mm .

14.‘

 

[23.1111
3...... .r .
.

iiufistvt.
» 4.1

 

      

«753...»: ..'| . I ‘ .
.«Lttv . . A "44... . J .14! ‘AI’I Atfi

f zafiéfié g§§¢

    

o
i.

5
;

     

F

 

 

 

.W3
A
at

kl

k)

C
_.Q

LIBRARY
Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

X-RAY CRYSTALLOGRAPHIC STUDIES OF CELLULAR
RETINOIC ACID-BINDING PROTEIN II MUTANTS DESIGNED AS
RHODOPSIN MIMICS

presented by

XIAOFEI JIA

has been accepted towards fulfillment
of the requirements for the

doctoral degree in chemistry

 

 

 

Major Professor’s Signature V/

ll/ll/Og

Date

MSU is an affinnetive-action, equal-opportunity employer

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K'IProj/AccsPres/ClRC/DateDue.indd

X-RAY CRYSTALLOGRAPHIC STUDIES OF CELLULAR RETINOIC ACID-
BINDING PROTEIN II MUTANTS DESIGNED AS RHODOPSIN MIMICS

By

Xiaofei J ia

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Chemistry

2008

ABSTRACT

X-RAY CRYSTALLOGRAPHIC STUDIES OF CELLULAR RETINOIC ACID-
BINDING PROTEIN II MUTANTS DESIGNED AS RHODOPSIN MIMICS

By

Xiaofei Jia

Color vision is one of the most fascinating mechanisms of the human anatomy. It is
achieved by the collaboration of four visual pigments in the retina: the rod, red, green and
blue rhodopsins. Upon binding of 11-cis-retinal, the four rhodopsins exhibit distinct
absorption maxima, which altogether cover the entire visible spectrum and enable color
vision. The underlying opsin-chromophore interactions within each rhodopsin are
therefore of great interest and significance. Elucidation of these mechanisms, however,
has been hampered by difficulties associated with the overexpression, purification, and
routine crystallization of the membrane-bound rhodopsins. To overcome these
difliculties and advance our understanding of the wavelength regulation mechanism in
color vision, we have established a model system for rhodopsin, employing a mimic
protein, cellular retinoic acid-binding protein II (CRABPII).

Upon binding of a shorter retinal analog, Cls-aldehyde, several rationally designed
CRABPII mutants exhibited altered absorption maxima of the bound chromophore.
These mutants were characterized both biochemically and structurally. Chapter II of this
dissertation focuses on the structural insights obtained for four such holo-CRABPII
mutants based on the high-resolution crystal structures. The protonated Schiff base
formation between Lysl32 and the chromophore has been confirmed in all structures.

Two distinctive binding orientations have been observed in different C's-aldehyde-

CRABPII complexes. A single mutation at 121 is shown to play a key role in
determining the binding mode of the ligand. Based on the available high-resolution
structures, further rational design of CRABPII has led to the discovery that placing
negatively charged residues along the polyene chain of the chromophore leads to a clear
trend in wavelength regulation.

High-resolution crystal structures of six ape-CRABPII mutants have also been
determined. Chapter III details the discussion of how different mutations affect the
structural integrity of the protein. A water-mediated hydrogen-bonding network has been
observed in all apo-CRABPII structures, including both wild-type and mutants.
Mutations usually lead to altered water-mediated networks and yet do not disrupt the
overall structure of the protein. Mutations on the helical portal, however, tend to have
detrimental structural effects.

Chapter IV presents an attempt toward elucidating the direct interaction between
CRABPII and the retinoic acid receptor (RAR), which is responsible for the retinoic acid
delivery.

Chapter V discusses the efforts toward obtaining the first crystal structure of the
cellular retinaldehyde-binding protein (CRALBP), another important protein in vision.
CRALBP functions in the regeneration process of 11-cis-retinal from all-trans-retinol in
the retina pigment epithelium. Initial efforts toward crystallizing the full length
CRALBP were unsuccessful. Subsequent mutagenesis combined with biochemical
studies have helped identify suitable truncations at the N-terminal of this protein. Recent
crystallization trials with a truncated CRALBP complexed with a synthetic ligand have

shown encouraging results.

Copyright by
XIAOFEI JIA
2008

To my father

ACKNOWLEDGMENTS

First and foremost, I would like to thank my two advisors, professor James H.
Geiger and professor Babak Borhan. My journey of pursuing a doctoral degree in
chemistry would have been interrupted if not for Jim and Babak, who provided me the
opportunity to continue my study and advance myself in science. Jim and Babak share
some of the finest qualities as scientists and mentors. They both have great passion for
science and possess tremendous knowledge. Besides, what made them very special is
their joyful attitude toward science. They both put their interests and passion to science
in such a positive and contagious way that people around them are easily inspired. I
consider myself very lucky to work in the labs of these two great scientists/mentors and
learn a great deal from each of them.

I also feel very grateful to my colleagues working in the same project. Dr.
Chrysoula Vasileiou has been with this project since the very beginning and has been the
biggest factor for its success. Chrysoula not only contributes excellent scientific insights
into the research, but also coordinates this group of people in a very enjoyable, cordial
way. I especially thank her for all the suggestions she gave to me. Dr. Soheila
Vaezeslami worked on the structure determination of CRABPII mutants before me. Her
success certainly paved the way for me. I owe a very special thanks to Kin-Sing Lee.
Sing and I had the closest collaboration. We exchanged results and ideas on a regular
basis. His intellectual inputs as well as his hard work have greatly aided the growth of
the project and also helped me in doing my part of the research. Many of his data have

been incorporated and discussed in this dissertation. I would also like to thank Calvin

vi

Grant, Wenjing Wang, Camille Watson and Rafida Nossori who are working on other
aspects of this project. I appreciate all your collaboration and support over the years.

I would like to thank my labmates in the Geiger group. Dr. Stacy Hovde was the
one who helped me in everything when I first started in the lab. Her intelligence,
diligence, friendliness and her super love for Spartan hockey all impressed me very
much. It was very a pleasant experience working with her. Dr. Sara Cnude and Dr. Fang
Sheng helped me a great deal on getting started with crystallography. Their kindness and
patience had made my learning process very enjoyable. I also want to thank Lei Feng for
the discussions and friendship over the years. I owe a special thanks to Susan Wortas for
her taking so much time in proofreading my various writings. Her kindness and patience
are deeply appreciated. I also want to thank Remie Touma for bringing her energy and
positive attitude into the lab.

I would also like to thank the Borhan group members, especially Dr. Tao Zheng,
Dr. Jun Yan and Xiaoyong Li. I could not list all the names, but I certainly appreciate the
their friendship and support over the years.

I am deeply gratefirl to my friends, Dr. Jian Yi, Dr. Ningqing Ran, Dr. Yana Cen,
Dr. Jinsong Yang, Dr. Gang Hu, Dr. Justas Jancauskas and Xianying Jiang. I am truly
blessed by your friendship and support through all these years.

Last but not least, I thank my family for their unconditional love and support
throughout my life. My father, mother, grandma and my sister have always been there
for me whenever I needed them. My life would not be the same without everything they
have done for me. This dissertation is dedicated to my father, Ping J ia, who passed away

on December 6'", 2008, two days before my defense.

vii

TABLE OF CONTENTS

 

 

 

LIST OF TABLES
LIST OF FIGURES
ABBREVIATIONS
CHAPTER I. Introduction _ - - - -1
1.1. The Mystery of Vision ............................................................................................... 1
1.2. Rhodopsins as mammalian visual pigments: structure and function in
photoactivation. ............................................................................................................... 9
1.3. Other rhodopsins .................................................................................................... 12
1.4. Color vision ............................................................................................................. 15
1.5. Designing CRABPII mutants as rhodopsin mimics ................................................ 30
1.5.1. CRABPII general background. ........................................................................ 31
1.5.2. Designing CRABPII mutants as rhodopsin mimics. ....................................... 35
1.6. Cellular retinaldehyde-binding protein and the 11-cis-retinal regeneration. ....... 38
Reference ....................................................................................................................... 44
CHAPTER 11. Crystal Structures of Ligand-bound Cellular Retinoic Acid Binding
Protein 11 Mutants 62
2.1. Introduction. ........................................................................................................... 62
2.1.1. Need for a ligand that can be totally buried inside the CRABPII mutants. ..... 62
2.1.2. Engineered CRABPII mutants as switchable colorimetric fusion proteins.....65
2.2. C I 5-aldehyde-bound CRABPII mutant structures ................................................... 67
2.2.3. Protonated Schiff base stabilization. ............................................................... 74
2.2.4. Cis versus trans protonated Schiff base. ......................................................... 77
2.2.5. Flexibility of the helix-tum-helix lid. .............................................................. 79
2.2.6. An evolutionary correlation between CRABPII and plasma retinol-binding
protein? ...................................................................................................................... 82
2.2.7. Testing the point charge theory based on the C‘s-aldehyde-bound KLE-
R59W-CRABPII structure. ........................................................................................ 86
2.2.8. Conclusion ....................................................................................................... 87
2.3. Merocyanin-bound CRABPII mutant structure. ..................................................... 89
References ...................................................................................................................... 93
CHAPTER 111. Crystal Structures of the Apo-CRABPII Mutants -_ 98
3.1. Introduction ............................................................................................................ 98

viii

3.2. Results ................................................................................................................... 100

3.2.1. Mutations of conserved and non-conserved residues .................................... 100
3.2.2. The structure ofapo-R132KzR111L2L121EzR59E-CRABPII (apo-KLE-R59E)
................................................................................................................................. 102

3.2.3. The structure ofapo-R132KzY134F:R111L:L121E-CRABPII (apo-KF LE) 105
3.2.4. The structure ofapo-Rl32K:Y134F:R111L:L121D-CRABPII (apo-KFLD)

................................................................................................................................. 107
3.2.5. The structure ofapo-R132K:Y134F:R111L:L121D:T54V-CRABPII (apo-
KFLD-T54V) ........................................................................................................... 108
3.2.6. The structure ofapo-R132KzR111L:T54E-CRABPII (apo-KL-T54E) ........ 109
3.2.7. The structure of apo-R132KzR1 1 1L:A32E-CRABPII (apo-KL-A32E) ........ 111

3.3. Discussion ............................................................................................................. 112
3.3.1. Rigid overall fold of the apo-CRABPII mutant structures ............................ 112

3.3.2. A32E mutation is detrimental to the a2 helix in apo-KL-A32E ................... 114
3.3.3. Water-mediated network acts as a pillar inside the cavity ............................ 116
3.3.4. The hydrophobic cavity allows binding of BTP ............................................ 117

3.5. Conclusion ............................................................................................................ 118
Reference ..................................................................................................................... 1 19

CHAPTER IV. Investigation of the Interaction between Cellular Retinoic Acid
Binding Protein 11 and Retinoic Acid Receptor Gamma Ligand Binding Domain.122

 

4.1. Background ........................................................................................................... 122
4.2 Results and discussion ........................................................................................... 128
4.2.6. Circular dichroism of RARy-LBD and fluorescence titration of RARy-LBD
with retinoic acid. .................................................................................................... 133
4.2.7. Titration of KLE-CRABP II with retinal followed by UV. ........................... 135
4.2.8. In vitro binding assay using RARy-LBD directly from Ni-NTA resin. ........ 135
4.3. Conclusions .......................................................................................................... 13 7
References .................................................................................................................... 138
CHAPTER V. Preliminary X-Ray Crystallographic Studies of Cellular
Retinaldehyde Binding Protein _ - - 140
5.1. Background ........................................................................................................... 140
5.2. Crystallization of the full-length CRALBP ........................................................... 140
5.2.1. His-tagged full-length CRALBP crystallization ............................................ 140
5.2.2. Non-fusion full-length CRALBP crystallization. .......................................... 142
5.2.3. Full-length CRALBP with a cleavable His-tag. ............................................ 144
5.3. Crystallization of the N-terminal-truncated CRALBP .......................................... 145
5.3.1. Truncation at 43/44 on the N-terminal of CRALBP. .................................... 145
5.3.2. Discovery of the ideal N-terminal truncation site .......................................... 147
5.3.3. Enhancing the expression and purification of functional CRALBP .............. 158
5.3.4. Crystallization trials. ...................................................................................... 159

ix

5.4. Conclusion. ........................................................................................................... 160

 

Reference ..................................................................................................................... 1 61
CHAPTER VI. Experimental - ...... -- ..... 162
6.1. Cloning, mutagenesis, expression and purification .............................................. 162
6.1.1. Cloning, mutagenesis, expression and purification of CRABPII mutants. ...162
6.1.2. Cloning, expression and purification of RARy-LBD. ................................... 164
6.1.3. Cloning, mutagenesis, expression and purification of CRALBP. ................. 165
6.2. Protein crystallization .......................................................................................... 1 70
6.2.1. Co-crystallization of CRABPII mutants bound with ligands. ....................... 170
6.2.2. Crystallization of apo-CRABPII mutants. ..................................................... 173
6. 3. Data processing and structure determination ...................................................... I 75
6.3.1. Structures of the Cls-aldehyde-bound CRABPII mutants. ............................ 175
6.3.2. Structure of the merocyanin-bound KLE-R59W-CRABPII mutant. ........... 176
6.3.3. Structures of the apo-CRABPII mutants. ..................................................... 177
6.4. Protein characterization ....................................................................................... 193
6.4.1. General procedure for UV-vis titrations of CRABPII/CRALBP with retinoids.
................................................................................................................................. 193
6.4.2. Determination of the extinction coefficients of proteins. .............................. 193
6.4.3. Reductive amination procedure. .................................................................... 195
6.6.4. Circular Dichroism of protein samples. ......................................................... 195
6.4.5. Fluorescence. ................................................................................................. 196
6.4.6. General protocol followed for Fluorescence Quenching and K, determination
experiments. ............................................................................................................. 198
References .................................................................................................................... 203

LIST OF TABLES

Table 1-1. Rhodopsins of various species. ......................................................................... 17
Table II-l. CRABPII mutants bound with C1 5-aldehyde. ................................................. 64
Table II-2. Residues around the ionone rings. ................................................................... 86

Table II-3. Absorption maxima observed for the KLE-X-CRABPII series of mutants ....87

Table II-4. UV maximum absorption of the merocyanin-bound CRABPII mutants. ....... 89
Table VI-l. X-ray data collection statistics of the ligand-bound CRABPII mutant crystals.

....................................................................................................................... 178
Table VI-2. Structure refinement statistics of the ligand-bound CRABPII mutant crystals.

....................................................................................................................... 179
Table VI-3. X-ray data collection statistics of the apo-CRABPII mutant crystals .......... 180
Table VI-4. Structure refinement statistics of the apo-CRABPII mutant crystals. ......... 18]

xi

LIST OF FIGURES

Figure H. The human eye ................................................................................................... 2

Figure [-2. A. rod and cone cells in retina; B. rhodopsins embedded in the disks in
photoreceptor cells. ............................................................................................ 3

Figure I-3. Schematic representation of the intermediates observed in the isomerization of

11-cis-retinal to all-trans-retinal in rhodopsin ................................................... 5
Figure I-4. Photoactivation cascade ..................................................................................... 7
Figure I-5. Crystal structure of bovine rod rhodopsin. ...................................................... 10
Figure I-6. Absorption profiles of different human rhodopsins. ....................................... 16
Figure I-7. Model compounds used to study wavelength shifts controlled by distance

between the PSB and the counter anion. ......................................................... 20
Figure I-8. The point charge model ................................................................................... 21
Figure [-9. Hypothesis of steric interactions leading to wavelength regulation ................ 23
Figure I-10. Vicinity of retinal in cone pigments as shown in the homology models ....... 24

Figure I-l 1. Illustration of the sites of mutations for converting the red rhodopsin to a

green-rhodopsin-like pigment .......................................................................... 27
Figure I-12. Illustration of the sites of mutations for converting the green rhodopsin to a
red-rhodopsin-like pigment. ............................................................................ 27
Figure [-1 3. Illustration of the sites of mutations for converting the rod rhodopsin to a
red-shifted pigment .......................................................................................... 28
Figure I-14. Crystal structure of retinoic acid bound in WT-CRABPII ............................ 32
Figure I-15. Water-mediated hydrogen bonding interactions observed in the molecule A
of the apo-WT—CRABPII crystal structure ............................................ 34
Figure I-16. Crystal packing observed in the structure of apo-WT-CRABPII. ................. 35
Figure I-17. All-trans-retinal bound in KLE-CRABPII. ................................................... 37
Figure [-18 The visual cycle. ........................................................................................... 39
Figure I-19. Stereoview of the structure models built for CRALBP. ................................ 42

xii

Figure II-l. Retinoic acid bound in wild-type-CRABPII. ................................................. 63
Figure II-2. Structures of the related compounds. ............................................................. 63
Figure II-3. Ovelay of the ligands bound in different binding orientations ....................... 69

Figure II-4. 2Fo-Fc electron density map calculated at 1.0 o for bound ligands. A. C15-
aldehyde bound to Lysl32 (upper) and bis-tris-propane (lower) in KLE-
R59W; B. Cis-aldehyde bound to Lysl32 in KL-A32E. ................................. 69

Figure II-5. Interactions between the PSB and the residues nearby in different C15-
aldehyde-bound CRABPII mutants. ................................................................ 70

Figure II-6. A. The structures of KLE-Rt and Cls-aldehyde-KL-A32E are overlaid.
Glul2l from KLE-Rt (magenta) and Cl5-aldehyde from hoIo-KL-A32E
(green) are shown; B. Overlay of the C15-aldehyde bound in three CRABPII
mutants: KL-A32E (magenta), KL-T54E (green), and KFLDV (yellow) ....... 72

Figure II-7. A. Proposed nucleophilic attack of lysine amino group to the aldehyde
following the Bilrgi-Dunitz model; B. Surface representation of CI5-aldehyde
(yellow) bound in KL-A32E (gray). ................................................................ 78

Figure II-8. Stereoview of the overlay between holo-KL-A32E (green) and halo-KLE-
R59W (magenta) .............................................................................................. 80

Figure II-9. Overlay between C‘s-aldehyde-bound KLE-R59W (blue) and apo-KLE-
R59E (green). A. Model of C15-aldehyde in apo-KLE-R59E; B. Crystal

structure of Cis-aldehyde bound in KLE-R59W. ............................................ 81
Figure II-10. Stereoview of the overlay of Cls-aldehyde-bound KL-A32E (green) and

retinol-bound RBP (pink) ................................................................................ 84
Figure II-l 1. C15-aldehyde bound in the KLE-R59W-CRABPII mutant. ......................... 87
Figure II-12. Structures of merocyanin and MES. ............................................................ 89
Figure II-13. Merocyanin bound in KLE-R59W-CRABPII mutant .................................. 91

Figure III-1. Stereoview of the mutation sites involved in these CRABPII mutants ...... 101

Figure III-2. Structure of theapo-KLE-R59E-CRABPII ................................................ 104
Figure III-3. Structure of the apo-KFLE-CRABPII ........................................................ 106
Figure III-4. Structure of the apo—KFLD-CRABPII. ....................................................... 107
Figure III-5. Structure of the apo-KFLD-T54V-CRABPII ............................................. 109
Figure III-6. Structure of the apo-KL-T54E-CRABPII ................................................... 110

xiii

Figure III-7. Structure of the apo-KL-A32E-CRABPII .................................................. 111

Figure III-8. Stereoview of the overlay of all apo structures together ............................. 113
Figure IV—l. all-trans retinoic acid and 9-cis retinoic acid. ............................................ 122
Figure IV-2. Structural organization of nulear receptors ................................................. 123

Figure IV-3. Comparison of the crystal structures of the apo-RXRa (A, PDB: lLBD) and
holo-RARy (B, PDB: 2LBD) ligand-binding domains reveals the ligand-

induced conformational change ..................................................................... 124
Figure 1V-4. The synthetic retinoid analog CD270. ........................................................ 127
Figure IV-5.1n vitro binding assay .................................................................................. 129
Figure IV-6. Detergent titration of the protonated Schiff base formed between KLE-

CRABPII and retinal ...................................................................................... 131
Figure IV-7. Fluorescence titration of RARy-LBD with all—trans-retinoic acid. ............ 133
Figure IV-8. Titration of KLE-CRABPII with retinal. .................................................... 134
Figure IV-9. In vitro binding assay using RARy-LBD directly from Ni-NTA resin. ..... 136
Figure V-l. Pure fractions of His-rCRALBP eluted after Ni—NTA purification. ............ 141
Figure V-2. Non-fusion CRALBP purified. .................................................................... 142
Figure V-3. UV titration of non-fusion CRALBP with ll-cis-retinal. ........................... 143
Figure V—4. Full length CRALBP expressed in pET28a and purified with his-tag

cleaved ........................................................................................................... 144
Figure V-5. N-43/44-truncated CRALBP expressed from pET28a and purified by Ni-

column ........................................................................................................... 147
Figure V-6. Expression of the truncated CRALBPs ........................................................ 148
Figure V-7. UV titration of full-length and truncated CRALBPs with l 1-cis-retinal.....l49
Figure V-8. Gel filtration separation of full-length-CRALBP. ....................................... 152

Figure V-9. UV titration of different fractions from gel filtration separation of A20-
CRALBP. ....................................................................................................... 153

Figure V-lO. CD spectra of different fractions from gel filtration separation of A20-
CRALBP ........................................................................................................ 155

xiv

Figure V-l 1. Gel filtration purification of the A20-CRALBP grown at different
temperatures ................................................................................................... 1 57

Figure V-12. Trypsin digestion performed for both full length and A20-CRALBP under

4°C. ................................................................................................................ 1 58
Figure V-13. Crystals of merocyanin-bound proteins. A. merocyanin-bound A20-

CRALBP; B. merocyanin-bound KLE-R59W-CRABPII. ............................ 159
Figure VI-l. R132K:Y134F:R111L:L121E-CRABPII (KFLE) purified after source-ISQ

anion exchange column. ................................................................................ 162
Figure VI-2. Elution fractions of RARy-LBD from Ni-NTA purification. ..................... 164

Figure VI-3. Pure fractions of the his-tagged rCRALBP eluted off the Ni column. ....... 166
Figure VI-4. Non-fusion CRALBP after purification. .................................................... 166
Figure VI-5. Primers used for constructing different N-tenninal deletion CRALBP ..... 167

Figure VI-6. Typical size-exclusion purification profile of CRALBP represented by A20-

CRALBP grown. at 16°C. .............................................................................. 169
Figure VI-7. CRALBP after purification ......................................................................... 169
Figure VI-8. Pictures of crystals of the ligand-bound CRABPII mutants ....................... 171
Figure VI-9. Pictures of crystals of the apo-CRABPII mutants ...................................... 174

Figure VI-10. Ramachandran plot of the C15-aldehyde-bound-KLE-R59W-CRABPII. 182
Figure VI-l 1. Ramachandran plot of the C15-aldehyde-bound-KL-A32E-CRABPII. 1 83
Figure VI-12. Ramachandran plot of the C15-aldehyde-bound-KL-T54E-CRABPII. 1 84
Figure VI-13. Ramachandran plot of the C15-aldehyde-bound-KFLDV-CRABPII ........ 185

Figure VI-l4. Ramachandran plot of the merocyanin-bound-KLE-RS9W-CRABPII.... 186

Figure VI-15. Ramachandran plot of the apo-KLE-R59E-CRABPII. ............................ 187
Figure VI-16. Ramachandran plot of the apo-KFLE—CRABPII. ..................................... 188
Figure VI-17. Ramachandran plot of the apo-KFLD-CRABPII. .................................... 189
Figure VI-18. Ramachandran plot of the apo-KFLDV-CRABPII. ................................. 190
Figure VI-19. Ramachandran plot of the apo-KL-T54E-CRABPII. ............................... 191

XV

Figure VI-20. Ramachandran plot of the apo-KL-A32E-CRABPII ................................ 192

Images in this dissertation are presented in color.

xvi

Amino Acids

Ala, A
Arg, R
Asn, N
Asp, D
Cys, C
Gln, Q
Glu, E
Gly, G
His, H
Ile, I
Leu, L
Lys, K
Met, M
Phe, F
Pro, P
Ser, S
Thr, T
Trp, W
Tyr, Y
Val, V

ABBREVIATIONS

Alanine
Arginine
Asparagine
Aspartic acid
Cysteine
Glutamine
Glutamic acid
Glycine
Histidine
Isoleucine
Leucine
Lysine
Methionine
Phenylanaline
Proline

Serine
Threonine
Tryptophan
Tyrosine
Valine

CRABPII Mutants Abbreviations

KFLD
KFLDV
KFLE
KFLEV
KL

KL-A32E
KL-T54E

KLE

KLE-R59E
KLE-R59W

R132K:Y134F:R111L:L121D
R132K:Y134F:R111L:L121D:T54V
R132K:Y134F:R111L:L121E
R132KzYl34F2R111L:L121E:T54V
R132K1R111L
R132K:R111L:A32E
R132K2R111LzT54E
R132K:R111L:L121E
R132K:R111L:L121E:R59E
R132K1R1 11L:L121E:R59W

Other Symbols and Abbreviations

A
Abs

Angstrom
Absorption

xvii

BTP

Cor

CCP4

CD

cGMP
Chlor.
CRABPII
CRALBP
CRBP
C-terminal

Da
DEAE
DNA
dNTP
DTT

8
E. coli

FABP
FPLC
FTIR

Gar
GDP
GMP
GPCR
GTP

h
HTP

iLBP
IPTG

Kr)

L
LB

Amax

Ampicillin
Advanced Photon Source

Base pair
Bis-tris-propane

The alpha carbon in the peptide bond
Collaborative computational project, number 4
Circulat dichroism

Guanisine 3’, 5’-cyclic monophosphate
Chlroamphenicol

Cellular retinoic acid-binding protein 11
Cellular retinaldehyde-binding protein
Cellular retinol binding protein

Carboxy terminal

Dmmn

Diethylarninoethyl cellulose
Deoxyribonucleic acid
Deoxynucleotide triphosphate
Dithiothreitol

Extinction coefficient
Escherichi coli

Fatty acid binding protein
Fast protein liquid chromatography
Fourier transform infra red

Transducin, activated
Guanisine diphosphate
Guanisine monophosphate
G-protein coupled receptor
Guanisine triphosphate

Hour
Hydroxyapatite

Intracellular lipid binding protein
Isopropyl-l-thio- B-D-galactopyranoside

Dissociation constant

Liter
Luria broth

Wavelength of the maximum absorption

xviii

MES

ul
ml

MOPS
MPD
MW

nm
nM

NMR
N-terminal

PCR
PDB
PDE
PEG
PSB

R1!

RA
RAR
RARE
RBP
R-factor
Rmsd
RNA
RPE
RXR

SB
SDS-PAGE

Tris
UV

vdw
vis

WT

Molar

2-Morpholinoethanesulfonic acid, monohydrate
microliter

milliliter

millimeter

3-(N-Morpholino)-propanesulfonic acid
2-Methyl-2,4-pentanediol

Molecular weight

Nanometer

Nanomolar

Nuclear Magnetic Resonance
Amino terminal

Polymerase chain reaction
Protein Data Bank
Phosphodiesterase
Polyethylene glycol
Protonated Schiff base

Rhodopsin, activated
Retinoic acid

Retinoic acid receptor
RAR-response element
Retinol-binding protein
Reliability factor

root mean square deviation
Ribonucleic acid

Retinal pigment epithelium
Retinoid X receptor

Schiff base
Sodium dodecyl sulfate — polyacrylamide gel electrophoresis

2-Amino-2-(hydroxymethyl)-1 ,3-propanediol
Ultraviolet light

Van der Waals
Visible light

wild type

xix

CHAPTER I. Introduction

1.1. The Mystery of Vision

Vision is a precious gift that deeply enriches our lives. Through vision, one is able
to perceive the order and beauty of the world. The visual process is one of the most
delicate and efficient mechanisms of the human anatomy.

Given the fascinating nature of vision, it is not surprising that people have been
intrigued by the question of how one is enabled to see since the early days of human
history. Ancient Greek philosophers, Plato (ca. 450 BC), Democritus (ca. 350 BC),
Aristotle (ca. 330 BC) and Euclid (ca. 300 BC), were the first known to seek the
explanations of vision (1). However, only Aristotle's theory, which was later extended by
his student Theophrastus (ca. 300 BC), was close to capturing the true nature of this
process: light is reflected by the objects, travels through the medium (air), and is then
received by the eye (2). Although not popular at its time, this theory was improved
throughout the centuries and eventually revisited by Johannes Kepler (1571-1630) in
1604 (3). He postulated the first complete description of the optics of the eye, refuting
the once-popular theory that vision is enabled when the objects are being touched by the
emissions sent out from the eyes. In 1704, Isaac Newton (1642-1727) published the book
Optic/rs, which revolutionized the understanding of the laws that govern light (4). Most
of his insights on the physical and physiological nature of color remain valid today. In
the 18303 it was discovered that the perception of light happens in the rod and cone cells

of the retina, the inner part of the eye. In 1878 the photosensitive protein in the rod cells,

now known as rhodopsin, was isolated by F. W. Kuhne (1837-1900) (5). The active
research of the subject over many years finally led to the revolutionary discovery by
Wald (1906-1997) and coworkers that greatly advanced our understanding of the first

steps in the vision process (6-8).

Vitreous gel

 

Figure [-1. The human eye (9).

In a simplified picture, the current model of the vision process can be depicted as
follows. When light enters the eye through the pupil, it is immediately focused by the
lens before being sensed by the photoreceptor cells in the retina (Figure I-l) (10).
Through a cascade of complicated events, the absorption of photons is converted into
specific electrical signals in the photoreceptor cells. These signals are then transmitted in
specific patterns out of the photoreceptor cells, through the optic nerves, and onto certain

parts of the brain.

2pm

rod

rhodopsins \

 
    

cone CHO

Disc\ 1 1-cis-retinal

diskb«~

   

__-- cytoplasm _-,

  

35 um (~1000 disks)

  

membrane

o
n
-——-————-—-—-———0
n
* a
‘ r .1 ' 9.

‘Iv‘
.,
.r-.v

{HHH‘H ”Y1

rhodopsin molecule

Figure [-2. A. rod and cone cells in retina (11); B. rhodopsins embedded in the disks
in photoreceptor cells (12).

Through decades of hard work by many scientists, a more thorough and complete
picture of this process has been achieved. For vertebrates, in the retina where vision
initiates, two types of photoreceptor cells, rod and cone, exist (Figure I-2A) (13). The
human eye contains approximately 100 million rod cells and about 6 million cone cells
(14). Both types of the photoreceptor cells contain the visual pigments, known as
rhodopsins. Rhodopsins refer to the specific proteins, called opsins, when they are bound
to retinal via a Schiff base linkage with the active site lysine. In human eyes, each rod

cell uses rod rhodopsins as its sole visual pigments, which are responsible for dim vision,

while each cone cell contains its own specific rhodopsins: blue, green or red opsins,
respectively. The blue, green and red cones are responsible for color vision. In the dark
state, all four opsin proteins bind to the same natural substrate, ll-cis-retinal. Although
the four rhodopsins differ in their absorption spectrum, they share the same mechanism in
the process called phototransduction (15-1 7).

As shown in Figure I-2A, a photoreceptor cell, either rod or cone, contains an outer
segment and an inner segment (18). Taking the rod cell for example, in the outer
segment hundreds of phospholipid bilayer disks are stored in an ordered, parallel fashion.
Rhodopsins, the 7-helical trans-membrane protein, are embedded inside these disks with
an orientation that is mostly perpendicular to the disk (Figure I-2B). The rhodopsins
occupy ~50% of the disk surface area. Although it varies between species, each rod cell
typically contains 1700 disks and each disk contains up to 1.5 million rhodopsin
molecules (19). The inner segment contains the mitochondria, the nucleus and the
synaptic end.

The human vision process starts with the photolysis of rhodopsin, which involves a
Cis-trans isomerization of the bound ll-cis-retinal. This activation-deactivation process
of rhodopsin involves several thermal/dark steps characterized by intermediates identified
at very low temperatures (Figure I-3) (20-25). In the dark state, ll-cis-retinal binds
within rhodopsin through a protonated Schiff base (PSB) with Lys296 (26-29). Within
femtoseconds after radiation, photorhodopsin is formed as a highly distorted "all-trans"
PSB moiety (30), which then relaxes to the first isolatable intermediate, bathorhodopsin
(31, 32). The formation of lumirhodopsin immediately follows within nanoseconds,

followed by metarhodopsin I and then metarhodopsin II. The formation of

metarhodopsin II happens milliseond after light absorption. Metarhodopsin 11 turns out
to be the activated form of rhodopsin, which triggers the subsequent phototransduction
cascade. The Schiff base is found to be unprotonated at metarhodopsin II. Following the
metarhodopsin II activation, the imine bond is protonated to afford metarhodopsin 111.

Finally, dissociation of all—trans-retinal from rhodopsin is achieved through hydrolysis.

    

all- Inns, 385 nm

 

opsin

1 1 -cls, 498 nm N/H
\ 6) {>0' C min.
hv <200 is “*1“

 

meta-ll. 0' C. 380 nm . activates transducin

 

 

 

 

 

Unprotonated —" begins the visual
.11/334‘3/ Schiff 3859 transduction process
+H+ ‘H+ rns
Photo. 570 nm
“all-trans“ highly distorted meta-l, -15° C
480 nm
ps

us
11 (+3 ns .
12 H 497 nm

Bathe, -140' C, 543 nm
“distorted trans”

Figure [-3. Schematic representation of the intermediates observed in the
isomerization of ll-cis-retinal to all-trans-retinal in rhodopsin (21).

The electric signal is generated as follows (Figure 1-4). In vertebrate, a dark current

carried by Na+ and Ca2+ enters the outer segment of the rod cell through cGMP-gated

channels keeping the cytoplasmic membrane relatively depolarized (33-35). The

activation of rhodopsin through photon absorption generates metarhodopsin II (36, 37),
which is able to bind to transducin, a G-protein consisting of three subunits: Tor, TB and
Ty. It is believed that this interaction is achieved when the a-subunit of transducin binds
to amino acids located in the second and third cytoplasmic loops of metarhodopsin (38-
41). This interaction triggers a conformational change in Ta, which involves the opening
of the nucleotide-binding pocket and release of the guanosine-5'-diphosphate (GDP)
bound (42). Transducin is then activated through binding to guanosine-5'-triphosphate
(GTP) and subsequently the dissociation of To from TB and Ty, which eventually enables
the signal transduction going from transducin to the next entity, the cyclic guanosine-5'-
monophosphate (cGMP)-specif1c phosphodiesterase (PDE).

PDE is a membrane-bound protein consisting of four subunits: Pa, PB and 2Py.
The catalytic activity of PDE is enabled by the two subunits, Pa and PB. In the inactive
state, the two inhibitory Py subunits block the activity of PDE. Activation of PDE is
achieved through binding with Tor-GTP and subsequent removal of the two inhibitory Py
subunits (43, 44). The activated PDE then hydrolyzes the cGMP into 5'-GMP, reducing
the cellular concentration of cGMP. The lowered concentration of cGMP signals the
closing of the cGMP-gated ion channels, which then results in interruption of the dark
current. In this way, the absorption of light is transformed by the rod cells to a
hyperpolarized electrical signal which is subsequently transmitted to the synaptic end and

sent out to the brain through the optic nerves.

.annEoE «Emma 05 5 £8530 :2 cucumdfiOo
mo 2320 05 E 3:52 33:55 3285:0280 A2200 mo _o>o_ 533 dong—588 came—m8? m: mics»: £200 Co max—83:
05 DRESS Aim—Q5 man— uoumzuoa BE. .3833. > 05 mo coufioowmmu E 330:8 .3833 > m: 55 mason—BE t3 mam magnum
382.333 *0 GPO a c8 mQOAVgon a mammfifioxo .3 9.0V 3338a 0583 WESESE 05 :25; com: £38355 3208::
808:8 :05 .& scam 6.5 Emmovofi 8558 5 9 mauso— =Qofi mo 53.8%? AWE cage nouguoaSesm .vA 2:»:—

l QEGU 12.0

 

 

Phototransduction in vertebrate involves a light-activated isomerization of a double
bond, which triggers a cascade of events and eventually leads to vision. This process is
one of the most efficient and sensitive biological processes known, with a quantum yield
of ~O.67 (46). The energy amplification of the visual process is calculated to be in the
order of 100,000 based on quantitative analysis of the electrical response to a single
photon (47, 48). On the molecular level, the amplification is achieved in primarily two
steps. The first step lies in the interaction between metarhodopsin II (R*) with transducin
(T). Within a fraction of a second, one molecule of R* is able to contact hundreds of
transducins. The second stage of amplification is provided by the activity of PDE which
catalyzes ~1000 molecules of cGMP per second (49, 50). The total amplification
therefore equals to ~100 x 1000 = 100,000, which is in agreement with the calculations
based on the electrical response. This explains the excellent sensitivity of vision that
allows us to see even in the darkest environments.

The above steps are involved in the rising phase of the light-signal response. The
inactivation of this response is also important and involves mainly three mechanisms.
The first mechanism is the guanylate cyclase-catalyzed generation of cGMP from GTP.
This process compensates the loss of cGMP caused by light-activation. The decreased
cGMP concentration signals closing of the ion channels, which leads to a decrease in the

concentration of calcium (35, 51). The activity of guanylate cyclase is stimulated by the
2+ . . . .
lowered Ca concentration in the rod cell. The cGMP concentration then increases so

that the cGMP-gated channels are open and the dark current is restored.

The second mechanism of inactivation is through phosphorylation of rhodopsin.
The major role of rhodopsin phosphorylation, mediated by rhodopsin kinase and protein
kinase C, is to promote arrestin binding and decrease transducin binding (52, 53).

The third mechanism involves the transducin-catalyzed hydrolysis of GTP to GDP
(54, 55). The GTPase activity is greatly accelerated during the rising phase of the light
activation, which later on leads to reduced concentration of activated transducin, Tor-
GTP.

Phototransduction in invertebrates shares a great deal of similarity with that of
vertebrates in the overall strategy. Light activation of rhodopsin triggers a series of
intermediates leading to metarhodopsin. Similarly, this metarhodopsin also triggers a
cascade of events leading to an electrical response. One difference is that the
metarhodopsin in invertebrates is stable and inter-convertible with rhodopsin. Another
interesting difference of the phototransduction in invertebrates is that the electrical signal
is generated by depolarization of the plasma membrane, which is caused by the light-
activated closing of the ion channels (56). Overall, the vision process in invertebrates

shares several important features and similar sensitivity with that in vertebrates (5 7, 58).

1.2. Rhodopsins as mammalian visual pigments: structure and
function in photoactivation.

Rhodopsin is a 39KD protein, consisting of 348 amino acids and two oligosacchride
chains (59, 60). It belongs to the family of G protein-coupled receptors (GPCR), which is
by far the largest class of cell surface signal transducing receptors. Similar to the other

membrane-bound GPCR family proteins, rhodopsin features a typical 7-helical structure.

Y268

E113

 

1 1 -cis-natinal

 

Figure [-5. Crystal structure of bovine rod rhodopsin. A. Overall fold; B. Binding of
11-cis-retinal in rhodopsin.

Our understanding of rhodopsin was greatly advanced with the determination of the
crystal structure of bovine rod rhodopsin at 2.8 A in the year 2000 by Palczewski and
coworkers (28). This first-ever crystal structure of rhodopsin was able to define 95% of
the amino acids as well as the posttranslational modifications (Figure I-5A). Continued
efforts from different research groups has since yielded rhodopsin crystal structures in
different crystal forms and improved resolutions to 2.2 A (61, 62). As revealed by the
crystal structure, ll-cis-retinal is bound to the transmembrane portion of rhodopsin
through a covalent trans-imine bond with Ly5296 (Figure I-SB). This confirms the
protonated Schiff base (PSB) formation between the chromophore and Ly5296 that was
suggested previously based on other biochemical evidence (26, 27). Glu113 is identified

as the counter anion that stabilizes the PSB 3.45 A away. This electrostatic interaction

10

between the PSB and the carboxylate counter anion is found to be conserved only among
visual pigments and is considered to be important for wavelength shifi of the bound
chromophore as well as for maintaining the inactive, dark state conformation of
rhodopsin (63-66). Prior to the determination of the crystal structure of rhodopsin, the
role of Glu113 acting as a counter anion had been identified through mutagenesis studies

(65). Mutation of Glu113 to Gln113 leads to a dramatic change in the absorption

maximum from 500 nm to 380 nm, which is accompanied by a decrease in pKa from over

16 to around 6.

Rhodopsin also features a compact extracellular (intradiscal) portion, part of which
folds inward to enclose the bound chromophore (67). The two oligosacchrides are also
present in the extracellular part of the protein, linked through two Asn residues, Asn2 and
AsnlS, at the N-terrninus. The cytoplasmic portion of rhodopsin is not as rigid as the
extracellular portion. In fact, the B-factors for this region vary considerably among
different ground—state crystal structures, indicating the flexible nature of this part of
rhodopsin (68). The transmembrane (TM) portion of rhodopsin constitutes as much as
60% of its secondary structure (69, 70). This region also contains the majority of the
conserved residues in rhodopsin and other GPCRs, many of which are suggested to play
important roles in maintaining the stability of rhodopsin.

As mentioned earlier, the initial light activation of rhodopsin leads to several
photoproducts. The crystal structures of some of these photoproducts have been
successfully obtained after trapping these photolyzed states at low temperatures (71-73).
Three-dimentional structures have been obtained for bathorhodopsin and lumirhodopsin.

A structural model for metarhodopsin I has also been obtained based on electron

11

crystallography of two-dimentional crystals. Even though these structures are only at
moderate to low resolutions (2.7 - 5.5 A), comparison of them with the dark-state
structure of rhodopsin reveals that the double bond between C11 and C12 in retinal
adopts the trans conformation upon light-activation (74). From bathorhodopsin to
lumirhodopsin, a positional/confonnational change of the [ii-ionone ring of retinal has
been suggested, but only by low-resolution structures. A structural model of an
intermediate containing a deprotonated Schiff base has been obtained from
lumirhodopsin crystals (75). This metarhodopsin II-like photoproduct seems to be
similar to the dark state rhodopsin in overall structure. However, parts of the second and
third loops in the cytoplasmic portion are found to be disordered in this photoactivated
rhodopsin. In addition, residues Phe212, Trp265, Leu266 and Tyr268, were found to
make direct contact with the B-ionone ring of the chromophore, which is consistent with

observations from solid state NMR studies.

1.3. Other rhodopsins

Rhodopsins exist in all three domains of life, Archaea, Bacteria and Eukarya (76).
However, not all rhodopsins fimction as visual pigments as they do in higher animals. In
fact, microbial rhodopsins represent a diverse group of photochemically reactive proteins,
whose functions involve converting light energy for either ion transport or photosensory
movement of cells (77). A generally accepted hypothesis that unifies microbial

rhodopsin transport and signaling is that the accessibility of the Schiff base is switched

12

from the periplasmic side to the cytoplasmic side through structural changes, which
opens the cytoplasmic channel for ion pumping and contributes to signal relay in the
sensors (78).

Bacteriorhodopsin, isolated from the haloarchaea Halobacterium halobium, was the
first microbial rhodopsin discovered (79). It functions as a proton pump through
converting light energy into a proton gradient across the bacterial cell wall (80).
Bacteriorhodopsin binds all-trans-retinal through a protonated Schiff base with Ly3216

(81). The PSB is stabilized by two Asp residues through water-mediated hydrogen

bonding interactions. The pKa of this PSB was measured to be over 12 (82, 83). With

the chromophore bound, bacteriorhodopsin is able to absorb at 570 nm (84). Similar
rhodopsin proton pumps have been observed in systems other than Haloarchaea,
including marine Archaea (euryarchaeotic rhodopsins) (85) and Bacteria
(proteorhodopsin and xanthorhodopsin) (86, 87).

A similar case is halorhodopsin which is found in haloarchaea (88). Halorhodopsin
functions as a chloride pump, transporting chloride ions into the cytoplasm and
hyperpolarizing the cell membrane (78). In halorhodopsin, the PSB is stabilized by a
tightly bound chloride ion which is translocated upon photoactivation. This unique
structural feature enables the specific function of halorhodopsin. Halorhodopsin has also
been found in the halophilic bacterium Salinibacter ruber (89).

Another major class of microbial rhodopsins consists of the sensory rhodopsins
(90). The first sensory rhodopsin, sensory rhodopsin I (SR1), was discovered in a search
for receptors responsible for phototaxis in the haloarchaeal prokaryote Halobacterium

salinarum (91, 92). SR] can signal the cells to be either attracted or repelled by the light,

13

depending on the light-activation mechanisms. Sensory rhodopsin II (SRII) is functional
only as a repellent phototaxis receptor in haloarchaea. Both SRI and SRII have been well
characterized both structurally and functionally. As revealed in the atomic structure of
SRII, the PSB in SRII is stabilized by water-mediated hydrogen bonding interactions
with two adjacent Asp residues, Asp85 and Asp212 (93). SRII absorbs at ~500 nm,
similar to the rod rhodopsin in humans. The photoactivated intermediates of SRII have
also been characterized using Fourier transformed infrared spectroscopy (FTIR) (94-97).
Upon activation, both SRI and SRII form 2:2 complexes in haloarchaeal membranes with
their cognate transducers, HtrI and HtrII, respectively. Formation of the complexes
initiates the signal relay and is now under intense investigation.

A unique sensory rhodopsin, Anabaena sensory rhodopsin (ASR), has been
identified from the freshwater cyanobacterium Anabaena sp. PCC7120 (98). The crystal
structure of ASR has been solved at 2.0 A resolution (99). The overall membrane-
embedded seven helical structure of ASR resembles the other haloarchaeal rhodopsins.
However, there are also striking differences separating ASR from other haloarchaeal
rhodopsins. First, the interior of the cytoplasmic side of ASR is rather hydrophilic,
compared to other haloarchaeal rhodopsins. These hydrophilic residues are stabilized
through a water-mediated hydrogen bonding network, which extends toward the
cytoplasmic surface where ASR interacts with its transducer. Second, ASR exists in a
mixture of both all-trans and 13-cis retinal isomeric forms, with absorption maxima at
550 nm and 537 nm, respectively. Each of the two forms exhibits efficient light-induced
conversion to the other, which results in the ratio of the two being dependent on the

wavelength of illumination. This mechanism resembles that of the phytochrome in

14

plants, which led to the suggestion that ASR might be responsible for modulating
chromatic adaptation in photosynthesis by triggering differential biosynthesis of light-
absorbing pigments based on the quality of light.

An interesting discovery reported recently shed light on the relationship between
the transport rhodopsins and the sensory rhodopsins (100). Spudich and coworkers
successfully converted the bacteriorhodopsin (BR) to a SRII-like sensor by replacing
three residues in BR with the corresponding residues in SRII. The resulting protein is
capable of relaying the retinal photoisomerization signal to the transducer HtrII, which
then induces robust phototaxis responses. This finding not only identified the key
residues in determining the functions of the two proteins, but also demonstrated the
elegant and yet simple design by nature to achieve functional diversity from the same

protein scaffold and similar ligand-protein complexation.

1.4. Color vision

Rod rhodopsin is the most studied and best understood rhodopsin among the over
1000 visual rhodopsins discovered so far due to its importance in animal vision and also
its abundance in the retina. The cone rhodopsins, however, are less understood despite
their equal or even greater importance in vision. As mentioned before, rod rhodopsins

are responsible for dim vision and cone rhodopsins are responsible for color vision.

15

 

 

 

420 498 534 564
100-
Q
0
C
N
'E
O
3 ,I
g 50“ ’1”
0 ’a
.2 , '
.5 .-
E
O
z
G r 11 IlirIirtrlrrrrlrlrilrrrrirrfiri

l
400 500 600 700
Wavelength (nm)

Figure [-6. Absorption profiles of different human rhodopsins (101).

In the human vision process, cone rhodopsins share similar photoactivation
cascades as rod rhodopsins. However, the maximal sensitivities of each rhodopsin differ
at different wavelengths. Human rod rhodopsin absorbs at 500 nm. The human cone
rhodopsins: red, green and blue, absorb at 560 nm, 530 nm and 420 nm, respectively
(102-105). All four rhodopsins exhibit red-shifted maximal absorption wavelengths afier
binding to the chromophore ll-cis-retinal, which by itself absorbs at 380 nm as a free
aldehyde. The combination of the four pigments constitutes the whole 400-700 nm

range, which is defined as the visible spectrum (Figure I-6).

16

Table [-1. Rhodopsins of various species (106).

 

 

 

Cone sensor group (nm)

Species SW81 SW82 M M/L L

Human 420 530 560

Dog 430 555

Horse 430 540

Squid 470/485 500

Black bream 475 560

Gold fish 360 450 540 625

Chicken 420 455 510 560
- Peking robin 355 455 500 570

Pacific shearwater 405 450 505 565

Butterfly 360 400 450 520 610

Housefly 355 460/49 530

Honey bee 350 445 550

 

Within animals, the requirement for color vision has been met in different ways,
suggesting that color vision has developed independently several times during evolution
(107). Depending on the number of types of cone photoreceptors present, some degree of
color vision becomes possible. Primates express three types of cone opsins, while many
animals have two and some insects, birds and fish have four (108). These are described
as being trichromats, dichromats and tetrachromats, respectively. As a general
classification, the cone rhodopsins can be roughly grouped into the following categories:
very short wavelength (i.e. UV)-sensitive (SWSl), short wavelength ("blue")-sensitive
(SWS2), medium wavelength ("green")-sensitive (MWS), and long wavelength ("red")-
sensitive (LWS) (106). Table [-1 shows how certain cone rhodopsins are classified based

on their maximal absorption wavelengths. Apparently, most of the cone rhodopsins

display red-shifted absorption maxima (Max): while other cone rhodopsins are blue-

shifted.

17

Though to different extents, the red shift of kmax observed in most of the

rhodopsins had been found to be partly due to the formation of the protonated Schiff base
(PSB) between 11-cis-retinal and the active site lysine. The discovery of the PSB started
with the isolation of retinal from rhodopsin in 1933 by Wald (109). In 1944, Morton and
Goddwin showed that retinal is the aldehyde form of vitamin A (110), which was
confirmed to be 11-cis-retinal in 1950 (111). As early as 1955, the formation of the PSB
between retinal and a lysine residue in rhodopsin was proposed to account for the
bathochromic shift in visual pigments. During the late 1960's, Bownds and Akhtar,
through independent studies, provided evidences for the formation of a Schiff base
between retinal and a lysine within rhodopsin (112, 113). It was not until 1974 that
protonation of the Schiff base was unambiguously confirmed by Callender through
resonance Raman studies (114).

As a free aldehyde, ll-cis-retinal absorbs at 380 run. When the Schiff base is
formed with n-butylamine, a slight blue shifl in absorbance can be observed and the
deprotonated Schiff base absorbs at 365 nm. However, when the Schiff base is
protonated there is a significant red shifi of the chromophore’s absorbance to 440 nm
(115, 116). With controlled concentration of the chromophore and careful choice of
solvent this value may shift up to ~500 nm (117, 118). Protonation of the Schiff base
alone, however, is probably not responsible for the spectral tuning in color vision as 440
nm is just the beginning of the region of the electromagnetic spectrum that we can
actually see (~400 —- 700 nm). The differences in the absorbance observed for different

visual pigments as compared to the PSB of 11-cis-retinal in solution are defined as opsin

18

shifts and are believed to result from specific interactions of each individual opsin protein
with the retinylidene chromophore (119, 120).

Several hypotheses have been proposed to explain the opsin shifts. Although
progress has been made, a complete and satisfactory explanation for color vision is still
missing. The following is a brief description of the hypotheses and accomplishments
achieved in this field of research.

In 1958, Hubbard made the first attempt to explain the opsin shift (12]). He not
only suggested formation of a protonated Schiff base, but also proposed that specific
electrostatic interactions between the chromophore and the protein could induce a
delocalization of the PSB charge along the polyene backbone, leading to altered
spectroscopic behavior of the bound chromophore. Later, in the 1970's, theoretical
calculation studies suggested that a pigment protein should be able to regulate the
wavelength shifts by placing polar or charged residues around the chromophore (122-
126). In 1976, Honig proposed, through theoretical calculation, that the PSB is closely
associated with a counterion and additional negatively charged or polar groups are
positioned by the protein in the vicinity of the ring half of the chromophore (123). Later,
Honig provided firrther support for this external point-charge model using theoretical
calculations based on spectroscopic data of model compounds ([23). Despite the lack of
absolute validity, additional support from Nakanishi and others made this model
generally accepted at the time (127-129).

The confirmation of the PSB counter anion was achieved in 1989 when three
independent studies indicated that Glu113, a highly conserved residue within vertebrate

visual pigments (130), functions as the counter anion in bovine rhodopsin (63-65).

19

Replacement of Glu113 with Gln113 resulted in two species that absorb at 490 nm and
380 nm, respectively. These two species were identified as the protonated and
unprotonated forms of the Schiff base which exist in a pH-dependent equilibrium. The
formation of the Schiff base between the chromophore and Lys296 and the presence of
the Glu113 as counter anion were both confirmed with the determination of the crystal

structure of bovine rhodopsin in 2000 (28).

R R R

T: h h

N fcoort N 000” N COOH
n

. R] COOH h
choorr N\/' N\/\/

\\ \
R=Wsfl

Figure [-7. Model compounds used to study wavelength shifts controlled by distance
between the PSB and the counter anion (131).

The distance between the PSB nitrogen and the stabilizing counter anion was also
shown to affect the red shift of the bound chromophore. In 1972, Blatz suggested that the
different absorption maxima in different opsins were caused by differed distances
between the PSB and the counter anion, with increased distance leading to greater red

shifi (132). This hypothesis was supported by the discovery in 1989 that replacement of

Glu113 with Asp resulted in a red shift of rod rhodopsin Xmax to 505 nm, a 5 nm more

20

red shified pigment than the native rod rhodopsin (64). Furthermore, the effect of
varying both distance and angle between the PSB and the counterion was also
investigated. Sheves and coworkers synthesized a series of model compounds (Figure
I-7), each adopting a unique arrangement of the PSB and a nearby carboxylate (131).

The measured wavelength shifts as well as pKa values were shown to vary according to

each molecule. However, the maximum difference in red shifted )‘vmax among these

synthetic model molecules was only 10 nm, which is much smaller than the maximum

difference observed between the visual pigments (I33, 134). Therefore, although the

position and orientation of the counterion could affect the Max, it could not be the cause

for the whole spectrum shift in different rhodopsins.

 

Figure [-8. The point charge model. Positioning negative charges at different places
along the polyene chain of the chromophore may modulate the maximum absorption
wavelength.

As discussed earlier Honig suggested in 1976 that additional charged or polar

residues are placed close to the polyene chain (123). Following this hypothesis Nathans

21

suggested, by examining and comparing the gene sequences of all four human opsins,
that the placement of different negatively charged amino acids along the polyene
backbone in different proteins could result in extended or disrupted conjugation of the
positive charge, which could be nature’s way of tuning the wavelength (Figure I-8) (130).

In addition to the point charge theories that mostly focus on the electrostatic or
polar interactions between the chromophore and the opsin protein is another series of
hypotheses that emphasize the steric interactions between the two parties. An early
proposal by Blatz suggested that twisting the planes about the single bonds of the
chromophore would lead to different levels of conjugation and therefore wavelength
regulation (Figure [-9) (135). Although the polyene system in retinal tends to adopt the
planar conformation in order to achieve maximum conjugation, steric constraints in 11-
Cis-retinal in fact results in a non-planar molecule. NMR studies had shown that the C6-
C7 single bond of 11-cis-retinal bound in rhodopsin adopts a Cis-like conformation,
namely 6-S-cis (128). Further investigation of the system revealed that, due to the steric
interactions between the C5-Me and C8-H, the 6-S-cis conformation is indeed twisted to
be non-planar. A similar situation is suggested for the non-planar 12-S-trans
conformation, where the steric interaction between C13-Me and CIO-H forces a twisting
in the C12-C13 single bond (136, 137). According to this theory, red rhodopsin should

adopt the retinal in a more planar conformation, optimizing the orbital overlap and

leading to a more red-shifted kmax. However, theoretical studies combined with NMR

have also shown that the induced degree of conjugation does not appear to be solely

responsible for the opsin shifts (I38).

22

 

 

Figure [-9. Hypothesis of steric interactions leading to wavelength regulation.
Twisting the single bonds will reduce the degree of conjugation and therefore lead to
different absorption maxima.

Another hypothesis in explaining the opsin shift involves excitonic coupling
between the chromophore’s excited states and nearby aromatic residues. Based on
mathematical analysis of the excited state energies of 11-cis-retinal, the bathochromic
shift in visual pigments were proposed to be caused by excitonic coupling, which could
lead to the formation of superconducting pairs in the pigments (9). Light-induced
perturbation of aromatic residues in bovine rhodopsins and bacteriorhodopsins (139) as
well as photoexcited energy transfer in frog rhodopsins (11) were observed in support of
the existence of such coupling. Recent spectral studies demonstrated the presence of
excitonic coupling between retinal and tryptophan residues in the visual pigments. The
coupling induced by alignment of dipole moments of the chromophore and a nearby
tryptophan has been suggested to be responsible for driving the tortional motion of the
chromophore and leading to the Cis-trans isomerization (12). However, no correlation
between the excitonic coupling and spectral tuning in the visual pigments has been

observed.

23

 

Figure I- 10. Vicinity of retinal in cone pigments as shown in the homology models
(140). Charged amino acids are indicated in red. (A) Blue cone pigment; (B) Green; (C)
Red.

The determination of the first crystal structure of bovine rod rhodopsin in 2000 by
Palczewski marked one of the most remarkable milestones in the field (28). This 2.8 A
structure not only provided absolute evidence for many previous results and suggestions,
but also paved the way for future research on other visual pigments. Continued efforts on
the crystallography of rhodopsin has since yielded improved resolutions as good as 2.2 A
(62).

As shown in Figure [-58, the formation of the PSB between ll-cis-retinal and
Ly5296 is confirmed. Glu113 acts as the counter anion at 3.45 A away from the PSB
nitrogen. A negatively charged residue is observed to be over 5 A away from the double
bond between C11 and C12, which is further away than what is expected based on both
the hypothesis of Honig and the results from related NMR studies. However, two polar
residues, Thr118 and Tyr268, are found to be in the vicinity of the central part of the
polyene. Twisting of the single bonds is also observed, suggesting the contribution from
the steric-induced non-planarity of the chromophore on wavelength regulation. The
double bond between C11 and C12 has also been shown to adopt a twisted conformation
as revealed by the 2.2 A-resolution structure, which is consistent with the predictions
based on theoretical calculations. Furthermore, Trp265 is found in the structure to be
near the ionpne ring as well as the C13-Me of the chromophore, suggesting possible
excitonic coupling effects.

The determination of the crystal structure also led to construction of homology
models for the three human cone pigments using the rod rhodopsin structure as a template
(140). The amino acid sequence identities for the blue, green and red cone pigments with

rhodopsin are 41%, 38% and 37%, respectively. According to the homology models, the

25

secondary structures of all three cone opsins are nearly identical to that of the rod
rhodopsin, exhibiting a seven helical transmembrane portion together with one
cytoplasmic helix and two beta strands. The only difference, the same for all three cone
opsins, is the one-amino-acid extension at the N-tenninal l3 strand. Similar to rod
rhodopsin, the retinal binding environment is highly hydrophobic in all three cone
rhodopsins (Figure I-10). Red and green rhodopsins share almost identical binding
probably due to their high homology (96% homologous, differ by only 15 amino acids)
(130). Both contain a Trp182 in the vicinity of the central part of the polyene. The two
opsins also share an additional glutamate at 102, besides the counter anion Glu129 that is
analogous to Glu113 in rod rhodopsin. The blue opsin, however, displays a much
different arrangement of amino acids around the retinal. As shown in the model, the
central residue forming the cavity in the blue cone pigment is Tyr262. Because of its
proximity to the retinal B-ionone ring, Tyr262 is believed to be the major factor in the
blue shift of this pigment. The additional glutamate observed in red and green rhodospin

models is missing in the blue pigment model structure, which is suggested to be

responsible for additional blue shift of the Amax.

In fact, prior to the rhodopsin crystal structure, mutational studies involving specific
key residues in visual pigments also yielded very interesting results. Inspired by the high
sequence homology between red and green rhodopsins, Oprian and coworkers showed
that the red cone can be converted into a green-rhodopsin-like pigment by implanting

seven specific mutations, S116Y:Sl80A:123OT:A233S:Y277F:T285A:Y309F (Figure

I-1l) (102). The artificial pigment is able to constitute a nearly 40 nm shift in the Xmax.

Within the seven sites, three were identified to play more significant roles in the spectrum

26

I“ 8180

$116

1 1-cis-retinal

K... .2... )1
Y309
Y2" a< T285

Figure [-1]. Illustration of the sites of mutations for converting the red rhodopsin to
a green-rhodopsin-like pigment (102).

A233 ’

 

Figure [-12. Illustration of the sites of mutations for converting the green rhodopsin
to a red-rhodopsin-like pigment (141).

27

tuning. A triple mutant of the green opsin, F 277Y:A285T:A180$ (Figure I-12), is able to
account for a majority of the red shift, going from green rhodopsin to a red-rhodopsin-
like pigment (141). Interestingly, all three mutations are in fact conversions from
hydrophobic residues to polar residues around the beta-ionone ring of the chromophore.
A similar trend was also observed in mutagenesis studies of the rod rhodopsin. A rod
rhodopsin mutant, F 261Y:A269T:A164S (Figure I-13), red shifts to 520 nm, achieving an

amazing 200 nm red shift (142).

A164

E113

   
 

1 1 -cls-retinal

K296

 

Figure [-13. Illustration of the sites of mutations for converting the rod rhodopsin to
a red-shifted pigment (142).

In recent years, quantum mechanics (QM) combined with molecular mechanics
(MM) (denoted as QM/MM) techniques have been used intensively in studying the
spectral tuning in the visual pigments (143). Electrostatic interactions have been

suggested to affect the wavelength shifts between different rhodopsins due to the fact that

28

the HOMO-LUMO transition is actually an intramolecular electron transfer from the
ionone ring to the PSB (144-147). The charge distributions within the ll-cis-retinilidene
were shown to be different for the ground states and excited states (148).

In a most recent study done by Nakatsuji and co-workers, QM/MM calculations
were performed based on the homology models of the cone pigments. The results
suggested that the color tuning is regulated by amino acids at specific positions (145).
More specifically, a common Ser (186 in rod rhodopsin) was shown to discriminate
human blue rhodopsins from other human rhodopsins presumably due to its unique
orientation relative to the It system of the chromophore. In the human red rhodopsin,
Tyr277 and Thr285 (replacing Phe and Ala, respectively, in other human rhodopsins),
were shown to provide negative electrostatic potential around the ionone ring, which
differentiates the red human rhodopsins from other rhodopsins.

Studies from Goddard and co-workers, however, indicated that conformational
twisting is what causes major blue shift of the absorption maxima in blue pigments (149).
They also suggested the red-shifting role of dipolar side chains in the binding pocket as a
fine adjustment to the opsin shift. Besides, electronic polarization effects of the protein
environment were also suggested to play a predominant role in causing bathochromic
shifts in bacteriorhodopsins (150, 151).

In addition, a counterion switch mechanism was proposed by Mathies and co-
workers (152). Glu113 was shown earlier to be the counterion that stabilizes the PSB in
the ground state. Mutation E181Q resulted in nearly unchanged Raman vibrational

specctra in the ground state. However, it was shown that the same E118Q mutation led to

a dramatic shift of the pKa of the PSB in the metarhodopsin II state. Mathies then

29

proposed that a counterion switch occurs within the photoactivation process through a
proton transfer from Glu18l to Glu113. This is accompanied by a specific
conformational change of the rhodopsin protein, leading to its activation. This model has
been validated by several theoretical calculations (153, 154). Such mechanism is
possibly responsible for the shifted absorption maxima observed for certain photo
intermediates.

In conclusion, environmental differences around the ionone ring and the PSB of the
bound chromophore among different visual pigments seem to play major roles in the
wavelength regulation mechanism. Other factors, including distortion of the planar
conformation of retinal as well as excitonic coupling, might also contribute to the
wavelength changes. The real spectrum tuning mechanism very possibly is achieved by a
combination of these factors. The wavelength regulation mechanism is indeed much
more complicated than anticipated. Alternative approaches are therefore desired in trying

to provide a clearer picture of the mechanism.

1.5. Designing CRABPII mutants as rhodopsin mimics

As concluded above, the hypotheses regarding the wavelength regulation in visual
pigments need to be further tested. However, difficulties associated with working with
the membrane-bound protein rhodopsins, such as complication in mutagenesis, low

heterologous expression level, and difficulties in obtaining 3-D structures, still pose

30

problems. This is why our research team opted for a model system that can be used to
elucidate the wavelength regulation mechanism.

Our goal is to develop a model protein that mimics rhodopsin in the binding of the
chromophore, retinal. Through protein engineering techniques, we created different
situations for the bound retinal, such as placing charged or polar residues at various
positions along the chromophore backbone, introducing steric interactions to force the
non-planarity of the molecule, or placement of aromatic residues in the vicinity of the
chromophore. For this particular purpose, this protein mimic should first show good
affinity for retinal and also be very tolerant towards mutagenesis. Additionally, this
system should also be more friendly than rhodopsin in heterologous expression,
purification and crystallization, which would facilitate the determination of the 3-D
structures and therefore provide valuable information in terms of both binding details and

directions for future rational redesign.

1.5.1. CRABPII general background.

Retinoic acid (RA), the biologically active metabolite of vitamin A, functions as a
morphogen during embryonic morphogenesis (155) and has been suggested to play an
indispensable role in regulating cell differentiation and growth (156-158). RA and
synthetic retinoids have been used successfully for treatment of acute promyelocytic
leukemia (APL), skin disorders, and epithelial malignancies, such as skin cancer and

cervical cancer (155, 158-1 63).

31

     

Retinoic acid

R111

R132

Figure [-14. Crystal structure of retinoic acid bound in WT-CRABPII (164). A.
Stereoview of the overall fold of the protein; B. Binding of retinoic acid in the active site.

RA exerts its regulatory function in the nucleus by association with the
transcriptional activator, retinoic acid receptor (RAR) (155, 165-170). Upon
complexation with RA, the RAR/RXR (retinoid X receptor) heterodimer recognizes the
RA-response elements (RARE) in the promotor region of certain genes and regulates
their transcription.

The association of RA to RAR is facilitated by cellular retinoic acid binding
proteins, CRABPI and CRABPII (171-174). CRABPs belong to the intracellular Lipid

Binding Protein (iLBP) family and are present in virtually all vertebrates (175). The

32

overexpression of CRABPs has been reported in a wide variety of human cancers,
suggesting important roles in cancer development (176-183). CRABPs specifically bind
to RA and modulate its cellular concentration in the cytoplasm. The RA delivery
mechanism has been suggested to be different between CRABPI and CRABPII. While
CRABPI shows no evidence of interacting with RAR, CRABPII shuttles RA to RAR
through direct protein-protein interaction (184, 185). Nuclear localization of CRABPII,
but not CRABPI, upon ligand binding has also been observed, providing further evidence
for its interaction with the nuclear receptor RAR (186).

Recently we reported the first crystal structure of apo-WT CRABPII (187). The
crystal structure of WT-CRABPII-RA has been reported previously by other groups
(188), however a higher resolution structure of this complex (Figure I-14), determined at
1.48 A, is now available (187). In the latter work, using three different data sets collected
on apo-WT CRABPII we showed that apo- and holo-CRABPII share very similar
structures. Binding of RA appeared to increase the overall rigidity of the structure,
although the induced structural changes were not as pronounced as previously thought
(1 89). The enhanced structural rigidity may be an important determinant for the
enhanced nuclear localization of the RA-bound protein. Interestingly, water-mediated
interactions are observed in the cavity of apo-WT-CRABPII MolA, connecting the
Argl 11 deep inside the cavity through Arg132 to Ala36 and Val33 both located on the
(12 helix (Figure I-15). The water-mediated network is not observed in MolB probably
due to crystal packing. The difference in crystal packing environments between MolA
and MolB at the (12 helix region is significant (Figure [-16), which is apparent from the

number of intermolecular interactions that each a2 helix makes with the neighboring

33

molecules: MolA has 0 hydrogen bond and 8 vdw contacts and MolB has 7 hydrogen

bonds and 43 vdw contacts.

2.9..

2.7 .0‘

Acetate ,o' 204
R111

R132

 

Figure [-15. Water-mediated hydrogen bonding interactions observed in the
molecule A of the apo-WT-CRABPII crystal structure (164).

Similar water-mediated networks have been observed in other members of the iLBP
family, such as CRABPI (190) and intestinal fatty acid-binding protein (191), and were
suggested to play important roles in maintaining the stability of these proteins as well. In
the rat intestinal fatty acid-binding protein (FABP) structure (IIFC, 1.2 A), the water-
mediated network connects the Arg106 (conserved residue of Argl 11 in CRABPII)
through Arg126 (conserved residue of Arg132 in CRABPII) to Asp34 and Ala32 on the
(12 helix at the portal. In the CRABPI structure (lCBI), the network is observed in part,
but the lower resolution of the structure (2.7 A) probably leads to an incomplete
observation of all the ordered water molecules in the cavity. In the high resolution
structure of cellular retinol-binding protein 11 (CRBPII, 1.2 A, PDB ID: 2RCQ), which is

also an iLBP family member, the water-mediated network is again observed, which also

34

resembles the one in apo-WT-CRABPII. In the apo-CRBPII structure, the network starts
at the Glu108 (conserved residue of Arglll in CRABPII based on similarity) and
extends to Ala36 and Val33 (both conserved in CRABPII based on identity). These
observations further emphasize the structural importance of these water-mediated

networks for the iLBP family proteins.

a

 

Figure [-16. Crystal packing observed in the structure of apo-WT-CRABPII. A.
Molecule A; B. Molecule B.

1.5.2. Designing CRABPII mutants as rhodopsin mimics.

CRABPII was chosen as our model protein for mimicking rhodopsin. First, all-
trans retinoic acid, the natural substrate of CRABPII, resembles retinal that is bound in
rhodopsin. This similarity therefore eases our way in introducing retinal into our model
systems, minimizing the risks in redesigning an entire binding site. Second, CRABPII is

a small and soluble protein that is easily expressed and purified in large quantities using

35

recombinant expression systems, which represents a huge advantage over the membrane-
bound rhodopsin. Third, CRABPII and other iLBP family proteins are well known for
their excellent tolerance of multiple mutations (192). The stability of CRABPII in the
presence of multiple mutations provides us additional freedom in introducing mutations
for protein redesign. Fourth, CRABPII mutants can be readily crystallized routinely
using established conditions, leading to three-dimensional structures usually at atomic
resolution (189, 193, 194). This feature contrasts drastically with that of rhodopsin,
which is more difficult to crystallize. The excellent tendency of CRABPII toward
crystallization gives us routine access to detailed structural information, leading to more
rapid data acquisition/analysis.

Previously, through rational design, CRABPII mutants were obtained that were
capable of forming protonated Schiff bases (PSB) between all-trans-retinal (Rt) and an
active site lysine introduced through site-directed mutagenesis (195, 196). A triple
mutant, R132K:R111L:L121E (KLE), is of particular interest due to its excellent ability
to shift the absorption maxima of the chromophore as well as its nanomolar dissociation
constant toward Rt (Figure I-17) (196). Besides the active site Lysl32, it has been
proven that the Glu121 that provides the counterion stabilization for the PSB is also
required for the bathochromic shift. The mutation of Arg to Leu at position 111 is also
required in that an ordered water molecule is removed upon this mutation, which helps
place the aldehyde group of Rt in a favorable position for nucleophilic attack and

eventually leads to PSB formation.

36

 
   

All-trans-retinal

Figure I-17. All-trans-retinal bound in KLE-CRABPII (196).

The triple mutant KLE-CRABPII served as a great example, demonstrating the
importance of formation of the PSB as well as a stabilizing counter anion nearby.
However, although the red shift to 447 nm observed for the all-trans-retinal-bound KLE-
CRABPII complex is impressive, the mechanism identified within it does not seem to
fully explain the wavelength regulation. As shown earlier, the PSB formed between
retinal and n-butylarnine absorbs at 440 nm and the red cone rhodopsin absorbs at 564
nm. Apparently, other mechanisms besides the PSB formation and the counter anion
interaction must be fimctioning in order to achieve the dramatic red shift in red
rhodopsin. Our next mission then is to further engage our model system into testing the
hypotheses proposed for the wavelength regulation mechanism in color vision. Most

recent work in this project features manipulation of the absorption maxima of rationally

designed CRABPII mutants when bound to a shorter retinal analog, CI5-aldehyde.

37

Chapter II of this dissertation details about the X-ray crystallographic studies of such

C15-aldehyde-CRABPII mutant complexes. Chapter II also includes brief descriptions of

the results from the biochemical studies performed by other colleagues. Chapter [[1
describes the structural insights derived from the crystal structures of the holo-CRABPII
mutants in terms of structural integrity upon mutations and the structural role of the
water-mediated network. Chapter IV describes an effort that we took toward
understanding the direct channeling of retinoic acid from CRABPII to RAR through

protein-protein interactions.

1.6. Cellular retinaldehyde-binding protein and the 1 1-cis-retinal
regeneration.

Cellular retinaldehyde binding protein (CRALBP) is a water-soluble protein found
in the retina and pineal gland (197). This 36 KD protein is remarkably stereoselective,
carrying as endogenous ligands 11-cis-retinol and 11-cis-retinaldehyde (ll-cis-retinal)
that are only known to function in vision. Notably, ll-cis-retinal bound to CRALBP is
less susceptible to photo-isomerization than when bound to rhodopsin (198). Mouse
knockout studies have shown that CRALBP gene defects impair regeneration of visual

pigments (I 99).

38

 

\ \ \ \ \ \QRhodopsin
O <—2— | H

all tr ans- retinal all-trans-retinilidene
all-trans- RDH/ \\photon

W0,

all- trans-retinol

    

nod outeroegment

 

 

E? Rhodopsin
H

 

 

 

 

 

  

 

 

 

E
2 \ \ \
2 I \ OH
g
g all- trans- retinol 11-cis-retinal CHO
° \LRAT /f 1-cis-RDH
'3.
.7; npees
3
all- trans-retinyl ester 11-cis-retinol OH
L J

 

Figure I-18. The visual cycle.

Retinoids undergo intercellular transport and several enzymatic transformations
during the course of the visual cycle (Figure I-18). All-trans-retinal, the product of visual
cycle bleaching, is reduced to all-trans-retinol within photoreceptor cells and transported
through an extra-cellular compartment to the retinal pigment epithelium (RPE) where it is
esterified by lecithin-retinol acyltransferase (LRAT) and then isomerized to 11-cis-retinol
by the isomerohydrolase (RPE65). A dehydrogenase-mediated oxidation in RPE in
mammals produces ll-cis-retinal, which must then traverse the extracellular
compartment to gain access to the photoreceptors where regeneration of visual pigments
occurs. CRALBP functions in retinal pigment epithelium (RPE) as a major acceptor of

ll-cis-retinol in the isomerization step of the rod visual cycle. It also serves as a

39

substrate carrier for ll-cis-retinol dehydrogenase (RDH5), facilitating the oxidation of
ll-cis-retinol to ll-cis-retinal (197, 198, 200-203). Direct CRALBP interactions have
been demonstrated in vitro with ll-cis-retinol dehydrogenase and with ERM (ezrin,
radixin, moesin)-binding phosphoprotein 50 (EBPSO), also known as sodium hydrogen
exchanger regulator factor type 1 (NHERF-l) (204). Interactions with EBP50 have been
suggested as a mechanism for localizing CRALBP to the apical RPE plasma membrane
for export of 11-cis-retinal to the adjacent rod photoreceptor cells for visual pigment
regeneration. Ongoing proteomic studies also support the existence of a RPE retinoid
processing protein complex containing CRALBP (205).

Due to the lack of crystal structures of this protein, the interaction of human
recombinant CRALBP (rCRALBP) with retinoids has been investigated by other
methods. Based on preliminary NMR studies (206), Crabb and co-workers first targeted
Met and Trp residues as important sites for retinoid interaction (203). They created five
mutants, M208A, M222A, M225A, W165F, W244F, and evaluated retinoid binding
properties and substrate carrier functions. It turned out that all mutants bound ll-cis-
retinal and 9-cis-retinal and therefore were not grossly misfolded. However, altered UV-
visible spectra and lower retinoid binding affinities were observed, supporting modified
ligand interactions. Altered kinetic parameters were also observed for RDH5 oxidation
of 11-cis-retinol bound to rCRALBP mutants M222A, M225A, and W224F, suggesting
impaired substrate carrier function. Heteronuclear single quantum correlation NMR
analyses confirmed localized structural changes upon photoisomerization of rCRALBP-
bound ll-cis-retinal and demonstrated ligand-dependent conformational changes for

residues Met-208, Met-222, Trp-165, and Trp-244. From this study, all five residues

40

(Met-208, Met-222, Met-225, Trp-165, and Trp-244) have been implicated to be
components of the rCRALBP ligand binding cavity.

Later, Crabb and co-workers further investigated the binding cavity using
photoaffrnity labeling and hydrogen/deuterium exchange (207). Photo-induced covalent
labeling employing synthetic ligand 3-diazo-4-keto-1l-cis-retinal, followed by LC-MS
analysis, successfully identified eight residues being photoaffinity-modified: Tyr-179,
Phe-197, Cys-198, Met-208, Lys-221, Met-222, Val-223, and Met-225. Among these
eight, three of them (Met-208, Met-222, and Met-225) had already been identified in their
previous investigation. In addition, topological analysis of apo- and holo-rCRALBP by
hydrogen/deuterium exchange and mass spectrometry demonstrated that residues
198~255 incorporated significantly less deuterium when the retinoid binding pocket is
occupied with ll-cis-retinal (203). This further confirmed that the identified residues
were inside the retinoid binding pocket of rCRALBP.

Based on sequence information, CRALBP has been classified as a member of the
CRAL-TRIO lipid-binding protein family (Pfam entry: PF00650) (208). Crystal
structures have been determined for three CRAL-TRIO lipid-binding proteins, including
human or-tocopherol transfer protein (ATTP) (209, 210), yeast Sec14, a
phosphatidylinositol-binding protein (211), and human supernatant protein factor (SPF)
(212). CRALBP shares 20~33% sequence identity and ~50% homology with these three
proteins. Structure models of CRALBP have been built based on the available crystal
structures of the three family members (207, 213). Both open and closed conformation
models have been produced (Figure I-l9). The model of the closed conformation agreed

well with the residues identified earlier, which are involved in interactions with retinoid.

41

It has also been suggested based on MD simulations of the model that the open
conformation of CRALBP participates in the ligand association and dissociation
processes; the closed form serves as a relatively stable carrier of retinoids, protecting

them from the cellular hydrophilic environment.

 

Figure [-19. Stereoview of the structure models built for CRALBP (213). A. Closed
form; B. Open form. N—temrinal starts from 66 (blue) in the rainbow presentations.

Despite all the efforts of elucidating the structural mechanism of CRALBP
interacting with retinoids, the lack of a crystal structure still poses serious problems for
research associated with CRALBP. More reliable answers are needed for questions like
how retinoid is released from the high affinity CRALBP binding cavity, which remains
an unresolved but important issue for understanding how ll-cis-retinal is exported from
the RPE for visual pigment regeneration. Besides that, the function of CRALBP in

tissues other than the RPE remain to be determined (i.e. in retinal Muller cells, ciliary

42

epithelium, iris, cornea, pineal gland, and a subset of oligodendrocytes of the optic nerve
and brain). To answers these questions, determination of the three-dimensional structure
of CRALBP will be essential.

In addition to elucidating the physiological role of CRALBP, another reason that we
are interested in determining the crystal structure of CRALBP is its potential as a model
for rhodopsin. As discussed earlier, the qualifications of a good rhodopsin mimic should
include affinity for retinoids, ease in molecular biology manipulations, high expression
levels, great solubility, and routine access to high-resolution crystal structures. CRALBP
fits in almost all categories except for the crystal structures. Beyond these features,
CRALBP is superior over CRABPII in mimicking rhodopsin in that retinal in CRALBP
is completely buried as in rhodopsin, while in CRABPII the ionone ring of the
chromophore is partially solvent-exposed. With these two reasons, we are driven to
pursue the crystallization of CRALBP. Chapter V details about our effort toward

obtaining the first crystal structure of CRALBP.

43

Reference

10.

11.

12.

13.

14.

15.

Wade, N. J. (1998) A Natural History of Vision, Bradford Books.

Woodbridge, F. J. E. (1983) Aristotle's Vision of Nature, Greenwood Press
Reprint.

Kepler, J ., and Donahue, W. H. (2000) Optics, Green Lion Press.

Newton, 1., Cohen, B. J., Einstein, A., and Whittaker, E. (1952) Opticks: Or a
Treatise of the Reflections, Refractions, Inflections & Colours of Light-Based on
the Fourth Edition London, 1 730, Dover Publications.

Crescitelli, F. (1977) Friedrich Wilhelm Kuhne. The Centennial of Rhodopsin,
Arch. Opthalmol. 95, 1766.

Wald, G., and Brown, P. K. (1958) Human Rhodopsin, Science 127, 222-226.
Wald, G. (1968) Molecular Basis of Visual Excitation, Nature 219, 800-&.
Wald, G. (1968) Molecular Basis of Visual Excitation, Science 162, 230-&.

Dunne, L. J. (1974) Excitonic Coupling in Visual Pigments, Phys Lett A A 48, 13-
13.

Tiffany, J. M., Noorjahan, P., Abdel-Latif, A. A., Harding, J. J., Mayne, R.,
Brewton, R. G., and Ren, Z. X. (1997) Biochemistry of the Eye, Chapman and
Hall, London, UK.

Shirane, K. (1975) Conversion of Photoexcitation Energy in Rhodopsin, Nature
254, 722-723.

Schenkl, S., van Mourik, F., van der Zwan, G., Haacke, S., and Chergui, M.
(2005) Probing the ultrafast charge translocation of photoexcited retinal in
bacteriorhodopsin, Science 309, 917-920.

Crescitelli, F. (1972) The visual cells and visual pigments of the vertebrate eye.,
in Handbook of Sensory Physiology (Dartnell, H. J. A., Ed.), pp 245-363,
Springer-Verlag, New York , NY.

Wissinger, B., and Sharpe, L. T. (1998) New aspects of an old theme: The genetic
basis of human color vision, Am J Hum Genet 63, 1257-1262.

Kefalov, V., Fu, Y. B., Marsh-Armstrong, N., and Yau, K. W. (2003) Role of
visual pigment properties in rod and cone phototransduction, Nature 425, 526-
531.

44

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

Sakurai, K., Onishi, A., Imai, H., Chisaka, O., Ueda, Y., Usukura, J., Nakatani,
K., and Shichida, Y. (2007) Physiological properties of rod photoreceptor cells in
green-sensitive cone pigment knock-in mice, J Gen Physiol 130, 21-40.

Shi, G., Yau, K. W., Chen, J ., and Kefalov, V. J. (2007) Signaling properties of a
short-wave cone visual pigment and its role in Phototransduction, J Neurosci 27,
10084-10093.

Pugh, E. N. (2001) Rods are rods and cones cones, and (never) the twain shall
meet, Neuron 32, 375-376.

Hoffmann, R. (1995) The Same and Not the Same, Columbia University Press,
New York, NY.

Bennett, N., Michelvillaz, M., and Kuhn, H. (1982) Light-Induced Interaction
between Rhodopsin and the th-Binding Protein Metarhodopsin-Ii Is the Major
Photoproduct Involved, Eur J Biochem 127, 97-103.

Borhan, B., Souto, M. L., Imai, H., Shichida, Y., and Nakanishi, K. (2000)
Movement of retinal along the visual transduction path, Science 288, 2209-2212.

Emeis, D., Kuhn, H., Reichert, J., and Hofrnann, K. P. (1982) Complex-
Formation between Metarhodopsin-Ii and th-Binding Protein in Bovine
Photoreceptor-Membranes Leads to a Shift of the Photoproduct Equilibrium, F ebs
Lett 143, 29-34.

Kliger, D. S., and Lewis, J. W. (1995) Spectral and kinetic characterization of
visual pigment photointermediates, Israel J Chem 35, 289-307.

Kliger, D. S., Lewis, J. W., and Szundi, I. (1999) Time-resolved spectroscopy of
visual pigment intermediates, Abstr Pap Am Chem S 21 7, U323-U323.

Souto, M. L., Borhan, B., and Nakanishi, K. (2000) Low-temperature
photoaffinity labeling of rhodopsin and intermediates along transduction path,
Vertebrate Phototransduction and the Visual Cycle, Pt B 316, 425-435.

Collins, F. D. (1953) Rhodopsin and indicator yellow, Nature 1 71, 469-472.

Morton, R. A., and Pitt, G. A. J. (1955) Rhodopsin. IX. pH and the hydrolysis of
indicator yellow., Biochem. J. 59, 128-134.

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., and

Miyano, M. (2000) Crystal structure of rhodopsin: A G protein-coupled receptor,
Science 289, 739-745.

Rando, R. R. (1996) Polyenes and vision, Chem Biol 3, 255-262.

45

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

Shichida, Y., Matuoka, S., and Yoshizawa, T. (1984) Formation of

Photorhodopsin, a Precursor of Bathorhodopsin, Detected by Picosecond Laser
Photolysis at Room-Temperature, Photobioch Photobiop 7, 221-228.

Shieh, T., Han, M., Sakmar, T. P., and Smith, S. O. (1997) The steric trigger in
rhodopsin activation, J Mol Biol 269, 373-384.

Yan, B., Nakanishi, K., and Spudich, J. L. (1991) Mechanism of Activation of
Sensory Rhodopsin-I - Evidence for a Steric Trigger, P Natl Acad Sci USA 88,
9412-9416.

Detwiler, P. (2002) Open the loop: Dissecting feedback regulation of a second
messenger transduction cascade, Neuron 36, 3-4.

Graykeller, M. P., and Detwiler, P. B. (1994) The Calcium Feedback Signal in the
Phototransduction Cascade of Vertebrate Rods, Neuron 13, 849-861.

Schnetkamp, P. P. M. (1995) How Does the Retinal Rod Na-Ca+K Exchanger
Regulate Cytosolic-Free Ca2+, J Biol Chem 270, 13231-13239.

Abdulaev, N. G., and Ridge, K. D. (1998) Light-induced exposure of the
cytoplasmic end of transmembrane helix seven in rhodopsin, P Natl Acad Sci
USA 95, 12854-12859.

Farrens, D. L., Altenbach, C., Yang, K., Hubbell, W. L., and Khorana, H. G.
(1996) Requirement of rigid-body motion of transmembrane helices for light
activation of rhodopsin, Science 274, 768-770.

Konig, B., Welte, W., and Hofinann, K. P. (1989) Photoactivation of Rhodopsin
and Interaction with Transducin in Detergent Micelles - Effect of Doping with
Steroid Molecules, F ebs Lett 25 7, 163-166.

SHI, W., Osawa, S., Dickerson, C. D., and Weiss, E. R. (1995) Rhodopsin
Mutants Discriminate Sites Important for the Activation of Rhodopsin Kinase and
G(T), J Biol Chem 270, 2112-2119.

Ernst, O. P., Hofrnann, K. P., and Sakmar, T. P. (1995) Characterization of
Rhodopsin Mutants That Bind Transducin but Fail to Induce th Nucleotide

Uptake - Classification of Mutant Pigments by Fluorescence, Nucleotide Release,
and F lash-Induced Light-Scattering Assays, J Biol Chem 270, 10580-10586.

Franke, R. R., Sakmar, T. P., Graham, R. M., and Khorana, H. G. (1992)
Structure and Function in Rhodopsin - Studies of the Interaction between the
Rhodopsin Cytoplasmic Domain and Transducin, J Biol Chem 267, 14767-14774.

Onrust, R., Herzrnark, P., Chi, P., Garcia, P. D., Lichtarge, O., Kingsley, C., and

Boume, H. R. (1997) Receptor and beta gamma binding sites in the alpha subunit
of the retinal G protein transducin, Science 275, 381-3 84.

46

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

Granovsky, A. E., Natochin, M., and Artemyev, N. O. (1997) The gamma subunit
of rod cGMP-phosphodiesterase blocks the enzyme catalytic site, J Biol Chem
272, 11686-11689.

Stryer, L. (1996) Vision: From photon to perception, P Natl Acad Sci USA 93,
557-559.

Leskov, I. B., Klenchin, V. A., Handy, J. W., Whitlock, G. G., Govardovskii, V.
I., Bownds, M. D., Lamb, T. D., Pugh, E. N., and Arshavsky, V. Y. (2000) The
gain of rod phototransduction: Reconciliation of biochemical and
electrophysiological measurements, Neuron 27, 525-537.

Birge, R. R. (1990) Nature of the Primary Photochemical Events in RhodOpsin
and Bacteriorhodopsin, Biochim Biophys Acta 1016, 293-327.

Pepe, I. M. (2001) Recent advances in our understanding of rhodopsin and
phototransduction, Progress in Retinal and Eye Research 20, 733-759.

Rieke, F ., and Baylor, D. A. (1998) Origin of reproducibility in the responses of
retinal rods to single photons, Biophys J 75, 1836-1857.

Lamb, T. D. (1996) Gain and kinetics of activation in the G-protein cascade of
phototransduction, P Natl Acad Sci USA 93, 566-570.

Miki, N., Keims, J. J., Marcus, F. R., Freeman, J., and Bitensky, M. W. (1973)
Regulation of Cyclic Nucleotide Concentrations in Photoreceptors - Atp-
Dependent Stimulation of Cyclic Nucleotide Phosphodiesterase by Light, P Natl
Acad Sci USA 70, 3820-3824.

Graykeller, M. P., and Detwiler, P. B. (1994) Intracellular Calcium Measurements
in Isolated Rod Photoreceptors, Invest Ophth Vis Sci 35, 1486-1486.

Palczewski, K., Buczylko, J., Vanhooser, P., Carr, S. A., Huddleston, M. J., and
Crabb, J. W. (1992) Identification of the Autophosphorylation Sites in Rhodopsin
Kinase, J Biol Chem 26 7, 18991-18998.

Palczewski, K., Rispoli, G., and Detwiler, P. B. (1992) The Influence of Arrestin
(48k Protein) and Rhodopsin Kinase on Visual Transduction, Neuron 8, 117-126.

Lamb, T. D., and Matthews, H. R. (1988) External and Internal Actions in the
Response of Salamander Retinal Rods to Altered External Calcium-
Concentration, Journal of Physiology-London 403, 473-494.

Sagoo, M. S., and Lagnado, L. (1997) G-protein deactivation is rate-limiting for
shut-off of the phototransduction cascade, Nature 389, 392-395.

Bacigalupo, J., and Lisman, J. E. (1983) Single-Channel Currents Activated by
Light in Limulus Ventral Photoreceptors, Nature 304, 268-270.

47

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

Lott, J. S., Wilde, J. I., Came, A., Evans, N., and Findlay, J. B. C. (1999) The
ordered visual transduction complex of the squid photoreceptor membrane,
Molecular Neurobiology 20, 61-80.

Oday, P. M., Bacigalupo, J ., Vergara, C., and Haab, J. E. (1997) Current issues in
invertebrate phototransduction - Second messengers and ion conductances,
Molecular Neurobiology 15, 41-63.

Hargrave, P. A. (1977) Amino-Terminal Tryptic Peptide of Bovine Rhodopsin -
Glycopeptide Containing 2 Sites of Oligosaccharide Attachment, Biochim
Biophys A cta 492, 83-94.

Hargrave, P. A., Mcdowell, J. H., Curtis, D. R., Wang, J. K., Juszczak, E., Fong,
S. L., Rao, J. K. M., and Argos, P. (1983) The Structure of Bovine Rhodopsin,
Biophysics of Structure and Mechanism 9, 23 5-244.

Okada, T., Fujiyoshi, Y., Silow, M., Navarro, J., Landau, E. M., and Shichida, Y.
(2002) Functional role of internal water molecules in rhodopsin revealed by x-ray
crystallography, P Natl Acad Sci USA 99, 5982-5987.

Okada, T., Sugihara, M., Bondar, A. N., Elstner, M., Entel, P., and Buss, V.
(2004) The retinal conformation and its environment in rhodopsin in light of a
new 2.2 angstrom crystal structure, J Mol Biol 342, 571-583.

Nathans, J. (1990) Determinants of Visual Pigment Absorbency - Identification of

the Retinylidene Schiff-Base Counterion in Bovine Rhodopsin, Biochemistry-Us
29, 9746-9752.

Sakmar, T. P., Franke, R. R., and Khorana, H. G. (1989) Glutamic Acid-113
Serves as the Retinylidene Schiff-Base Counterion in Bovine Rhodopsin, P Natl
Acad Sci USA 86, 8309-8313.

Zhukovsky, E. A., and Oprian, D. D. (1989) Effect of Carboxylic-Acid Side-
Chains on the Absorption Maximum of Visual Pigments, Science 246, 928-930.

RAO, V. R., Cohen, G. B., and Oprian, D. D. (1994) Rhodopsin Mutation G90d
and a Molecular Mechanism for Congenital Night Blindness, Nature 367, 639-
642.

Schertler, G. F. X. (1998) Structure of rhodopsin, EYE 12, 504-510.

Palczewski, K. (2006) G protein-coupled receptor rhodopsin, Annu Rev Biochem
75, 743-767.

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

48

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

Unger, V. M., and Schertler, G. F. X. (1995) Low-Resolution Structure of Bovine
Rhodopsin Determined by Electron Cryomicroscopy, Biophys J 68, 1776—1786.

Nakamichi, H., and Okada, T. (2006) Local peptide movement in the
photoreaction intermediate of rhodopsin, P Natl Acad Sci USA 103, 12729-12734.

Nakamichi, H., and Okada, T. (2006) Crystallographic analysis of primary visual
photochemistry, Angewandte Chemie-International Edition 45, 4270-4273.

Ruprecht, J. J., Mielke, T., Vogel, R., Villa, C., and Schertler, G. F. X. (2004)
Electron crystallography reveals the structure of metarhodopsin I, Embo J 23,
3609-3620.

Ridge, K. D., and Palczewski, K. (2007) Visual rhodopsin sees the light: Structure
and mechanism of G protein signaling, J Biol Chem 282, 9297-9301.

Salom, D., Lodowski, D. T., Stenkamp, R. E., Le Trong, I., Golczak, M.,
Jastrzebska, B., Harris, T., Ballesteros, J. A., and Palczewski, K. (2006) Crystal

structure of a photoactivated deprotonated intermediate of rhodopsin, P Natl Acad
Sci USA 103,16123-16128.

Spudich, J. L., Yang, C. S., Jung, K. H., and Spudich, E. N. (2000) Retinylidene
proteins: Structures and functions from archaea to humans, Annu Rev Cell Dev Bi
16, 365-+.

Sharma, A. K., Spudich, J. L., and Doolittle, W. F. (2006) Microbial rhodopsins:
functional versatility and genetic mobility, Trends in Microbiology 14, 463-469.

Spudich, J. L. (1998) Variations on a molecular switch: transport and sensory
signalling by archaea] rhodopsins, Mol Microbiol 28, 1051-1058.

Oesterhe.D, and Stoecken.W. (1973) Functions of a New Photoreceptor
Membrane, P Natl Acad Sci USA 70, 2853-2857.

Stoeckenius, W., and Bogomolni, R. A. (1982) Bacteriorhodopsin and Related
Pigments of Halobacteria, Annu Rev Biochem 51, 587-616.

Luecke, H., Schobert, B., Richter, H. T., Cartailler, J. P., and Lanyi, J. K. (1999)
Structure of bacteriorhodopsin at 1.55 angstrom resolution, J Mol Biol 291, 899-
911.

Doukas, A. G., Pande, A., Suzuki, T., Callender, R. H., Honig, B., and Ottolenghi,
M. (1981) On the Mechanism of Hydrogen-Deuterium Exchange in
Bacteriorhodopsin, Biophys J 33, 275-279.

Ehrenberg, B., Lewis, A., Porta, T. K., Nagle, J. F., and Stoeckenius, W. (1980)
Exchange Kinetics of the Schiff-Base Proton in Bacteriorhodopsin, P Natl Acad
Sci-Biol 77, 6571-6573.

49

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

94.

Degroot, H. J. M., Harbison, G. S., Herzfeld, J., and Griffin, R. G. (1989) Nuclear
Magnetic-Resonance Study of the Schiff-Base in Bacteriorhodopsin - Counterion
Effects on the N-15 Shift Anisotropy, Biochemistry-Us 28, 3346-3353.

Frigaard, N. U., Martinez, A., Mincer, T. J., and DeLong, E. F. (2006)
Proteorhodopsin lateral gene transfer between marine planktonic Bacteria and
Archaea, Nature 439, 847-850.

Balashov, S. P., Imasheva, E. S., Boichenko, V. A., Anton, J., Wang, J. M., and
Lanyi, J. K. (2005) Xanthorhodopsin: A proton pump with a light-harvesting
carotenoid antenna, Science 309, 2061-2064.

Beja, 0., Aravind, L., Koonin, E. V., Suzuki, M. T., Hadd, A., Nguyen, L. P.,
Jovanovich, S., Gates, C. M., Feldman, R. A., Spudich, J. L., Spudich, E. N., and
DeLong, E. F. (2000) Bacterial rhodopsin: Evidence for a new type of
phototrophy in the sea, Science 289, 1902-1906.

Matsunoyagi, A., and Mukohata, Y. (1977) 2 Possible Roles of Bacteriorhodopsin
- Comparative-Study of Strains of Halobacterium-Halobium Differing in

Pigmentation, Biochemical and Biophysical Research Communications 78, 237-
243.

Mongodin, E. F., Nelson, K. E., Daugherty, S., DeBoy, R. T., Wister, J., Khouri,
H., Weidman, J., Walsh, D. A., Papke, R. T., Perez, G. S., Sharma, A. K., Nesbo,
C. L., MacLeod, D., Bapteste, E., Doolittle, W. F ., Charlebois, R. L., Legault, B.,
and Rodriguez-Valera, F. (2005) The genome of Salinibacter ruber: Convergence

and gene exchange among hyperhalophilic bacteria and archaea, P Natl Acad Sci
USA 102,18147-18152.

Spudich, J. L. (2006) The multitalented microbial sensory rhodopsins, Trends in
Microbiology 14, 480-487.

Bogomolni, R. A., and Spudich, J. L. (1982) Identification of a 3rd-Rhodopsin-
Like Pigment in Phototactic Halobacterium-Halobium, P Natl Acad Sci-Biol 79,
6250-6254.

Spudich, J. A., Kuczrnarski, E. R., Pardee, J. D., Simpson, P. A., Yamarnoto, K.,
and Stryer, L. (1981) Control of Assembly of Dictyostelium Myosin and Actin-
Filaments, Cold Spring Harbor Symposia on Quantitative Biology 46, 553-561.

Ren, L., Martin, C. H., Wise, K. J., Gillespie, N. B., Luecke, H., Lanyi, J. K.,
Spudich, J. L., and Birge, R. R. (2001) Molecular mechanism of spectral tuning in
sensory rhodopsin II, Biochemistry— Us 40, 13906-13914.

Bergo, V., Spudich, E. N., Spudich, J. L., and Rothschild, K. J. (2003)
Conformational changes detected in a sensory rhodopsin II-transducer complex, J
Biol Chem 278, 36556-36562.

50

95.

96.

97.

98.

99.

100.

101.

102.

103.

104.

105.

106.

107.

Furutani, Y., Kamada, K., Sudo, Y., Shimono, K., Kamo, N., and Kandori, H.
(2005) Structural changes of the complex between pharaonis phoborhodopsin and

its cognate transducer upon formation of the m photointermediate, Biochemistry-
Us 44, 2909-2915.

Hein, M., Radu, 1., Klare, J. P., Engelhard, M., and Siebert, F. (2004)
Consequences of counterion mutation in sensory rhodopsin II of
Natronobacterium pharaonis for photoreaction and receptor activation: An FTIR
study, Biochemistry-Us 43, 995-1002.

Sudo, Y., Furutani, Y. J., Shimono, K., Kamo, N., and Hidekori, K. (2003)
Hydrogen bondin alteration of Thr-204 in the complex between pharaonis
phoborhodopsin and its transducer protein, Biochemistry- Us 42, 14166-14172.

Jung, K. H., Trivedi, V. D., and Spudich, J. L. (2003) Demonstration of a sensory
rhodopsin in eubacteria, Mol Microbiol 47, 1513-1522.

Vogeley, L., Sineshchekov, O. A., Trivedi, V. D., Sasaki, J., Spudich, J. L., and
Luecke, H. (2004) Anabaena sensory rhodopsin: A photochromic color 0 sensor
at 2.0 angstrom, Science 306, 1390-1393.

Sudo, Y., and Spudich, J. L. (2006) Three strategically placed hydrogen-bonding
residues convert a proton pump into a sensory receptor, P Natl Acad Sci USA 103,
16129-16134.

Bowmaker, J. K., and Dartnall, H. J. A. (1980) Visual Pigments of Rods and
Cones in a Human Retina, Journal of Physiology-London 298, 501-511.

Asenjo, A. B., Rim, J., and Oprian, D. D. (1994) Molecular Determinants of
Human Red/Green Color Discrimination, Neuron 12, 1131-1138.

Merbs, S. L., and Nathans, J. (1992) Absorption-Spectra of Human Cone
Pigments, Nature 356, 433-435.

Oprian, D. D., Asenjo, A. B., Lee, N., and Pelletier, S. L. (1991) Design,
Chemical Synthesis, and Expression of Genes for the 3 Human Color-Vision
Pigments, Biochemistry- Us 30, 1 1367-1 1372.

Stockman, A., Sharpe, L. T., Merbs, S., and Nathans, J. (2000) Spectral
sensitivities of human cone visual pigments determined in vivo and in vitro,
Vertebrate Phototransduction and the Visual Cycle, Pt B 316, 626-650.

Dawson, T. L. (2006) Colour and colour vision of creatures great and small,
Coloration Technology 122, 61-73.

Bowmaker, J. K. (1991) The evolution of vertebrate visual pigments and
photoreceptors, The Macmillan Press, Houndsmills, Basingstoke, Hampshire,
UK.

51

108.

109.

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

Kelber, A., Vorobyev, M., and Osorio, D. (2003) Animal colour vision -
behavioural tests and physiological concepts, Biological Reviews 78, 81-118.

Wald, G. (1933) Vitamin A in the retina., Nature 132, 316-317.

Morton, R. A., and Goodwin, T. W. (1944) Preparation of retinene in vitro.,
Nature 153, 405-406.

Wald, G., and Brown, P. K. (1950) The Synthesis of Rhodopsin from Retinene, P
Natl Acad Sci USA 36, 84-92.

Akhtar, M., Blosse, D., and Dewhurst, P. (1968) Studies on vision. The nature of
the retinal-opsin linkage, Biochem. J. 110, 693-702.

Bownds, D. (1967) Site of attachment of retinal in rhodopsin, Nature 216, 1178-
1181.

Oseroff, A. R., and Callende.Rh. (1974) Resonance Raman-Spectroscopy of
Rhodopsin in Retinal Disk Membranes, Biochemistry- Us 13, 4243-4248.

Blatz, P. E., Baumgartner, N., Balasubramaniyan, B., Balasubramaniyan, P., and
Stedman, F. (1971) Wavelength regulation in visual pigment chromophore. Large
induced bathochromic shifts in retinol and related polyenes., Photochem.
Photobiol. 14, 531-549.

Pitt, G. A. J., Collins, F. D., Morton, R. A., and Stok, P. (1955) Rhodopsin. VIII.
N-Retinylidenemethylamine, an indicator yellow analog, Biochem. J. 59, 122-
128.

Blatz, P. E., Mohler, J. H., and Ahmed, W. (1991) Spectroscopic Observation of
Solvent Interaction with Selected Retinal Schiff-Bases, Photochem Photobiol 54,
255-264.

Erickson, J. O., and Blatz, P. E. (1968) N-retinylidene-1-amino-2-propanol: a
Schiff base analog for rhodopsin, Vision Res. 8, 1367-1375.

Motto, M. G., Sheves, M., Tsujimoto, K., Baloghnair, V., and Nakanishi, K.
(1980) Opsin Shifts in Bovine Rhodopsin and Bacteriorhodopsin - Comparison of
2 External Point-Charge Models, J Am Chem Soc 102, 7947-7949.

Nakanishi, K., Baloghnair, V., Amaboldi, M., Tsujimoto, K., and Honig, B.
(1980) An External Point-Charge Model for Bacteriorhodopsin to Account for Its
Purple Color, J Am Chem Soc 102, 7945-7947.

Hubbard, R. E. (1958) Paper 4, in Proc. Nat. Phys. Lab. Symp. No. 8, London.

52

122.

123.

124.

125.

126.

127.

128.

129.

130.

131.

132.

133.

Blatz, P. E., and Mohler, J. H. (1975) Effect of Selected Anions and Solvents on

Electronic Absorption, Nuclear Magnetic-Resonance, and Infrared-Spectra of N-
Retinylidene-N-Butylammonium Cation, Biochemistry- Us 14, 2304-2309.

Honig, B., Greenberg, A. D., Dinur, U., and Ebrey, T. G. (1976) Visual-Pigment
Spectra - Implications of Protonation of Retinal Schiff-Base, Biochemistry- Us 15,
4593-4599.

Kliger, D. S., Milder, S. J., and Dratz, E. A. (1977) Solvent Effects on Spectra of
Retinal Schiff-Bases .1. Models for Bathochromic Shift of Chromophore
Spectrum in Visual Pigments, Photochem Photobiol 25, 277-286.

Suzuki, H., Komatsu, T., and Kitajirna, H. (1974) Theory of Optical Property of
Visual Pigment, Journal of the Physical Society of Japan 3 7, 177-185.

Waddell, W. H., Schaffer, A. M., and Becker, R. S. (1977) Visual Pigments .7.
Experimental and Theoretical Investigations of Absorption Spectral Properties of
Protonated Retinal Schiff-Bases and Implications for Bathochromic Shift in
Visual Pigments, J Am Chem Soc 99, 8456-8460.

Amaboldi, M., Motto, M. G., Tsujimoto, K., Baloghnair, V., and Nakanishi, K.
(1979) Hydro-Retinals and Hydro-Rhodopsins, J Am Chem Soc 101, 7082-7084.

Mollevanger, L. C. P. J., Kentgens, A. P. M., Pardoen, J. A., Courtin, J. M. L.,
Veeman, W. S., Lugtenburg, J ., and Degrip, W. J. (1987) High-Resolution Solid-
State C-13-Nmr Study of Carbons C-5 and C-12 of the Chromophore of Bovine

Rhodopsin - Evidence for a 6-S-Cis Conformation with Negative-Charge
Perturbation near C-12, Eur J Biochem 163, 9-14.

Nakanishi, K. (1985) Bioorganic Studies with Rhodopsin, Pure Appl Chem 57,
769-776.

Nathans, J., Thomas, D., and Hogness, D. S. (1986) Molecular-Genetics of
Hmnan Color-Vision - the Genes Encoding Blue, Green, and Red Pigments,
Science 232, 193-202.

Gat, Y., and Sheves, M. (1993) A Mechanism for Controlling the Pk(a) of the
Retinal Protonated Schiff-Base in Retinal Proteins - a Study with Model
Compounds, J Am Chem Soc 115, 3772-3773.

Blatz, P. E., Mohler, J. H., and Navangul, H. V. (1972) Anion-Induced
Wavelength Regulation of Absorption Maxima of Schiff-Bases of Retinal,
Biochemistry-Us 11, 848-&.

Livnah, N., and Sheves, M. (1993) The Schiff-Base Bond Configuration in
Bacteriorhodopsin and in Model Compounds, Biochemistry- Us 32, 7223-7228.

53

134.

135.

136.

137.

138.

139.

140.

141.

142.

143.

144.

Livnah, N., and Sheves, M. (1993) Model Compounds Can Mimic Spectroscopic
Properties of Bovine Rhodopsin, J Am Chem Soc 115, 351-353.

Blatz, P. E., and Liebman, P. A. (1973) Wavelength regulation in visual
pigments, Exp. Eye Res. 17, 573-580.

Tan, Q., Lou, J. H., Borhan, B., Karnaukhova, E., Berova, N., and Nakanishi, K.
(1997) Absolute sense of twist of the C12-C13 bond of the retinal chromophore in
bovine rhodopsin based on exciton-coupled CD spectra of 11,12-dihydroretinal
analogues, Angew Chem Int Edit 36, 2089-2093.

Verdegem, P. J. E., Bovee-Geurts, P. H. M., de Grip, W. J., Lugtenburg, J., and de
Groot, H. J. M. (1999) Retinylidene ligand structure in bovine rhodopsin,
metarhodopsin-I, and lO-methylrhodopsin from internuclear distance
measurements using C-l3-labeling and 1-D rotational resonance MAS NMR,

Biochemistry- Us 38, 1 1316-1 1324.

Honig, B., Hudson, B., Sykes, B. D., and Karplus, M. (1971) Ring Orientation in
Beta-Ionone and Retinals (Theoretical/Nmr/Semi-Empirical
Approach/Overhauser Effect/S-Cis Conformation), P Natl Acad Sci USA 68,
1289-&.

Rafferty, C. N. (1979) Light-Induced Perturbation of Aromatic Residues in
Bovine Rhodopsin and Bacteriorhodopsin, Photochem Photobiol 29, 109-120.

Stenkamp, R. E., Filipek, S., Driessen, C. A. G. G., Teller, D. C., and Palczewski,
K. (2002) Crystal structure of rhodopsin: a template for cone visual pigments and
other G protein-coupled receptors, Bba-Biomembranes 1 5 65 , 168-182.

Yokoyama, R., and Yokoyama, S. (1990) Convergent Evolution of the Red-Like

and Green-Like Visual Pigment Genes in Fish, Astyanax-Fasciatus, and Human,
P Natl Acad Sci USA 87, 9315-9318.

Chan, T., Lee, M., and Sakmar, T. P. (1992) Introduction of Hydroxyl-Bearing
Amino-Acids Causes Bathochromic Spectral Shifts in Rhodopsin - Amino-Acid
Substitutions Responsible for Red-Green Color Pigment Spectral Tuning, J Biol
Chem 26 7, 9478-9480.

Altun, A., Yokoyama, S., and Morokuma, K. (2008) Quantum
mechanical/molecular mechanical studies on spectral tuning mechanisms of visual
pigments and other photoactive proteins, Photochem Photobiol 84, 845-854.

Fujimoto, K., Hasegawa, J. Y., Hayashi, S., and Nakatsuji, H. (2006) On the
color-tuning mechanism of Human-Blue visual pigment: SAC-CI and QM/MM
study, Chem Phys Lett 432, 252-256.

54

145.

146.

147.

148.

149.

150.

151.

152.

153.

154.

155.

156.

Fujimoto, K., Hasegawa, J. Y., and Nakatsuji, H. (2008) Origin of color tuning in
human red, green, and blue cone pigments: SAC-CI and QM/MM study, Chem
Phys Lett 462, 318-320.

Birge, R. R., Murray, L. P., Pierce, B. M., Akita, H., Baloghnair, V., Findsen, L.
A., and Nakanishi, K. (1985) 2-Photon Spectroscopy of Locked-1 l-Cis-

Rhodopsin - Evidence for a Protonated Schiff-Base in a Neutral Protein-Binding
Site, P Natl Acad Sci USA 82, 4117-4121.

Mathies, R., and Stryer, L. (1976) Retinal Has a Highly Dipolar Vertically
Excited Singlet-State - Implications for Vision, P Natl Acad Sci USA 73, 2169-
2173.

Wanko, M., Hoffrnann, M., Frauenheim, T., and Elstner, M. (2006)
Computational photochemistry of retinal proteins, J Comput Aid Mol Des 20,
511-518.

Trabanino, R. J., Vaidehi, N., and Goddard, W. A. (2006) Exploring the
molecular mechanism for color distinction in humans, J Phys Chem B 110,
17230-17239.

Houjou, H., Inoue, Y., and Sakurai, M. (2001) Study of the opsin shift of
bacteriorhodopsin: Insight from QM/MM calculations with electronic polarization
effects of the protein environment, J Phys Chem B 105, 867-879.

Warshel, A., and Chu, Z. T. (2001) Nature of the surface crossing process in
bacteriorhodopsin: Computer simulations of the quantum dynamics of the primary
photochemical event, J Phys Chem B 105, 9857-9871.

Yan, E. C. Y., Kazrni, M. A., Ganim, Z., Hou, J. M., Pan, D. H., Chang, B. S. W.,
Sakmar, T. P., and Mathies, R. A. (2003) Retinal counterion switch in the

photoactivation of the G protein-coupled receptor rhodopsin, P Natl Acad Sci
USA 100, 9262-9267.

Kholmurodov, K. T., Fel'dman, T. B., and Ostrovsky, M. A. (2007) Interaction of
chromophore, ll-cis-retinal, with amino acid residues of the visual pigment
rhodopsin in the region of protonated Schiff base: A molecular dynamics study,
Russ Chem 3+ 56, 20-27.

Martinez-Mayorga, K., Pitrnan, M. C., Grossfield, A., Feller, S. E., and Brown,
M. F. (2006) Retinal counterion switch mechanism in vision evaluated by
molecular simulations, J Am Chem Soc 128, 16502-16503.

Giguere, V. (1994) Retinoic Acid Receptors and Cellular Retinoid-Binding
Proteins - Complex Interplay in Retinoid Signaling, Endocrine Reviews 15, 61-79.

Charnbon, P. (1996) A decade of molecular biology of retinoic acid receptors,
Faseb J 10, 940-954.

55

157.

158.

159.

160.

161.

162.

163.

164.

165.

166.

167.

168.

Lotan, R. (1980) Effects of Vitamin-a and Its Analogs (Retinoids) on Normal and
Neoplastic-Cells, Biochimica Et Biophysica Acta 605, 33-91.

DeLuca, L. M. (1991) Retinoids and their receptors in differeciation,
embryogenesis, and neoplasia., Faseb J 5, 2924-2933.

Chomienne, C., F enaux, P., and Degos, L. (1996) Retinoid differentiation therapy
in promyelocytic leukemia, Faseb J 10, 1025-1030.

Delva, L., Bastie, J. N., Rochette-Egly, C., Kraiba, R., Balitrand, N., Despouy, G.,
Chambon, P., and Chomienne, C. (1999) Physical and functional interactions
between cellular retinoic acid binding protein [1 and the retinoic acid-dependent
nuclear complex, Molecular and Cellular Biology 1 9, 7158-7167.

Fisher, G. J ., and Voorhees, J. J. (1996) Molecular mechanisms of retinoid actions
in skin, Faseb J 10, 1002-1013.

Tallman, M. S. (1996) Differentiating therapy in acute myeloid leukemia,
Leukemia 10, 1262-1268.

Hong, W. K., Itri, L. M. (1994) Retinoids and human cancer, in The Retinoids:
Biology, Chemistry, and Medicine (Sporn, M. B., Roberts, A. B., Goodman, D. S.,
Ed.) 2nd ed., pp 597-630, Raven Press, New York.

Vaezeslami, S., Mathes, E., Vasilelou, C., Borhan, B., and Geiger, J. H. (2006)
The structure of apo-wild-type cellular retinoic acid binding protein [1 at 1.4 A
and its relationship to ligand binding and nuclear translocation, J Mol Biol 363,
687-701.

Kliewer, S. A., Umesono, K., Mangelsdorf, D. J ., and Evans, R. M. (1992)
Retinoid X-Receptor Interacts with Nuclear Receptors in Retinoic Acid, Thyroid-
Hormone and Vitamin-D3 Signaling, Nature 355, 446-449.

Kreutz, M., Fritsche, J., Andreesen, R., and Krause, S. W. (1998) Regulation of
cellular retinoic acid binding protein (CRABP 11) during human monocyte
differentiation in vitro, Biochem Bioph Res Co 248, 830-834.

Leid, M., Kastner, P., Lyons, R., Nakshatri, H., Saunders, M., Zacharewski, T.,
Chen, J. Y., Staub, A., Gamier, J. M., Mader, S., and Chambon, P. (1992)
Purification, Cloning, and er Identity of the Hela—Cell Factor with Which Rar or
Tr Heterodimerizes to Bind Target Sequences Efficiently, Cell 68, 377-395.

Leroy, P., Krust, A., Kastner, P., Mendelsohn, C., Zelent, A., Chambon, P. (1992)
Retinoic acid receptors, in Retinoids in Normal Development and T eratogenesis
(Morris-Kay, G., ed., Ed.), pp 7-25, Oxford University Press, Oxford.

56

169.

170.

171.

172.

173.

174.

175.

176.

177.

178.

179.

180.

Mangelsdorf, D. J., Thummel, C., Beato, M., Herrlich, P., Schutz, G., Umesono,
K., Blumberg, B., Kastner, P., Mark, M., Chambon, P., and Evans, R. M. (1995)
The Nuclear Receptor Superfamily - the 2nd Decade, Cell 83, 835-839.

Morrisskay, G. (1993) Retinoic Acid and Craniofacial Development - Molecules
and Morphogenesis, Bioessays 15, 9-15.

Bailey, J. S., and SIU, C. H. (1988) Purification and Partial Characterization of a
Novel Binding-Protein for Retinoic Acid from Neonatal Rat, J Biol Chem 263,
9326-9332.

ONG, D. E., and Chytil, F. (1975) Retinoic Acid-Binding Protein in Rat Tissue -
Partial-Purification and Comparison to Rat Tissue Retinol-Binding Protein, J Biol
Chem 250, 6113-6117.

Sani, B. P. H., D.L. (1974) Retinoic acid-binding-protein in chick-embryo
metatarsal skin, Biochem. Biophys. Res. Commun. 61, 1276-1282.

Willmert, L. J., and Noy, N. (2003) Involvement of CRABP-II in transcriptional
activity of RAR, Faseb J 17, A315-A3l6.

Ong, D. E., Newcomer, M. E., & Chytil, F. (1994) Cellular retinoid-binding
proteins, in The Retinoids: Biology, Chemistry, and Medicine (Sporn, M. B.,
Roberts, A. B., Goodman, D. S., Ed.) 2nd ed., pp 283-318, Raven Press, New
York.

Bertucci, F., Houlgatte, R., Benziane, A., et al. (2000) Gene expression profiling
of primary breast carcinomas using arrays of candidate genes, Hum. Mol. Genet.
9, 2981-2991.

Bertucci, F., Van Hulst, S., Bernard, K., Loriod, B., Granjeaud, S., Tagett, R.,
Starkey, M., Nguyen, C., Jordan, 8., and Birnbaum, D. (1999) Expression
scanning of an array of growth control genes in human tumor cell lines, Oncogene
18, 3905-3912.

Delva, L., Comic, M., Balitrand, N., Guidez, F., Miclea, J. M., Delmer, A.,
Teillet, F., Fenaux, P., Castaigne, S., Degos, L., and Chomienne, C. (1993)
Resistance to All-Trans-Retinoic Acid (Atra) Therapy in Relapsing Acute
Promyelocytic Leukemia - Study of in-Vitro Atra Sensitivity and Cellular
Retinoic Acid-Binding Protein-Levels in Leukemic-Cells, Blood 82, 2175-2181.

Hibbs, K., Skubitz, K. M., Pambuccian, S. E., Casey, R. C., Burleson, K. M.,
Oegema, T. R., Thiele, J. J., Grindle, S. M., Bliss, R. L., and Skubitz, A. P. N.
(2004) Differential gene expression in ovarian carcinoma - Identification of
potential biomarkers, American Journal of Pathology 165, 397-414.

Li, C. M., Guo, M. R., Borczuk, A., Powell, C. A., Wei, M., Thaker, H. M.,
Friedman, R., Klein, U., and Tycko, B. (2002) Gene expression in Wilms' tumor

57

181.

182.

183.

184.

185.

186.

187.

188.

189.

190.

mimics the earliest committed stage in the metanephric mesenchymal-epithelial
transition, American Journal of Pathology [60, 2181-2190.

Tsibris, J. C. M., Segars, J., Coppola, D., Mane, S., Wilbanks, G. D., O'Brien, W.
F., and Spellacy, W. N. (2002) Insights from gene arrays on the development and
growth regulation of uterine leiomyomata, Fertility and Sterility 78, 114-121.

Vo, H. P., and Crowe, D. L. (1998) Transcriptional regulation of retinoic acid
responsive genes by cellular retinoic acid binding protein-II modulates RA
mediated tumor cell proliferation and invasion, Anticancer Research 18, 217-224.

Zhou, D. C., Hallam, S. J., Lee, S. J., Klein, R. S., Wiemik, P. H., Tallman, M. S.,
and Robert, E. (1998) Constitutive expression of cellular retinoic acid binding
protein 11 and lack of correlation with sensitivity to all-trans retinoic acid in acute
promyelocytic leukemia cells, Cancer Research 58, 5770-5776.

Budhu, A. S., and Noy, N. (2002) Direct channeling of retinoic acid between
cellular retinoic acid-binding protein 11 and retinoic acid receptor sensitizes

mammary carcinoma cells to retinoic acid—induced growth arrest, Molecular and
Cellular Biology 22, 2632-2641.

Dong, D., Ruuska, S. E., Levinthal, D. J., and Noy, N. (1999) Distinct roles for
cellular retinoic acid-binding proteins I and II in regulating signaling by retinoic
acid, J Biol Chem 274, 23695-23698.

Sessler, R. J ., and Noy, N. (2005) A ligand-activated nuclear localization signal in
cellular retinoic acid binding protein-II, Molecular Cell 18, 343-353.

Vaezeslami, S., Mathes, E., Vasileiou, C., Borhan, B., and Geiger, J. H. (2006)
The Structure of Apo-wild-type Cellular Retinoic Acid Binding Protein II at 1.4 A
and its Relationship to Ligand Binding and Nuclear Translocation, J Mol Biol.

Kleywegt, G. J ., Bergfors, T., Senn, H., Lemotte, P., Gsell, B., Shudo, K., and
Jones, T. A. (1994) Crystal-Structures of Cellular Retinoic Acid-Binding Protein-I
and Protein-Ii in Complex with All-Trans-Retinoic Acid and a Synthetic Retinoid,
Structure 2, 1241-1258.

Chen, X., Tordova, M., Gilliland, G. L., Wang, L. C., Li, Y., Yan, H. G., and Ji,
X. H. (1998) Crystal structure of apo-cellular retinoic acid-binding protein type 11
(R1 11M) suggests a mechanism of ligand entry, J Mol Biol 278, 641-653.

Thompson, J. R., Bratt, J. M., and Banaszak, L. J. (1995) Crystal-Structure of
Cellular Retinoic Acid-Binding Protein-I Shows Increased Access to the Binding
Cavity Due to Formation of an Intermolecular Beta-Sheet, J. Mol. Biol. 252, 433-
446.

58

191.

192.

193.

194.

195.

196.

197.

198.

199.

200.

201.

Scapin, G., Gordon, J. I., and Sacchettini, J. C. (1992) Refinement of the Structure
of Recombinant Rat Intestinal Fatty Acid-Binding Apoprotein at 1.2-a Resolution,
.1. Biol. Chem. 26 7, 4253-4269.

Wang, L. C., and Yan, H. G. (1998) NMR study suggests a major role for Arg] 11
in maintaining the structure and dynamical properties of type [1 human cellular
retinoic acid binding protein, Biochemistry- Us 3 7, 13021-13032.

Chaudhuri, B. N., Kleywegt, G. J., Broutin-L'Hermite, 1., Bergfors, T., Senn, H.,
Le Motte, P., Partouche, 0., and Jones, T. A. (1999) Structures of cellular retinoic

acid binding proteins 1 and II in complex with synthetic retinoids, Acta
Crystallogr D 55, 1850-1857.

Kleywegt, G. J., Bergfors, T., Senn, H., Le Motte, P., Gsell, B., Shudo, K., and
Jones, T. A. (1994) Crystal structures of cellular retinoic acid binding proteins I
and II in complex with all-trans retinoic acid and a synthetic retinoid, Structure
2,1241-1258.

Crist, R. M., Vasileiou, C., Rabago-Smith, M., Geiger, J. H., and Borhan, B.
(2006) Engineering a rhodopsin protein mimic, J Am Chem Soc 128, 4522-4523.

Vasileiou, C., Vaezeslami, S., Crist, R. M., Rabago-Smith, M., Geiger, J. H., and
Borhan, B. (2007) Protein design: Reengineering Cellular Retinoic Acid Binding
Protein 11 into a rhodopsin protein mimic, J Am Chem Soc 129, 6140-6148.

Saari, J. C., Bredberg, L., and Garwin, G. G. (1982) Identification of the
Endogenous Retinoids Associated with 3 Cellular Retinoid-Binding Proteins from
Bovine Retina and Retinal-Pigment Epithelium, J Biol Chem 25 7, 3329-3333.

Saari, J. C., and Bredberg, D. L. (1987) Photochemistry and Stereoselectivity of
Cellular Retinaldehyde-Binding Protein from Bovine Retina, J Biol Chem 262,
7618-7622.

Saari, J. C., Nawrot, M., Kennedy, B. N., Garwin, G. G., Hurley, J. H, Huang, J.,
Possin, D. E., and Crabb, J. W. (2001) Visual cycle impairment in cellular
retinaldehyde binding protein (CRALBP) knockout mice results in delayed dark
adaptation, Neuron 29, 739-748.

Golovleva, 1., Bhattacharya, S., Wu, Z. P., Shaw, N., Yang, Y. W., Andrabi, H.,
West, K. A., Burstedt, M. S. I., Forsman, K., Holmgren, G., Sandgren, O., Noy,
N., Qin, J., and Crabb, J. W. (2003) Disease-causing mutations in the cellular

retinaldehyde binding protein tighten and abolish ligand interactions, J Biol Chem
278, 12397-12402.

Stecher, H., Gelb, M. H., Saari, J. C., and Palczewski, K. (1999) Preferential
release of ll-cis-retinol from retinal pigment epithelial cells in the presence of
cellular retinaldehyde-binding protein, J Biol Chem 274, 8577-8585.

59

202.

203.

204.

205.

206.

207.

208.

209.

210.

211.

212.

Winston, A., and Rando, R. R. (1998) Regulation of isomerohydrolase activity in
the visual cycle, Biochemistry- Us 3 7, 2044-2050.

Wu, Z. P., Yang, Y. W., Shaw, N., Bhattacharya, S., Yan, L., West, K., Roth, K.,
Noy, N., Qin, J., and Crabb, J. W. (2003) Mapping the ligand binding pocket in
the cellular retinaldehyde binding protein, J Biol Chem 278, 12390-12396.

Nawrot, M., West, K., Huang, J., Possin, D. E., Bretscher, A., Crabb, J. W., and
Saari, J. C. (2004) Cellular retinaldehyde-binding protein interacts with ERM-

binding phosphoprotein 50 in retinal pigment epithelium, Invest Ophth Vis Sci 45 ,
393-401.

Bhattacharya, S. K., Wu, Z., Jin, Z., Yan, L., Miyagi, M., West, K., Nawrot, M.,
Saari, J. C., and Crabb, J. W. (2002) Proteomic approach to identification of a
mammalian visual cycle protein complex., Faseb J 16, Al4-A14.

Crabb, J. W., Carlson, A., Chen, Y., Goldflarn, S., Intres, R., West, K. A.,
Hulmes, J. D., Kapron, J. T., Luck, L. A., Horwitz, J., and Bok, D. (1998)
Structural and functional characterization of recombinant human cellular
retinaldehyde-binding protein, Protein Sci 7, 746-757.

Wu, Z. P., Hasan, A., Liu, T. Y., Teller, D. C., and Crabb, J. W. (2004)
Identification of CRALBP ligand interactions by photoaffmity labeling,
hydrogen/deuteritun exchange, and structural modeling, J Biol Chem 279, 27357-
27364.

Panagabko, C., Morley, S., Hernandez, M., Cassolato, P., Gordon, H., Parsons,
R., Manor, D., and Atkinson, J. (2003) Ligand specificity in the CRAL-TRIO
protein family, Biochemistry- Us 42, 6467-6474.

Meier, R., Tomizaki, T., Schulze-Briese, C., Baumann, U., and Stocker, A. (2003)
The molecular basis of vitamin E retention: Structure of human alpha-tocopherol
transfer protein, J Mol Biol 331, 725-734.

Min, K. C., Kovall, R. A., and Hendrickson, W. A. (2003) Crystal structure, of
human alpha-tocopherol transfer protein bound to its ligand: Implications for
ataxia with vitamin E deficiency, P Natl Acad Sci USA 100, 14713-14718.

Sha, B. D., Phillips, S. E., Bankaitis, V. A., and Luo, M. (1998) Crystal structure
of the Saccharomyces cerevisiae phosphatidylinositol-transfer protein, Nature
391, 506-510.

Stocker, A., and Baumann, U. (2003) Supernatant protein factor in complex with
RRR-alpha-tocopherquuinone: A link between oxidized vitamin E and
cholesterol biosynthesis, J Mol Biol 332, 759-765.

60

213. Liu, T. Y., Jenwitheesuk, E., Teller, D. C., and Samudrala, R. (2005) Structural
insights into the cellular retinaldehyde-binding protein (CRALBP), Proteins-
Structure Function and Bioinformatics 61 , 412-422.

61

CHAPTER II. Crystal Structures of Ligand-bound
Cellular Retinoic Acid Binding Protein ll Mutants

2. 1. Introduction.

2.1.1 . Need for a ligand that can be totally buried inside the CRABPII

mutants.

The triple mutant KLE-CRABPII was a successful design in that it proved that the
protonated Schiff base and the stabilizing counter anion are two essential factors in
inducing the red shift of the bound retinal (1 ). However, there are still limitations of the
KLE-CRABPII-all-trans-retinal system in terms of its application towards investigation
of the point charge theory. According to the point charge theory, the placement of
counter anions along the polyene chain would potentially stabilize the positive charge of
the bound ligand at different positions and therefore lead to stabilization of different
resonance structures, which would result in differed red shift abilities. To test this theory,
the ideal situation is that the ligand is fully buried inside the protein and able to
experience various protein-ligand interactions generated by binding site mutations.
However, as shown in Figure II-l, the all-trans-retinal-bound CRABPII mutants all
feature a binding mode with the ionone ring of the retinoid partially solvent-exposed. To
circumvent this less-than-ideal situation associated with the Rt-bound CRABPII model
systems, we decided to explore the possibility of introducing a shorter ligand into the

CRABPII mutants. By docking simulations, we have identified a shorter retinal analog,

62

C15-aldehyde (Figure II-2), which is predicted to be totally buried inside the protein once

bound to the CRABPII mutants.

 

Figure [I-l. Retinoic acid bound in wild-type-CRABPII.

OH H H OH
I \ \ \O HO\%’N\/\’N]<OH
HO HO
C15-aldehyde Bls-trIs-propane (BTP)
| \ \ \ \ ‘0

all-trans retlnal

Figure [1-2. Structures of the related compounds.

63

Table [1-1. CRABPII mutants bound with Cls-aldehyde.

 

 

Mutant kmgxa red. am.b quen. fluor.C
1 C15—aldehyde (buffer) 340 N/A N/A
2 C15-aldehyde PSB (EtOH) 330 N/A N/A
3 R132K2R1 I 1L:L121E 380 yes 0.80
4 R132K1R1 1 1L:L121E:R59W 404 yes 0.81
5 R132K2R111LzL121EzR59E 424 yes 0.81
6 R132K:R111L:L121E:R59D 414 yes 0.87
7 R132K:R111L:L121E:A36E 410 yes 0.89
8 RI32K1R1 l 1L:L121E:V76Y 388 yes 0.85
9 R132K:R111L:L121Q 413 yes 0.80
10 R132K:R111L:A32E 407 yes 0.61
1 1 R132K:R1 11L:T54E 402 yes 0.67
12 R132KIY134F:R1 1 1L:L121D1T54V 415 yes 0.78
13 R132K:R111L:R59L 406 yes 0.57
14 R132K:R111L:F15D 390 yes 0.69

 

b
a Deconvolusion of overlapping UV-vis spectra is detailed in reference (1). Yes/no refer to the
+
results obtained from MALDI-TOF analysis (presence of [M+204] ) of protein-C1 5-aldehyde

. . . . . . C
complex that has been subjected to reductive amrnatron conditions. End fluorescence level after
fluorescence titration.

The C15-aldehyde was then synthesized by Kin-Sing Lee and subsequently used
for binding assays with different CRABPII mutants. Some of the results are shown in

Table II-l (results from Kin-Sing Lee). The Cis-aldehyde itself absorbs at 340 nm as a

free ligand in buffer. Red shift to 380 nm is observed upon formation of the protonated

Schiff base when C15-aldehyde is incubated with tributylamine. Further red shift is

observed when the protonated Schiff base is formed between the CRABPII mutants and

the C15-aldehyde, which is referred to as "opsin shift" since it is induced by protein-

chromophore interactions. The most dramatic opsin shift is observed in the CRABPII

mutant, KLE-R59E, where an opsin shift of 44 nm is obtained. Compared to the Rt-

64

KLE-CRABPII complex where the opsin shift is ~7 nm, the C15-aldehyde-bound

CRABPII mutants seem to possess a better wavelength red shift ability.

The focus now shifts toward understanding of the details of how C15-aldehyde is

bound in the CRABPII mutants in order to further explain the observed binding data as

well as to guide our future rational re-engineering of CRABPII. The determination of the

C15-aldehyde-bound CRABPII mutant structures was therefore pursued.

2.1.2. Engineered CRABPII mutants as switchable colorimetric fusion

proteins.
The ability of the CRABPII mutants to form protonated Schiff bases with the

aldehydic ligands and afford a red shift of the ligands Amax suggests a natural advantage

of the system in terms of generating ligand-dependent fluorescence. By introducing
highly conjugated chromophores to the designed CRABPII mutants, we could envision a
situation where a particular protein-ligand complex is able to produce fluorescence. Such
a scenario, combined with the fact that CRABPII is a small and soluble cellular protein,
suggests a potential of the designed CRABPII mutants being used as ligand-inducible
fluorescent fusion tags.

Green Fluorescence Protein (GFP) has been the top choice in the recent decade
(2-8). GFP is a spontaneously fluorescent protein from the jelly fish Aquorea victoria.
The wide usage of GFP and its genetic variants as fusion tags has addressed many

important biological questions (5, 6). The small size and the solubility of GFP are two

65

other important factors that make GFP suitable as fusion tags. However, despite the
popularity of GFP as fluorescence fusion tags, there are some limitations with GF P. First

of all, molecular oxygen is required for the formation of the chromophore. This also

leads to the formation of the by-product, H202, which could be toxic during the

overexpression of GFP in the organisms (7). Second of all, the formation of GP P is slow
(up to 2 hrs) and is problematic for the study of proteins with short half-lives (9, 10).
Thirdly, raising the growth temperature always results in reduced fluorescence levels in
GFP. Finally, although GF P variants featuring different excitation and emission spectra
has provided the possibility of simultaneous detections, problems still exist both in the
high photobleaching rates and instability of some mutants and in the overlapping
excitation and emission spectra of the mutants (9, 11-14).

Our approach in using the CRABPII or possibly other lipocalin family proteins
could potentially provide an alternative option of flubrescence tagging. One of the
advantages of designing CRABPII into fluorescence fusion tags is its tolerance toward
multiple mutations. It is proven that the lipocalin family proteins are highly amenable
toward substantial modifications without destroying their structural integrity (15-18).
Family members of the lipocalins share less than 10% sequence identity and yet all have
a highly conserved overall fold. Other advantages that CRABPII has in terms of being

used as a fusion tag include its even smaller size (137 aa vs ~240 aa for GF P) as well as

its great solubility. The binding cavity of the CRABPII is huge (600 A3), which possibly

allows adopting chromophores in a decent range of sizes and conformations. The
combination of different mutations and different chromophores introduced can be

expected to lead to color tuning of the fusion tags and therefore simultaneous detection of

66

multiple proteins of interests. The most attractive feature of the CRABPII fluorescence
fusion tags is its tunable nature using small molecules, which could potentially provide a
temporal control of the fluorescence signal that can not be achieved using GFP and its

variants.

2. 2. C15-aldehyde-bound CRABPII mutant structures.

2.2.1. Two distinctively different binding orientations of C15-aldehyde

in different CRABPII mutants.

The crystal structure of CRABPII mutant R132K:R111L:L121E:R59W (KLE-

R59W) bound to C‘s-aldehyde was determined at 1.95 A. This mutant was designed

based on the determinants of PSB formation in CRABPII mutants derived from our
previous work. These determinants include three aspects: (1) the nucleophilic Lys must
be properly positioned for proper nucleophilic attack; (2) the aldehyde carbonyl must also
be properly oriented and activated for nucleophilic attack; (3) a counter ion must be
provided to stabilize the positively charged protonated form of the Schiff base. As stated,
the first three mutations in the KLE-R59W mutant, R132K:R111L:L121E, turned out to
be key for the formation of a stable PSB from Retinal (Rt) in CRABPII. The fourth

mutation is at a position where the corresponding residue is presumably within contact

distance of the ionone ring of the bound C15-aldehyde based on our energy minimized

model (19).

67

By looking at the overlay of the crystal structures of the C15-aldehyde-bound

KLE-R59W complex and the Rt-bound R132K:R111L:L121E (KLE) complex, it is clear

that the C15-aldehyde adopts a similar trajectory to Rt in KLE (Figure II-3). There are
also distinct differences between the two structures. Besides C 15-aldehyde being shorter

in length, the two structures also differ in that the PSB formed in the C15-a1dehyde-bound

structure is a trans imine, whereas in the KLE-Rt complex a cis double bond is formed.

The 2Fo-Fc electron density map for the C15-aldehyde in KLE-R59W is shown in Figure

II-4A and the detailed binding site is portrayed in Figure II-5A. Glu121, which was
established as the counter anion stabilizing the PSB in the KLE-Rt complex, is now 4.8 A
away from the PSB nitrogen. A buffer molecule, bis-tris-propane (BTP, Figure II-2), is
also present inside the binding pocket. The -O8-H hydroxy group of BTP makes a 2.9 A
hydrogen bond with the Schiff base nitrogen, which turns out to be the only interaction
that the PSB makes. BTP also interacts with Glu121 through four hydrogen bonding
interactions, with two of them involving the same hydroxy that interacts with the PSB. In
other words, the stabilizing effect of Glu121 for the PSB is now indirect and mediated by
the hydroxy group of BTP through hydrogen bonding interactions. These hydrogen
bonding interactions further secure the position of BTP inside the pocket. BTP is
important in obtaining ligand-bound crystals of this KLE-R59W mutant. Crystals of this
mutant obtained from co-crystallization experiments without having BTP in the growing
conditions all turned out to be apo, with no electron density for ligand evident after

refinement of the structures.

68

 

Figure II-3. Ovelay of the ligands bound in different binding orientations. Green:
all-trans-retinal in KLE; Yellow: C15-aldehyde in KLE-R59W; Magenta: Cls—aldehyde
in KL—A32E; KLE-CRABPII Cor trace is shown in lime.

 

Figure “-4. 2Fo-Fc electron density map calculated at 1.0 a for bound ligands. A.
C15-aldehyde bound to Lysl32 (upper) and bis-tris-propane (lower) in KLE-R59W; B.
C15—aldehyde bound to Lysl32 in KL-A32E.

69

   

Cls-aldehyde
...4..or (
1b 537 ,
W59
F15 ‘
L121 W109
C. D

s: 2:44“ «a ”‘1' ‘1‘“

  

     
 

‘3.
~“ 0 .‘.?.74 ‘.
3.03%:
----- C aldeh de
.. 3.64 15' Y

     

C‘s-aldehyde 4,65“ '3'4'1’ ,’ 3.22.
:2.80
* W109 ( r
K132 L121 312 K132 D121

   

W109

  

Cls-aldehyde
132 (.0121

Figure [1-5. Interactions between the PSB and the residues nearby in different C15-
aldehyde-bound CRABPII mutants. A. KLE-R59W; B. KL-A32E; C. KL—T54E,
MolA; D. KFLDV, MolA; E. KFLDV, MoIB. Distances are in Angstrom (A).

 

As expected, C15-aldehyde is now totally buried inside the protein. Trp59 is

sitting right on top of the ionone ring with its shortest distance to the ligand just over 4 A.

70

This binding mode provided us with the possibility of engineering around any part of the

bound C15-aldehyde analog.

The structure of the C15-aldehyde-bound R132K:R111L:A32E CRABPII mutant

(KL-A32E) at 1.22 A resolution revealed a very surprising result. As shown in the

overlay between hoIo-KLE-R59W and hoIo-KL-A32E, the C15-a1dehyde in KL-A32E

swings about 90° away from the C15-aldehyde binding site seen in the KLE-R59W

mutant and all other retinal-binding mutants as well (Figure II-3). The ligand is now
buried even further within the binding pocket (Figure II-4B), approximately where the
BTP was bound in the KLE-R59W structure. In this “alternative” binding trajectory, the

Schiff base adopts a cis conformation and is stabilized through hydrogen bonding with

two backbone carbonyl oxygens, Ala36 and Ser37 (Figure II-5B). The C15-aldehyde

resides in a very hydrophobic environment in this structure. Note that the wild-type Leu,
instead of Glu as in KLE-R59W, resides at the 121 position. Mutation at this position
turns out to be key in determining the binding orientation. Mutation A32E was meant to
introduce a negatively charged residue next to the ligand bound in the orthogonal

orientation, but is indeed far away from the ligand due to the new binding mode.

2.2.2. A single mutation at position 121 is key in determining the

binding orientation of C15-aldehyde.

By overlaying the C 15-aldehyde binding site in KL-A32E with that in the KLE-Rt

complex, it is clear that the L121E mutation sterically interferes with the alternative

71

binding trajectory of C15-aldehyde (Figure II-6A), while Leu121 in KL-A32E allows the

alternative binding mode. Two additional C15-aldehyde-bound crystal structures were

obtained, which confirmed that smaller residues at 121 result in the ligand binding to this
alternative, more hydrophobic cavity. The two mutants are R132K:R111LzT54E (KL-
T54E) and R132K:Y134F:R111L:L121D:T54V (KFLDV). In these two mutants, either a
Leu (KL-T54E) or an Asp (KFLDV) is present at the 121 position. Despite the fact that
Leu is neutral and Asp is negatively charged, mutants containing either of the two

residues at 121 all allow the ligand to bind in the alternative orientation (Figure II-6B).

  

A. B
E121
, 55* 2.04 T54 0r E54 or V54 1! Q
1}!" E95 P39
/\\ _‘ g . ‘3;

   

C‘s-aldehyde r. e)?“ . K132

Cls-aldehyde
é/VLIZI or D121

Figure [1-6. A. The structures of KLE-Rt and C's—aldehyde—KL—A32E are overlaid.
Glu121 from KLE—Rt (magenta) and C15-aldehyde from holo-KL-A32E (green) are

shown; B. Overlay of the C1 5-aldehyde bound in three CRABPII mutants: KL-A32E
(magenta), KL-T54E (green), and KFLDV (yellow).

In addition to the crystal structures, we have another line of evidence to support

the role of residue 121 in directing the ligand binding orientation. Fluorescence titration

experiments were used to measure the Cl5-aldehyde binding affinity of the CRABPII

72

mutants by measuring ligand-induced fluorescence quenching of protein tryptophan
residues. The quenching efficiency at saturation of the various mutants fell into two
categories (19). The fluorescence of mutants containing Glu or Gln at 121 levels off
between 0.8 and 0.9, while that of mutants having smaller residues (Leu, Asp, etc.) at 121

levels off usually between 0.5 and 0.7 (Table II-l). This observation again suggests two

different types of binding for C15-aldehyde. If we compare the two binding orientations

in terms of their relative distances to the nearby tryptophan(s), we can see that the new

binding orientation positions the ionone ring of the C's-aldehyde right next to Trp109

with the shortest distance between the two residues being only 3.7 A in KL-A32E (Figure
II-SB). In the orthogonal binding trajectory, except for the introduced Trp at 59 in KLE-
R59W, all of the three native Trp’s are far away from the C15-aldehyde, with the closest
one being over 10 A. The alternative binding trajectory features a much closer distance
to Trp and therefore should lead to more efficient fluorescence quenching, which should

result in decreased fluorescence levels at saturation. Our fluorescence titration results are

consistent with the observation in our crystal structures, both indicating that the single

mutation at 121 is key in determining the different binding orientations of C15-aldehyde

bound in CRABPII mutants. Glu121 forces the ligand to bind in the orthogonal
orientation and smaller residues at 121 prefer the ligand to go into the alternative binding

cavity.

73

2.2.3. Protonated Schiff base stabilization.

The PSB in bovine rhodopsin is stabilized by the carboxylate oxygen of Glu113
3.4 A away (20-22). The counterion interaction has been shown to be responsible for the
majority of the red shift of the chromophore upon binding in the protein’s active site (23).
We have previously established a similar counterion interaction in our CRABPII model,
the KLE-Rt complex, where the carboxylate of Glu121 is positioned 2.6 A away from the
PSB nitrogen (1). Stabilization of the PSB through hydrogen bonding has also been seen
in certain rhodopsin crystal structures. In the sensory rhodopsin 11, two Asp residues are
sitting 3.8 A and 4.0 A away from the PSB nitrogen, respectively, and a water molecule
makes contacts with both the PSB nitrogen (2.6 A) and the two Asp residues (24). Very
similar arrangements have been observed in both bacteriorhodopsin (25) and
halorhodopsin (26) where, besides weak counterion interactions, the PSB is also

stabilized by polarized water molecules through hydrogen bonding.

For our C15-aldehyde-bound CRABPII mutants, the formation of the PSB and

Schiff base has been confirmed by the shift of the chromophore’s Max and by reductive

amination followed by MALDI analysis, respectively (Table II-l) (19). Each of the four
mutant structures displays a unique mechanism for PSB stabilization. As described
earlier, the PSB in KLE-R59W (Figure II-5A) is stabilized only by the hydroxy group of
BTP through hydrogen bonding and the same hydroxy group is polarized through
hydrogen bonding interactions with the carboxylate of Glu121. This is an arrangement
similar to that seen in sensory rhodopsin II, bacteriorhodopsin and halorhodopsin where
the carboxylate can indirectly interact with the PSB through water-mediated hydrogen

bonding.

74

In the case of KL-A32E, however, the PSB stabilization involves hydrogen
bonding of the PSB nitrogen with two carbonyl oxygens from the helix backbone (Figure
II-SB). This stabilization mechanism is interesting in that it is the very opposite of the
classical “oxyanion hole” interactions present in many enzymes where the negatively
charged oxygen of the tetrahedral intermediate is hydrogen bonded with two backbone
amide -NH groups (2 7-30).

In the crystal structure of mutant KL-T54E, there are two molecules in the
asymmetric unit and they share the same mechanism for PSB stabilization. The PSB

nitrogen makes hydrogen bonding interactions with a nearby water molecule (3.0 A

distance) (Figure II-5C). For the C15-aldehyde-bound KFLDV mutant, the two

molecules in the asymmetric unit exhibit slightly different binding situations. In
molecule A, the PSB nitrogen makes hydrogen bonds with two nearby residues, a water
molecule at 3.2 A and the carbonyl group of Ala36 at 3.5 A (Figure II-5D). In molecule
B, a sole hydrogen bonding interaction between the PSB and a non-polarized water 2.7 A
away is stabilizing the PSB. Ala36 is moved away due to the crystal-packing-induced
partial unwinding of the (12 helix at its C-terminus (Figure II-5E). What is unique in our

system is that the PSB can be stabilized by interacting only with water molecules as
shown in both molecules of the holo-KL-T54E mutant and M01 B of the halo-KFLDV

mutant.

The pKa values of known rhodopsins have been reported to be between 9.3 and

~16 (31-35). The pK,,1 values of the PSB in the C15-aldehyde-bound CRABPII mutants,

KLE-R59W, KL-A32E, KL-T54E and KFLDV, were measured to be 7.7, 8.6, 7.8 and

75

8.4, respectively (19). Overall, these values are much lower than that of the bovine

rhodopsin (pKa >16) (35) and that of the bacteriorhodopsin (pKa 13.3) (31). The striking

difference might be contributed to whether negatively charged residues are placed in the
vicinity of the PSB. In all of the rhodopsin structures solved so far, one or two negatively
charged residues are always present in the vicinity of the PSB and interact with the PSB
nitrogen either directly (Glu113 in bovine rhodopsin and Asp238 in halorhodopsin) or
through water-mediated hydrogen bonding (Asp85, Asp212 in bacteriorhodopsin and

Asp75 in sensory rhodopsin). Although having a negatively charged residue next to the

PSB does not guarantee a greatly enhanced pKa as demonstrated in the all-trans-retinal-

bound KLE-CRABPII mutant (pKa 8.7), it is reasonable to rationalize that the lack of

negatively charged residues next to the PSB could possibly make the protonation of the

Schiff base harder and therefore results in a lower pKa of the PSB. Among the four C 15-

aldehyde-bound CRABPII mutants, KL-A32E, KL-T54E and KFLDV all lack negatively

charged residues in the vicinity of the PSB and are therefore not surprising to have much
lower pKa values compared to the rhodopsins.

In the C's-aldehyde-bound KLE-R59W-CRABPII mutant, the counterion,
Glu121, is close to the PSB (4.8 A) and is making indirect hydrogen bonding with the
PSB. However, similar to the all-trans-retinal-bound KLE-CRABPII mutant, the C15-
aldehyde-bound KLE-R59W-CRABPII also shows relatively lower pKa (7.7). It has

been proven that distances between the PSB nitrogen and the counter ion as well as the

geometry of the hydrogen bonding also play important roles in determining the pKa of

76

the PSB (36-38). Having a negatively charged residue next to the PSB is therefore not a
guarantee for increased pKa. While the relatively short distance between the counterion
and the PSB (2.6 A) in the Rt-bound KLE-CRABPII is probably responsible for its

relatively lower pKa (8.7), the apparent pKa of the C15-aldehyde-bound KLE-R59W-
CRABPII (7.7) is harder to interpret. The pKa of the Cl5-aldehyde-bound KLE-R59W-

CRABPII was measured in the absence of BTP. Attempts to measure the pKa in the

presence of 100 mM BTP were not successful due to precipitation of the protein probably

caused by the high concentration of BTP.

2.2.4. Cis versus trans protonated Schiff base.

As revealed in the crystal structures of different rhodopsins, the PSB always
adopts a trans conformation (20, 21, 24-26). In our KLE-R59W-C15-aldehyde complex,
a trans double bond is also observed for the PSB. However, in all three of the CRABPII
mutants where the C15-aldehyde adopts the alternative binding mode, a cis double bond

was observed. A cis imine double bond has also been observed in our KLE-Rt complex
previously reported (1 ). Given the fact that trans double bonds are thermodynamically
more favored in solution, the formation of the cis PSB is apparently due to the
positioning of the ligand by specific residues inside the binding cavities of these

particular proteins.

77

 

Figure “-7. A. Proposed nucleophilic attack of lysine amino group to the aldehyde

following the Biirgi-Dunitz model; B. Surface representation of C1 5-aldehyde
(yellow) bound in KL-A32E (gray).

In an overly simplified model, let us first assume that the nucleophilic attack
happens following the Biirgi-Dunitz trajectory (39, 40), where the lysine nitrogen adopts
a favorable 107° attacking angle with the terminal C=O bond of the chromophore and the

N-C-O plane bisects the H-C(O)-C angle (Figure II-7A). We could also assume that the
C15—aldehyde is not moving much from its original binding position due to the tight
interactions of the residues around it (Figure lI-7B), which is supported by the three
CRABPII mutant crystal structures with C15-aldehyde bound in the alternative binding

pocket (Figure II-6B). Under these two assumptions we would see that during the
reaction the lysine side chain would have to either swing in a way to facilitate the
formation of a cis double bond or swing in another way facilitating formation of a trans

PSB. Given that the energy difference between a cis and a trans protonated Schiff base is

7-12 kJ mol-I (15-17 kJ mol_1 for the unprotonated imine) (41), if the energy gained by

the lysine swinging into the “cis” conformation versus “trans” is greater than 12 kJ mol-

78

I, the PSB formed would be cis. Otherwise, a trans double might be preferred. In the

case where the chromophore is less tightly packed, the increased freedom of the ligand

would probably favor the more thermodynamically stable trans double bond, as seen in

the C15-aldehyde-bound KLE-R59W mutant structure.

2.2.5. Flexibility of the helix-turn-helix lid.

When overlaying different CRABPII mutant structures, one can clearly see that
the B-barrel portion of the protein is very rigid, whereas the helix-tum-helix shows
flexibility. The flexibility of the helix-tum-helix portion of CRABPII has been

previously observed (42). Based on the comparison of the crystal structures we have

obtained for CRABPII mutants, the largest movement is observed between the C15-

aldehyde-bound KL-A32E mutant and the C15-aldehyde-bound KLE-R59W mutant, with

the longest distance being over 10 A between the same Ala35 C01 in the two different
structures (Figure [1-8).

As shown in holo-KLE-R59W, where the largest outward movement of 012 is
observed, the C-terrninal of this helix (blue color) loses one helical turn and becomes a
loop-like structure. Similar unwinding has been observed in several apo mutants as well
and therefore is not induced by ligand binding. Unfortunately, up to this point, the apo

structure of KLE-R59W has not been obtained probably due to the decreased structural

stability of this protein when there is no ligand bound. This partial unwinding in the C 15-

aldehyde-bound KLE-R59W results in the biggest outward movement among all our

79

CRABPII mutant crystal structures. Overlay of the C15-aldehyde-bound KLE-R59W and

the apo structure of KLE-R59E that shares the same partial unwinding shows that the

ligand would bump into the unwound portion of apo-KLE-R59E (Figure II-9A). In holo-
KLE-R59W the bound C15-aldehyde indeed pushed out the unwound C-terminal portion

of the 012 helix even more (Figure II-9B). In other words, the flexible helix lid allows the

ligand to bind in this particular trajectory.

   

Figure II-8. Stereoview of the overlay between holo-KL-A32E (green) and holo-
KLE-R59W (magenta). Blue portion indicates the unwinding of the 012 helix. Distance
(A) is measured between the C01 of Ala35 in the two structures.

The same partial unwinding has been observed in M01 B of the C15-aldehyde-

bound KFLDV mutant. However, the outward movement is more moderate compared to
the holo-KLE-R59W. In Mol A of halo-KFLDV, the 0.2 helix is well maintained and
adopts a “median” position in terms of movement by comparison with all the other

CRABPII mutant structures solved so far. The fact that the two molecules in the same

80

asymmetric unit are displaying fairly large structural differences indicates that crystal
packing could be playing a role in this partial unwinding of the helix. Both molecules in
the asymmetric unit of the holo-KL-T54E mutant exhibit well-maintained helices. The
positions of the helices in both molecules could also be described as “median” in the

context of all the known CRABPII mutant structures.

 

Figure [1-9. Overlay between C1 5-aldehyde-bound KLE-R59W (blue) and apo-KLE-
R59E (green). A. Model of C15-aldehyde in apo-KLE-R59E; B. Crystal structure of
C 1 5-aldehyde bound in KLE-R59W.

81

The most inward movement of the (12 helix is observed in the KL-A32E mutant,
where the helical structure is well maintained (Figure 11-8). The inward movement of the
helix is possibly achieved through hydrogen bonding interactions between the two
backbone carbonyl oxygen atoms on the helix and the PSB nitrogen (Figure II-5B).
Crystal packing effects from the tight arrangement of neighboring molecules in this

crystal, which is not observed in the other three holo-CRABPII mutant structures, could

also contribute to the movement in the structure of the C15-aldehyde-bound KL-A32E

mutant. Nevertheless, the results from our crystal structures confirm and clearly

demonstrate the great flexibility of the helix-turn-helix motif in CRABPII.

2.2.6. An evolutionary correlation between CRABPII and plasma

retinal-binding protein?

Plasma retinol-binding protein (RBP) is a 21kD protein with 184 amino acid
residues and is the first structurally characterized lipocalin protein (16, 17, 43-47). RBP
is synthesized in hepatocytes and transports vitamin A (retinol) from the liver, its storage
site, to a variety of target tissues (44, 48-50). After being secreted out of the liver and
into the extracellular fluid, RBP circulates the plasma by forming a complex with
transthyretin (TTR), a thyroid-homone-transporting protein (51, 52). The complexation
protects the relatively small RBP from being lost through glomerular filtration. RBP
exhibits a typical lipocalin scaffold (43, 45-4 7, 53). A hydrophobic binding cavity is
formed within an 8-strand B-barrel. Two helices are attached at each terminus of the

polypeptide. Unlike CRABPs, the RBP helices are not positioned at the entrance of the

82

binding site, but instead swing outward into an almost parallel orientation with
neighbouring B-barrel strands.

Both CRABPs and RBP belong to the calycin superfamily, which includes fatty
acid binding proteins, lipocalins, avidin, metalloproteinase inhibitors, and triabin (I 7).
The calycin proteins all contain a basic B-barrel scaffold to form the binding cavity,
which binds small, lipophilic ligands. An evolutionary correlation between cellular
retinoid-binding proteins and RBP has been previously suggested, though the sequence
homology between the two is very low (10% identity). Through site-directed
mutagenesis, Schmidt and co-workers have created a point mutation, Q1171, deep inside
the binding pocket of RBP (54). This Gln to Ile mutation established a better
hydrophobic interaction with the ionone ring of retinol, which resulted in more than
three-fold increase of the expression of the recombinant protein. The conversion into a
hydrophobic residue from a hydrophilic one in a hydrophobic environment was believed
to lead to further stability of the protein and therefore higher expression level. The same
mutant also showed slightly enhanced binding affinity for retinol due to a higher
association rate. The authors then raised the question of why nature would place a
hydrophilic Gln residue at that critical position when it has better choices. They then
compared RBP with cellular retinol-binding protein (CRBP). In CRBP, retinol binds in a
completely opposite orientation and a Gln is also present inside the hydrophobic binding
pocket, interacting with the hydroxy group of the ligand (55, 56). Interestingly,
replacement of Gln with Arg in CRBP has been shown to diminish retinol affinity (57).

They therefore speculated that “the Gln117 residue in RBP constitutes an evolutionary

83

remnant from ancient times when vitamin A might have been bound to the B-barrel in a

different manner” (5 4).

 

 

 

 

 

A.
\7
'74:! v
I a
B.
C-l a-Idehyde
I 121
YIJ—l
- . rL-rinol HUN
37 L l 157
\\ 2.1 Xvitl ‘\ 24

 

«4

Figure “-10. Stereoview of the overlay of C1 5-aldehyde-bound KL-A32E (green)
and retinol-bound RBP (pink). A. Overall scaffold; B. Binding sites overlay.

What we have observed in our C15-aldehyde-bound CRABPII system favors the

same argument. Although both CRABPs and RBP are retinoid-binding proteins, they
differ dramatically in the way that ligands sit inside the [ii-barrel pockets. Like its iLBP
family member CRBP, WT-CRABPII positions its retinoid ligand, RA, with the
carboxylate end going inside the cavity and the ionone ring portion solvent-exposed at
the entrance. RBP, however, positions retinol in the opposite way with the ionone ring

totally buried inside the cavity and the hydroxy group pointing outside, exposed to the

84

solvent. With the discovery of the alternative binding trajectory of C] 5-aldehyde in KL-
A32E and the other two CRABPII mutants, we noticed the similarity between the way

that C15-aldehyde is bound in KL-A32E and the retinol in RBP.

By overlaying the two structures together through secondary structure alignment
using the online program “secondary-structure matching” (SSM) (58), 70% of RBP and
58% of CRABPII are aligned, respectively, with an rrnsd of 2.62 A (Figure II-lOA). In
the [ii-barrel region, the overall folds of the two proteins very much resemble one another.

Helical portions of the two, however, do not overlay at all. When looking into the

binding cavities of the two, we were happy to find that the ionone ring of C15-aldehyde

and that of retinol occupy relatively similar positions (Figure 11-10). Certain residues
around the ionone rings seem to be conserved (Figure II-lOB and Table II-2). Phe65 in
CRABPII and Phe77 in RBP are close to each other in space and both are relatively close
to the ionone rings of the ligands (6.2 A and 4.4 A, respectively). Tyr134 in CRABPII
and Phe137 are also close (6.2 A and 4.6 A, respectively), approaching the ionone rings
from another direction. Trp7 in CRABPII and Trp24 in RBP are finther away (beyond 8
A), but are very well conserved. Several other nearby residues, mostly hydrophobic, are
also listed in Table 3, with each pair making similar interactions with the ionone rings.
Based on this observation, we propose that an evolutionarily conserved, hydrophobic, yet
under-used binding cavity is present in CRABPII. This observation favors the argument
that an evolutionary correlation exists between the intracellular Lipid-Binding Proteins

(iLBPs) and the lipocalin protein family.

85

Table [1-2. Residues around the ionone rings.

 

 

 

 

 

 

 

 

 

 

 

 

Entry CRABPII RBP
1 F65 (6.21) F77 (4.44)
2 Y134 (6.25) F137 (4.62)
3 W7 (9.64) W24 (8.40)
4 T54 (3.74) A57 (3.45)
5 152 (3.66) A55 (3.99)
6 L121 (3.67) 9117 (4.60L
7 V41 (4.86) A43 (4.69)

 

Numbers in parenthesis refer to the closest distances (A) between the specific residues to the
corresponding ionone rings of the ligands.

2.2.7. Testing the point charge theory based on the C15-aldehyde-
bound KLE-R59W-CRABPII structure.

The high-resolution crystal structures of the C15-aldehyde-bound CRABPII

mutants revealed detailed binding interactions between the chromophore and the nearby
residues. Such information not only gave us confidence in interpreting the wavelength
shifts results, but also helped us to identify options for firrther rational design. We then
set out to design out experiments to test the wavelength shifts upon placing negative
charged residues next to the polyene chain, referred to as the “point charge theory” (59).

Table II-3 shows a series of KLE-X-CRABPII mutants. As shown in discussed

earlier, the three mutations, KLE, not only ensure formation of the PSB between C15-

aldehyde and Lysl32, but also render the ligand to adopt only one binding mode. In
addition to the KLE mutations, these mutants were designed in a way that a negatively
charged residue was introduced as a fourth mutation at places close to the polyene chain
of the chromophore (Figure II-l 1). As shown in Table II-3, the greatest red shift was
observed when a Glu is placed at position 59, which is right next to the ionone ring. As

the negative charge moves gradually toward the PSB, the absorption maxima also drop

86

gradually to about 400 nm. This observation agrees well with the proposed theory in that
negative charges placed further away from the PSB should stabilize the positive charge
more remotely, leading to more extended conjugation and therefore greater red shift of

the absorption maxima.

Table [[-3. Absorption maxima observed for the KLE-X-CRABPII series of mutants

 

 

Mutant Mu (mm)
C] s-aldehyde (buffer) 340
C15-aldehyde PSB (EtOH) 380
R132K:R111L:L121E 330

R132K:R111L:L121E:R59E 424

R132K:R111L:L121EzA32E 416

R132K:R111L:L121EzA36E 410

R132K:R111L:L121EzF15D 397

R132K:R111L:L121E2V76D 399
* Results from Kin-Sing Lee.

F15 6""
A32

N
C15-aldOhYd9 W59
A36 \- K

Figure II-ll. C1 5-aldehyde bound in the KLE-R59W-CRABPII mutant.

 

2.2.8. Conclusion

The crystal structures of four different C15-aldehyde-bound CRABPII mutants

were determined. Protonated Schiff base formation was observed between the active site

87

Lys at position 132 and the chromophore. Each of the four mutants contains a unique
hydrogen bonding pattern that stabilizes the PSB. Two dramatically different binding
orientations were observed in the different mutants with a single mutation at position 121

being key in determining which binding pocket the ligand occupies. Through steric

interactions, Glu or Gln at position 121 forces the C15-aldehyde into the orthogonal

binding mode where the ionone ring of the ligand extends towards the entrance of the
cavity. A smaller residue (Leu or Asp) at position 121 allows the ligand to swing into an
alternative, deeper, and more hydrophobic cavity. The constrained binding interactions
in the alternative binding mode prefers the formation of a cis PSB, while the less tightly
packed ligand environment in the orthogonal binding mode results in trans imine

formation. Flexibility of the helix-turn-helix motif was also observed, featuring a

dramatic 10 A movement of Ala35 C01 in two Cis-aldehyde-bound CRABPII mutants.

By overlaying the alternative binding mode in the mutant CRABPII with the retinol-
bound RBP through secondary structure alignment, we have observed structural
resemblances between the two proteins both in the overall B-barrel fold and in the
binding environments around the ionone rings of the ligands. Taken together with the
previous literature, we proposed that a conserved, yet underused binding pocket is present
in CRABPII and that an evolutionary correlation exists between the intracellular Lipid-
Binding Proteins (iLBP) and the lipocalin protein family.

A series of KLE-X-CRABPII mutants featuring negatively charged residues
placed next to the polyene chain of the chromophore displayed interesting wavelength
regulation properties. Negative charge placed further away from the PSB led to a greater

red shift of the absorption maximum of the bound chromophore. This result agrees well

88

with the proposed point charge theory and presumably shed more light to the wavelength

regulation mechanism in color vision.

2.3. Merocyanin-bound CRABPII mutant structure.

 

H0.S.0
K‘ NW 0
O
Merocyanin MES

Figure II-12. Structures of merocyanin and MES.

Table “-4. UV maximum absorption of the merocyanin-bound CRABPII mutants.

 

 

Mutant Amax
1 Merocyanin in EtOH 461
2 Merocyanin in buffer 493
3 Merocyanin PSB in EtOH 578
4 Merocyanin PSB in buffer 573
5 R132K:R111L:L121E 588
6 R132K:R111L:L121E:R59E 596
7 R132K:R111L:L121E:R59W 596
8 R132K:R111L:L121EzT56W 591
9 R132K:R111L:L121E2T56R 591
10 R132K:R111L:L121EzV76E 593
11 R132K:R111L:L121E:V76W 570
12 R132K:R111L:L121EzV76R 586

 

* Results from Kin-Sing Lee.

As an effort toward designing CRABPII mutants as fluorescence tags, Kin-Sing

Lee and Calvin Grant in the Borhan lab have been working on synthesizing putative

89

chromophores. An interesting compound, merocyanin, was synthesized and showed

excellent spectroscopic properties (Table II-4). By feeding this molecule to a series of

CRABPII mutants, red shift of the Amax was observed in almost all mutants tested. The

largest shifts were observed in mutants KLE-R59E and KLE-R59W, both affording an

opsin shift of over 20 nm. However, the binding of merocyanin into CRABPII mutants

appeared to be much slower compared to C15-aldehyde or all-trans-retinal.

The crystallization of merocyanin-CRABPII mutant complexes was pursued.
Structure of the KLE-R59W CRABPII mutant bound with merocyanin was determined at
2.60 A. The merocyanin is bound the orthogonal way with its aromatic ring extending
toward the entrance and positioning itself right under the Trp59 (Figure 11-13). The
Schiff base formed appears to be in a trans conformation and is stabilized by a weak
hydrogen bonding interaction from a water molecule that is 3.4 A away (Figure II-13C).

It is interesting to see the trans imine formation in this mutant especially when we

compare this mutant to the C15-aldehyde-bound-KLE-R59W-CRABPII where the ligand

binds also in the orthogonal way and also in trans imine form. The similarity of binding
in the two complexes seems to suggest that shorter chromophores that bind the
orthogonal way in CRABPII tends to favor a trans imine formation probably due to the
ligands’ freedom in this spacious binding cavity. A buffer molecule, 2-(N-
morpholino)ethanesulfonic acid (MES), is found to be bound inside the cavity as well
(Figure II-13B), occupying a space approximately where BTP molecules are found in

some other CRABPII mutant structures. The overall fold of this merocyanin-bound-

9O

KLE-R59W-CRABPII resembles the C15-aldehyde-bound-KLE—R59W-CRABPII

structure, with similar partial unwinding at the C-terminal end of the 012 helix.

 

Figure [I-13. Merocyanin bound in KLE-R59W-CRABPII mutant. A. Stereoview of
the overall fold of the protein; B. 2Fo-Fc map calculated at 1 o; C. Detailed binding
interactions between merocyanin and the protein.

This merocyanin molecule shows excellent fluorescence ability after being
complexed with CRABPII mutants. In particular, the complex between merocyanin and

KLE-R59W-CRABPII is capable to emit fluorescence at 614 nm after being excited at

91

570 nm (to be published by Kin-Sing Lee). This serves as the proof-of-concept for
designing CRABPII mutants as ligand-inducible fluorescence fusion tags. The structural
information obtained from this complex will also be instructional for directing our future

ligand design and protein engineering efforts.

92

References

10.

ll.

12.

13.

Vasileiou, C., Vaezeslami, S., Crist, R. M., Rabago-Smith, M., Geiger, J. H., and
Borhan, B. (2007) Protein design: Reengineering Cellular Retinoic Acid Binding
Protein 11 into a rhodopsin protein mimic, J Am Chem Soc 129, 6140-6148.

Gerisch, G., and Muller-Taubenberger, A. (2003) GFP-fusion proteins as
fluorescent reporters to study organelle and cytoskeleton dynamics in chemotaxis
and phagocytosis, Method Enzymol 361, 320-337.

Hoffman, R. M. (2005) The multiple uses of fluorescent proteins to visualize
cancer in vivo, Nat Rev Cancer 5, 796-806.

Lippincott-Schwartz, J., Snapp, E., and Kenworthy, A. (2001) Studying protein
dynamics in living cells, Nat Rev M01 Cell Bio 2, 444-456.

Muller-Taubenberger, A., and Anderson, K. I. (2007) Recent advances using
green and red fluorescent protein variants, Appl Microbiol Biot 77, 1-12.

Shaner, N. C., Patterson, G. H., and Davidson, M. W. (2007) Advances in
fluorescent protein technology, J Cell Sci 120, 4247-4260.

Tsien, R. Y. (1998) The green fluorescent protein, Annu Rev Biochem 67, 509-
544.

VanEngelenburg, S. B., and Palmer, A. E. (2008) Fluorescent biosensors of
protein function, Curr Opin Chem Biol 12, 60-65.

Heim, R., Prasher, D. C., and Tsien, R. Y. (1994) Wavelength Mutations and
Posttranslational Autoxidation of Green Fluorescent Protein, P Natl Acad Sci USA
91, 12501-12504.

Phillips, G. J. (2001) Green fluorescent protein - a bright idea for the study of
bacterial protein localization, Fems Microbiol Lett 204, 9-18.

Baumann, C. T., Lim, C. S., and Hager, G. L. (1998) Simultaneous visualization
of the yellow and green forms of the green fluorescent protein in living cells, J
Histochem Cytochem 46, 1073-1076.

Chalfie, M., Tu, Y., Euskirchen, G., Ward, W. W., and Prasher, D. C. (1994)
Green Fluorescent Protein as a Marker for Gene-Expression, Science 263, 802-
805.

Cormack, B. P., Valdivia, R. H., and Falkow, S. (1996) FACS-optimized mutants
of the green fluorescent protein (GFP), Gene 1 73, 33-38.

93

14.

15.

16.

17.

l8.
19.

20.

21.

22.

23.

24.

25.

26.

Hein, R., and Tsien, R. Y. (1996) Engineering green fluorescent protein for
improved brightness, longer wavelengths and fluorescence resonance energy
transfer, Curr Biol 6, 178-182.

Akerstrom, B., Flower, D. R., and Salier, J. P. (2000) Lipocalins: unity in
diversity, Bba-Protein Struct M 1482, 1-8.

Flower, D. R. (2000) Experimentally determined lipocalin structures, Bba-Protein
Struct M 1482, 46-56.

Flower, D. R., North, A. C. T., and Sansom, C. E. (2000) The lipocalin protein
family: structural and sequence overview, Bba-Protein Struct M 1482, 9-24.

Skerra, A. (2000) Lipocalins as a scaffold, Bba-Protein Struct M 1482, 337-350.
Lee, K. S., Jia, X., Vasileiou, C., Geiger, J. H., and Borhan, B., To be published.

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

Palczewski, K., Kumasaka, T., Hori, T., Behnke, C. A., Motoshima, H., Fox, B.
A., Le Trong, 1., Teller, D. C., Okada, T., Stenkamp, R. E., Yamamoto, M., and
Miyano, M. (2000) Crystal structure of rhodopsin: A G protein-coupled receptor,
Science 289, 739-745.

Kholmurodov, K. T., Fel'dman, T. B., and Ostrovsky, M. A. (2007) Interaction of
chromophore, ll-cis-retinal, with amino acid residues of the visual pigment
rhodopsin in the region of protonated Schiff base: A molecular dynamics study,
Russian Chemical Bulletin 56, 20-27.

Sekharan, S., Sugihara, M., and Buss, V. (2007) Origin of spectral tuning in
rhodopsin - It is not the binding pocket, Angewandte Chemie-International
Edition 46, 269-271.

Luecke, H., Schobert, B., Lanyi, J. K., Spudich, E. N., and Spudich, J. L. (2001)
Crystal structure of sensory rhodopsin II at 2.4 angstroms: Insights into color
tuning and transducer interaction, Science 293, 1499-1503.

Facciotti, M. T., Rouhani, S., Burkard, F. T., Betancourt, F. M., Downing, K. H.,
Rose, R. B., McDermott, G., and Glaeser, R. M. (2001) Structure of an early
intermediate in the M-state phase of the bacteriorhodopsin photocycle, Biophys J
81 , 3442-3455.

Kolbe, M., Besir, H., Essen, L. O., and Oesterhelt, D. (2000) Structure of the
light-driven chloride pump halorhodopsin at 1.8 angstrom resolution, Science 288,
1390-1396.

94

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

Brady, L., Brzozowski, A. M., Derewenda, Z. S., Dodson, E., Dodson, G., Tolley,
S., Turkenburg, J. P., Christiansen, L., Hugejensen, B., Norskov, L., Thim, L., and
Menge, U. (1990) A Serine Protease Triad Forms the Catalytic Center of a
Triacylglycerol Lipase, Nature 343, 767-770.

Jaeger, K. E., Dijkstra, B. W., and Reetz, M. T. (1999) Bacterial biocatalysts:
Molecular biology, three-dimensional structures, and biotechnological
applications of lipases, Annual Review of Microbiology 53, 315-+.

Kraut, J. (1977) Serine Proteases - Structure and Mechanism of Catalysis, Annu
Rev Biochem 46, 331-358.

Menard, R., and Storer, A. C. (1992) Oxyanion Hole Interactions in Serine and
Cysteine Proteases, Biological Chemistry Hoppe-Seyler 373, 393-400.

Druckmann, S., Ottolenghi, M., Pande, A., Pande, J., and Callender, R. H. (1982)
Acid-Base-Equilibriurn of the Schiff-Base in Bacteriorhodopsin, Biochemistry- Us
21 , 4953-4959.

Koutalos, Y. (1992) High-Ph Form of Bovine Rhodopsin, Biophys J 61 , 272-275.

Koutalos, Y., Ebrey, T. G., Gilson, H. R., and Honig, B. (1990) Octopus
Photoreceptor-Membranes - Surface-Charge Density and Pk of the Schiff-Base of
the Pigments, Biophys J 58, 493-501.

Liang, J., Steinberg, G., Livnah, N., Sheves, M., Ebrey, T. G., and Tsuda, M.
(1994) The Pk(a) of the Protonated Schiff-Bases of Gecko Cone and Octopus
Visual Pigments, Biophys J 6 7, 848-854.

Steinberg, G., Ottolenghi, M., and Sheves, M. (1993) Pk(a) of the Protonated
Schiff-Base of Bovine Rhodopsin - a Study with Artificial Pigments, Biophys J
64, 1499-1502.

GAT, Y., and Sheves, M. (1993) A Mechanism for Controlling the Pk(a) of the
Retinal Protonated Schiff-Base in Retinal Proteins - a Study with Model
Compounds, J Am Chem Soc 115, 3772-3773.

Scheiner, S., and Hillenbrand, E. A. (1985) Modification of Pk Values Caused by
Change in H-Bond Geometry, P Natl Acad Sci USA 82, 2741-2745.

Taylor, R., Kennard, O., and Versichel, W. (1983) Geometry of the N-H=/=C
Hydrogen-Bond .1. Lone-Pair Directionality, J Am Chem Soc 105, 5761-5766.

Burgi, H. B., and Dunitz, J. D. (1983) From Crystal Statics to Chemical-
Dynamics, Accounts of Chemical Research 16, 153-161.

95

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

Burgi, H. B., Dunitz, J. D., and Shefter, E. (1973) Geometrical Reaction
Coordinates .2. Nucleophilic Addition to a Carbonyl Group, J Am Chem Soc 95,
5065-5067.

Hammerum, S., and Solling, T. I. (1999) The proton affinities of imines and the
heats of formation of immonium ions investigated with composite ab initio
methods, J Am Chem Soc 121, 6002-6009.

Wang, L. C., Li, Y., Abildgaard, F ., Markley, J. L., and Yan, H. G. (1998) NMR
solution structure of type [1 human cellular retinoic acid binding protein:
Implications for ligand binding, Biochemistry- Us 37, 12727-12736.

Cowan, S. W., Newcomer, M. E., and Jones, T. A. (1990) Crystallographic
Refinement of Human Serum Retinol Binding-Protein at 23 Resolution, Proteins
8, 44-61.

Kanai, M., Raz, A., and Goodman, D. S. (1968) Retinol-Binding Protein -
Transport Protein for Vitamin a in Human Plasma, Journal of Clinical
Investigation 4 7, 2025—&.

Newcomer, M. E., Jones, T. A., Aqvist, J., Sundelin, J., Eriksson, U., Rask, L.,
and Peterson, P. A. (1984) The 3-Dimensional Structure of Retinol-Binding
Protein, Embo Journal 3, 1451-1454.

Zanotti, G., Berni, R., and Monaco, H. L. (1993) Crystal-Structure of Liganded
and Unliganded Forms of Bovine Plasma Retinol-Binding Protein, J Biol Chem
268, 10728-10738.

Zanotti, G., Ottonello, S., Berni, R., and Monaco, H. L. (1993) Crystal-Structure
of the Trigonal F orm of Human Plasma Retinol-Binding Protein at 2.5-Angstrom
Resolution, Journal of Molecular Biology 230, 613-624.

Blomhoff, R. (1994) Transport and Metabolism of Vitamin-A, Nutrition Reviews
52, 13-23.

Blomhoff, R., Green, M. H., Berg, T., and Norum, K. R. (1990) Transport and
Storage of Vitamin-A, Science 250, 399-404.

Quadro, L., Blaner, W. S., Salchow, D. J., Vogel, S., Piantedosi, R., Gouras, P.,
Freeman, S., Cosma, M. P., Colantuoni, V., and Gottesman, M. E. (1999)
Impaired retinal firnction and vitamin A availability in mice lacking retinol-
binding protein, Embo Journal 18, 4633-4644.

Goodman, D. S., and RAZ, A. (1972) Extraction and Recombination Studies of
Interaction of Retinol with Human Plasma Retinol-Binding Protein, Journal of
Lipid Research 13, 338-&.

96

52.

53.

54.

55.

56.

57.

58.

59.

Monaco, H. L., Rizzi, M., and Coda, A. (1995) Structure of a Complex of 2
Plasma-Proteins - Transthyretin and Retinol-Binding Protein, Science 268, 1039-
1041.

Grzyb, J., Latowski, D., and Strzalka, K. (2006) Lipocalins - a family portrait,
Journal of Plant Physiology 163, 895-915.

Schmidt, A. M., Bloss, 1., and Skerra, A. (1998) Improved folding of apo-retinol-
binding protein in the periplasm of Escherichia coli: positive influences of dst
coexpression and of an amino acid exchange in the vitamin A binding site,
Protein Engineering 11, 601-607.

Calderone, V., Folli, C., Marchesani, A., Berni, R., and Zanotti, G. (2002)
Identification and structural analysis of a zebrafish apo and holo cellular retinol-
binding protein, Journal of Molecular Biology 321 , 527-535.

Cowan, S. W., Newcomer, M. E., and Jones, T. A. (1993) Crystallographic
Studies on a Family of Cellular Lipophilic Transport Proteins - Refinement of P2
Myelin Protein and the Structure Determination and Refinement of Cellular
Retinol-Binding Protein in Complex with All-Trans-Retinol, Journal of

Molecular Biology 230, 1225-1246.

Stump, D. G., Lloyd, R. S., and Chytil, F. (1991) Site-Directed Mutagenesis of
Rat Cellular Retinol-Binding Protein - Alteration in Binding-Specificity Resulting
from Mutation of Glutamine-108 to Arginine, J Biol Chem 266, 4622-4630.

Krissinel, E., and Hendrick, K. Protein structure comparison service SSM at
European Bioinformatic Institute (http://www.ebi.ac.uk/msd-srv/ssm).

Nathans, J., Thomas, D., and Hogness, D. S. (1986) Molecular-Genetics of
Human Color-Vision - the Genes Encoding Blue, Green, and Red Pigments,
Science 232, 193-202.

97

CHAPTER III. Crystal Structures of the Apo-CRABPII
Mutants

3. 1. Introduction

Retinoic acid (RA) is the most important metabolite of vitamin A (retinol) in
vertebrates. It has been proven to play indispensable roles during cell grth and
differentiation by regulating the transcription of a variety of target genes (1-4). RA also
possesses antitumor activity. RA and other retinoid analogs have been successfully
employed in treatment again several types of cancers, including acute promyelocytic
leukemia (2), skin cancer (5) and cervical cancer (6).

The transcription effects of RA are exerted through binding of RA to the retinoic
acid receptor (RAR). RAR is a nuclear transcriptional activator, which forms a
heterodimer with retinoids X receptor (RXR) upon association with RA (7-10). The
RAR/RXR dimer then recognizes the RA-response elements on specific target genes and
regulates their transcription.

The delivery of RA to RAR is facilitated by cellular retinoic acid binding proteins
(CRABPS), members of the intracellular lipid-binding protein family (iLBP) (1 1-13).
CRABPs are cytosolic, small (~16 kD), and soluble proteins that are found in all
vertebrates (14). There are two types of CRABPs, CRABPI and CRABPII. CRABPI
and II are highly conserved (both >90% identity) across species, respectively (15, 16).
The identity between CRABPI and CRABPII in the same specie, however, is relatively
lower (74% in human). Therefore, the two proteins probably diverged at a fairly early

stage during evolution and have distinct cellular functions. CRABPs modulate the RA

98

effective concentration through multiple ways, including binding to excess RA and/or
metabolizing the excess RA in the cell through interaction with metabolizing enzymes.
CRABPI and II deliver RA to RAR through different mechanisms (17-19). While
CRABPI shows no interaction with RAR, CRABPII has been shown to localize into the
nucleus upon RA complexation and delivers RA to RAR through protein/protein
interactions (16, 20).

The crystal structures of the wild-type CRABPI and II have been solved earlier
(21, 22). Both proteins adopt a typical iLBP fold with a rigid "ti-barrel" forming a huge
binding cavity and a flexible helix-turn-helix motif functioning as a "lid" at the entrance
(Figure III-1). Interestingly, water-mediated interactions are observed in the cavity of
apo-WT-CRABPII MolA, connecting the Argl 11 deep inside the cavity through Arg132
to the Ala36 and Val33 on the 012 helix. The water-mediated network is not observed in
MolB probably due to the crystal packing. The difference in crystal packing
environments between MolA and MolB at the (12 helix region is dramatic, which is
apparent from the number of the intermolecular interactions that each 012 helix makes
with the neighboring molecules: MolA has 0 hydrogen bond and 8 vdw contacts and
MolB has 7 hydrogen bonds and 43 vdw contacts.

Similar water-mediated networks have been observed in other members of the
iLBP family, such as CRABPI (23) and intestinal fatty acid-binding protein (24), and
suggested to play important roles in maintaining the stability of these proteins. In the rat
intestinal fatty acid-binding protein (FABP) structure (lIF C, 1.2 A), the water-mediated
network connects the Arg106 (conserved residue of Argl 11 in CRABPII) through

Arg126 (conserved residue of Arg132 in CRABPII) to Asp34 and Ala32 on the (12 helix

99

at the portal. In the CRABPI structure (lCBI), the network is observed, but not in a
continuous manner. This is probably a result of certain water molecules not being
observed due to the relatively low resolution of the structure (2.7 A). In the high
resolution structure of cellular retinol-binding protein [1 (CRBPII, 1.2 A, PDB ID:
2RCQ), which is also an iLBP family member, the water-mediated network is again
observed, which also resembles the one in apo-WT-CRABPII. In the apo-CRBPII
structure, the network starts at the Glu108 (conserved residue of Arglll in CRABPII
based on similarity) and extends to Ala36 and Val33 (both conserved in CRABPII based
on identity). These observations further emphasize the structural importance of these
water-mediated networks for the iLBP family proteins.

In the course of our study, we have generated many different CRABPII mutants.
The crystallization trials of these mutants have resulted in both ligand-bound and apo-
structures of certain mutants. In this chapter, the apo-CRABPII mutant structures will be
reported. Interesting structural information that is obtained from these structures will be

discussed.

3.2. Results

3.2.1. Mutations of conserved and non-conserved residues

Through site-directed mutagenesis, we have created a series of CRABPII mutants,
each of which carry multiple mutations involving both conserved and non-conserved
residues (25). All of the mutation sites that are involved in these mutants are shown in

Figure 111-1.

100

Arg132 resides deep inside the binding pocket and is a fully conserved residue in
the basis of identity in the intracellular binding protein family. In the WT-CRABPII,
Arg132 directly interacts with the incoming carboxylate group of the retinoic acid (RA).

The R132K mutation therefore greatly decreases the binding affinity of CRABPII toward

 

Figure [II-1. Stereoview of the mutation sites involved in these CRABPII mutants.
Tertiary structure of CRABPII is represented by retinoic acid-bound CRABPII (PDB
code: 2FR3).

Tyr134 and Arglll are also positioned deep in the pocket and are conserved
residues based on similarity. Y134F is a very conserved mutation and ~20% of the iLBP
family members actually have Phe at 134 as the native residue. R1 1 1L mutation,
however, introduces a dramatic change from the hydrophilic Arg to a hydrophobic Leu,
which enhances the hydrophobicity of the cavity and also creates an extra space due to
the size differences of the two residues.

Arg59, Ala32 and Thr54 are also conserved residues based on similarity. Among

the three, Arg59 and Ala32 are fully conserved residues and Thr54 is conserved in 46 out

101

of 52 iLBP family members. These three residues are not deep inside the cavity and are
instead at positions where the linear part of RA interacts with. R59E mutation, which
happens at the entrance, switches the positively charged Arg to a negatively charged Glu.
T54E or T54V mutations meant to decrease or increase the hydrophobicity of the cavity,
respectively. A32E mutation is unique in that it is the only mutation listed here that
involves a mutation of a residue on the heilx. The change induced by this mutation is
also unique in that the whole 012 helix in MolB of apo-KL-A32E mutant is completely
disordered.

The Leu121 is not a conserved residue. The mutations, L121E or L121D, both
introduce negatively charged residues deep inside the cavity replacing a neutral residue,
which decreases the hydrophobicity of the cavity.

Generally, combinations of these mutations do not change the overall fold of the

CRABPII greatly, except for the R132K:R111LzA32E (KL-A32E) mutant where the 012

helix is completely disordered in MolB.

3.2.2. The structure of apo-R132K:R111L:L121E:R59E-CRABPII (apo-
KLE-R59E)

The structure of apo-KLE-R59E is solved at 1.85 A with final Rwork and Rfi-ee of

18.6% and 24.8%, respectively. The Ramachandran plot indicated that all of the residues
except Asp126 are in the favored/allowed region. Asp126 in both MolA and M018 is in
the disallowed region. However, the electron density map suggests the correct
assignment of the residue. The same residue has always been found in the disallowed

region in previously solved CRABPII structures (22), which added to our confidence for

102

the assignment for this residue. The overall fold of this mutant is approximately the same
as WT-CRABPII (PDB ID: 1CBS, 2FR3) with RMSD of 1.117 for MolA (Figure III-2A)
and 0.463 for MolB (Figure III-2B). The MolBs of the two molecules are very similar
which is also obvious from the RMSD value of 0.463. The two MolAs are, however,
experiencing greater differences. The main difference lies in the region where the C-
terrninal of the (12 helix meets the BC-BD loop. A partial unwinding of the 012 helix is
clearly observed in MolA of the apo-KLE-R59E mutant structure, which is not observed
in MolB and in either of the molecules of the WT-CRABPII structure. To accommodate
this movement, the [SC-8D loop in apo-KLE-R59E correspondingly adopts an outward
movement, compared to the WT-CRABPII MolA. Similar movements have been
observed in previously solved apo-KLE-CRABPII mutant structure (PDB ID: 3D97).

The Arg59 in the WT-CRABPII resides on the (BC-8D loop and guards the
entrance of the protein. In the WT-CRABPII structure, Arg59 interacts with Gln74
through two water-mediated hydrogen bonding (Figure III-2C), which is disrupted upon
the R59E mutation in the apo-KLE-R59E mutant (Figure III-2D). Clearly, the disruption
of this interaction does not affect the overall rigidity of the CRABPII. The other three
mutations, R132K:R111L:L121E, are all deep inside the active site. A buffer molecule,
bis-tris-propane (BTP), is present inside the binding cavity in both MolA and MolB
(Figure III-2E). In both molecules, BTP is close to Trp109 and partly fills the
hydrophobic hole created by R111L mutation as the overlay between this mutant and the
WT-CRABPII shows that Arg] 11 side chain would compete with the BTP for the same

space. The atomic positions of BTP in MolA are well defined with an average B factor

of 37.8 A2, reasonably higher than the overall B factor (28.9 A2). The BTP in MolB,

103

 

 

 

Figure III-2. Structure of the apo-KLE-RS9E-CRABPII. (a) Overlay of MolA of apo—
KLE—R59E-CRABPII (gray) and MolA of apo—WT-CRABPII (green); (b) Overlay of
MolB of apo-KLE—R59E-CRABPII (purple) and MolB of apo-WT-CRABPII (cyan); (c)
Salt bridge between Arg59 and Gln74 in apo-WT-CRABPII MolA; (d) Disrupted salt
bridge in apo-KLE—R59E-CRABPII MolA; (e) 2Fo-Fc map calculated at 1.0 0 for bis-
tris-propane bound in apo-KLE-R59E-CRABPII MolA; (f) Water-mediated interactions
in MolA of apo-KLE-R59E-CRABPII.

104

however, is poorly defined (average B: 65.2 A2) even though the bulk electron density

clearly suggests its presence. Overlaying the MolA and MolB shows that the two BTP
molecules do not overlap with each other. In MolA, the BTP makes three hydrogen
bonding interactions with the nearby protein residues. In MolB, however, the hydrogen
bonding interactions are harder to assign due to our lack of confidence for each atomic
position of the molecule.

A water-mediated network is clearly shown, involving polar residues, BTP and
the ordered water molecules and extending from deep down in the cavity to the 012 helix
at the entrance (Figure III-2F). Similar water-mediated interactions have been observed
in previously solved CRABPII mutant structures and have been proposed to be important
in maintaining the structural integrity of CRABPII by acting as a pillar inside the huge,

hydrophobic binding cavity of CRABPII.

3.2.3. The structure of apo-R132K:Y134F:R111 L:L121E-CRABPII (apo-
KFLE)

The structure of apo-KFLE was solved at a resolution of 1.78 A with Rwork and

Rfree of 18.0% and 23.8%, respectively. The Ramachandran plot indicates again that

Asp126 is the only residue sitting outside of the allowed region in both molecules of the
asymmetric unit. The protein also maintains the rigid overall structure. Overlay of this
mutant structure with the WT-CRABPII gives RMSD of 1.137 for MolA (Figure III-3A)

and 0.450 for MolB (Figure III-3B), which is similar to the ones observed for KLE-

105

 

2'9st 3_45 K132
" ‘ 2.78
2.8.4... J -W
' ..... .o‘” ~

‘8 .~ .' 2.56 ‘
w96 .g3.‘90’ 2.7 2.93 ‘ 3.26 E121

' ”2.74 "=:
337 mu ,/ 2.77

"223 W260

Figure III-3. Structure of the apo-KFLE-CRABPII. (a) Overlay of MolA of apo-
KFLE-CRABPII (purple) and MolA of apo-WT-CRABPII (green); (b) Overlay of MoIB
of apo-KFLE-CRABPII (orange) and MolB of apo-WT-CRABPII (cyan); (c) Water-
mediated interaction in apo-KFLE—CRABPII MolA.

R59E. In fact, this apo-KFLE mutant structure greatly resembles that of the KLE-R59E
mutant including the same partial unwinding at the 012 helix region in MolA. The RMSD
between the two structures is 0.200 for MolA and 0.276 for MolB. This resemblance is
not very surprising. Except for the three common mutations, R132K:R111L:L121E, that
are shared between KF LE and KLE-R59E, the fourth mutation in the KFLE mutant,

Y134F, is a conserved mutation inside the binding site. Therefore, minimal overall

106

structural changes should be expected upon this mutation. There is no BTP bound in
either of the molecules even though the growth condition for the crystal also contains
BTP. In MolA of apo-KFLE, the water-mediated interactions are as follows: Glu121-
W260-Lys132-W274-W223-carbonyl group of Ala32/carbonyl group of Ser37 (Figure

III-3C).

 

 

Figure III-4. Structure of the apo-KFLD-CRABPII. (a) Overlay of MolA of apo-
KFLD—CRABPII (magenta) and MolA of apo-WT-CRABPII (green); (b) Overlay of
MolB of apo-KFLD-CRABPII (yellow) and MolB of apo-WT-CRABPII (cyan).

3.2.4. The structure of apo-R132K:Y134F:R111L:L121D-CRABPII (apo-
KFLD)

The structure of apo-KFLD was solved at 1.94 A with Rwork and Rfree of 21.0%

and 27.7%, respectively. Again, the overall structure of this mutant resembles those of
KLE-R59E (0.271 for MolA and 0.256 for MolB) and KFLE (0.210 for MolA and 0.188

for MolB). The same partial unwinding is also observed at the C-terminal of the 012 helix

107

region in MolA (Figure 111-4). The helical structures are well maintained in MolB. The
crystal was also grown in conditions with BTP as the buffer and yet no BTP was found in
the crystal structure of this mutant. Surprisingly, the water-mediated interactions are not

observed in this mutant.

3.2.5. The structure of apo-R132K:Y134F:R111L:L121D:T54V-CRABPII
(apo-KFLD-T54V)

The structure of apo-KFLD-T54V was solved at 1.51 A. Final Rwork and Rfi-ee

after refinement are 17.1% and 21.2%, respectively. The Ramachandran plot showed
again that Asp126 is the only residue that is outside of the favored/allowed region. There
is no buffer molecule bound inside the binding cavity. Interestingly, no partial
unwinding is observed for the 012 helix in MolA of this mutant (Figure III-5A,B). Since
this KFLD-T54V mutant only differs from the KFLD mutant by a single mutation, T54V,
and this mutation did not change the size, shape or conformation of the residue, it may
suggest that the appearance of the partial unwinding of the 012 helix C-terrninal is
random. Due to the fully maintained helical turns, the structure of this mutant now shares
more structural similarity with the apo-WT-CRABPII structure (RMSD: 0.302 for MolA
and 0.363 for MolB) than with that of the apo-KFLD mutant (RMSD: 1.136 for MolA

and 0.344 for M018). The water-mediated interactions are shown in Figure III-5C.

108

 

 

Figure [II-5. Structure of the apo-KFLD-T54V-CRABPII. (a) Overlay of MolA of
apo-KFLD-T54V-CRABPII (pink) and MolA of apo-WT-CRABPII (green); (b) Overlay
of MolB of apo-KFLD-T54V-CRABPII (purple) and MolB of apo-WT-CRABPII (cyan);
(c) Water-mediated interactions in the apo-KFLD-T54V-CRABPII.

3.2.6. The structure of apo-R132K:R111L:T54E-CRABPII (apo-KL-
T54E)

The structure of apo-KL-T54E was solved at 1.85 A resolution. Final Rwork and

Rfree after refinement are 18.8% and 24.4%, respectively. Asp126 is still the only

residue that sits outside of the favored/allowed region in the Ramachandran plot. There
is no partial unwinding on the 012 helix in MolA. The RMSD values between this mutant
structure and that of the WT-CRABPII are 0.369 for MolA (Figure III-6A) and 0.344 for

M013 (Figure III-6B). BTP molecules are found in the cavities of both MolA and MolB

109

of this mutant structure (Figure HI-6C). It need to be pointed out that the B factors for

both BTP molecules in this structure (56.3 A2 in MolA and 84.0 in MolB) are much

higher than the average B factor of the whole structure (24.0 A2). Long-range water-

mediated interactions are absent in this mutant probably due to the absence of charged
residues (Argl 11 or Glu121) deep inside the cavity. Short-ranged water-mediated
interactions are found between the Glu54 and the 012 helix: carboxylate group of Glu54-
W187-W172-carbonyl group of Ala36 (Figure III-6D). The Glu54 also forms hydrogen

bonds with the bound BTP.

 

Figure [II-6. Structure of the apo-KL-T54E-CRABPII. (a) Overlay of MolA of apo-
KL-T54E-CRABPII (hot pink) and MolA of apo-WT-CRABPII (green); (b) Overlay of
MolB of apo-KL—T54E—CRABPII (yellow) and MolB of apo—WT-CRABPII (cyan); (c)
2Fo-Fc map of the BTP bound calculated at 1.0 o; (d) Water-mediated interactions in
MolA of the apo-KL-T54E-CRABPII.

110

3.2.7. The structure of apo-R132K:R111L:A32E-CRABPII (apo-KL-
A32E)

K132

 

2.76 '1" 3.02
\2137

U
s‘ 2.25, ..~‘zs1
. ‘ W277 ‘~
W331 ‘35

Figure [II-7. Structure of the apo-KL-ASZE-CRABPII. (a) Overlay of MolA of apo-
KL—A32E-CRABPII (purpleblue) and MolA of apo-WT-CRABPII (green); (b) Salt
bridge between Arg59 and Glu32 in the apo-KL-A32E—CRABPII MolA; (c) Water-
mediated interactions in MolA of the apo-KL—A32E—CRABPII.

The crystal structure of apo-KL-A32E was solved at 1.56 A resolution. Final
Rwork and Rfi-ee after refinement are 15.3% and 19.9%, respectively. The Ramachandran

plot showed that Asp126 is the only residue outside of the favored/allowed region as

observed before. The helical structure in MolA is well maintained with no partial

111

unwinding at the C-terrninal of the 012 helix. The RMSD between this MolA and MolA
of the WT-CRABPII is calculated to be 0.414, consistent with the observation that these
two structures are very similar (Figure III-7A). The introduced Glu32 on the (12 helix
forms a salt bridge with Arg59 on the BC-BD loop (Figure III-7B). In MolB of the apo-
KL-A32E, however, the 012 helix is completely disordered. The rest of the KL-A32E
MolB still folds normally and overlays perfectly with MolB of other apo-CRABPII
mutants. Given the fact that the structures listed earlier all contain R132K:R111L
mutations and are all capable of maintaining the structural integrity of the protein, A32E
mutation in KL-A32E is clearly detrimental to the helical structure of the 012 helix. No
buffer molecule is found in the cavity in either MolA or MolB. The water-mediated

network in MolA is shown in Figure III-7C.

3.3. Discussion

3.3.1. Rigid overall fold of the apo-CRABPII mutant structures

By overlaying all six apo-CRABPII mutant structures together (Figure III-8), it is
evident that most of the B-barrel portion of the protein remains rigid in every structure
(red structures in Figure 111-8). The flexibility can be seen for the helix-turn-helix region,
which serves as a lid governing the entry and exit of the ligand, and the BC-BD loop that
is next to the 012 helix. This is consistent with what have been observed in the NMR

solution structure (26) of apo-WT-CRABPII and in the crystal structures of other

CRABPII mutants (2 7).

112

 

unwinding unwinding

Figure [II-8. Stereoview of the overlay of all apo structures together. Red: MolBs of
WT, KLE-R59E, KFLE, KFLD, KFLD-T54V, and KL—T54E; green: MolAs of WT,
KFLD-T54V, KL—TS4E, and KL—A32E; blue: MolAs of KLE-R59E, KFLE, and KFLD.

Three different conformations of the helices are present in different molecules.
The first conformation is shared by all of the MolBs, except for that of the apo-KL-A32E
mutant structure which lost its 012 helix completely. The MolBs are shown in red colors
in Figure III-8. All of these MolB molecules adopt almost an identical conformation with
the RMSD against WT-CRABPII MolB between 0.347 and 0.484.

Some of the MolAs adopt the second conformation (WT, KFLD-T54V, KL-A32E
and KL-T54E), which is shown in the green structures in Figure III-8. In this
conformation, the helices are clearly shifted compared to the MolB structures, together
with the neighboring B-tum regions.

The third conformation is present in the rest of the MolA molecules (KLE-R59E,
KFLE, KFLD, blue colors in Figure 111-8). The interesting feature of this conformation is
the partial unwinding at the C-terrninal of the 012 helix. Similar unwinding has also been
observed in crystal structures of other CRABPII mutants, such as apo-KLE-CRABPII

(PDB ID: 3D97) mutant and apo-KE-CRABPII.

113

The occurrence of multiple conformations of the helices results from their unique
flexibility as discussed previously. The reason that all these MolBs display a uniform
conformation is probably because of the tight crystal packing. In other words, the tighter
interactions from the neighboring molecules that the MolB helices are experiencing
probably constrain the helices in these MolBs and leave them with only one favorable
conformation to adopt. The helices of the MolAs, however, are in a much less
constrained environment in the crystals and therefore are able to adopt different and more
relaxed conformations given their flexible nature.

It is not certain what causes the partial unwinding at the C-terrninal of the 012
helix. It seems that the R132K and L121E/D mutations are present in all of the structures
that share the unwinding (KE, KLE, KLE-R59E, KFLE, and KFLD). However, at the
same time there are other mutants that share the same two mutations and yet do not have
the unwinding, such as KFLD-T54V and KFLE-T54V (PDB ID: 2FSO). Therefore the
partial unwinding is probably favored by the combination of R132K and L121E/D
mutations, but not guaranteed. Also, the T54V mutation seems to be able to reverse the

unwinding through some unclear interactions.

3.3.2. A32E mutation is detrimental to the 012 helix in apo-KL-A32E

Considering that the 012 helices in the apo-CRABPII mutant structures reported
here are all well ordered except for the KL-A32E mutant, we are able to see that the
A32E mutation is probably responsible for the loss of helical structure in MolB of the
apo-KL-A32E CRABPII mutant. Similarly, the 012 helix has been seen to be disordered

in a previously solved CRABPII mutant structure, apo-FISW-CRABPII (PDB ID: 2F RS)

114

(22). Phe15 is located at the 011 helix. The F15W mutation clearly destabilizes the 012
helix and results in a totally disordered 012 helix in MolA of the apo-F15W-CRABPII
mutant. Steric hindrance between Trp15 and the neighboring 012 helix residues is
believed to be responsible for the structural loss of the helix. When it comes to the KL-
A32E mutant, the loss of ordered helical structural is probably not because of steric
reasons since the Glu32 is close to the entrance and can clearly swing its side chains to
avoid bumping into neighboring residues. More possibly, the difference in favored phi
angles between Glu and Ala might be the reason for the detrimental effects of Glu32.

It is also interesting to notice that the 012 helix in MolA of the apo-KL-A32E

mutant is well preserved and the average B factor for the MolA 012 helix is 20.6 A2,

which is even lower than the overall B factor of 26.3 A2. In MolA, a salt bridge is

established between Glu32 and Arg59 (Figure III-7B). The same salt bridge is
maintained in another crystal structure of the same KL-A32E mutant when complexed
with a retinoid analog, suggesting a preference of this salt bridge of this mutant. When
overlaying the MolA and MolB of different apo mutant structures together, we can see
that the BC-BD loop in M013 is shifted further away from the 012 helix than that in MolA
(Figure III-8), which causes the Arg59 to move away from the 012 helix. In the structure
of MolB of KL-A32E, this movement might cause disruption of the salt bridge between
A32E and Arg59. We then speculate that, while the A32E mutation is detrimental to the
structural integrity of the 012 helix, the salt bridge between Arg59 and Glu32 observed in
MolA of apo-KL—A32E possibly compensate for the reduced structural integrity

introduced by having Glu32 on the 012 helix and hold the ordered structure of the helix.

115

The disruption of the salt bridge in MolB due to crystal packing then lead to the

disordered 012 helix.

3.3.3. Water-mediated network acts as a pillar inside the cavity

The water-mediated interactions have been observed in other apo-CRABPII
mutant structures and were suggested to play an essential role in preserving the structural
integrity of the protein. Similar situations are seen in the apo-mutant structures reported
here. Water-mediated interactions are observed in all six apo structures, except for apo-
KFLD. The appearance of these ordered waters that constitute these hydrogen bonding
networks in all of these crystal structures again suggest their possible structural roles. If
we overlay the water-mediated interactions together, we are able to see that these
different network patterns do not overlap with each other. This is due to the different
mutations that each mutant is carrying. However, at places where no mutation is
involved, we could be able to see some "conserved" water molecules in different mutant
structures.

However, contrary to what have been suggested before, the lack of the continuous
water-mediated network in apo-KFLD does not seem to break down the overall fold of
the mutant. Similarly, the lack of long-range water-mediated interactions in apo-KL-
A32E and apo-KL-T54E due to the absence of charged residues inside the cavities of the
two mutants does not lead to collapsed structures either. It is therefore more reasonable
to postulate that the ordered water molecules appear due to the stabilizing effect of the

polar residues inside the cavity of CRABPII.

116

3.3.4. The hydrophobic cavity allows binding of BTP

BTP contains six hydroxy groups with three on each end of the molecule, which
are linked through a hydrocarbon chain. Out of the total of six apo-CRABPII mutant
structures reported here, two of them, apo-KLE-R59E and apo-KL-T54E, showed
binding of BTP in the cavity of the protein. These binding situations differ in their
average B factors, indicating a difference in terms of how comfortable the molecule is
bound inside a specific mutant. By overlaying the structures that have BTP in the
binding sites, it is clear that these BTP molecules do not overlap. The freedom of binding
for BTP is due to lack of highly specific interactions between BTP and the binding
environment, which is consistent with the relatively high B-factors observed for some of
the bound BTPs. All BTP molecules occupy approximately the same hydrophobic hole
created by the R1 1 1L mutation and at the same time tend to adopt multiple conformations
with no particular preference energetically. That could possibly explain the less
definitive electron densities of specific atoms of BTP in apo-KL-T54E. There are also
cases where there is no BTP bound even when the crystal also grew in conditions that
contain BTP, indicating a low binding affinity of BTP toward CRABPII mutants.
Binding of BTP is not believed to alter the conformation of the protein, which is apparent
from the fact that different mutants display minimal structural changes with or without
BTP in the binding sites, such as BTP-bound KL-T54E vs. WT-CRABPII and BTP-

bound KLE-R59E vs KFLE-CRABPII.

117

3.5. Conclusion

The structural information deduced from six high resolution apo-CRABPII
mutant structures are presented and discussed. The B-barrel portion of CRABPII is again
demonstrated to be very rigid and resistant to multiple mutations involving either
conserved or non-conserved residues. The helix-turn-helix is proven to be very flexible,
with three conformations observed in different mutants. A partial unwinding at the C-
terminal of the 012 helix is seen in three mutant structures. This unwinding is believed to
be caused by a combination of helices' flexibility as well as mutations of R132K and
L121E/D. Water-mediated interactions linking residues deep inside the cavity with the
ones at the helical entrance are found in almost all mutants, indicating their important

roles in maintaining the overall fold of these apo-CRABPII mutants.

118

Reference

10.

11.

12.

13.

Chambon, P. (1996) A decade of molecular biology of retinoic acid receptors,
F aseb J 10, 940-954.

Chomienne, C., F enaux, P., and Degos, L. (1996) Retinoid differentiation therapy
in promyelocytic leukemia, Faseb J 10, 1025-1030.

DeLuca, L. M. (1991) Retinoids and their receptors in differeciation,
embryogenesis, and neoplasia., Faseb J 5, 2924-2933.

Lotan, R. (1980) Effects of Vitamin-a and Its Analogs (Retinoids) on Normal and
Neoplastic-Cells, Biochimica Et Biophysica Acta 605, 33-91.

Fisher, G. J ., and Voorhees, J. J. (1996) Molecular mechanisms of retinoid actions
in skin, Faseb J 10, 1002-1013.

Hong, W. K., Itri, L. M. (1994) Retinoids and human cancer, in The Retinoids:
Biology, Chemistry, and Medicine (Sporn, M. B., Roberts, A. B., Goodman, D. S.,
Ed.) 2nd ed., pp 597-630, Raven Press, New York.

Giguere, V. (1994) Retinoic Acid Receptors and Cellular Retinoid-Binding
Proteins - Complex Interplay in Retinoid Signaling, Endocrine Reviews 15, 61 -79.

Kliewer, S. A., Umesono, K., Mangelsdorf, D. J ., and Evans, R. M. (1992)
Retinoid X-Receptor Interacts with Nuclear Receptors in Retinoic Acid, Thyroid-
Horrnone and Vitamin-D3 Signaling, Nature 355, 446-449.

Mangelsdorf, D. J ., Thummel, C., Beato, M., Herrlich, P., Schutz, G., Umesono,
K., Blumberg, B., Kastner, P., Mark, M., Chambon, P., and Evans, R. M. (1995)
The Nuclear Receptor Superfamily - the 2nd Decade, Cell 83, 835-839.

Morrisskay, G. (1993) Retinoic Acid and Craniofacial Development - Molecules
and Morphogenesis, Bioessays 15, 9-15.

Bailey, J. S., and SIU, C. H. (1988) Purification and Partial Characterization of a
Novel Binding-Protein for Retinoic Acid from Neonatal Rat, J Biol Chem 263,
9326-9332.

ONG, D. E., and Chytil, F. (1975) Retinoic Acid-Binding Protein in Rat Tissue -
Partial-Purification and Comparison to Rat Tissue Retinol-Binding Protein, J Biol
Chem 250, 6113-6117.

Sani, B. P. H., D.L. (1974) Retinoic acid-binding-protein in chick-embryo
metatarsal skin, Biochem. Biophys. Res. Commun. 61 , 1276-1282.

119

14.

15.

16.

17.

l8.

19.

20.

21.

22.

23.

24.

Ong, D. E., Newcomer, M. E., & Chytil, F. (1994) Cellular retinoid-binding
proteins, in The Retinoids: Biology, Chemistry, and Medicine (Sporn, M. B.,
Roberts, A. B., Goodman, D. S., Ed.) 2nd ed., pp 283-318, Raven Press, New
York.

Budhu, A., Gillilan, R., and Noy, N. (2001) Localization of the RAR interaction
domain of cellular retinoic acid binding protein-II, Journal of Molecular Biology
305, 939-949.

Budhu, A. S., and Noy, N. (2002) Direct channeling of retinoic acid between
cellular retinoic acid-binding protein 11 and retinoic acid receptor sensitizes

mammary carcinoma cells to retinoic acid-induced growth arrest, Molecular and
Cellular Biology 22, 2632-2641.

Boylan, J. F., and Gudas, L. J. (1991) Overexpression of the Cellular Retinoic
Acid Binding Protein-I (Crabp-I) Results in a Reduction in Differentiation-
Specific Gene-Expression in F9 Teratocarcinoma Cells, J Cell Biol 112, 965-979.

Boylan, J. F., and Gudas, L. J. (1992) The Level of Crabp-I Expression Influences
the Amounts and Types of All-Trans-Retinoic Acid Metabolites in F9
Teratocarcinoma Stem-Cells, J Biol Chem 26 7, 21486-21491.

Napoli, J. L. (2000) Retinoic acid: Its biosynthesis and metabolism, Prog Nucleic
Acid Re 63, 139-188.

Dong, D., Ruuska, S. E., Levinthal, D. J ., and Noy, N. (1999) Distinct roles for
cellular retinoic acid-binding proteins I and II in regulating signaling by retinoic
acid, J Biol Chem 274, 23695-23698.

Kleywegt, G. J ., Bergfors, T., Senn, H., Lemotte, P., Gsell, B., Shudo, K., and
Jones, T. A. (1994) Crystal-Structures of Cellular Retinoic Acid-Binding Protein-I
and Protein-Ii in Complex with All-Trans-Retinoic Acid and a Synthetic Retinoid,
Structure 2, 1241-1258.

Vaezeslami, S., Mathes, E., Vasilelou, C., Borhan, B., and Geiger, J. H. (2006)
The structure of apo-wild-type cellular retinoic acid binding protein 11 at 1.4 A
and its relationship to ligand binding and nuclear translocation, Journal of
Molecular Biology 363, 687-701.

Thompson, J. R., Bratt, J. M., and Banaszak, L. J. (1995) Crystal-Structure of
Cellular Retinoic Acid-Binding Protein-I Shows Increased Access to the Binding

Cavity Due to Formation of an Intermolecular Beta-Sheet, J. Mol. Biol. 252, 433-
446.

Scapin, G., Gordon, J. 1., and Sacchettini, J. C. (1992) Refinement of the Structure
of Recombinant Rat Intestinal Fatty Acid-Binding Apoprotein at 1.2-a Resolution,
.1. Biol. Chem. 267, 4253-4269.

120

25.

26.

27.

Gunasekaran, K., Hagler, A. T., and Gierasch, L. M. (2004) Sequence and
structural analysis of cellular retinoic acid-binding proteins reveals a network of
conserved hydrophobic interactions, Proteins 54, 179-194.

Wang, L. C., Li, Y., Abildgaard, F ., Markley, J. L., and Yan, H. G. (1998) NMR
solution structure of type [1 human cellular retinoic acid binding protein:
Implications for ligand binding, Biochemistry- Us 3 7, 12727-12736.

Wang, L. C., Li, Y., and Yan, H. G. (1997) Structure-function relationships of
cellular retinoic acid-binding proteins - Quantitative analysis of the ligand binding
properties of the wild-type proteins and site-directed mutants, J Biol Chem 272,
1541-1547.

121

CHAPTER lV. Investigation of the Interaction between
Cellular Retinoic Acid Binding Protein ll and Retinoic
Acid Receptor Gamma Ligand Binding Domain

4.1. Background

As stated earlier, the effects of retinoic acid (RA) are mediated by at least two
families of proteins, nuclear hormone receptors and cellular retinoic acid binding proteins
(CRABPs). The transcriptional activities of RA are mediated by two members of the
nuclear hormone receptor superfamily: Retinoic Acid Receptor (RAR) and Retinoid X
Receptor (RXR). This superfamily is composed of nuclear ligand-activated

transcriptional regulators that also include steroid hormone, thyroid hormone and vitamin

D3 receptors. The RARs and the RXRs each consist of three isotypes (or, B, and y)

encoded by separate genes. Each isotype has several isoforrns, arising from the
differential usage of two promoters and alternative splicing (I, 2). The RAR family
(RARa, B, and y) binds to and is therefore activated by all-trans (tRA) and 9-cis RA

(9cRA), whereas the RXR family (RXRot, B, and y) is activated only by 9cRA (Figure

IV-l).
l \ \
\ O \
\ \ \ OH
I \
HO O
all-trans-retinoic acid 9-cis-retinoic acid

Figure IV-l. All-trans retinoic acid and 9-cis retinoic acid.

122

N-terminal Hinge C-terminal
domain region domain

       

DNA binding Ligand binding
' domain (DBD) doman (LDB)

3D

 

Figure [V-2. Structural organization of nulear receptors (3).

The RARs and the RXRs exhibit the conserved modular structure of nuclear
receptors and their amino acid sequence can be divided into six regions (A-F) based on
homology among themselves and with other members of the nuclear receptor superfamily
(Figure IV-2) (4). Region C, consisting of 66 amino acids, contains two zinc finger
motifs and is the sequence specific DNA binding domain (DBD). Within a given species,
region C is highly conserved between the three RAR isotypes (94-97%) and between the
three RXR isotypes (91-97%). Region B consists of 220 amino acids and is the ligand
binding domain (LBD) which is also highly conserved between RAR isotypes (84-90%)
and RXR isotypes (88-95%). This region is functionally complex, including the ligand
binding domain, the ligand-inducible transcriptional activation function AF -2, and a
dimerization surface. The interspecies conservation of the DBD and the LED of a given

RAR type or of a given RXR type is even greater than the similarity found when

123

comparing the three RAR types or the three RXR types within a given species, which
suggests that the DNA and ligand binding properties of the various RARs and RXRs
could exhibit subtle functional differences.

Region D is located as a hinge between the DBD and the LBD and is conserved
both among RARs or RXRs within a given species and is conserved between species for
a given RAR or RXR. The C-terminal F domain, which exists in the RARs, is missing in
the RXRs and its function is still unknown. However, the amino end A/B regions of
RARs and RXRs have both been shown to contain a second transcriptional activation

function AF - 1 .

   

Figure [V-3. Comparison of the crystal structures of the apo-RXRa (A, PDB:
lLBD; (5)) and holo-RARy (B, PDB: 2LBD; (6)) ligand-binding domains reveals the
ligand-induced conformational change. N-terminal is labeled in blue and C-tenninal in
red in the rainbow representation in both structures.

124

The crystal structures of the LBDs of both RARs and RXRs have been
determined (Figure 3) (5, 6). The LBDs are formed by 12 conserved alpha-helices and a
beta-tum (between H5 and H6) which are folded into a three-layered, anti-parallel helical
sandwich with H4, H5, H8, H9 and H11 situated between H1, H2 and H3 on one side and
H6, H7 and H10 on the other. In this structure, the C-terrninal helix H12, which
encompasses the AF-2 activation domain, points away from the LBD core. By
comparing the crystal structures of apo-RXRa LBD and holo-RARy LBD, the major
conformational change before and after ligand binding seems to be the relocation of H12,
which after ligand binding folds back towards the LBD core and “seals” the ligand entry
site.

In the absence of the ligand, retinoic acid receptors are found primarily in the
nucleus. They bind as asymmetric, oriented RAR/RXR heterodimers to specific DNA
sequences, RA response elements (RAREs), composed typically of two direct repeats of a
core hexameric motif, (A/G)G(G/T)TCA (2, 7). The classical RARE is a 5-bp-spaced
direct repeat (DRS). However, RAR/RXR heterodimers also bind to direct repeats
separated by 1 bp (DRl) or 2 bp (DR2). Unliganded and DNA-bound retinoid receptors
repress transcription through recruitment of the corepressors NcoR and SMRT (8, 9).
After ligand binding, the resulting conformational changes cause the release of the
corepressors and recruit coactivators which upon association with larger complexes with
chromatin modifying and remodeling activities will decompact repressive chromatin and
facilitate the positioning of the transcriptional machinery at the promoter region (9-11).

CRABP I and II, are the other family of proteins that mediate the biological

effects of RA (12-15). CRABPs belong to a family of proteins called intracellular lipid-

125

binding proteins (iLBP), including fatty acid-binding proteins, lipid-binding proteins, and
cellular retinoid-binding proteins (16). CRABPs are responsible for modulating the
intracellular existence of RA. The sequences of CRABPs are highly conserved in
vertebrates (17, 18). CRABP I and II are closely related, with an identity of 76% in
human. CRABPs are believed to act as carriers of RA, serving to solubilize and protect
RA in the aqueous environment of the cytosol. In addition to this general role, CRABPs
have also been shown to facilitate RA modification upon interaction with other proteins
(12-15).

Despite the sequence similarity between CRABPI and CRABPII, the two proteins
have been shown to have distinct roles in regulating RA (17). Professor Noy and
colleagues demonstrated that the rate constant for movement of RA from CRABPII, but
not CRABPI, to RAR strongly depends on the concentration of RAR (19). They
therefore suggested that transfer of RA from CRABPI to RAR involves dissociation of
RA from CRABPI, followed by association of RA with the acceptor. In contrast,
movement of RA from CRABPII to RAR is facilitated by a mechanism involving direct
interactions between CRABPII and RAR. However, the interaction between the two is
short-lived according to their observation as no stable complex could be trapped in spite
of several attempts including chemical crosslinking, electrophoresis under denaturing
conditions, fluorescence anisotropy titrations and electrophoretic mobility shift assays.

Their later investigations revealed three spatially aligned residues E75Q, K81P,
and E102K (CRABPI/CRABPII), displaying dramatic changes in electrostatic surface
potential of the two proteins (1 7). Substituting the corresponding CRABPII residues onto

CRABPI conferred upon the protein the ability to channel RA to RAR and to enhance the

126

transcriptional activity of RAR in cells. Conversely, substituting the three residues on
CRABPII into their CRABP I counterparts resulted in loss of the ability to interact with
RAR and to enhance the activity of the receptor.

Professor Chomienne and co-workers also worked on the interaction between
RAR and CRABPII. By attaching CRABPII with fluorescence proteins they have
demonstrated the distribution of CRABPII in the nucleus (20). However, contradictory to
what had been suggested by Professor Noy, they have shown that CRABPII was
associated with MR0 and RXRa in a ligand-independent manner through a GST pull-
down assay using GST-tagged RARa or RXRor. Moreover, they have demonstrated that,
in the presence of retinoids that bind both the nuclear receptors and CRABPII,

transcription by RARot-RXRot heterodimers is enhanced by the presence of CRABPII,

suggesting CRABP II as a novel transcriptional regulator involved in RA signaling.

COOH

 

Figure [V-4. The synthetic retinoid analog CD270.

To further investigate the problem, Budhu and Noy performed a similar in vitro
pull-down assay using immobilized His-tagged RARor-LBD incubated with CRABPII in
the presence of RA or a synthetic analog CD270 (Figure IV-4) which had been reported

to interact with CRABPII with a high affinity but to be a poor ligand for RARor (18).

127

The results are consistent with their earlier statement that CRABPII interacts with RAR
in a ligand-dependent manner and that the complex is not stable. No CRABPII was
found to co-precipitate with RARor-LBD in the absence of ligand or in the presence of
RA. However, CD270 stabilized the complex in a dose-dependent manner and CRABPII
was co-precipitated with RARy-LBD in the presence of CD270.

As we have shown earlier, among our mutants of CRABPII, the triple mutant

KLE-CRABPII shows an excellent retinal-binding property (Kd = 1.4 :1: 4.9 nM). In this

particular protein, retinal forms a protonated Schiff base with the lysine residue in
position 132, which can be followed by the UV red shift from 380 run (free retinal) to
447 nm (protonated Schiff base). This covalent protein-ligand complex gave us an
opportunity that resembles the situation in Noy’s research involving the synthetic CD270.
In both cases, ligands prefer to stay with the CRABPII proteins instead of being
transferred into RAR, which has been suggested by Noy and colleagues as key in order to
achieve a stable complex between the two proteins. We therefore set out to detect the
protein-protein interaction between CRABPII and RAR employing the all-trans-retinal-

bound KLE-CRABPII mutant.

4.2 Results and discussion

4.2.1. In vitro binding assay.
With both RARy-LBD and CRABPIIs (WT-CRABPII and KLE-CRABPII) in
hand, the binding assay was performed. His-tagged RARy-LBD was first immobilized

on Ni-NTA resin. WT-CRABPII and KLE-CRABPII were then incubated with the

128

immobilized RARy-LBD either in the absence of ligand or in the presence of either RA
(for wild type) or retinal (for triple mutant) at 4°C for 1 h. The resin was then washed
three times with Tris buffer (10 mM Tris, 100 mM NaCl, pH 8.0). SDS gel loading
buffer was then added to the beads and proteins attaching to the resin were analyzed by
SDS electrophoresis. The results are shown in Figure IV-S. For wild-type CRABPII, no
interaction between RARy-LBD and WT-CRABP II was observed either in the absence
(lane 6) or in the presence of RA (lane 7). The same results were obtained for the triple
mutant CRABP II: no KLE-CRABPII was co-precipitated with RARy-LBD either in the

absence (lane 8) or in the presence of RA (lane 9) or retinal (lane 10).

A.
12345678910

 

Figure IV-5. In vitro binding assay. 1. MW; 2. WT-CRABPII; 3. KLE-CRABPII; 4.
Ni-NTA beads + WT-CRABPII; 5. Ni-NTA beads + KLE-CRABPII; 6. Ni-NTA beads +

His6-RAR-LBD + WT-CRABPII; 7. Ni-NTA beads + His6-RAR-LBD + WT-CRABPII
+ RA; 8. Ni-NTA beads + His6-RAR-LBD + KLE-CRABPII; 9. Ni-NTA beads + His6-

RAR-LBD + KLE-CRABPII + RA; 10. Ni-NTA beads + His6-RAR-LBD + KLE-
CRABPII + retinal. A. Coomassie blue stain; B. Silver stain.

129

4.2.2. Following the protonated Schiff base formation by UV.

As mentioned before, there is a major red shift in the UV spectrum from 380 nm
to 447 run before and after retinal binds to KLE-CRABPII. In order to investigate
whether RARy-LBD will prevent the formation of the protonated Schiff base, RARy-
LBD was added into the complex between KLE-CRABPII and retinal. A UV scan was
performed and it turned out that the protonated Schiff base is no longer existent after
RAR-LBD was added, based on the fact that the UV absorbance went back to around 375
nm. Two possibilities were proposed: one is that the Schiff base might have been
hydrolyzed and retinal was transferred into RARy-LBD; the other possibility is that the
Schiff base is not completely hydrolyzed, but instead of being protonated, the Schiff base

is now unprotonated, which could also lead to an absorption at 375 nm.

4.2.3. In vitro binding assay using reductively aminated Schiff base
complex.

Since a secondary amine base is much more stable than a protonated Schiff base

in the aqueous environment, a reductive amination was conducted using Na(CN)BH3.
The excess of retinal in the reaction was removed from the reduced protein-ligand
complex by buffer exchange using a concentrator. Similar in vitro binding assay was

performed. However, again, no co-precipitation with RARy-LBD was observed for

reduced KLE-CRABPII-retinal complex.

4.2.4. Presence of Tween20 leads to hydrolysis of the protonated
Schiff base.

130

Tween20 (0.1%) was present in the purified RARy-LBD solution. It was
introduced during the dialysis after Ni-NTA purification (see detailed procedures in
Chapter VI). The function of this detergent is to prevent aggregation of the protein.
Since detergent is well known to be able to change the solubility of proteins in buffers, it
is suspected that the presence of Tween20 might lead to certain conformational change of

CRABPII, which will eventually lead to the hydrolysis of the PSB.

 

Abs

 

 

 

 

300 400 500 600
Wavelength (nm)

Figure IV-6. Detergent titration of the protonated Schiff base formed between KLE-
CRABPII and retinal.

By adding the Tween20 containing buffer into the Schiff base complex, it was
observed that the absorption went back to 375 nm. A detergent titration experiment was

also performed (Figure IV-6). To the Schiff base formed complex was added Tris buffer

131

containing different concentrations of Tween20. The concentration of Tween20 that
starts to destroy the protonated Schiff base was found out to be between 0.001% and

0.005% (v/v).

4.2.5. Purification of RARy-LBD without detergent.

Since the detergent was the reason for disrupting the protonated Schiff base,
RARy-LBD was then purified without detergent. After Ni column, the protein solution
was dialyzed against 10 mM Tris, 100 mM NaCl, pH 8.0. It was then concentrated,
followed by gel filtration purification. However, the spectrum of the gel filtration looked
quite different from the previous run when Tween20 was used in the buffer. RARy-LBD
purified this way came off the column quite early and covered a much broader range of
fractions. What’s more, the UV of RARy-LBD after purification does not look normal.
Absorption at 260 nm is strongly interfering the protein absorption at 280 nm, suggesting
possible nucleotide contamination.

The fact that RARy-LBD came off the gel filtration column earlier as well as the
fact that the UV spectrum of RARy-LBD showed an elevated baseline indicate possible
aggregation of the protein. To test this possibility, detection for inclusion bodies was
performed and insoluble RARy-LBD was detected in the cell pellet. To reduce the

inclusion bodies and aggregation problems, several things were modified in the RARy-
LBD purification. The optimized purification included the following steps. After OD600

reached 0.6, IPTG was added and the growth was continued at 16°C for 20 h (previously
5 h at r.t.). The proteins were eluted off the Ni-NTA column by 250 mM imidazole, 100

mM NaCl, 10 mM Tris, pH 8.0 (previously 100 mM EDTA, 500 mM NaCl, 20 mM Tris,

132

pH 8.0). After the Ni column, each individual fraction was examined by UV and the best
fraction(s) were immediately loaded onto the Sephadex G-50 gel filtration column. It
was then eluted with 10 mM Tris, 100 mM NaCl, 10% glycerol, pH 8.0. The fractions
were again examined under UV and it turned out that most of the fractions obtained this
way showed maximum absorption very close to or exactly at 280 nm, suggesting much

less or even no contamination from nucleotides.

4.2.6. Circular dichroism of RARy-LBD and fluorescence titration of
RARy-LBD with retinoic acid.

The purified RARy-LBD was then characterized first with CD. Both apo- and
holo-RARy-LBD were measured and the obtained spectra resemble the literature data
(21 ). a-Helix content was indicated by the negative ellipticity with minima at 222 and

208 nm. There are no major changes before and after tRA binding, which is also

consistent with the published data.

 

800000
750000

g 700000 -

g 650000 «

8

E 600000 .
550000 -

 

 

500000 " ___AT TIA T I I I l
0 0.0000002 0.0000004 0.0000006 0.0000008 0.000001
Concentration of RA

 

 

 

 

Figure [V-7. Fluorescence titration of RARy-LBD with all-trans-retinoic acid.

133

0.2 A.

—-0
——0.1 eq

 

0.3 B.

--0.1 eq

0.25 _o.2 eq
0.3 eq

— 0.4 eq

——0.5 eq

—0.6 eq

—0.7 eq

b‘ —0.8 eq
// \ —-0.9 eq

1.0 eq

   

250 300 350 400 450 500 550 600
um I

Figure IV-8. Titration of KLE-CRABPII with retinal. A. Using KLE-CRABPII
purified earlier; B. Using fully active KLE-CRABPII.

134

Fluorescence titration of RARy-LBD with all-trans-retinoic acid was also

performed to verify the ligand binding ability of the purified protein. The extinction

coefficient of RARy-LBD was measured to be 11,972 chM-l. Titration data is shown

in Figure lV-7. The calculated Kd is 0.5 nM, which is very close to the 0.6 nM binding

affinity reported in the literature.

4.2.7. Titration of KLE-CRABP II with retinal followed by UV.

A perfect KLE-CRABPII (R132K:R111L:L121E) should bind retinal up to 1
equivalent and still maintain the maximum absorption at around 440 nm corresponding to
the protonated Schiff base. However, the KLE-CRABPII that was purified in the
previous batch failed to give a decent titration (Figure IV-8A). The retinal could only be
added up to 0.5 eq and after that the absorption started to blue shift. Later the same
titration using KLE-CRABPII purified in a different batch turned out to be fully active

(Figure IV-8B). This batch of protein was later used in the in vitro binding assays.

4.2.8. In vitro binding assay using RARy-LBD directly from Ni-NTA
resin.

Since we had concerns on loss of activity of RARy-LBD due to dialysis and other
handling of the protein, we decided to try the binding assay one more time using the
RARy-LBD directly from the Ni resin. This way we would save the effort of eluting the

protein off the Ni resin and later on binding it again onto the resin and also exert

135

minimum handling of the protein. Moreover, since the purity of the protein was proved
to be fairly good after Ni-NTA purification, this seemed to be a method worth trying.
Besides RARy-LBD, retinal solution was also freshly prepared. The KLE-
CRABPII used here is the one proved to be fully functional based on the titration
experiments described earlier. In this assay the interaction between KLE-CRABPII and
RARy-LBD was investigated and that between WT-CRABPII and RARy-LBD was not.
The procedure is the same as described earlier in the first in vitro binding assay. The
results are shown in Figure IV-9. KLE-CRABPII showed up together with RARy-LBD
on the gel either in the absence (lane 3) or presence of retinal (lane 4), however, in the
same intensity as the control (lane 2), indicating again lack of interaction between the two
proteins. Lane 5, 6 and 7 showed the protein content of the supematants corresponding
to lane 2, 3 and 4 respectively. As shown, the same amount of KLE-CRABPII appeared
on the gel, suggesting the same degree of binding for KLE-CRABPII both in control and

in real assays and therefore no interaction between the two proteins.

Lane12345678910

Figure [V-9. In vitro binding assay using RARy-LBD directly from Ni-NTA resin. 1.

RAR-LBD on Ni beads; 2. Ni-NTA beads + KLE-CRABPII; 3. Ni-NTA beads + Hisé-RAR-LBD
+ KLE-CRABPII; 4. Ni-NTA beads + His6-RAR-LBD + KLE-CRABPII + retinal 5. Supernatant
of lane 2; 6. Supernatant of lane 3; 7. Supernatant of lane 4; 8. Washed off lane 2; 9. Washed off
lane 3; 10. Washed off lane 4.

136

4.3. Conclusions

Despite several efforts, no in vitro protein-protein interaction between CRABPII and
RAR-LBD is detected. CD270 should be used to set up a positive control for future

binding experiments.

137

References

10.

11.

12.

Gronemeyer, H., and Laudet, V. (1995) Transcription factors .3. Nuclear
receptors, Protein Profile 2, 1173-1308.

Leid, M., Kastner, P., and Chambon, P. (1992) Multiplicity Generates Diversity in
the Retinoic Acid Signaling Pathways, Trends Biochem Sci 1 7, 427-433.

Hoffmeier. http://cn.wikipcdia.org/wiki/Image:Nuclcar Receptor Structurepng.

 

Steinmetz, A. C. U., Renaud, J. P., and Moras, D. (2001) Binding of ligands and
activation of transcription by nuclear receptors, Annu Rev Bioph Biom 30, 329-
359.

Renaud, J. P., Roche], N., Ruff, M., Vivat, V., Chambon, P., Gronemeyer, H., and
Moras, D. (1995) Crystal-Structure of the Rar-Garnma Ligand-Binding Domain
Bound to All-Trans-Retinoic Acid, Nature 378, 681-689.

Bourguet, W., Ruff, M., Chambon, P., Gronemeyer, H., and Moras, D. (1995)
Crystal-Structure of the Ligand-Binding Domain of the Human Nuclear Receptor
er-Alpha, Nature 375, 377-382.

Mangelsdorf, D. J., and Evans, R. M. (1995) The er Heterodimers and Orphan
Receptors, Cell 83, 841-850.

Aranda, A., and Pascual, A. (2001) Nuclear hormone receptors and gene
expression, Physiol Rev 81, 1269-1304.

Glass, C. K., and Rosenfeld, M. G. (2000) The coregulator exchange in
transcriptional functions of nuclear receptors, Gene Dev 14, 121-141.

Nagy, L., Kao, H. Y., Love, J. D., Li, C., Banayo, E., Gooch, J. T., Krishna, V.,
Chatterjee, K., Evans, R. M., and Schwabe, J. W. R. (1999) Mechanism of

corepressor binding and release from nuclear hormone receptors, Gene Dev 13,
3209-3216.

Perissi, V., Staszewski, L. M., Mclnemey, E. M., Kurokawa, R., Krones, A.,
Rose, D. W., Lambert, M. H., Milburn, M. V., Glass, C. K., and Rosenfeld, M. G.

(1999) Molecular determinants of nuclear receptor-corepressor interaction, Gene
Dev 13, 3198-3208.

Boylan, J. F ., and Gudas, L. J. (1992) The Level of Crabp-I Expression Influences
the Amounts and Types of All-Trans-Retinoic Acid Metabolites in F9
Teratocarcinoma Stem-Cells, J Biol Chem 26 7, 21486-21491.

138

13.

14.

15.

16.

17.

18.

19.

20.

21.

F iorella, P. D., and Napoli, J. L. (1991) Expression of Cellular Retinoic Acid
Binding-Protein (Crabp) in Escherichia-Coli - Characterization and Evidence That
Holo-Crabp Is a Substrate in Retinoic Acid Metabolism, J Biol Chem 266, 16572-
16579.

Fiorella, P. D., and Napoli, J. L. (1994) Microsomal Retinoic Acid Metabolism -
Effects of Cellular Retinoic Acid-Binding Protein (Type-I) and C18-
Hydroxylation as an Initial Step, J Biol Chem 269, 10538-10544.

Napoli, J. L., Boerrnan, M. H. E. M., Chai, X., Zhai, Y., and Fiorella, P. D. (1995)
Enzymes and Binding-Proteins Affecting Retinoic Acid Concentrations, J Steroid
Biochem 53, 497-502.

Ong, D. E., Newcomer, M. E., & Chytil, F. (1994) Cellular retinoid-binding
proteins, in The Retinoids: Biology, Chemistry, and Medicine (Spom, M. B.,
Roberts, A. B., Goodman, D. S., Ed.) 2nd ed., pp 283-318, Raven Press, New
York.

Budhu, A., Gillilan, R., and Noy, N. (2001) Localization of the RAR interaction
domain of cellular retinoic acid binding protein-II, Journal of Molecular Biology
305, 939-949.

Budhu, A. S., and Noy, N. (2002) Direct channeling of retinoic acid between
cellular retinoic acid-binding protein 11 and retinoic acid receptor sensitizes

mammary carcinoma cells to retinoic acid-induced growth arrest, Molecular and
Cellular Biology 22, 2632-2641.

Dong, D., Ruuska, S. E., Levinthal, D. J., and Noy, N. (1999) Distinct roles for
cellular retinoic acid-binding proteins I and II in regulating signaling by retinoic
acid, J Biol Chem 274, 23695-23698.

Delva, L., Bastie, J. N., Rochette-Egly, C., Kraiba, R., Balitrand, N., Despouy, G.,
Chambon, P., and Chomienne, C. (1999) Physical and functional interactions
between cellular retinoic acid binding protein [1 and the retinoic acid-dependent
nuclear complex, Molecular and Cellular Biology 1 9, 7158-7167.

Lupisella, J. A., Driscoll, J. E., Metzler, W. J., and Reczek, P. R. (1995) The
Ligand-Binding Domain of the Human Retinoic Acid Receptor-Gamma Is
Predominantly Alpha-Helical with a Trp Residue in the Ligand-Binding Site, J
Biol Chem 270, 24884-24890.

139

CHAPTER V. Preliminary X-Ray Crystallographic Studies
of Cellular Retinaldehyde Binding Protein

5.1. Background

As stated earlier, the lack of a three-dimensional structure of the cellular
retinaldehyde binding protein (CRALBP) has hindered many research efforts in
addressing important biological questions associated with this protein. So we decided to

tackle this problem.

5. 2. Crystallization of the full-length CRALBP

The CRALBP (rCRALBP) gene was generously provided by professor John W.
Crabb. The rCRALBP gene was then cloned into both His-tagged (pETl9b) and non-

fusion (pET17b) plasmids and each was expressed respectively in BL21(DE3) cells.

5.2.1. His-tagged full-length CRALBP crystallization.

The His-tagged rCRALBP was purified using Ni-NTA affinity chromatography to
homogeneity (Figure V-l). The binding property of this His-tagged rCRALBP was
tested by observing UV shifts while titrating the protein with one of its natural ligands,
l l-cis-retinal.

The purified His-rCRALBP was then tested at room temperature for

crystallization. Primary screen of the apo protein at different concentrations (25 mg/mL,

140

10 mg/mL, and 5 mg/mL) were performed and results obtained were compared. None of

these trials gave protein crystals.

 

Figure V-l. Pure fractions of His-rCRALBP eluted after Ni-NTA purification.

Since room temperature crystallization did not work well. Crystallization in the
cold room (4°C) was then pursued. Primary screen of this protein (9 mg/mL) at 4°C gave
several hits, which later on were all proven to be salt crystals.

Since the apo protein failed to crystallize, we turned to the holo-protein based on
the knowledge that ligand-bound CRALBP was more soluble than the protein itself (I).
The protein was concentrated to 10 mg/mL and ll-cis-retinal was added up to 2.5 eq.
The primary screen was set up in the cold room (4°C) under dim red light. Boxes were
wrapped with aluminum foil and kept in the dark. Unfortunately, no crystal was

observed in any conditions.

141

5.2.2. Non-fusion full-length CRALBP crystallization.

Since the His-tagged protein failed to give any crystals either in the apo form or

bound to ll-cis-retinal, we decided to try to crystallize the protein without the His-tag.
CRALBP was expressed using the pETl7b vector in BL21(DE3) codon+ cells as a tag-

free protein. The purification largely follows the procedure developed by Crabb and
colleagues (I). In the first step of purification, a DEAE column was employed. The
roughly purified protein mixture was then applied to a hydroxyapatite (HTP) column and
eluted with a gradient of sodium phosphate buffer. This step eliminated most of the
impurities and gave moderately pure CRALBP. The last step of purification employed
the FPLC source Q column, after which the CRALBP was purified to homogeneity

(Figure V-2).

l
I
v
V
i
1‘

Figure V-2. Non-fusion CRALBP purified. This gel displays the fractions after the
FPLC Q column purification. Fractions containing CRALBP are 38 through 43. Lane 1.
Molecular Weight Standard; 2. fraction 29; 3. fraction 30; 4. fraction 31; 5. fraction 38; 6.
fraction 39; 7. fraction 40; 8. fraction 41; 9. fraction 42; 10. fraction 43.

142

The binding activity of purified CRALBP with its natural ligand 11-cis-retinal
was tested using the UV titration method. When low equivalents of ligand (up to 0.2 eq)
were added, the complex showed a distinct red shift from 380 nm to ~420 nrn. However,
when more ligand was added, the absorption started to shift back toward 380 nm (Figure
V-3). By looking thoroughly in the original paper of Crabb, I found that the protein they
purified was also not fully active (1). The only graph that showed the red shift of the
bound chromophore in the paper is actually looking very similar to our titration with 0.2
eq of ligand being added.

Primary crystallization trials of the apo form of this protein were performed at
4°C. However, again, no crystals were observed for this non-fusion full-length

CRALBP.

— 1.0 eq
~—0.9 eq
——-0.8 eq
—0.7 eq
-—0.6 eq
-——0.5 eq
—0.4 eq
0.3 eq
——0.2 eq
—0.1 eq
—0 eq

500

 

Figure V-3. UV titration of non-fusion CRALBP with ll-cis-retinal.

143

5.2.3. Full-length CRALBP with a cleavable His-tag.

In previous attempts, the purified non-firsion CRALBP showed reduced activity,
probably due to the prolonged purification steps. In order to overcome this possible
problem, we decided to introduce a cleavable tag to the protein so that the purification

could be done in much shorter time and hopefully with similar or even better purity.

A. Overnight cutting B. 3 hour cutting

-. . -

w~~~-mmw‘ 4

Figure V-4. Full length CRALBP expressed in pET28a and purified with his-tag
cleaved. A. Overnight cutting followed by FPLC resource Q purification; B. 3 hour
cutting followed by FPLC resource Q purification.

The full-length gene encoding CRALBP was successfully cloned into the pETQ8a
expression vector. The expressed protein contains a poly-histidine tag at the N-terminal
followed by a thrombin-cutting site. The full length CRALBP was first over-expressed
and purified by Ni-column. The purified His-tagged protein was then treated with
thrombin to cleave the tag. It turned out that this step could be non-specific if the cutting
time is too long. The overnight cutting gave two bands, which could not be separated
after resource Q FPLC or gel filtration (Figure V-4A). Cutting for 3 hours gave a much

better result. The full length protein turned out to be the major component, with very

144

little over-cut protein present (Figure V-4B). However, crystallization of this construct

did not yield any crystals either.

5.3. Crystallization of the N-terminal-truncated CRALBP.

Attempts to crystallize the native, full-length CRALBP all turned out to be not
successful. This then lead us to question whether floppy domains exist in native
CRALBP, which hinder the crystallization of this protein. According to the literature, the
N-temimal of CRALBP is subjected to limited trypsin digestion (2). When we looked
back at our trypsin cutting experiment (Figure V-4), we could also identify that the over-
digested peptide is only slightly smaller than the full-length protein. This result possibly
suggests that a floppy N-terminal exists in CRALBP. What's more, one of the family
members of CRALBP, a-tocopherol transfer protein (aTTP), was crystallized with a N-
terminal truncation (3). These facts all suggest that N-terminal deletion of CRALBP

might help the crystallization of the protein.

5.3.1. Truncation at 43144 on the N-terminal of CRALBP.

It is vital to make the right decision on where to make the truncation. Certainly,
cutting should not disrupt the globular structure of the protein. It is also important that
not a lot of unstructured amino acids are left hanging at the N-terminus after the deletion,
otherwise they will still hamper the crystal packing of the protein. To make the decision
on the cutting position, I first looked at the corresponding position where others have
made deletion to help crystallize aTTP. The first 20 amino acids were actually cleaved

off aTTP, while its first on helix starts at position 27 based on its crystal structure.

145

Another family member, Sec14, has its first or helix start at position 33 (4). Besides,
based on the model for CRALBP, the proposed first helix starts at around 66 (5). Two
sequence alignment methods, ClustalW and MAP, were used to compare the
corresponding position 20 of aTTP on CRALBP. However, the two disagree with each
other to a large extent especially at the N—tennini of this family of proteins. By
comparing the starting position of the first a-helix of each protein based on the secondary
structure prediction generated with the two methods, MAP seemed to give more accurate
estimations (1AA difference) over ClustalW (5AA difference). Therefore, from the
comparison of the MAP sequence alignment, position 43 on CRALBP turned out to be
the corresponding position 20 on aTTP. Since 43 is still far enough from position 66,
where the first a-helix starts according to the model built for CRALBP, it should be a
safe place to make the deletion. Therefore, it was decided that the truncated protein
would start from position 44.

The truncated gene was generated through PCR and cloned into the pET28a
vector. However, multiple attempts to express this truncated CRALBP resulted in no
expression at all. The inclusion body was later detected in the pellet of cells after
sonication, indicating misfolded proteins. To improve the situation, lower temperature
growth (16°C) and less IPTG induction were investigated. Fortunately, this time small
amounts of the target protein were obtained after Ni column purification (Figure V-5).
After cutting with thrombin, the protein was further purified by resource Q F PLC. The

protein was mostly pure except for a contaminant with a higher MW.

146

Lane 1 2 3 4 5 6 7 8

Figure V-S. N-43/44-truncated CRALBP expressed from pET28a and purified by
Ni-column. 1. MW; 2. Flow through; 3. First wash; 4. Second wash; 5~8. Elution
fractions.

However, the 43/44 truncation made at the N-terminus of CRALBP was later
proven to be not active for binding to ll-cis-retinal. The truncated protein failed to bind
to ll-cis-retinal based on the fact that no red shift of the absorption maxima was
observed when the ligand was introduced to the truncated protein. Although red shift of
the absorption maxima does not necessarily equal to ligand binding, it is still safe to say
that cleavage of 43 amino acids off the N-terminal is too much, which either results in a
completely non-functional protein or removes important protein-ligand interactions that

lead to no red shift of the bound chromophore.

5.3.2. Discovery of the ideal N-terminal truncation site.

To look further into truncation options, we decided to make several truncations at
the N-terminus (A10, A15, A20, A25, A30, A35). These mutants were successfully
cloned, expressed and purified using the same systems that were employed for the 43/44

truncation construct earlier. During expression of these truncation mutants, it turned out

147

that, as expected, not all the mutants could be expressed. Deletion of up to 25 amino
acids was able to maintain good expression of the protein. Deletion of 30 amino acids or
more resulted in very poor or no expression (Figure V-6).

MW A10 A15 A20 A25 A30

H

Figure V-6. Expression of the truncated CRALBPs.

To further test whether the well expressed proteins are all active, UV titration
experiments were carried out for each of the four mutants: A10, A15, A20 and A25. By
continuously adding 0.1 equivalent of the ligand ll-cis-retinal into the protein solutions,
red shifts from 380 nm to 410 nm were observed in all cases. The first three proteins
(A10, A15 and A20) bind to the ligand almost as well as the full length CRALBP. For
these three truncations, red shifts continued until up to 0.4~0.5 eq was added for all three
truncations (0.5~0.6 eq for the full length protein). Afterwards, the absorption started

going back toward 380 nm. However, the last one, A25, could only take up to 0.3~0.4 eq

148

of ligand before the absorption started going back toward 380 nm (Figure V-7). The

result suggested that A25 might not be as active as the other three truncations.

A. FL-CRALBP titrated wlth 11-cis-retinal

1.1 eq
1.0 eq
0.9 eq
0.8 eq
—0.7 eq
—-O.6 eq
—-0.5 eq
—0.4 eq
-0.3 eq
— 0.2 eq
0.1 eq
--0 eq
— 11-cis-retlnal

 

B. 10-CRALBP titrated with 11-cls-retinal

-~‘0.6 eq
0.5 eq
—0.4 eq
—0.3 eq
—0.2 eq
0.1 eq
—0 eq

ABS

.°
p—c

 

i

N
U1
0

 

Figure V-7. UV titration of full-length and truncated CRALBPs with ll-cis-retinal.

149

Figure V-7 (cont’d).

 

 

C. 15-CRALBP titrated with 11-cis-retlnal

 

D. 20-CRALBP titrated with 11-cis-retinal

 

150

—0.8 eq
--0.7 eq
—-~ 0.6 eq
~ , 0.5 eq
—0.4 eq
—0.3 eq
-—-O.2 eq
0.1 eq
—0 eq

—0.7 eq
-.._ 0.6 eq
0.5 eq
--0.4 eq
—0.3 eq
—0.2 eq
0.1 eq
—0 eq

Figure V-7 (cont’d).

E. 25-CRALBP titrated with 11-cis—retinal

—0.5 eq
—0.4 eq 2nd
~ 0.4 eq

0.3 eq 2nd
—0.3 eq
--0.2 eq
—0.1 eq 2nd

0.1 eq
—0 eq

 

 

 

F. 30-CRALBP titrated with 11-cis-retinal

0.1 eq
—0eq

 

250 350 450
nm

151

Further proof of A25 being different from other active truncations came from gel
filtration experiments. Gel filtration purification for each of the full length CRALBP as
well as the four truncations (A10, A15, A20, A25) were carried out, respectively, to
further purify the proteins as well as to observe whether they would purify as monomers
or dimers or even other association forms. A10, A15, and A20 all gave similar profiles to
the wild-type CRALBP for their gel filtration run (a huge peak at the void, a smaller,
sharper peak fractions later and a shoulder in between, Figure V-8), while A25 only gave

a peak at the void.

Figure V-8. Gel filtration separation of full-length-CRALBP.

152

A. Fractions 13-16 titrated with 11-cis-retinal

—0.5 eq
— 0.4 eq
— 0.3 eq

0.2 eq
--O.1 eq

ABS

 

 

B. Fractions 17-20 titrated with 11-cis-retlnal

— 1.0 eq
—-0.9 eq
—O.8 eq
---0.7 eq
.._ -0.6 eq
--0.5 eq
—O.4 eq
—0.3 eq
—0.2 eq
0.1 eq
..._0 eq

ABS

 

Figure V-9. UV titration of different fractions from gel filtration separation of A20-
CRALBP.

153

Figure V-9 (cont’d).

C. Fractions 21-23 titrated with 11-cis-retinal

 

 

 

0.04
U!
n
<
-o.o1 . , A
250 350 450
nm

This observation, together with the previous observation that A25 was not as
active as the other truncations, also led to the suspicion that the more active form of the
protein might reside in the fractions that showed up later in the gel filtration runs of A10,
A15, and A20. To test this suspicion, UV titration experiments were carried out for
different gel filtration fractions, respectively (Figure V-9). It turned out that the most
active protein did come from the later fractions corresponding to the smaller, sharper
peak with an approximate molecular weight of 120 KD (0.8 eq, tetramer?). The protein
coming out of the void had the worst binding property (0.2~0.3 eq) and the shoulder

fractions showed an average binding activity (~0.6 eq).

154

A. CD of void (13-16) fractions from 20-CRALBP gel filtration

150 q
100

50 j

 

 

190 200 210 220 230 240 250

B. CD of middle (17-20) fractions from 20-CRALBP gel filtration

 

190 200 210 220 230 240 250

Figure V-10. CD spectra of different fractions from gel filtration separation of A20-
CRALBP.

155

Figure V-10 (cont’d).

C. CD of late (21-23) fractions from 20-CRALBP gel filtration

 

 

30 J
l
-20 ..
l
-70 l
l
l
-120 éae — ——e _. —.— Te —“» ,._,_._ .g. -ej - —.A —. _~»—~— ,.
190 200 210 220 230 240 250

The circular dichroism (CD) absorptions of A20 out of different gel filtration
fractions were also measured (Figure V-IO). The CD spectrum of the later fractions
resembles the reported CD spectrum of active full length CRALBP. The CD of the void
fractions only showed a signal at ~220 nm and was missing the signal at ~208 nm. The
middle fractions showed a CD signal as an average of the previous two.

Based on the above results, it turned out that A20-CRALBP is the ideal construct,
which is hopefully free of floppiness and, at the same time, proven to be able to express

well as a functional protein.

156

 

 

 

 

 

 

 

 

Figure V-ll. Gel filtration purification of the A20-CRALBP grown at dim-rent

temperatures. A. 37°C; B. 16°C.

157

5.3.3. Enhancing the expression and purification of functional
CRALBP.

As shown in Figure V-8, the gel filtration purification of the A20-CRALBP
results in an elution profile with a majority of the protein being eluted at the void. This
suggests possible aggregation of the overexpressed protein. To tackle this problem,
different growth temperatures were tried and yielded exciting results. It turned out that
lower temperature did greatly favor the functional A20-CRALBP over the aggregated
form (Figure V-l l). The trypsin cutting conditions were also optimized to remove the
tag after the Ni-NTA purification (Figure V-12). The complete procedure for optimized

growth and purification of CRALBP (both truncated and the full length) is described in

Chapter VI.
Full length CRALBP A20 CRALBP
Trypsin:prot l :50 l :500 1550 15500
-. . 52% fl ‘ ‘4 fl #5.
Q t g ' . ‘Cb.“.”‘ ‘
" I- a c... unused
MW STD 05“ 1h 2" 4" 05h ”1 2“ 4*! MW STD 0.511 1h 2h 4h 0.5h lh 2h 4h

Figure V-12. Trypsin digestion performed for both full length and A20-CRALBP
under 4°C.

158

5.3.4. Crystallization trials.

The crystallization of the pure, active apo-A20-CRALBP (10 mg/ml and 20

mg/ml, respectively) at both room temperature and 4°C did not yield any hits.

 

Figure V-13. Crystals of merocyanin-bound proteins. A. merocyanin-bound A20—
CRALBP; B. merocyanin-bound KLE-R59W-CRABPII.

The crystallization efforts then turned to the holo-CRALBP complexes.
Merocyanin has been proven to be an excellent substrate for CRALBP. The merocyanin-
A20-CRALBP complex was then used in crystallization. Very interestingly, purple,
crystal-like entities were observed in a particular condition (Figure V-l3A). Based on
our previous experience with the merocyanin-bound-KLE-RS9W-CRABPII mutant
crystals (Figure V-l3B), we know that the purple color can result from the formation of
the protonated Schiff base between a protein Lys residue and the aldehyde fimctional
group on the ligand. Therefore, we are hopeful that this entity is formed by proteins
instead of salts. Unfortunately, this crystal-like entity failed to diffract X-rays, which
probably indicates its lack of a more ordered arrangement of the molecules

microscopically. Further efforts toward optimization of the quality of the crystals did not

159

improve the situation. However, the discovery so far gives us confidence that the A20-
CRALBP finally showed some tendency toward crystallization and that we might be at
the very edge of obtaining the first crystal of the CRALBP protein. The synthesis of 11-
cis-retinal and ring-locked-l l-cis-retinal is underway and we hope we will have better

luck trying to crystallize CRABLP bound with these new ligands.

5.4. Conclusion.

The crystallization of CRALBP in both full length and the N-terminally truncated
version has been performed. Despite various efforts, no crystals have been obtained so
far. However, by introducing retinoid analogs into the protein, we seem to be able to
improve the protein's tendency toward crystallization. It seems very promising that
future crystallization trials probably with new ligands might lead us to obtaining the first

crystal of the CRALBP.

160

Reference

1. Crabb, J. W., Chen, Y., Goldflam, S., West, K., Kapron, J. . (1998) Methods for
producing recombinant human cellular retinaldehyde-binding protein, Methods in
Molecular Biology 89, 91-104.

2. Crabb, J. W., Gaur, V. P., Garwin, G. G., Marx, S. V., Chapline, C., Johnson, C.
M., and Saari, J. C. (1991) Topological and Epitope Mapping of the Cellular
Retinaldehyde-Binding Protein from Retina, J Biol Chem 266, 16674-16683.

3. Min, K. C., Kovall, R. A., and Hendrickson, W. A. (2003) Crystal structure, of
human alpha-tocopherol transfer protein bound to its ligand: Implications for
ataxia with vitamin E deficiency, P Natl Acad Sci USA 100, 14713-14718.

4. Sha, B. D., Phillips, S. E., Bankaitis, V. A., and Lao, M. (1998) Crystal structure
of the Saccharomyces cerevisiae phosphatidylinositol-transfer protein, Nature
391, 506-510.

5. Wu, Z. P., Hasan, A., Liu, T. Y., Teller, D. C., and Crabb, J. W. (2004)
Identification of CRALBP ligand interactions by photoaffinity labeling,

hydrogen/deuterium exchange, and structural modeling, J Biol Chem 279, 27357-
27364.

161

CHAPTER VI. Experimental

6.1. Cloning, mutagenesis, expression and purification

6.1.1. Cloning, mutagenesis, expression and purification of CRABPII
mutants.

Mutagenesis. Site-directed mutagenesis was performed using the CRABPII-pET17b

plasmid following Stratagene’s Quikchange® protocol. The PCR products were

transfected into J M109 competent Escherichia coli cells for plasmid maintenance.

MW 18 19 21 23 25 28 31 33 35

Figure VI-l. R132K:Y134F:R111L:L121E-CRABPII (KFLE) purified after source-
15Q anion exchange column. The corresponding fraction number is listed on top of
each lane.

Protein Expression and Purification. Human recombinant CRABPII was expressed and

purified as described (1 ). The target gene (CRABPII/pETl7b), as isolated from JM109

162

using Qiagen’s Maxi Prep® DNA isolation kit, was transformed into E. coli strain

BL21(DE3)pLysS cells (Stratagene), according to standard protocols. The transformed
cells were grown at 37°C in an LB agar plate containing both ampicillin (100 ug/mL) and
chloramphenicol (25 ug/mL). A single colony was inoculated in 50 mL of LB containing
the same amount of the two antibiotics and grown overnight with shaking at 250 rpm at

37°C. This culture was then transferred to 1 L of LB with the same antibiotics. The
culture was incubated at 37°C until A600 reached between 0.6 and 0.8. The expression

was induced by addition of 0.5 mM isopropyl-l-B-D-galactopyranoside (IPTG). The
growth was continued for another 5 h at 30°C. The cells were harvested using
centrifugation (5000 rpm, 10 min) and frozen at -80°C. The frozen cells from 2 L of
growth culture were thawed and resuspended in 50 mL of 10 mM Tris (pH 8.0). The
suspended cells were lysed by four l-minute bursts of sonication (probe sonicator,
Biologics Inc., 60% power) on ice and centrifuged for 30 min at 4°C at 7000 rpm. The
supernatant was subjected to FastQ (Q-sepharose Fast Flow Resin) column
chromatography (100 mL bed volume; Amersham Biosciences). The column was
washed with 10 mM Tris (pH 8.0) and proteins eluted with 100 mM of NaCl in 10 mM
Tris (pH 8.0). The fractions were analyzed by SDS 16% polyacrylamide gel
electrophoresis and those of the highest purity were combined. The protein solution was
desalted by concentration using Centriprep 3 concentrators (Amicon, MWCO 3000),
followed by dilution with 10 mM Tris (pH 8.0). The proteins were further purified on a
BioRad system (BioLogics Duo Flow, BioRad) using a SourceISQ (Amersham
Biosciences) anion exchange column. Fractions were analyzed by UV absorption at 280

nm as well as SDS 16% PAGE. All the steps of purification were performed at 4°C. The

163

most pure fractions were combined and concentrated to ~20 mg/mL, as determined by

A280. The pure and concentrated protein solutions were aliquoted and stored at -80°C.

6.1.2. Cloning, expression and purification of RARy-LBD.

The pDONR vector (gentamycin resistant) containing the RARy-LBD gene (178-
423) with a hexa-His tag at the amino end was obtained from Dr. Marc Ruff in professor
Pierre Chambon’s lab. The gene was subcloned into the NdeI-BamHI sites of pETle
and the resulting plasmid was amplified in DHSor. Transformation into E. coli

BL21(DE3) provided the construct ready for overexpression.

M a. ...-....

Figure VI-2. Elation fractions of RARy-LBD from Ni—NTA purification.

Cells were grown in LB medium containing ampicillin as described (2, 3). Cell
culture from 6 L shake flasks was centrifuged and the cell pellet was resuspended in 50
mL buffer A (5 mM imidazole, 500 mM NaCl, 20 mM Tris, pH 8.0), followed by

sonication. The cell lysate, after centrifirgation, was mixed with Ni-NTA resin (20 mL)

164

and shaken at 4°C for about 1 hour. The flow through was collected. The resin was then
washed twice with 50 mL of buffer B (50 mM imidazole, 500 mM NaCl, 20 mM Tris, pH
8.0). The proteins were eluted with buffer C (100 mM EDTA, 500 mM NaCl, 20 mM
Tris, pH 8.0) as 5 mL fractions, yielding 14 mg of protein (Figure VI-2).

The protein solution was dialyzed against Dialysis buffer (10 mM DTT, 500 mM
NaCl, 10 mM Tris, 0.1% Tween20, pH 8.0). It was then concentrated to 2.5 mg/mL,
followed by gel-filtration chromatography (Superdex200, Pharmacia) using the same

Dialysis buffer. The RARy-LBD (28 KD) was obtained mainly as a homodimer as

expected.

6.1.3. Cloning, mutagenesis, expression and purification of CRALBP.

6.1.3.1. Initial constructs of CRALBP.

The rCRALBP gene was generously provided by professor John W. Crabb. The
rCRALBP gene was then cloned into both His-tagged (pETl9b) and non-fusion
(pETl7b) plasmids and expressed respectively in BL21(DE3) cells. The His-tagged
rCRALBP was purified using Ni-NTA affinity chromatography to homogeneity (Figure
VI-3). The binding of this His-tagged rCRALBP was tested by observing UV shifts
while titrating the protein with its natural ligand ll-cis-retinal.

Non-fusion CRALBP was expressed from the pETl7b vector in BL21(DE3)

codon® cells as a tag-free protein. The purification largely follows the procedure

developed by Crabb and colleagues (4). In the first step of purification, a DEAE column
was employed using a gradient of 10 mM to 400 mM of NaOAc in 25 mM Tris buffer

with 0.1 mM DTT, pH 7.0. The roughly purified protein mixture was then applied onto a

165

hydroxyapatite (HTP) chromatography and eluted with a gradient of sodium phosphate
(0-80mM) buffer (25 mM Tris, 300 mM NaOAc, 0.1 mM DTT, pH 7.0). This step
eliminated most of the impurities and gave moderately pure CRALBP. The last step of
purification employed the FPLC Q column (gradient from 50 mM to 500 mM of NaOAc
in 20 mM MOPS and 0.1 mM DTT, pH 7.0), after which the CRALBP was purified to

homogeneity (Figure VI-4).

 

Figure VI-3. Pure fractions of the his-tagged rCRALBP eluted off the Ni column.

12345678910

 

.-,. w

1!

g I I c}
6 ""‘~ r. mm... her.

Figure VI-4. Non-fusion CRALBP after purification. This gel displays the fractions
after the FPLC Q column purification. Fractions containing CRALBP are 38 through 43.
Lane 1. Molecular Weight Standard; 2. fraction 29; 3. fraction 30; 4. fraction 31; 5.
fraction 38; 6. fraction 39; 7. fraction 40; 8. fraction 41; 9. fraction 42; 10. fraction 43.

166

6.1.3.2. Cloning, mutagenesis, expression and purification of the newer constructs.

Molecular cloning. Inserts of full length, A10, A15, A20, A25, A30, A35-CRALBP (all
N-terrninal deletions) were prepared using sequential digestion with BamHI and HindIII.
The primers used for generating different deletions are listed in Figure VI-5. The

digested genes were then successfully ligated into pET28a vectors, respectively.

10 AAGGATCCgtacctgaagaggaacaggag
15 AAGGATCCcaggagctccgtgcccaa

20 AAGGAchcaactggagcagctcac

25 AAGGATCCacaaccaaggaccatggacc
30 AAGGATCCggacctgtctttggcccgt
35 AAGGATCcccgtgcagccagctgc

Figure VI-S. Primers used for constructing different N -terminal deletion CRALBP.
Capital letters represent non-matching bases. The BamHl cutting sites together with the
overhang bases are in purple.

Expression and purification. The plasmids were transformed into E. coli BL21 codon69

cells, respectively. Cells were plated on the LB plates containing both kanamycin and
chloramphenicol and grown at 37°C for about 12 hrs. Single colonies were used to
inoculate 50 mL LB/kan/chlor medium and grown overnight at 37°C. The overnight
culture was then used to inoculate 1L LB/kan/chlor medium. The cells were allowed to
grow first at 37°C till OD600 reached between 0.5 and 0.6 (roughly 1.5 hrs). The culture
was then taken out of the shaker and cooled down in the cold room (4°C) for 2-3 hrs.

IPTG was then added to 0.5 mM and the growth was continued at 16°C for ~20 hrs. The

167

cells were harvested after centrifugation at 4,000 rpm for 10 min at 4°C. If not used
immediately, cells were usually frozen using liquid nitrogen and stored at -20°C.

Cells were defrosted and re-suspended in 40 mL of buffer A (same as described in
RAR-LBD purification). Protease inhibitor cocktail tablet (complete, EDTA-free,
Roche) was then added. The cell suspension was lysed using four 1-minute bursts of
sonication (probe sonicator, Biologics Inc., 60% power) on ice and centrifuged for 30
min at 4°C at 7000 rpm. The supernatant was loaded onto Ni-NTA resin (bed volume of
~ 15 mL) and the flow-through was collected. The resin was then washed with 2x50 mL
of buffer B (same as used in RAR-LBD purification). The protein of interest was then
eluted with 5x10 mL of the elution buffer (250 mM imidazole, 500 mM NaCl, 20 mM
Tris, pH 8.0). The pure fractions were combined and thrombin was added to a molar
ratio of ~ 1:500 (thrombin : protein). The solution was dialyzed against 20 mM Tris, 50
mM NaCl, pH 8.0 for a total period of 6-9 hours, during which the buffer was changed 3
times. After dialysis, the protein solution was concentrated using Centriprep 3
concentrators (Amicon, MWCO 3000) to a volume of 1-2 mL. The sample was then

placed in an eppendorf tube and centrifuged at 14,000 rpm at -4°C for 10 min to remove

precipitate. The supernatant was further purified by a size-exclusion column (HiPrepTM

16/60, SephacrleM S-300, Pharmacia Biotech) on F PLC (Pharmacia Biotech) and eluted

with the same buffer as used in the dialysis. The elution profile features two peaks, with
one at the void and one that immediately follows (Figure VI-6). Although both peaks
show very pure CRALBP, only the later peak was verified to contain fully active protein
and therefore is collected. The pure, active protein (Figure VI-7) was then concentrated

to the desired concentration, frozen in liquid nitrogen and stored at -80°C.

168

 

Figure VI-6. Typical size-exclusion purification profile of CRALBP represented by
A20-CRALBP grown at 16°C.

Figure VI-7. CRALBP after purification. A. Full length CRALBP; B. A20-CRALBP.

169

6.2. Protein crystallization
6.2.1. Co-crystallization of CRABPII mutants bound with ligands.

Preparation of complexes of the C15-a1dehyde bound to the CRABPII mutants. The light

sensitive C15-aldehyde was synthesized by a colleague (5), dissolved in ethanol, stored at
-20°C, and always protected from light. The complex was prepared freshly by adding the

Cl 5-aldehyde stock solution (0.245 M in ethanol) to the protein solution (~ 20 mg/mL) in
a volume ratio of 1:16 (CU-aldehyde in ethanol : protein solution) to a final molar ratio

of ~10:1 (C15-aldehyde : protein). The complex was first incubated at room temperature

for 30 min and then 4°C for l~2 hrs before being used for crystallization. The complex
was only exposed to dim red light when necessary (crystallization, flash freezing or under

the microscope) and was kept in the dark for the rest of the time.

Preparation of complexes of the merocyanin bound to the CRABPII mutants. The light
sensitive merocyanin was synthesized by a colleague, dissolved in ethanol, stored at -
20°C, and always protected from light. The complex was prepared by adding the
merocyanin stock solution (0.032 M) to the protein solution (~ 20 mg/mL) in a volume
ratio of 1:10 (merocyanin in ethanol : protein solution) to a final molar ratio of 2.5:1
(merocyanin : protein). The complex was first incubated at 4°C for over 1 hr and then at
room temp for several more hrs before being used for crystallization. The complex was
only exposed to dim red light when necessary (crystallization, flash freezing or under the

microscope) and was kept in the dark for the rest of the time.

170

 

Figure VI-8. Pictures of crystals of the ligand-bound CRABPII mutants. A. C15-
aldehyde-bound KL-A32E-CRABP11; B. C‘s-aldehyde-bound KL-T54E-CRABPII; C.

C15-aldehyde-bound KFLDV-CRABPII; D. merocyanin—bound KLE-R59W—CRABPII;
E. merocyanin—bound KLE—R59W—CRABPII, taken while the crystal was being cryo
protected by the nitrogen stream when sitting on the goniometer at the LS-CAT at APS.

171

Crystallization and Cryoprotection. The hanging drop vapor diffusion method was used.
The crystallization was performed at 4°C under dim red light and the crystallization
boxes were covered with aluminum foil to avoid light exposure. Boxes were kept at 4°C.

Crystals of holo-KLE-R59W were grown in 0.1M bis-tris-propane (BTP), pH 7.5,

and 30% PEG 400. Crystals appeared within 24 hrs. On the fourth day crystals were

taken out of the drop and flash frozen by immersion in liquid N2.

Crystals of holo-KL-A32E were grown in 0.1M Na citrate, pH 7.5, 0.2M
ammonium acetate, and 26% PEG 4000 (Figure VI-8A). Crystals appeared within 24 hrs
and were immediately transferred to a soaking drop containing 4 uL of the reservoir
solution (0.1M Na citrate, pH 7.5, 0.2M ammonium acetate, and 20% PEG 4000), l uL
of protein solution (22 mg/mL), and 0.5 uL of ligand stock solution (0.245M in ethanol).

After 42 hrs of soaking, crystals were briefly immersed in a cryo solution composed of

0.1M Na citrate, pH 7.5, 0.2M ammonium acetate, 26% PEG 4000, and 20% glycerol and

flash frozen in liquid N2.

Crystals of holo-KL-T54E were grown in 0.1M NaOAc, pH 5.8, and 18% PEG
6000 (Figure VI-8B). Crystals appeared after 1 week. On the same day crystals were
soaked briefly in 0.1M NaOAc, pH 5.8, 18% PEG 6000, and 30% glycerol and flash
frozen.

Crystals of halo-KFLDV were grown and frozen in the same way as holo-KL-
T54E (Figure VI-8C). Crystals appeared after 40 hrs and were flash frozen right after
appearance.

Crystals of merocyanin-bound KLE-R59W were grown in 0.1M MES, pH 6.3,

0.2M ammonium sulfate, and 30% polyethylene glycol mono-methyl-ether (PEGMME)

172

5000 (Figure VI-8D,E). Crystals displayed a distinctive purple color and grew bigger
gradually for two weeks. Crystals were soaked briefly in 0.1M MES, pH 6.3, 0.2M

ammonium sulfate, and 30% PEGMME, and 20% glycerol and flash frozen.

6.2.2. Crystallization of apo-CRABPII mutants.

The hanging drop vapor diffusion method was used. The droplet consisted 1 ul of
concentrated protein (~20 mg/ml) and 1 ul of the reservior. The crystallization was
performed at room temperature. All crystals appeared within 24 hours and were frozen
within a week after appearance. Cryoprotectants were prepared with either 20% or 30%
glycerol in addition to the components of the corresponding growth conditions in the
same concentrations. All of these apo CRABPII mutant structures are in the P 1
spacegroup.

Crystals of the apo-R132K:R111L:A32E-CRABPII (KL-A32E) were grown in
0.1M Tris, pH 8.5, 0.2M NaOAc, and 26% PEG 4000 (Figure VI-9A). The crystals were
briefly immersed in a cryo solution containing 20% glycerol before they were flash
frozen in liquid nitrogen.

Crystals of the apo-R132K:R111L:T54E-CRABPII (KL-T54E) were grown in
0.1M bis-tris-propane (BTP), pH 7.5, and 30% PEG 4000. Two microliters of the
cryoprotectant (20% glycerol) was added to the crystal-containing droplet and the
crystals were then taken out of the drop after ~30 seconds, followed by flash freezing in
liquid nitrogen.

Crystals of the apo-R132K:Yl34F:R111L:L121D:T54V-CRABPII (KFLD-T54V

or KFLDV) were grown in 0.1M Tris, pH 7.5, 0.2M NaOAc, and 24% PEG 4000 (Figure

173

VI-9B). Two microliters of the cryoprotectant (20% glycerol) was added to the crystal-
containing droplet and the crystals were incubated in the new drop for ~30 sec before

being flash frozen in liquid nitrogen.

 

Figure VI-9. Pictures of crystals of the apo-CRABPII mutants. A. apo-KL-A32E-
CRABPII; B. apo-KFLDV-CRABPII; C. apo-KFLD—CRABPII; D. apo-KFLE—CRABPII.

The apo-R132K:R111L:L121E:R59E-CRABPII (KLE-R59E) was crystallized in
0.1M BTP, pH 7.5, and 22% PEG 8000. Two microliters of the cryoprotectant (30%
glycerol) was added to the crystal-containing droplet and the crystals were incubated in

the new drop for ~30 sec before flash frozen in liquid nitrogen.

174

 

The apo-R132K2Yl34FzR111L:L121D-CRABPII (KFLD) was crystallized in 0.1M
BTP, pH 8.1, and 22% PEG 4000 (Figure VI-9C). Cryoprotection and flash freezing
were performed in the same way as for apo-KLE-R59E.

The apo-R132K:Y134F:Rll1L:L121E-CRABPII (KFLE) was crystallized in 0.1M
BTP, pH 7.0-8.5, and 20-30% PEG 4000 (Figure VI-9D). Cryoprotection and flash

freezing were performed in the same way as for apo-KLE-R59E.

6.3. Data processing and structure determination

6.3.1. Structures of the C15-aldehyde-bound CRABPII mutants.

Data Collection. Datasets were collected at the Advanced Photon Source LS-CAT 21-ID

at Argonne National Laboratory. Processing and scaling was done using HKL2000 (6).

For the dataset of KLE-R59W (1.95 A, P3121), the crystal to detector distance was 200
mm and 120° of data were collected with an oscillation of 1°. For KL-A32E (1.22 A,

P21), the crystal to detector distance was 75 mm and 200° of data were collected with an

oscillation of 1°. For KL-T54E (2.00 A, P6122), the crystal to detector distance was 200
mm and 150° of data were collected with an oscillation of 05°. For KFLDV (1.54 A,
P212121), the crystal to detector distance was 150 mm and 360° of data were collected

with an oscillation of 1° (only the first 120° were scaled).

All datasets were collected at the Advanced Photon Source at Argonne National
Laboratory. Datasets for KL-A32E, KL-T54E, KFLD-T54V and KLE-R59E were

collected at LS-CAT 21-ID. The dataset for KFLD was collected at SBC-CAT 19-BM.

175

The dataset for KFLE was collected at COM-CAT 32-ID. All datasets were processed

with HKL2000 (6).

Structure solution and refinement. Structures of KLE-R59W and KL-A32E were solved

using Molrep (7) in CCP4 (8) with KLE-CRABPII as the search model (PDB ID:ZG7B).

Structures of KL-T54E and KFLDV were solved using Phaser (9) with the same model.
Model building was accomplished in both TURBO-FRODO (10) and COOT (11)

for each of the four structures. REFMACS (12-15) within the CCP4 suite of programs

was used to refine the structures against 95% of the data, while 5% of data was chosen
randomly for cross validation, Rfree. For KL-A32E, SHELXL-97 (16) was used in the

later stages of refinement. For KL-T54E, TLSMD analysis (17, 18) was performed and

TLS refinement in REFMACS with 3 groups for Chain A and 4 groups for Chain B

resulted in improved B factors and R/Rfree.

Asp126 was the only residue outside of the allowed region in the Ramachandran
plot in all four structures after refinement. In holo-KL-A32E it resides in the generously
allowed region, while in the other three structures in the disallowed region. The same
Asp126 was found to reside in disallowed regions in all previously determined CRABPII
structures (19). The good electron density maps at this position greatly enhanced our

confidence in the non-favorable geometry of this residue.

6.3.2. Structure of the merocyanin-bound KLE-R59W-CRABPII
mutant.

Data Collection. The dataset was collected at the Advanced Photon Source LS-CAT 21-

ID at Argonne National Laboratory. Processing and scaling was done using HKL2000

176

(6). The dataset was determined to be in P3121 spacegroup with the highest resolution at

2.60 A. The crystal to detector distance was 250 mm and 180° of data were collected
with an oscillation of 1°.
Structure solution and refinement. Structure of the merocyanin-bound KLE-R59W and
was solved using Molrep (7) in CCP4 (8) with KLE-CRABPII as the search model (PDB
ID:2G7B).

Model building was accomplished in both TURBO-FRODO (10) and COOT (11).
REFMACS (12-15) within the CCP4 suite of programs was used to refine the structure

against 95% of the data, while 5% of data was chosen randomly for cross validation,

Rfree. TLSMD analysis (I 7, 18) was performed and TLS refinement in REFMACS with 5

groups resulted in improved overall B-factor and R-factors.

6.3.3. Structures of the apo-CRABPII mutants.

All structures were solved using either Molrep (7) or rigid body refinement in
RefmacS (12-14) in CCP4 with apo-WT-CRABPII MolA as the search model (PDB ID:
2FS6). Model building was accomplished in both COOT (11) and TURBO-FRODO (10)
for every structure. Restrained refinement in Refinac5 was used to refine the structures
against 90-95% of the date, while 5-10% of data were chosen randomly for cross
validation, Rfree. Anisotropic refinement for temperature factors was employed for KL-
A32E and KFLD-T54V (13). For KL-T54E, TLSMD analysis was performed and TLS

refinement with 9 groups for both chain A and chain B resulted in improved B factors as

W611 as Rwork/Rfree (15, 17, 18).

177

dosm E5288 “mg 05 8 Home“ momoficocam E mos—«.5

 

 

838 888 88.8 488 88.8 488 88.8 £88 83.8 888 3 352m
8:; 3; 8.3V :8 88: 8.3 88b 28 8.5 8.3 $822868
:8 83. 8:2 as? 3.: assume 83::
2 am $88 :88 $32 $58 eczema a3
8.: a? 86 2... 8.3 3m 88 3 8d 2:. 0: omega.
8.8 No 8.9 3. 8.3 8.2 8.3 3 8i 3 858583.
83-38 2%. 7%.: 83-8.8 82-2. 5 88. 786
8.3%.? 3. 72.8 8.35.? mm. 72.3 3. 23a :3 owes 5:283.
8388 888.8 $888 8388 888.8 9: 58233
8.82 888 8.82 8.8 882 C s
8.8 8.8 8.8 28 8.8 C a
8.88 8.8 888 888 8.8 C a
$2 8.8 on a 2m 32 :3 o
8.3 8.3 m: m9. 2% 2V 8
8.3 3... 3» 3m 3m 9: a
mcommcoEmw :00 “ED
n N N _ s 85 .aeafioev N
a: €12 .888 8a ~22 9.888%
3833.9 3an 9.329 maid. 382-me
.5585: 8883.20 8883.20 8802920 8883-20

 

can? 23:8 53.5 225.2%: 2: a 8:88.... 838:8 as. 5.x .3.» 2%...

178

 

 

8888 8.888 9888 8808 8.888 888 858
8.8 8.8 8.8 8 8.8 8.88 88388885
8 8 8 8.8 8 888 883888 888288888
8.88 3. 88 8.8 8.8 8.88 88322
8.88 8.88 8.88 8.88 8.88 8.88 82888 882
SE gégnomsam
888.8 888._ 888.8 8.8 8888 C 2888 8888
888.8 38.8 88.8 :88 88.8 88 5882 8888
88888 888 9828
8888 88888 88: 8 88888 8882 8888 8888882 8888
88 888 88 88 _8_ sea 88 .882
00.8
8.88 2.8 8.88 88.8_ _.88 8.88 88
.53
8.88 8.8_ 8.8_ 88.: 8.88 8.88 s. 8
888.88 888.88 8 8.88 888.88 888.88 :8 8928.88 88282
3888-858 >388 8888.29 888.838 3888-838
888888282 88888288 _ o 8888288 J 88888288 G 8888288 J

 

8.58.8.5 .8385 58.55 8:88.858: 85 88 8888838 88858588.. 8.5258 8.5 8.88.8

179

£288 8532082 88— 05 8 88.888 88885583 5 8823/

 

 

2.88: 8.8 8.88: _.8 8.88: _.8 8.88: 8.8 2.8: 8.8 8.88: 8.8 8.8: seem
8.88: 8.88 8.88: 8.88 8.88: 8.88 88.88: 8.88 2.88: 8.88 8.88: 8.88 8888328880
88888 8888_ 88888 8888_ 88888 888 88888882 88885
:888_ 88888 8882 88888 :88: 88888 8888882 as...
88.8: 8.88 8.8: 8.88 8.8: 8.88 88.8: 8.8_ 88.8: 8.88 8.8: 8.88 88 8882.8.
8.8: 8.8 8.: 8.8 8.8: 8.8 8. : 8.8 8.8: 8.8 8.8: 8.8 8888888888
888-88.: 88.88.: 28.788: 888.788.: 888-88.: 88:88.:
8. 78.88 8. 78.88 8.8-8.88 8. 8.8.8 8. _.8.88 8.78.88 9: 8882 888288
8888.8 88888.8 88888.8 8888._ 888._ 88888.8 9: 58828883
8.88 8.88 8.88 8.88 8.88 8.88 C 8
8.82 8.88 8.88 8.88 8.888 8.88 C 8
8.88_ 8.88 8.88 8.88 _.88_ 8.88 C 8
8.88 8.88 888 8.88 8.88 8.88 8: 8
_.88 8.88 8.88 8.88 8.88 8.88 a: 8
8.88 8.88 8.88 8.88 8.88 8.88 9: 8
mzommaoemfl :00 3:3
8 8 8 8 8 8 888 88888888: 8
8 8 _ 8 Z _ 8 Z _ 8 88888888
8.88.838 8888-88 >388 388 88.88 8888.85.

 

8.8.88.5 Esau 5885.888 85 88 888838 88882.8 388 8838 8.5 8.888

180

 

888 858

 

2858 8888 8888 8.8.88 8.8.88 8E8
8.8 8.8 8.8 8.8 8.8 8.8 8.8: 88388885
8.88
8.8 8.8 8.8 8 8 8 883888 8888888:
8.8 8.8 8.8 8.8 8.8 8.8 88: 883888.
8.88 8.88 8.88 8.88 8.88 8.88 88: 82288 8882
SE 5585858838
88._ 88._ 88.8 82 88.8 88._ C 2888 888
88.8 888.8 88.8 888.8 88.8 888.8 9: 5882 888m
88882 858 8828
88888 88: 88:8 88888 88888 88888 8888 8888882 .988
888 :8 888 88_ 888 888 as; 88 .682
00..
8.8_ 8.88 8. 88 8.88 8.88 8.88 88: 88
.53
8.88 8.8_ 88. 8.88 8.8_ 8.88 8.8: 8 m
8.88 8.88 8.88 _.88 8.88 8.88 8: 8388.88 882838
888838 8.88888 >588 3.88 8388 8888-858

 

838.83 8:55:— =mm<m0698 2: .6 8388388 85:85.88.— eufiusbm .88-; 833.

181

 

 

    

 

Psi (degrees)

_90 —_ . g _—l
. ASP 126 (A)
I

-135—‘ CHI l—l—r ‘ l

‘l I
-180 -135 -9'0 -45 0 4% 9'0 1&5 go
Phi (degrees)

 

 

 

 

 

Figure VI-10. Ramachandran plot of the C15-aldehyde-bound-KLE-R59W-
CRABPII.

182

 

 

 

Psi (degrees)

 

 

 

 

 

 

4‘5 90 155 180
Phi (degrees)

 

Figure VI-ll. Ramachandran plot of the C ls-aldehyde-bound-KL-A32E-CRABPII.

183

Psi (degrees)

-l35~

gi-

 

 

 

 

 

 

-180 -135

-90

 

-457 0
Phi (degrees)

 

180

Figure VI-12. Ramachandran plot of the C15-aldehyde-bound-KL-T54E-CRABPII.

184

 

 

 

Psi (degrees)

 

 

 

 

 

-45' 0 90 1é5 180
Phi (degrees)

Figure VI-13. Ramachandran plot of the C15-aldehyde-bound-KFLDV-CRABPII.

185

 

 

 

Psi (degrees)

 

 

 

 

 

15:5 180

Phi (degrees)

Figure VI-14. Ramachandran plot of the merocyanin-bound-KLE-R59W-CRABPII.

186

 

 

 

Psi (degrees)

 

 

 

 

 

1
l
i

 

 

   

-45 i 0 45
Phi (degrees)

Figure VI-lS. Ramachandran plot of the apo-KLE-R59E-CRABPII.

187

 

180

Psi (degrees)

-135-

 

 

 

 

 

 

 

 

-180

1'
-135

-9'0

 

Phi (degrees)

Figure VI-16. Ramachandran plot of the apo-KFLE-CRABPII.

188

 

 

 

 

Psi (degrees)

 

 

 

 

 

-180 4‘35 -90 «is 0 4
Phi (degrees)

Figure VI-l7. Ramachandran plot of the apo-KFLD-CRABPII.

189

 

 

 

 

Psi (degrees)

  

-l35~

 

 

 

 

E r7 I l ' I
-180 -135 -90 -45 45
Phi (degrees)

8_

Figure VI-18. Ramachandran plot of the apo-KFLDV-CRABPII.

190

 

 
 

 

 

 

 

 

 

 

 

 

DD
is: . :i =
g ‘ l I

-45~ ‘* l l .

-90~— _ 4—-———a

—, r-J ' ASP 1.108%)
4351 ..l ‘ . i I l y . I J
I r g I J ‘ rim] ‘ '
430 -135 -90 -45 0 45 90 135 1 0

Figure VI-l9. Ramachandran plot of the apo-KL-T54E-CRABPII.

Phi (degrees)

191

Psi (degrees)

Figure VI-20. Ramachandran plot of the apo-KL-A32E-CRABPII.

 

 

 

 

 

 

 

-1358 l ‘ I
t l :l
I J —l A
-180 ' 435 do -is' 0 4'5 9'0 155
Phi (degrees)

192

 

180

6.4. Protein characterization

6.4.1. General procedure for UV-vis titrations of CRABPII/CRALBP
with retinoids.

A stock solution of the specific retinoid (retinoic acid, all-trans-retinal, ll-cis-

retinal or C15-aldehyde) was prepared in 95% spectroscopy grade EtOH. To a protein

solution, 0.1 equivalent of the retinoid was added as increments and spectra was recorded
at room temperature from 200 nm to 600 nm (Lambda 40, Perkin Elmer and Cary

WinUv, Varian). The secondary derivative of the spectra was calculated by using the

corresponding sofiware provided with the UV instrument. The determination of the Amax

of a UV peak using mathematical equations has been established as a valid method and

has been used in a variety of applications (20-24).

6.4.2. Determination of the extinction coefficients of proteins.

The absorption extinction coefficients (a) for the various proteins were determined
according to the method first described by Gill and von Hippel.(25) The Eden (8 of

denatured proteins in 6 M Guanidine HCl) is dependant primarily on the number of Tyr,

Trp and Cys residues present and can be calculated according to equation:
Eden = aetyr + ban-p + cacys
where a, b, and c represent the number of respective amino acids per molecule of protein,

and their 8 values have been determined experimentally (en-p = 5690 M.1 cm-l, em =

193

1280 M.1 cm—l, ecys = 120 M.1 cm-l, wild type CRABPII protein contains 3 Trp, 2 Tyr,

and 3 Cys residues).

Two protein solutions are prepared, at identical concentrations, one in the native

buffer (4 mM NaH2P04, 16 mM NazHPO4, 150 mM NaCl, pH = 7.3) and the other in a
denaturing buffer (6 M guanidine HCl, 4 mM NaH2P04, 16 mM NazHPO4, 150 mM
NaCl), and the Abszgo for each sample is measured.

The arm, the extinction coefficient for the native protein, can be determined
according to Beer’s law:
Abs den+ Eden 2 C den
Abs nat+ Enat = c nat

where Abs is the UV absorbance at 280 nm, under both native and denatured conditions.
Since the concentration of both samples is the same, we can equate the two equations and
solve for the extinction coefficient. The measurement must be repeated until a consistent

number is obtained.

Molar absorbance determined .°
RARy-LBD : 11,972 cm'lM'l

Full length CRALBP : 24,514 cm- i M-I

A20-CRALBP : 24,267 cm‘lM‘I

194

6.4.3. Reductive amination procedure.

CRABPII (0.3 mg) in PBS buffer (1 mL) was mixed with 1.2 equivalents of

retinal from an ethanol stock (final ethanol volume less than 35 uL). The mixture was
incubated at 25°C in darkness for 10 min after which 15 ML of 5 M NaCNBH3 in H20

were added. The reaction was left for 4 hours. Absorption at 330 nm was observed of

the reaction ixture, indicating a successful reduction.

6.6.4. Circular Dichroism of protein samples.

Circular dichroism (CD) spectra of protein samples were obtained in order to verify
the correct folding of each protein. The use of CD spectroscopy for protein
characterization depends mostly on empirical rules applied by comparing the obtained
spectrum with characteristic spectra of peptides with specific folding patterns.(26)

Instrument settings and sample concentrations were very important in acquiring
reproducible CD spectra. The protein concentrations used had to be in the range of l — 5
uM. Lower concentrations gave noisy spectra, while higher concentrations produced
spectra where the amino acid absorbance overshadowed the structural features. A similar
problem was also observed when the bandwidth of the instrument, which defines how
much light is allowed to enter the sample chamber, was set higher than 1 unit.

In a typical protocol, a 5 uM solution of protein in buffer was used for CD
measurement (JASCO 710). The spectra were scanned from 500 nm to 200 nm at 100
nm/min, with a bandwidth of 1 nm. Extended degassing of the sample chamber with

nitrogen must be performed to ensure quality spectra near the 200 nm region.

195

6.4.5. Fluorescence.

Fluorescence quenching is used for the determination of dissociation constants for
binding of both retina] and retinoic acid to CRABPII proteins.(1, 27) It is known that
proteins containing Trp residues absorb light at 280 nm and fluoresce at ~340 nm. This
fluorescence can be quenched in the presence of a chromophore that is close to Trp
residues in space. As the chromophore is titrated into the protein solution, some amount
of the Trp fluorescence will be quenched, until the pocket is fully saturated and a change
is no longer detected in the Trp fluorescence. Therefore, by monitoring the fluorescence
of Trp residues in the protein as a function of the concentration of added chromophore an
apparent dissociation constant for the ligand-protein complex can be calculated (detailed
description follows).

Due to the sensitive nature of the experiment, the optimization of the fluorescence
titration experiments has been particularly tedious. Air bubbles and small particles cause
noisy spectra, therefore buffers should be degassed and samples filtered before used (28).
In addition, it has been found that the oily residue from fingerprints can demonstrate
fluorescence, so care has to be taken not to touch the quartz cuvettes (29). When
handling the cuvettes lint free gloves should always be worn, as lint on the outside of a
cuvette interferes with the reading as well. All samples should be stored in glass
containers (29). In the case of protein samples, silanized glassware should be used, as
many plastic containers will leach fluorescent materials into the sample.

Sample concentrations are also very important. Literature reports suggest the use of

a sample concentration as close as possible to the expected Kd, maintaining an accurate

measurement. Considering that the Kd values would be in the nM range, protein

196

concentration of 500 nM was found to be the optimal for our experiments. Anything
lower would compromise the accuracy of our experiment.

At the same time, the low solubility of retinal in water can cause problems, since
precipitation when high numbers of equivalents are added can produce cloudy samples,
resulting in erroneous results. In addition, since the ligand is titrated into the protein
from a stock in EtOH one must make sure that the final EtOH concentration is kept below
2%, to prevent possible protein denaturation. It has also been shown that EtOH
concentrations between 2-8% can red shifi the maximum wavelength up to 40 nm (30).

Additionally, an optimum salt concentration must be used for protein analysis in
buffer. Polar solvents, including water have been shown to quench some amount of Trp
fluorescence (31). A high salt concentration (150 mM) potentially helps to shield the
protein and reduce the amount of unwanted quenching. A11 salts used in the buffer must
be of the highest grade available to ensure minimal background fluorescence.

Maintenance of temperature and pH of the measurement samples is also critical. If
the temperature rises above ambient, the wavelength of the Trp emission red-shifts and
the intensity of the peak decreases. At higher temperatures, the excited state can return to
the ground state via other means such as phosphorescence.(28, 32) Likewise, an optimal
pH is required for the protein sample, if it is either too acidic or too basic, the protein
sample may denature (33). Typical protein samples have a Trp emission in the range of
330 — 350 nm, but this varies with the protein, the number of Trp residues, and whether
or not they are solvent-exposed (34-36). Free Trp emits at 365 nm. A denatured protein

will red shift towards 365 nm and provide a false saturation point for the titration.

197

Another point of concern is the fact that proteins tend to stick to the surface of
glass. When storing the samples, silanized glassware should be used to avoid loss of
protein and a change in protein concentration. When performing the actual fluorescence
titration with low concentration of protein, a dummy protein, such as gelatin, should be
used to stabilize the protein and avoid loss due to binding with the surface of the cuvette.
Gelatin does not contain Trp residues, therefore it will not interfere with the fluorescence
measurements.

A final note to mention is the fact that retinal alone actually interferes to some
extend with the Trp fluorescence detection. The free chromophore can quench some of
the Trp emission. To correct for this, one must perform a blank titration with N-

acetyltryptophanamide and add back this lost fluorescence (37). Details of this process

are described below. Interestingly, retinoic acid, which has a closer XmaX value to the

Trp emission wavelength, does not have the same effect.

6.4.6. General protocol followed for Fluorescence Quenching and Kd
determination experiments.

The cuvette is allowed to sit with 3 mL of a 0.01% gelatin containing PBS (4 mM
NaH2P04, 16 mM NazHPO4, 150 mM NaCl, pH=7.3) for 30-60 min. This buffer is
discarded and the cell is rinsed once with distilled water. The cuvettes are all handled
while wearing gloves so as to avoid both fingerprints and lint. Then 3 mL of a room
temperature 500 nM protein is added to the cell. The sample is excited at 283 nm with a

slit width about 1.5 run (this varies in order to ensure that the intensity of each sample

remains below one million counts). The fluorescence is measured at the peak maximum,

198

about 345 nm (varies with different proteins). The chromophore is added at varying
equivalents from a ~1.5 mM stock (in 95% spectroscopy grade EtOH) sample maintained
in the dark. A measurement is taken after each addition at the same wavelength. The
results are plotted as concentration of chromophore versus relative fluorescence intensity.

When the curve levels off, the titration is complete.

If the chromophore is all-trans-retinoic acid (Sigma, 8 = 45,000 M“1 cm“, km... =

350 nm), the data can be plotted as it is. There is minimal effect of retinoic acid on the
Trp emission as seen in a blank titration run with N-acetyltryptophanamide.

If the chromophore is all-trans-retinal (Toronto Research Company, a = 48,000

M-1 cm-l, kmax = 380 nm), a blank titration must be run with 3 mL of 1.5 uM N-

acetyltryptophanamide. The peak maximum for the blank is at 365 nm. This curve is
plotted as concentration of retinal versus change in intensity. The retinal absorbs some of
the emission observed through the Trp residues. This will be added back into each of the
protein / chromophore titrations.
The correction for the retinal titrations involves a four step calculation after

completion of both the protein titration and the blank titration (3 7).

Fmax — F

Fm... - F0

 

F max = fluorescence upon saturation

F 0 = initial fluorescence

F = observed fluorescence

a = fraction of free binding sites

199

1. Determine the free ligand concentration, R.

R = R0 — nP0(1-a)

R0 = ligand concentration

n = number of binding sites / protein, assume n=1

P0 = protein concentration

2. Find fluorescence contribution of free ligand, FR, to be deduced

from blank (see solved example below).

3. Subtract fluorescence contribution of free ligand from actual

readings, plot vs. ligand concentration.

(F -F R) vs. R0

The dissociation constant is calculated with the program SigmaPlot through non-
linear least square regression analysis.(2 7, 38) Within SigmaPlot, the corrected relative
fluorescence intensities are plotted against the concentration of retinal. The protein
concentration and the final fluorescence intensity upon saturation are the other two values

needed to solve the equation for the dissociation constant:

200

 

 

5:“ le . 13+R,+Kd—,/(B+R,+Kd)2—413R,)
Ff Ff 2P:

F = observed fluorescence

F f: fluorescence of free protein
F b = fluorescence of bound protein
P, = total protein concentration

R, = total chromophore concentration

As input into SigmaPlot (for our specific example):
y = p1+p1*((p2/p1-1)*(p3+x+k-((p3+x+Kd)"2-4*p3 *x)"0.5))/(2*p3)

y = fluorescence value

x = chromophore concentration

p1 = initial fluorescence intensity = 1

p2 =fluorescence intensity upon saturation = 0.66

p3 = protein concentration = 5*10"(-7)

' Cleaning of Fluorescence C uvettes:
As both fingerprints and protein residue on the surface of the fluorescence cells may
cause interference in the measurements, they should be thoroughly cleaned regularly,

depending on volume of use. A proper cleaning procedure entails:

201

1. Soak 24 h in chromic acid, then rinse with water.

2. Soak 8 h in a KOH / MeOH solution (20 pellets / 100 mL), then rinse with

water.

3. Soak 4 h in HNO3, then rinse with water

202

References

10.

ll.

12.

Wang, L. C., Li, Y., and Yan, H. G. (1997) Structure-function relationships of
cellular retinoic acid-binding proteins - Quantitative analysis of the ligand binding
properties of the wild-type proteins and site-directed mutants, Journal of
Biological Chemistry 2 72, 1541-1547.

Renaud, J. P., Rochel, N., Ruff, M., Vivat, V., Chambon, P., Gronemeyer, H., and
Moras, D. (1995) Crystal-Structure of the Rat-Gamma Ligand-Binding Domain
Bound to All-Trans-Retinoic Acid, Nature 378, 681-689.

Rochel, N., Renaud, J. P., Ruff, M., Vivat, V., Granger, F ., Bonnier, D., Lerouge,
T., Chambon, P., Gronemeyer, H., and Moras, D. (1997) Purification of the
human RAR gamma ligand-binding domain and crystallization of its complex
with all-trans retinoic acid, Biochem Bioph Res Co 230, 293-296.

Crabb, J. W., Chen, Y., Goldflam, S., West, K., Kapron, J. . (1998) Methods for
producing recombinant human cellular retinaldehyde-binding protein, Methods in
Molecular Biology 89, 91-104.

Lee, K. S., Jia, X., Vasileiou, C., Geiger, J. H., and Borhan, B., To be published.

Otwinowski, Z., and Minor, W. (1997) Processing of X-ray diffraction data
collected in oscillation mode, Macromolecular Crystallography, Pt A 276, 307-
326.

Vagin, A., and Teplyakov, A. (1997) MOLREP: an automated program for
molecular replacement, J. Appl. Cryst. 30, 1022-1025.

Bailey, S. (1994) The Ccp4 Suite - Programs for Protein Crystallography, Acta
Crystallographica Section D-Biological Crystallography 50, 760-763.

Mccoy, A. J., Grosse-Kunstleve, R. W., Adams, P. D., Winn, M. D., Storoni, L.
C., and Read, R. J. (2007) Phaser crystallographic software, Journal of Applied
Crystallography 40, 658-674.

Roussel, A., and Cambillau, C. (1989) Turbo-Frodo, in Silicon Graphics
Geometry Partners Directory, Silicon Graphics, Mountain View.

Emsley, P., and Cowtan, K. (2004) Coot: model-building tools for molecular
graphics, Acta Crystallographica Section D-Biological Crystallography 60, 2126-
2132.

Murshudov, G. N., Vagin, A. A., and Dodson, E. J. (1997) Refinement of
macromolecular structures by the maximum-likelihood method, Acta
Crystallographica Section D-Biological Crystallography 5 3 , 240-255.

203

 

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

Murshudov, G. N., Vagin, A. A., Lebedev, A., Wilson, K. S., and Dodson, E. J.
( 1999) Efficient anisotropic refinement of macromolecular structures using F FT,
Acta Crystallographica Section D-Biological Crystallography 55 , 247-255.

Pannu, N. S., Murshudov, G. N., Dodson, E. J., and Read, R. J. (1998)
Incorporation of prior phase information strengthens maximum-likelihood
structure refinement, Acta Crystallographica Section D-Biological
Crystallography 54, 1285-1294.

Winn, M. D., Isupov, M. N., and Murshudov, G. N. (2001) Use of TLS
parameters to model anisotropic displacements in macromolecular refinement,
Acta Crystallographica Section D-Biological Crystallography 5 7, 122-133.

Sheldrick, G. M., and Schneider, T. R. (1997) Shelxl: High-Resolution
Refinement, Macromolecular Crystallography, Pt B 277, 319-343.

Painter, J ., and Merritt, E. A. (2006) Optimal description of a protein structure in
terms of multiple groups undergoing TLS motion, Acta Crystallographica Section
D-Biological Crystallography 62, 439-450.

Painter, J., and Merritt, E. A. (2006) TLSMD web server for the generation of
multi-group TLS models, Journal of Applied Crystallography 39, 109-111.

Vaezeslami, S., Mathes, E., Vasilelou, C., Borhan, B., and Geiger, J. H. (2006)
The structure of apo-wild-type cellular retinoic acid binding protein 11 at 1.4 A

and its relationship to ligand binding and nuclear translocation, Journal of
Molecular Biology 3 63, 687-701.

Cahill, J. E. (1979) Derivative Spectroscopy - Understanding Its Application, Am
Lab 11, 79-&.

Cahill, J. E., and Padera, F. G. (1980) Derivative Analysis of Uv-Visible Spectra,
Am Lab 12,101-&.

Horvath, M. P., Copeland, R. A., and Makinen, M. W. (1999) The second
derivative electronic absorption spectrum of cytochrome c oxidase in the Soret
region, Biophys J 77, 1694-1711.

LEE, D. C., Hayward, J. A., Restall, C. J., and Chapman, D. (1985) 2nd-
Derivative Infrared Spectroscopic Studies of the Secondary Structures of
Bacteriorhodopsin and Ca-2+-Atpase, Biochemistry-Us 24, 4364-4373.

Sehgal, A. C., Tompson, R., Cavanagh, J ., and Kelly, R. M. (2002) Structural and
catalytic response to temperature and cosolvents of carboxylesterase ESTl from

the extremely therrnoacidophilic archaeon Sulfolobus solfataricus P1, Biotechnol
Bioeng 80, 784-793.

204

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

Gill, S. C., and Vonhippel, P. H. (1989) Calculation of Protein Extinction
Coefficients from Amino-Acid Sequence Data, Analytical Biochemistry 182, 319-
326.

Venyaminov, S. Y., and Yang, J. T. (1996) Determination of protein secondary
structure, in Circular dichroism and the conformational analysis ofbiomolecules.
(Fasman, G. D., Ed.), pp 69-107, Plenum Press, New York, NY.

Norris, A. W., Cheng, L., Giguere, V., Rosenberger, M., and Li, B. (1994)
Measurement of Subnanomolar Retinoic Acid-Binding Affinities for Cellular
Retinoic Acid-Binding Proteins by F luorometric Titration, Biochim. Biophys.
Acta-Protein Struct. Molec. Enzym. 1209, 10-18.

Gally, J. A., and Edelman, G. M. (1962) The effect of temperature of the
fluorescence of some aromatic amino acids and proteins, Biochim. Biophys. Acta
60, 499-509.

Miller, J. N. (1981) Standards in fluorescence spectrometry, in Techniques in
visible and ultra-violet spectroscopy. (Miller, J. N., Ed.), pp 1-7, Chapman and
Hall, New York.

Kronman, M. J. (1967) Fluorescence studies of teh molecular conformation of or-
lactalbumin., Biochim. Biophys. Acta 133, 19-32.

Steiner, R. F., and Edelhoch, H. (1962) Effect of thermally induced structural
transitions on the ultra-violet fluorescence of proteins, Nature 193, 375-376.

Valeur, B. (2002) Characteristics of fluorescence emission, in Molecular
Fluorescence, pp 34-71, Wiley-VCH, Weinham, Germany.

White, A. (1959) Effect of pH on fluorescence of tyrosine, tryptophan and related
compounds, Biochem. J. 71, 217-220.

Chen, R. F. (1967) Extrinsic and intrinsic fluorescence in the study of protein
structure, in Fluorescence: Theory, Instrumentation, and Practice (Guilbault, G.
G., Ed.), pp 443-509, Marcel Dekker, Inc., New York.

Eftink, M. R. (2000) Intrinsic fluorescence of proteins, in Topics in Fluorescence
Spectroscopy (Lakowicz, J. R., Ed.), pp 1-15, Kulwer Academic / Plenum
Publishers, New York.

Konev, S. V. (1967) Fluorescence and phosphorescence of proteins and nucleic
acids. , Plenum Press, New York.

Cogan, U., Kopelman, M., Mokady, S., and Shinitzky, M. (1976) Binding
affinities of retinol and related compounds to retinal binding proteins, Eur. J.
Biochem. 65, 71-78.

205

38. Wang, L., Li, Y., and Yan, H. (1997) Structure-function relationships of Cellular
Retinoic Acid-binding proteins, J. Biol. Chem. 2 72, 1541-1547.

206