer. After the flow through was collected the column was washed
using the same buffer. 30 µL of sample is collected for running polyacrylamide gel
electrophoresis (SDS-PAGE). The concentration of the washed solutions was measured
with biorad protein assay. Finally the protein was eluted with 40 mL of 10mM TRIS_HCl,
100 mM NaCL pH 8.0 buffer. All of the samples were analyzed on 16% SDS-PAGE to
test for purity of sample. The protein samples pooled together and desalted using
Centriprep® centrifugal filter (10,000 MWCO) for the next step of purification. For the
last step of purification FPLC system ion exchange (BioLog Due Flow, Biorad) using
Source15Q (Amersham Biosciences) has been used.
Extinction Coefficient Determination of HCRBPII: For calculation the extinction
coefficient of hCRBPII Gill and Von Hippel method has been used.
Theoretical extinction coefficient for denatured protein εd =
5690 x Trp + 1280 x Try + 120 x Cys
in wt hCRBPII : 4 Trp, 4 Tyr, 2 Cys
and therefore the theoretical εd for hCRBPII is :
5690x4+1280x4+120x2 = 28,120
193

The absorption of the native protein and denatured protein (6M guanidine hydrogen
chloride) at 280 nm were measured at exact same protein concentration.
The extinction coefficient for the native protein is derived from the following equation:
εn = εd x (A280 native) ⁄ (A280 denatured).

V-2: Crystallization and Data Collection

All the crystals were grown using the hanging drop (vapor diffusion method). In
this method a drop (2-5µL) containing the mixture of protein and reservoir solution is
equilibrated against the reservoir solution (1 mL). At first around 200 different conditions
have been tried and the best crystals of the wt hCRBPII were grown in one of the
Hampton research crystal screen solution containing PEG 4000, ammonium acetate
and sodium acetate. The best crystals were obtained by screening around this condition
(30% PEG 4000, 1 M ammonium acetate and 1M sodium acetate pH 4.5) and both apo
and holo crystals were grown in the same condition. Since retinal is very sensitive to the
light all of the crystal boxes were wrapped in the aluminum foil to protect them against
light. Expect the crystals were grown after 3 weeks and reached to their maximum size
in one week. For preventing of degradation of retinal the crystals were flash frozen after
one week. The crystals were cryo-protected by soaking in the cryo-protectant solution
consisting the reservoir solution with 20% glycerol. The frozen crystals were taken to
the Advanced Photon Source (APS) synchrotron facility (Argonne, IL) for data
collection. All of the diffraction collections were collected under steam of nitrogen gas.
194

The collected data were integrated and scaled using the HKL2000 package at APS. The
processing and refinement were done at home.
Crystallization

of

all

Holo-mutants

Q108K:K40L:T51V:T53C:R58W:T29L:Y19Y:Q4R

HCRBPII

Expect
and

Q108K:K40L:T51V:T53C:R58W:T29L:Y19Y:Q4A:
All-trans-retinal was purchased from Sigma-Aldrich. The complex of protein and
retinal were prepared fresh before setting the crystallization box. Since retinal is a light
sensitive compound the stock solution of that was kept in black eppendorf tube in
freezer and all the process of incubation of that was done in dark and under the red
light. HCRBPII mutants cannot tolerate more than 10% of ethanol. Therefore the
saturated retinol solution in ethanol was prepared ~ 10-3 M. WT CRBP and the mutants
of this protein are very tolerant for concentration and can be concentrated more than 20
mg/mL but usually concentration of 15 mg/mL has been chosen. The lower
concentration of protein leads to longer for crystallization and also the smaller crystals
were obtained. The crystals with appear at room temperature after three days and
reached to the maximum size in a week. The procedure of the crystallization does not
need to be in dark and the crystals can be flash freeze at the normal light. The crystals
are colorful and the colors depend on the absorbance of the protein. All of the crystals
without R58 mutations are perfectly rectangular, while the one with R58F, R58Y and
R58W mutations form cluster. The diffraction data were collected at sector 21-LS-CAT
beam line D at Argonne National lab (IL). Since the crystals are in P1 space group for
each crystal 360 frames were collected with 1° oscillation. The distance of the detector
to the crystals was from 100 mm to 150 mm.
195

Crystallization

of

Holo-Q108K:K40L:T51V:T53C:R58W:T29L:Y19Y:Q4R

and

Q108K:K40L:T51V:T53C:R58W:T29L:Y19Y:Q4A mutants: These two mutants with low
pKa did not crystallized at the same crystal condition at room temperature and the
crystals grow at cold room (4°C) after more than two weeks. Because of the low pKa of
these two mutants the color of the crystals were light orange- yellow. The crystals were
flash frozen in with liquid nitrogen and cryo protectant solution. The diffraction data were
collected at Argonne National Lab sector 21, beam line D. for each crystal 360
diffraction data were collected with 1° oscillation.
Structure solution and refinement: All of the structure determinations and refinement
were performed using CCP4 (Collaborative Computational Project, Number 4, 1994)
package. The structure of the WT CRBP was determined using molecular replacement
(rigid body refinement) using crystal structure of wt hCRBPII (2RCQ) as the initial model
to produce the initial electron density map. The structures of the rest of mutants were
solved using the high resolution of the wt hCRBPII. All of the mutation and rebuilding
the structures and adding the waters were performed in COOT. All refinements were
carried out using REFMAC5 in CCP4 package. JLIGAND was used for generating the
ligands. The refinements statistics are reported in Table IV-2.

196

Table V-2: X-ray crystallography data and refinement statistics
Q108K:K40L:T51V

Space group

Q108K:K40L:T53C

Q108K:K40L:T51V
:R58F

Q108K:K40L:T51V:
T53S

P1

P1

P1

P2

a(Å)

29.739

29.947

30.057

53.917

b(Å)

35.931

35.890

35.882

36.608

c (Å)

64.187

63.866

64.163

65.651

α(° )

90.78

91.01

90.91

90.00
90.10

β(° )

92.56

92.24

91.21

δ(° )

113.22

113.59

113.76

Resolution (Å)

50-1.77(1.77-1.73)

a

50-1.64(1.64-1.59)

a

90.00

50-1.51(1.51-1.47)

a

50-1.57(1.57-1.53)

Rwork/ Rfree (%)

24.68/21.75

28.20/21.20

27.46/21.17

22.7/18.57

Total reflection

82430

132782

147350

119972

Unique Reflection

31572

59796

69283

Completeness (%)
Molecules per
Asymmetric Unit
I/σ
Rmerge(%)

92.9(94.1)

a

2
44.23(5.94)
4.8(29.0)a

97.3(84.0)

a

97.4 (82.5)

2
a

2

53.68(4.17)
3.3(23.2)

71252
a

a

a

53.18(2.9)

97.3(79.3)
2

a

4.1(26.9)

a

42.25(1.81)
5.1(40.8)

RMSD from Ideal Values
Bond Length (Å)

0.0332

0.0341

0.0257

0.0274

Bond Angle (°)
Number of water
molecules

2.687
174

2.730
251

2.410
287

2.384
380

Note: Values in the parenthesis refer to the last resolution shell

197

a

a

a

a

Table V-2 (cont’d)
Q108K:K40L:T51V:
Y19W:R58Y

Q108K:K40L:T51V:
Y19W:R58W

Q108K:K40L:T51V:T5
3S:Y19W:R58W:T20L

Q108K:K40L:T51V:T53
C:Y19W:R58W:T20L:Q
4H

P1

P1

P1

P1

a(Å)

29.748

29.964

29.870

30.918

b(Å)

36.030

35.914

35.691

35.978

c (Å)

64.782

64.568

64.915

64.483

α(° )

87.67

90.63

90.816

86.01

β(° )

88.06

91.28

88.726

86.56

Space group

δ(° )

64.72

113.89
a

66.106

65.17

Rwork/ Rfree (%)

28.87/22.36

31.04/23.16

27.00/20.85

21.98/18.18

Total reflection

91114

52210

68699

216594

Unique Reflection

41402

50012

36782

Completeness (%)

98.5(80.8)

Molecules per
Asymmetric Unit
I/σ
Rmerge(%)

98.2(59.6)

2

91.7(79.8)

2

51.56(3.07)
4.1(36.1)

a

50-1.56(1.56-1.53)

a

50-1.74(1.74-1.70)

a

50-1.68(1.68-1.64)

a

Resolution (Å)

a

a

2

35.70(2.33)
5.9(48.6)

a

a

a

96209
62925
a
96.5(61.8)
2

35.66(1.23)
4.3(44.6)

50-1.29(1.29-1.26)

a

a

48.61(3.6)
4.2(30.6)

RMSD from Ideal Values
Bond Length (Å)

0.0118

0.0231

0.0218

0.0325

Bond Angle (°)
Number of water
molecules

1.449
270

2.1733
207

2.080
282

2.556
380

Note: Values in the parenthesis refer to the last resolution shell

198

a

a

a

Table V-2 (cont’d)
Q108K:K40L:Y19W:R
58F:T53V

Q108K:K40L:T51V:T5
3C:R58W-Apo

Q108K:K40L:T51V:T5
3C:R58L

Q108K:K40L:T51V:R5
8Y:Y19W:T53L

P1

P3121

P1

P1

a(Å)

29.617

76.261

29.785

30.022

b(Å)

36.339

76.261

36.038

35.844

c (Å)

88.16

66.884

63.875

64.513

α(° )

90.67

90.00

89.21

90.67

β(° )

88.37

90.0

92.01

91.42

δ(° )

65.17

120.00

66.57

113.87

Space group

Resolution (Å)
Rwork/ Rfree (%)

50-1.77(1.77-1.72)

50-1.73(1.73-1.70)

a

50-1.84(1.84-1.76)

a

50-1.58(1.58-1.54)

24.17/21.45

29.13/21.91

27.71/20.93

Total reflection

87733

305066

88609

117570

Unique Reflection

36653

25252

60134

Completeness (%)
Molecules per
Asymmetric Unit
I/σ
Rmerge(%)

32.83/24.63

a

97.7(66.0)

a

2

a

a

a

2

81.08(4.3)
3.5(50.6)

48728

98.9(76.0)

1

47.58(3.8)
4.5(22.7)

100(100)

a

a

6.6(25.2)

a

a

52.26(3.15)
4.5(29.4)

RMSD from Ideal Values
Bond Length (Å)

0.0217

0.0324

0.0315

0.0226

Bond Angle (°)
Number of water
molecules

2.136
123

2.978
190

2.683
200

2.118
240

Note: Values in the parenthesis refer to the last resolution shell

199

a

2

48.49(3.21)

a

96.8(66.8)

a

a

a

Table V-2 (cont’d)
Q108K:K40L

Q108K:K40L:T51V:T5
3C:R58W:T29L:Y19W
:Q4R

Q108K:K40L:T51V:T
553C:R58W:T29L:Y
19W

P1

P1

P1

a(Å)

29.81

30.92

30.11

b(Å)

35.94

35.76

35.91

c (Å)

63.98

64.376

64.726

α(° )

90.83

86.40

90.93

β(° )

92.88

86.44

91.21

δ(° )

113.19

Space group

64.94

114.13

Molecules per
Asymmetric Unit
Total reflection

2

2

2

73367

145141

95388

Unique Reflection

27686

35955
a

92.2(72.5)

Average I/σ

40.38(1.27)

a

53.92(4.62)

a

47.27(4.71)

Rmerge(%)

0.036(0.55)

a

0.052(0.24)

a

0.034(0.213)

Rwork/ Rfree (%)

33.03-1.61(1.641.61)
21.25/26.57

32.36-1.5 (1.53-1.50)
17.07/20.08

91.0(79.8)

19.40/24.14

Bond Length (Å)

0.0205

0.0275

0.023

Bond Angle (°)
Average B factor
Number of water
molecules
PDB ID

1.95
30.86
164

2.38
17.17
303

2.09
24.53
238

4EEJ

4EFG

Values in the parenthesis refer to the last resolution shell

200

a
a

28.91-1.58(1.61-1.58)

RMSD from Ideal Values

4EXZ

62925

a

Completeness (%)

Resolution (Å) (Last Shell)

98.7(89.3)

30286
a

Table V-2 (cont’d)
Q108K:K40L:T51V:R
58W:Y19W:L77T:T53
L

Q108K:K40L:T51V:T5
3C:R58W:T29L:Y19W
:A33W

P1

P1

a(Å)

30.338

31.35

b(Å)

35.996

36.45

c (Å)

129.166

64.26

α(° )

89.92

90.04

Space group

β(° )

90.22

94.55

δ(° )

114.92

116.36

Resolution (Å)

50-1.46(1.46-1.40)

Rwork/ Rfree (%)

23.09/18.70

Total reflection

622288

Unique Reflection

246158

Completeness (%)
Molecules per
Asymmetric Unit
I/σ
Rmerge(%)

97.7(93.1)

20.04/24.24
142087
44292

a

93.5(80.4)

4

a

2

50.11(1.5)
4.7(10.6)

32.64-1.40(1.42-1.40)

a

a

47.93(2.70)
0.082(0.358)

a

a

RMSD from Ideal Values
Bond Length (Å)

0.0304

0.028

Bond Angle (°)
Number of water
molecules

2.558
530

2.38
279

Note: Values in the parenthesis refer to the last resolution shell

201

Figure V-1: The Ramachandran plot of Q108K:K40L mutant

202

Figure V-2: The Ramachandran plot of Q108K:K40L:T53C mutant

203

Figure V-3: The Ramachandran plot of Q108K:K40L:T51V:T53S mutant

204

Figure V-4: The Ramachandran plot of Q108K:K40L:T51V:R58F mutant

205

Figure V-5: The Ramachandran plot of
Q108K:K40:T51V:T53S:R58W:T29L:Y19W mutant

206

Figure V-6: The Ramachandran plot of Q108K:K40L:T51V:R58Y:Y19W mutant

207

Figure V-7: The Ramachandran plot of Q108K:K40L:Y60W dimer

208

Figure V-8: The Ramachandran plot of Q108K:K40L:Y60W monomer

209

FigureV-9: The Ramachandran plot of
Q108K:K40L:T51V:T53C:R58W:T29L:Y19W:Q4H mutant

210

Figure V-10: The Ramachandran plot of wt hCRBPII bound to all-trans-retinal

211

Figure V-11: The Ramachandran plot of Y60W dimer

212

Figure V-12: The Ramachandral plot of wt hCRBPII bound to all-trans-retinol

213