Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Visual arrestin interactions with clathrin adaptor AP-2 regulate photoreceptor survival
(USC Thesis Other)
Visual arrestin interactions with clathrin adaptor AP-2 regulate photoreceptor survival
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
VISUAL ARRESTIN INTERACTIONS WITH CLATHRIN ADAPTOR AP-2
REGULATE PHOTORECEPTOR SURVIVAL
by
Hormoz Moaven
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(NEUROSCIENCE)
August 2012
Copyright 2012 Hormoz Moaven
ii
Dedication
To my wonderful family - my dad, mom, and sister - whose love and support has guided
me along the path leading to all my successes. And wholeheartedly to my best friend,
Marjon Raji, for her invaluable patience and encouragement.
iii
Acknowledgements
I would like to acknowledge my dissertation committee: Dr. Jeannie Chen, Dr.
Alapakkam Sampath, and Dr. David Hinton. First and foremost, I owe many thanks to
Dr. Jeannie Chen, who is largely responsible for my Ph.D. achievement. I extend my
sincerest gratitude for her mentorship, guidance, encouragement, and most importantly,
the opportunity to conduct research and contribute to the advancement of scientific
knowledge. Furthermore, Dr. Alapakkam Sampath provided me with great advice and a
reliable support system that ushered me along my Ph.D. studies, for which I am very
grateful. Lastly, Dr. David Hinton offered a much-needed outside perspective that
strengthened my scientific thinking. His constructive criticism and kindness is greatly
appreciated.
I must also acknowledge all my laboratory members that I have worked with
throughout my graduate research career. Their help, friendship, and support contributed
so much to my research. The collaborations with Dr. Brian Soreghan, Chia-Ling Hsieh,
Tian Wang, Yun Yao, Helen He, Valeria Mancino, Yukihiro Koike, and Dr. Wen Mao
deserve extra thanks.
Last but not least, I give immeasurable thanks to my family and to all my friends,
past and present. I have been lucky enough to be surrounded by their love and
camaraderie throughout my life. Without them, I would not have been able to accomplish
all that I have.
iv
Table of Contents
Dedication ii
Acknowledgments iii
List of Figures vi
Abstract viii
Chapter 1 Introduction 1
1.1. Photoreceptor and Phototransduction 1
1.1.1. Rods and Cones 1
1.1.2. Phototransduction Activation 3
1.1.3. Dark/Light Current 6
1.1.4. Phototransduction Deactivation – Dark Adaptation Pt. 1 7
1.1.5. Retinoid Cycle & Rhodopsin Regeneration – Dark
Adaptation Pt. 2 9
1.1.6. Light Adaptation & Translocation 11
1.2. Retinal Degeneration 14
1.2.1. Genetic Mutations Lead to Retinal Degeneration 14
1.2.2. Rhodopsin Mutations (Class I-III) 15
1.3. K296E Mouse Model of Autosomal Dominant Retinitis Pigmentosa 18
1.3.1. Previous K296E Results 19
1.4. Insight from Drosophila Studies 21
1.5. Arrestins 22
1.5.1. Beta-Arrestin & Endocytosis 22
1.5.2. Endogenously Expressed Arrestin Splice Variant, p44 23
1.6. Dissertation Outline 25
Chapter 2 Visual Arrestin Interactions with Clathrin Adaptor AP2
Regulate Photoreceptor Cell Survival 27
2.1. Abstract 27
2.2. Introduction 28
2.3. Results 30
2.3.1. Rescue of Retinal Degeneration in K296E
+
/p44
+
/Arr1
-/-
Retinas 33
2.3.2. Recovery of Visual Function in K296E
+
/p44
+
/Arr1
-/-
Mice 36
2.3.3. Long Lasting Rescuing Effect of p44 38
2.3.4. Increased Level of Endocytic Proteins in K296E Rod Outer
Segments 40
2.4. Discussion 43
2.5. Materials and Methods 50
2.5.1. Generation of Mouse Lines 50
v
2.5.2. Interaction Between AP-2 and Different Arrestin Peptides
Monitored by Electron Paramagnetic Resonance 50
2.5.3. Retinal Morphometry 52
2.5.4. Western Blot Analysis 52
2.5.5. Immunocytochemistry 53
2.5.6. Electroretinogram 54
2.5.7. Statistics 54
2.6. Acknowledgements 55
Chapter 3 Roles of Arrestin and Rhodopsin Phosphorylation in Regulating
Photoreceptor Sensitivity 56
3.1. Introduction 56
3.2. Results 59
3.2.1. Rhodopsin Regeneration Rate is slower in Arrestin
-/-
Mice than
in wildtype Mice 61
3.2.2. Different Phosphorylation Aspects Between Serine-only and
Threonine-only Rhodopsin Mutations 63
3.3. Discussion 65
3.4. Material and Methods 72
3.4.1. Light Exposure Parameters 72
3.4.2. Rhodopsin Quantification 74
3.4.3. Isoelectric Focusing (IEF) 74
Chapter 4 Potential Retinal Degeneration Protection by Cyclic Nucleotide-
Gated Channel Blockers in rd1 and rd10 Mouse Models 76
4.1. Introduction 76
4.2. Results 78
4.2.1. Varying Degrees of Rescuing Effect on Treated rd1 Retinal 79
4.2.2. Intraperitoneal Injections of CNG Blockers are Less Effective
in rd10 Mice 81
4.2.3. Subretinal Injections in rd1 Mice Show Promising Signs of
Rescue 82
4.3. Discussion 83
4.4. Materials and Methods 85
4.4.1. Retinal Explant Culture 85
4.4.2. Intraperitoneal Injections 86
4.4.3. Subretinal Injections 86
4.4.4. Retinal Morphometry 87
4.4.5. Tetracaine Derivative Formation 87
Chapter 5 Concluding Remarks 89
Bibliography 95
vi
List of Figures
Figure 1.1 Comparison of photoreceptor structure in vertebrates and
invertebrates. 3
Figure 1.2 Schematic of the activation steps of the vertebrate
phototransduction cascade. 4
Figure 1.3 Schematic of activation of the invertebrate phototransduction
cascade. 6
Figure 1.4 Comparison of phototransduction deactivation in vertebrates
and invertebrates. 9
Figure 1.5 Schematic of the retinoid cycle. 11
Figure 1.6 Light-dependent protein translocation in vertebrates and
invertebrates. 14
Figure 1.7 Spidergram analysis of retinal degeneration in various K296E. 20
Figure 1.8 Rescue of retinal degeneration in drosophila double mutants. 22
Figure 1.9 Schematic model of GPCR internalization via β-arrestin complex
formation. 23
Figure 1.10 Amino acid sequence of the C-terminus of p44 and its functional
properties. 24
Figure 2.1 The interacting domain of arrestins and binding to the
β−appendage of AP-2. 32
Figure 2.2 Retinal degeneration in K296E mice is rescued by ARR1 splice
variant, p44. 35
Figure 2.3 Visual function is restored in K296E mice expressing p44. 37
Figure 2.4 The rescuing effect of p44 on K296E is long-lasting. 39
Figure 2.5 AP-2 and clathrin are targeted to the outer segment of K296E
+
retina. 41
vii
Figure 2.6 Endophilin is more abundant in the outer segments of K296E
+
retina. 43
Figure 2.7 Models for the role of ARR1 in K296E-induced rod photoreceptor
cell death. 45
Figure 3.1 Slower rhodopsin regeneration rates in Arrestin
-/-
mice than in
wildtype mice. 62
Figure 3.2 Phosphorylated rhodopsin species persists longer in Arrestin
-/-
mice than in wildtype mice. 63
Figure 3.3 Comparison of phosphorylated species of rhodopsin in wildtype,
TTM, and STM mutant mice. 64
Figure 3.4 Response to single active wild-type and mutant rhodopsins. 69
Figure 4.1 Retinal morphology of retinal explants treated with tetracaine
derivatives. 80
Figure 4.2 Comparison of retinal morphology of rd10 mice injected with
vehicle control versus compound 9. 81
Figure 4.3 Retinal morphology after subretinal injections of tetracaine
derivatives. 82
viii
Abstract
Autosomal dominant retinitis pigmentosa (ADRP) is a blinding disorder whose
most frequent causes are rhodopsin mutations. Of the more than 100 rhodopsin mutations
that have been found, some disrupt the intramolecular interactions that constrain the
molecule in an active conformation. This leads to a detrimental effect on the
phototransduction cascade, which ultimately leads to retinal degeneration. One such
mutation is Lys296Glu (K296E), in which case a lysine residue is replaced with a
glutamic acid residue. It has been previously shown that opsin mutations in Drosophila
form stable opsin/arrestin complexes that causes retinal degeneration via photoreceptor
cell death. Similarly, mice expressing the K296E rhodopsin mutation also show
progressive degeneration due to formation of a stable K296E/arrestin1 complex that is
toxic to mammalian photoreceptors. The degeneration in Drosophila has been attributed
to endocytosis of the rhodopsin/arrestin complexes via AP-2/clathrin binding. By
effectively altering the genetics of K296E mice, I investigated whether the underlying
causes and mechanisms that give rise to retinal degeneration associated with the K296E
mutation are comparable to those found in Drosophila. This was achieved by substituting
arrestin1 with a naturally occurring arrestin1 splice variant, p44, in which case the C-
terminus is truncated at residue 370. This truncation eliminates the AP-2 binding domain
and in turn prevents endocytosis. Expression of p44 in arrestin1 knockout mice reveals
normal binding to rhodopsin and phototransduction activity. Accordingly, expressing p44
in K296E
+
/Arr
-/-
mice should indicate whether endocytosis is mediating retinal
ix
degeneration and support our hypothesis that cell death is resultant of AP-2-mediated
endocytosis of the K296E/Arrestin1 complex.
Furthermore, I provide insight into another form of retinal degeneration, Oguchi
disease, which is hallmarked by an inability to effectively dark-adapt from bright light
conditions. The underlying cause of the disease has been identified as a mutation in
arrestin and/or rhodopsin kinase, where a defect in either protein would prolong recovery
from a light response. A further investigation of the roles of arrestin and rhodopsin kinase
is still needed. I show that arrestin knockout mice exhibit a reduced rate of rhodopsin
regeneration, thus reducing rod sensitivity. I go on to investigate whether specific
phosphorylation residues on the C-terminus of rhodopsin regulate proper recovery
kinetics and present their possible roles in dark adaptation.
Lastly, a mutation affecting the rod cGMP-phosphodiesterase causes a form of
retinitis pigmentosa. The rd1 mouse model of this disease exhibits an accumulation of
cGMP concentrations inside the rod photoreceptor, which causes cell death due to the
increase in cytosolic Ca
2+
influx through cGMP-gated ion channels. I investigated the
possible therapeutic treatment of CNG channel blockers in the rd1 mouse model.
Preliminary treatments show promising results as retinal degeneration was slowed in
several cases.
1
Chapter 1
Overview of Photoreceptors, Phototransduction, and Possible
Mechanism of Retinal Degeneration
1.1 Photoreceptors And Phototransduction
1.1.1 Rods and Cones
The mouse visual system has evolved to focus light to the rear of the eye where the
retina lies. The retina is made up of several cell layers organized to convert light into
electrical stimulation that is then transmitted to the visual cortex to be processed into a
perception. The layers are comprised of seven types of cells: retinal epithelium cells,
photoreceptors, bipolar cells, ganglion cells, horizontal cells, amacrine cells, and Müller
cells. Specifically, the core of phototransduction resides within the photoreceptor cells.
The mouse contains two types of photoreceptors. The first being the rod, which is
specialized for dim light intensities and is the most abundant cell type accounting for
~97% of all photoreceptors. The second type is the cone, which is specialized for bright
light intensities and color, and accounts for the remaining ~3% of photoreceptors. While
both types share similar functionality and structure, this review will mainly focus on the
rod cell type.
The rod photoreceptor can be described as having three sections - the outer
segment, the inner segment, and the cell body. The phototransduction cascade is localized
to the specialized ciliary outer segment, which is made up of many transmembrane discs
spanning the entire section. Each disc membrane contains the phototransduction
2
machinery, while the ion channels/pumps responsible for the rod’s membrane potential
localize along the outer segment membrane. The inner segment houses other key proteins
involved in the phototransduction cascade as well as the cellular organelles such as the
golgi complex, endoplasmic reticulum, and mitochondria. The cell body contains the
nucleus and the synapse where voltage-gated ion channels regulate the release of
neurotransmitter into the synaptic terminal (Fig. 1.1 A).
The drosophila visual system follows the same general principles as the mouse yet
the structure of its photoreceptors is vastly different. The drosophila eye itself is actually
a compound eye made up of about 800 simple eyes. Each simple eye is a hexagonal
structure, known as an ommatidium, and contains its own cornea and is made up of 20
cells - eight of which are photoreceptors cells, while the remainder consists of secondary
and tertiary pigments cells and mechanosensory bristle cells (Fig. 1.1 C, D).
Each photoreceptor cell contains a microvillar structure known as the rhabdomere.
The rhabdomere is analogous to the mouse rod and cone outer segments, however instead
of lining up with the rest of the cell, it is adjacent to the cell body. The microvilli
comprise the plasma membrane on which most of the phototransduction machinery
resides, equivalent to the membranous discs located in the rod and cone outer segments in
mice. The cell body contains the nucleus and the axon that advances the converted light
signal forward eventually reaching the brain (Fig. 1.1 E, F).
3
Fig. 1.1. Comparison of photoreceptor structure in vertebrates and invertebrates. The mammalian
photoreceptors come in two types: rods (a) and cones (b). Each photoreceptor contains an outer
segment, inner segment, and cell body – the outer segment is separated from the rest of the cell by
cilia. The outer segments are comprised of transmembrane discs, which house the
phototransduction machinery. The drosophila ommatidium (c) consists of 8 photoreceptor cells
(d). Each photoreceptor cell (e) contains stacks of microvillar rhabdomeres, which are comprised
of actin and house the phototransduction machinery (f). The rhabdomeres are adjacent to the cell
body and are not separated by cilia. (Figure adapted from Wright 2010 and Wang & Montell
2007, respectively)
1.1.2 Phototransduction Activation
The phototransduction cascade involves a series of biochemical steps that convert
photons of light into a downstream chemical signal (Fig. 1.2). The activation of this
cascade begins with the absorption of a photon of light by rhodopsin, the G-protein
coupled receptor bound to rod outer segment membrane. Rhodopsin contains a
photosensitive chromophore, 11-cis-retinal, which isomerizes into all-trans retinal upon
photon absorption (Fu and Yau 2007). This conversion results in a conformational
change in rhodopsin and leaves it in the active state known as Metarhodopsin II (R*). R*
4
then binds to its respective G-protein, transducin, causing an exchange of GTP for GDP
on the transducin alpha subunit (Trα*) and its subsequent dissociation from the
beta/gamma subunits and rhodopsin. R* is free to activate more transducins, while the
activated transducin alpha subunit binds to the cGMP-phosphodiesterase (PDE). This
interaction releases PDE inhibition and allows for hydrolysis of cGMP into GMP. As a
result, a decrease in cGMP concentration closes cGMP-gated channels and reduces the
concentration of Na
+
and Ca
2+
inside the cell. Thus, the rod becomes hyperpolarized,
leading to closure of downstream voltage-gated Ca
2+
channels and a decrease in
glutamate release in the synaptic cleft. The amount of glutamate release is responsible for
the continuation of the transduction signal further along the retina.
Fig. 1.2. Schematic of the activation steps of the vertebrate phototransduction cascade. Upon
absorption of a photon of light (hv), rhodopsin becomes activated (R*) and repeatedly binds the
G-protein, transducin (G). This interaction exchanges GDP for GTP, which dissociates transducin
into its subunits and activates the Gα subunit (G*α). G*α binds to phosphodiesterase (E) and
activates it (E*). E* then hydrolyzes intracellular cGMP (cG) into GMP. The decrease in cG
causes cyclic nucleotide channels to close, thus preventing an influx of Na
+
and Ca
2+
, which
results in the hyperpolarization of the membrane potential. The decrease of Ca
2+
triggers calcium
feedback mechanisms that restore the membrane potential, such as activation of Guanylyl
Cyclase (GC) via GCAP and the release of calmodulin (CM) from the ion channels. This restores
intracellular cG concentration and reopens channels. (Figure adapted from Lamb and Pugh 2000)
5
Phototransduction in the drosophila also begins with the absorption of light by
rhodopsin, however it contains a slightly different chromophore, 11-cis-3-hydroxyretinal
(Fig. 1.3). Upon light absorption, the chromophore also isomerizes to an all-trans state,
leaving rhodopsin in a stable active form, also referred to as metarhodopsin. However,
unlike mice, the drosophila chromophore does not dissociate from metarhodopsin after
light exposure. Similar to mice, drosophila metarhodopsin binds to its respective
heterotrimeric G-protein, G
q
, and causes an exchange of GTP for GDP. This also results
in the G
q
alpha subunit dissociation from its beta and gamma subunits. In contrast to
phototransduction in mice where cyclic-nucleotide signaling is observed, phosphoinositol
signaling is involved in drosophila phototransduction. Therefore, the active G
q
alpha
subunit activates the norpA encoded phospholipase C (PLC). Activated PLC hydrolyzes
phosphatidylinositol 4,5-bisphos-phate (PIP2) to generate inositol 1,4,5-trisphosphate
(IP3) and diacylglycerol (DAG). The concentration of PIP2, IP3, and DAG all play key
roles in the opening and closing of transient receptor potential (TRP) and TRP-like
(TRPL) channels. Similar to CNG channels in mice, TRP channels in drosophila regulate
the membrane potential through Na
+
/Ca
2+
influx.
6
Fig. 1.3. Schematic of the invertebrate phototransduction cascade. Upon light absorption (hv) by
11-cis-3-hydroxyretinal, rhodopsin is activated (metarhodopsin). Metarhodopsin then binds to the
G-protein (G
q
) exchanging GDP for GTP, thus activating the alpha subunit and causing it to
dissociate (Gα
q
). Activated Gα
q
binds to and activated phospholipase C (PLC) causing the
hydrolysis of PIP
2
into IP
3
and DAG. The increase in IP
3
and DAG cause the TRP channels to
open and increase the influx of Na
+
and Ca
2+
, resulting in the depolarization of the cell.
1.1.3 Dark/Light Current
In the absence of light, there is an intrinsic steady ‘dark’ current flowing through
the mouse rod photoreceptor. This current is maintained by a relatively large
concentration of cGMP present inside the cell, which keeps the CNG channels open and
allows a steady influx of Na
+
/Ca
2+
. The concurrent efflux of Na+ and Ca2+ by the
Na+/Ca2+, K+ exchanger and Na+/K+-ATPase, maintains a depolarized membrane
potential of about -40mV. The depolarized state of the cell causes voltage-gated Ca
2+
channels to open, which results in release of glutamate at the synaptic terminal. Upon
light absorption and ultimately the hydrolysis of cGMP into GMP by PDE, a decrease in
cGMP concentration results in closure of the CNG channels and a reduction of Na
+
/Ca
2+
inside the cell. Thus, the photoreceptor becomes hyperpolarized, leading to closure of
voltage-gated Ca
2+
channels and a decrease in glutamate release. This hyperpolarizing
7
current, the ‘light’ current, is maintained until deactivation of the phototransduction
cascade. Yet a major difference between mice and drosophila is that in the presence of
light, mice photoreceptors hyperpolarize due to closure of CNG channels, while in
drosophila, photoreceptors depolarize due to opening of TRP channels. Therefore, light
absorption leads to DAG-mediated opening of TRP channels and an increase in Ca
2+
concentration inside the cell, leading to depolarization.
1.1.4 Phototransduction Deactivation – Dark Adaptation Pt. 1
The termination of this cascade and restoration of its activated proteins involved is
essential in recovery of the photoreceptor back into a state ready to absorb light, this
process is known as ‘Dark Adaptation’. Deactivating rhodopsin and regenerating 11-cis-
retinal are the two most crucial in dark adaptation (the regeneration of 11-cis-retinal will
be discussed in detail in the next section, 1.1.5). Rhodopsin deactivation is a two-step
process. The first is rhodopsin phosphorylation by rhodopsin kinase, which occurs in both
mice and drosophila. However, only in mice, a calcium-bound protein, recoverin, is
bound to rhodopsin kinase in the dark-adapted state, which inhibits its phosphorylation of
activated rhodopsin. As the Ca
2+
concentration decreases in the presence of light,
recoverin is freed of calcium and dissociates from rhodopsin kinase. This allows the
phosphorylation of rhodopsin to occur. Once phosphorylated, rhodopsin loses most of its
activity, yet has still retains some residual activity. This residual activity is capped by the
protein arrestin (Fig 1.4), which binds to phosphorylated species of rhodopsin causing
subsequent release of all-trans retinal and inducing rhodopsin dephosphorylation. This
8
two-step process leaves rhodopsin in its chromophore-free state, opsin, and awaits
binding of a new chromophore to ready itself for further light absorption termed
rhodopsin regeneration.
With rhodopsin ready to absorb more light, there needs to be available transducin
and PDE to carry out the cascade. In mice, the deactivation of the active transducin-PDE
complex involves a few steps. The transducin alpha subunit remains active until bound
GTP is hydrolyzed into GDP. Along with transducinʼs intrinsic slow GTPase activity, a
GAP complex made up of RGS9, Gβ5, and R9AP speeds up hydrolysis. The hydrolysis
of GTP deactivates transducin alpha and dissociates it from PDE, which terminates
cGMP hydrolysis. In drosophila, the G
q
-PLC complex needs to be deactivated. G
q
-alpha
subunit also hydrolyzes GTP into GDP through intrinsic GTPase activity and is not well
understood yet whether a GAP complex is involved. This hydrolysis results in G
q
-alpha
subunit dissociating and in turn deactivating PLC activity, leading to an end to IP3 and
DAG synthesis (Fig. 1.4).
Along with deactivation of these key proteins, there needs to be a restoration of
cGMP and PIP2 in order to close the CNG channels and TRP channels, in mice and
drosophila respectively. In mice, guanylate cyclase is responsible for synthesizing cGMP
levels. Guanylate cyclase activity is regulated by guanylate cyclase activating protein
(GCAP), whose activity is mediated by Ca
2+
. When Ca
2+
is bound to GCAP, it inhibits
guanylate cyclase activity. So as a negative feedback mechanism, when Ca2+ levels drop
in response to light absorption, GCAP is free of Ca
2+
and is able to trigger guanylate
cyclase to synthesize cGMP back towards basal levels. In drosophila, PIP2 is synthesized
9
through a few steps. First, DAG kinase converts DAG into phosphatidic acid (PA). Then
through multiple enzymes, PA is converted into PIP2.
Fig. 1.4. Comparison of phototransduction deactivation in vertebrates and invertebrates. (A)
Vertebrate deactivation steps. Activated rhodopsin, Metarhodopsin (M), is deactivated when
recoverin dissociates from rhodopsin kinase (RK), allowing RK to phosphorylate metarhodopsin.
Upon phosphorylation, metarhodopsin’s activity is fully quenched by arrestin binding. In
addition, the Gα-PDE complex is deactivated when Gα subunit hydrolyzes GTP into GDP
through intrinsic GTPase activity and GAP activity of RGS9/Gβ5 and PDE. (B) Invertebrate
deactivation steps. In drosophila, metarhodopsin (M) is deactivated through phosphorylation by
rhodopsin kinase (RK) and subsequent binding of arrestin to quench activity. In addition, the G
q
-
PLC complex is deactivated when the G
q
-α subunit hydrolyzes GTP into GDP through intrinsic
GTPase activity. This hydrolysis results in G
q
-alpha subunit dissociating and in turn deactivating
PLC activity, leading to an end to IP3 and DAG synthesis. (Figure adapted from Hardie and
Raghu 2001)
1.1.5 Retinoid Cycle & Rhodopsin Regeneration – Dark Adaptation Pt. 2
As previously discussed, rhodopsin regeneration is the counterpart to rhodopsin
deactivation as the essential steps of dark adaptation. After deactivation of rhodopsin, it
must be reconstituted with new 11-cis-retinal chromophore to restore photosensitivity.
The regeneration process is known as the retinoid cycle, or visual cycle, in which the
release of all-trans-retinal after rhodopsin deactivation is converted into 11-cis-retinal
(Fig. 1.5).
The retinoid cycle begins with the photoisomerization of 11-cis-retinal upon
photon absorption, forming all-trans-retinal and thus activates rhodopsin. Then,
10
deactivation of rhodopsin by rhodopsin kinase phosphorylation and arrestin capping
initiates reduction of all-trans-retinal to all-trans-retinol by the enzyme all-trans-retinol
dehydrogenase (RDH). This reduction has been suggested to occur either when still non-
covalently bound to opsin or when introduced into the ROS cytoplasm by the ABCR
transporter. In either case, all-trans-retinol ends up in the ROS cytoplasm and is guided to
the RPE by the inter-photoreceptor retinoid binding protein (IRBP) found in the inter-
photoreceptor matrix (IPM). Once within the RPE, all-trans-retinol is guided by the
cellular retinol binding protein (CRBP) to become esterified by lecithin retinol acyl
transferase (LRAT). RPE65 then guides the now all-trans-retinyl to become isomerized
into 11-cis-retinol by the retinyl ester isomerohydrolase. Guided by cellular retinaldehyde
binding protein (CRALBP), 11-cis-retinol is then oxidized by 11-cis-retinol
dehydrogenase to form 11-cis-retinal. Finally, 11-cis-retinal diffuses back into the ROS
via IRBP guidance, where it ultimately reaches the ROS disc membrane and binds with
opsin to regenerate photosensitive rhodopsin.
While several of these proteins and enzymes have respective cone photoreceptor
counterparts, the process by which cone retinoid is released and opsin regenerated is
slightly different than rods. There is evidence indicating a complex series of events,
similar to rods, yet involve Müller cells instead of RPE cells. While this retinoid cycle
hallmarked by retinoid release is mostly conserved in all vertebrate visual systems, the
invertebrate retinoid cycle is vastly different in that the regeneration of opsin is
completely performed in the photoreceptor and the retinoid is notably covalently bound
11
to the opsin during the entire process. The regeneration is predominantly achieved by
photoisomerization of metarhodopsin back into rhodopsin.
Fig. 1.5. Schematic of the retinoid cycle. (A) The 11-cis-retinal chromophore inside rhodopsin
(Rh) absorbs a photon of light (hv) and photoisomerizes to all-trans-retinal, which activates
rhodopsin, termed ‘metarhodopsin’ (Meta). Upon rhodopsin phosphorylation and subsequent
arrestin binding, all-trans-retinal (AL) is reduced into all-trans-retinol (OL) by retinol
dehydrogenase (RDH). All-trans-retinol is released and transported across the plasma membrane
into the retinal pigment epithelium (RPE) by the inter-photoreceptor retinoid biding protein
(IRBP). Within the RPE, all-trans-retinol is chaperoned by cellular retinol binding protein
(CRBP) to become esterified (E) by lecithin retinol acyl transferase (LRAT). All-trans-retinyl
ester is then chaperoned by RPE65 to become isomerized by retinyl ester isomerohydrolase,
producing 11-cis-retinol. 11-cis-retinol is then oxidized by 11-cis retinol dehydrogenase (11-cis
RDH) to form 11-cis-retinal. 11-cis-retinal is then transported to the plasma membrane by cellular
retinaldehyde binding protein (CRALBP). Lastly, 11-cis-retinal is chaperoned across the inter-
photoreceptor matrix back into the photoreceptor by IRBP. Once inside the photoreceptor, 11-cis-
retinal is free to reconstitute with opsin to form phototsensitive rhodopsin. (Figure adapted from
Lamb and Pugh 2004)
1.1.6 Light Adaptation & Translocation
Light on earth spans a range of intensities over 11 orders of magnitude long. In
order to detect light over a wide range of those intensities, photoreceptors have been
12
designed to adjust their sensitivity and phototransduction kinetics according to the
amount of background light in the environment. This feature is known as light adaptation
and essentially increases the working range of the photoreceptors.
In general, intrinsic switching between rods and cones provides the sensitivity to a
wide range of light intensities, resulting from the different characteristics and kinetics of
rods and cones. The lower limit of this range is to detect single photons. The rods support
this ability and are sensitive to the first three orders of magnitude of light intensities.
With slight overlap with where rods begin to saturate, cones begin to respond and do so
until the brightest light condition in the range. The characteristics of each photoreceptor
allow for adaptation.
Adaptation can be parsed into two types, light and dark. Light adaptation is
adjustment to the presence of ambient light or 'background' light, while dark adaptation is
the return to basal 'dark' conditions after an intense bleaching illumination. In general,
light adaptation involves a decrease in sensitivity and an increase in phototransduction
kinetics. Ca2+ concentration also plays a role in how photoreceptors are able to account
for this range of intensities. There are a few mechanisms that contribute to the extension
of operating range during light adaptation.
In the presence of steady background light, cGMP-channels are partially closed and
in turn reduce calcium levels inside the cell. The drop in calcium leads to rhodopsin
kinase being freed from recoverin to in turn phosphorylate rhodopsin leading to
deactivation. This quickens turnover of rhodopsin and increases its availability to capture
additional photons of light, hence extending the operating range.
13
The decrease in calcium increases steady state guanylate cyclase activation by
GCAPs resulting in greater synthesis of cGMP. The rise in cGMP concentration increases
the amount of open CNG channels and helps restore the circulating current thereby
extending the operating range. Another form of range extension arises from a decrease in
calcium causing calmodulin to dissociate from CNG channels. This lowers the
concentration of cGMP needed to open channels, thus increasing the rate of circulating
current restoration.
However, at bright light intensities these mechanisms cannot account for the
bleaching of rhodopsin visual pigment. This is where the cone photoreceptor comes into
play. The cone photoreceptor has lower sensitivity and much faster recovery kinetics,
which allow for capturing light at bright intensities.
In contrast, an individual drosophila photoreceptor has the ability to respond
throughout the entire spectrum of light intensities. Light absorption leads to opening of
TRP channels and in turn a large increase in Ca
2+
inside the cell results. A steady
background light increases the steady state Ca
2+
concentration, which inhibits the TRP
channels conductivity through a calmodulin-mediated process. This inhibition results in
desensitizing the cell, requiring brighter light intensities to initiate the light current. The
increase of Ca
2+
is not large enough however to inhibit PLC activity. This allows for
further light absorption and the PLC activity is only inhibited by the subsequent transient
Ca
2+
influx.
Furthermore, translocation of transduction proteins provides light adaptation. Under
brighter light conditions in mice and drosophila, transducin-alpha and G
q
-alpha subunits
14
translocate from the outer segment and rhabdomere to the cell body, respectively. This
reduces the amount of G-protein present for rhodopsin to activate, thereby reducing
signal amplification, shifting the operating range once again to higher light intensities.
Arrestin translocates from the cell body to the outer segment and to the rhabdomere in
mice and drosophila, respectively (Fig. 1.6). This increases recovery kinetics as well as
reduces response amplitude.
Fig. 1.6. Light-dependent protein translocation in vertebrates and invertebrates. In mammalian
rods under dark conditions, arrestin is primarily found in the inner segment, while the G-protein,
Transducin (G
t
), resides in the outer segment. In light conditions, arrestin translocates to the outer
segment, while transducin moves to the inner segment. In Drosophila photoreceptors under dark
conditions, arrestin1 and arrestin2 (arrestin) is primarily found in the subrhabdomeral cell body,
while the G-protein (Gα
q
) resides in the rhabdomere. In light conditions, the arrestins translocate
to the rhabdomere, while Gα
q
moves to the subrhabdomeral cell body. (Figure adapted from
Wang and Montell 2007)
1.2 Retinal Degeneration
1.2.1 Genetic Mutations Lead to Retinal Degeneration
Retinitis Pigmentosa (RP) is one of the leading causes of blindness. Retinitis
pigmentosa affects roughly 15 million people worldwide (Bovolenta and Cisneros 2009).
15
RP patients experience progressive loss of their rod photoreceptors, which causes a
deficit in their peripheral visual field and renders them night blind. Over time central
vision is subsequently lost as cone photoreceptors begin to die, leading to a complete loss
of vision. As with many cases of blindness disorders, Retinitis Pigmentosa is caused by
genetic mutations, including autosomal dominant (ADRP), autosomal recessive, and X-
linked forms. To date, over 25 targets of genetic mutations have been identified that
cause retinitis pigmentosa, which is hallmarked by rod photoreceptor cell death. Each
genetic mutation may cause one of various functional defects, such as impairments with
phototransduction, photoreceptor structure, and the retinoid cycle, to name a few.
Mutations in arrestin or rhodopsin kinase genes lead to Oguchi disease, also known as
congenital stationary night blindness. These mutations result in an inability to deactivate
the phototransduction cascade, which leads to retinal degeneration in a light-dependent
manner (Chen 1999). While arrestin and rhodopsin kinase mutations account for the
autosomal recessive form, rhodopsin mutations seem to prevail in autosomal dominant
forms of RP.
1.2.2 Rhodopsin Mutations (Class I-III)
Strikingly, accounting for ~30% of all cases of ADRP, over 100 genetic mutations
affecting rhodopsin have been discovered. Previous studies have allowed for
classification of the mutations based on their biochemical characteristics and defects.
Class I rhodopsin mutations involve mistrafficking of near normal species of the protein,
Class II mutations cause improper folding of the protein, while Class III mutations
16
produce constitutively active forms of rhodopsin. All three classes exhibit distinct
phenotypes of retinitis pigmentosa.
1.2.2.1 Class I
Class I mutants are considered to be biochemically similar to rhodopsin in that
proper folding of the protein is maintained and normal catalytic activity is achieved.
Notably, the mutations seem to occur along the rhodopsin c-terminus tail, where domains
and residues involved in trafficking are located. Specifically, there is a deficit in the
QVAPA domain, which causes incorrect sorting of the rhodopsin to the rod outer
segment. The improper distribution of rhodopsin leads to retinal degeneration in both
light-independent and light-dependent manners. As seen in S334ter
+
/Rho
-/-
mutations of
rhodopsin, in which the QVAPA domain is disrupted, S334ter is maintained in the rod
inner segment and outer nuclear layer and failed to form rod outer segment. Moreover, in
Rho
+/+
and Rho
-/-
backgrounds, S334ter causes mislocalized normal rhodopsin species,
which is most likely due to disrupted cytoskeletal support and vesicular trafficking
caused by an over-accumulation of mislocalized S334ter in the IS and ONL.
Interestingly, mislocalized rhodopsin and rhodopsin mutants have the ability to be
activated by light in the IS. These activated species have access to chromophores, G-
proteins, and cyclases in the IS. A light activated cascade in the IS leading to increased
cAMP levels may induce a down stream cell death signal via caspase-3. In addition, some
Class I mutants lack phosphorylation sites necessary for rhodopsin shutoff. This can lead
to extended phototransduction and lead to instability of the rod photoreceptor cell.
17
1.2.2.2 Class II
Class II mutants are distinctly different than normal rhodopsin species in that they
do not fold correctly and are also left incapable of light activation. Various regions along
the entire span of the rhodopsin molecule have been identified to have such mutations,
such as P23H. The P23H mutation is a misfolded opsin species that also seems to
mislocalize in the photoreceptor. Retention of P23H molecules in the endoplasmic
reticulum causes elevated ER stress, which leads to disrupted protein
synthesis/degradation and may lead to apoptotic cell death signaling. As similarly
discussed in Class I mutants, mislocalized Class II mutants can over accumulate and
induce improper vesicular trafficking.
1.2.2.3 Class III
Class III mutants were classified by Chuang et al. in their 2004 publication that
focused on the rhodopsin mutation affecting the Arginine 135 residue, converting it to
Leucine (R135L). R135L was considered to be a Class IIb mutant, but Chuang et al.’s
work distinguished it from other mutants and thus required its own classification.
Specifically, Class III mutants, such as R135L, exhibit hyperphosphorylation and bind
with arrestin with high affinity. Stable R135L/arrestin complexes are formed and induce
endocytosis. The internalization of these complexes disrupts normal endosomal activities
of the cell and cause retinal degeneration. Other members of Class III mutations include
K296E and E113Q, which share similar characteristics as R135L, except for that they are
constitutively active forms of rhodopsin. Therefore, Class III mutants are characterized as
18
being hyperphosphorylated and tightly bound to arrestin forming complexes that are
internalized and disrupt receptor-mediated endocytic functions, and may be additionally
held in an active conformation.
1.2.3 K296E Mouse Model of Autosomal Dominant Retinitis Pigmentosa
The specific ADRP-associated rhodopsin mutation at the forefront of my research
is Lys296Glu (K296E), characterized by a naturally occurring point mutation in human
opsin at residue 296 at which lysine is replaced with glutamic acid. This mutation leaves
rhodopsin in a constitutively active form, incapable of binding chromophore, via loss of
the salt bridge between residue E113 and K296 (Robinson 1992). K296E mutants have
been previously shown to activate transducin (in vitro), become phosphorylated by
rhodopsin kinase, and bind to arrestin (Li 1995; Chen 2006). Previous notions attributed
photoreceptor cell death to constitutive activation of the phototransduction cascade.
However, not only has K296E mutants been shown to cause retinal degeneration
independent of light exposure, but K296E mutants only activated transducin upon
dephosphorylation and dissociation from arrestin (Li 1995). This shifted focus to the
notion that K296E is hyperphosphorylated by rhodopsin kinase and forms a stable
complex with arrestin. In both drosophila and mice, the stable rhodopsin/arrestin complex
has been shown to be toxic to the photoreceptor cell (Chen 2006; Alloway 2000). It has
been suggested that the endocytosis of this complex through AP-2 recruitment of
clathrin-coated pits is the cause for cell death in drosophila (Orem 2006), and this
concept remains to be confirmed in mice models of K296E.
19
Furthermore, apoptosis of the photoreceptor cells seem to be the means by which
retinal degeneration occurs in retinitis pigmentosa. Many mechanisms have been
proposed such as metabolic stress and oxidative stress through ʻlight damage’ (Sancho-
Pelluz 2008; Wenzel 2005). The exact mechanism leading to cell death remains to be
fully determined within each of these phenotypes. However, several pathways have been
identified which may be involved individually or in concert with each other in the
signaling process. Activation of calpains and cysteine as-partyl-specific proteases
(caspases) lead to mitochondrial leakage and can induce DNA fragmentation through
upregulation of endonucleases (Harwood 2005; Lohr 2006). Nevertheless, these
pathways trigger an apoptotic signal resulting in photoreceptor cell death. Among these
mechanisms, it is yet to be determined which one(s) are responsible for cell death in
K296E mutants.
1.2.4 Previous K296E Results
Similar experiments were performed in equivalent mouse models of retinal
degeneration, specifically the K296E mutation. The K296E mutation expressed in the
transducin
-/-
background undergoes retinal degeneration and can be attributed to stable
K296E/arrestin complex formation and thus independent of transducin activation of the
phototransduction cascade. Moreover, expressing K296E in the arrestin1
-/-
background
demonstrates severe retinal degeneration, which is a direct result of constitutively
activation of the phototransduction cascade through transducin. However, when K296E is
expressed in both the transducin
-/-
and arrestin1
-/-
background together, there is a
20
significant rescue from retinal degeneration (Fig. 1.7) (Chen 2006). The fact that K296E
does not activate transducin in the presence of arrestin, taken together with the
degeneration found in K296E
+
/transducin
-/-
, it can be suggested that the main cause of
retinal degeneration in K296E mutants is through stable complex formation with arrestin.
Fig. 1.7. Spidergram analysis of retinal degeneration in various K296E
+
mice. For each group
(n=4-8) of mice, the outer nuclear layer (ONL) thickness was measured at 20 equally spaced
positions along the retina. The 0 position represents the optic nerve. K296E
+
mice exhibit thinner
ONL than WT control, indicating retinal degeneration. K296E
Tr-/-
mice showed thinner ONL than
Tr
-/-
control mice, displaying retinal degeneration despite preventing excessive phototransduction.
K296E
arr1-/-
mice showed thinner ONL than arr1
-/-
control mice, showing retinal degeneration
despite blockade of K296E/Arr1 complex formation. However, in K296E
arr1-/-Tr-/-
mice, ONL
thickness is as normal as arr1
-/-
Tr
-/-
control mice, through blocking both light-dependent and
independent degeneration pathways. (Figure adapted from Chen 2006)
21
1.3 Insight from Drosophila Studies
It has been previously shown in Drosophila visual system, that some rhodopsin
mutations that lead to retinal degeneration involve the formation of rhodopsin/arrestin
complex. Specifically, light-reared norpA Drosophila mutants fail to dissociate arrestin
from activated rhodopsin and exhibit retinal degeneration (Alloway 2000). It is difficult
to distinguish whether this light dependent retinal degeneration is due to constitutive
activation of the cascade or the formation of stable rhodopsin/arrestin complexes. Yet,
removing the G-protein, Gq, in norpA mutants fails to prevent retinal degeneration and
provides evidence that the formation of stable complexes causes the retinal degeneration.
Moreover, in the arrestin null background, light-reared drosophila exhibit retinal
degeneration due to an inability to terminate the phototransduction cascade. However,
drosophila that exhibit the norpA mutation in the arrestin null background, exhibit a
rescue from retinal degeneration (Fig. 1.8) (Alloway 2000). Therefore, it can be said that
the direct formation of stable rhodopsin/arrestin complexes is responsible for retinal
degeneration and it has been further proposed that it is via endocytosis of those
complexes that seem to be the trigger for signaling photoreceptor cell death.
22
Fig. 1.8. Rescue of retinal degeneration in drosophila double mutants. (A) White-eyed control
retinal section. (B) norpA mutant retinal section reveals severe degeneration despite transduction
blockade. (C) arr2(null) mutant retinal section exhibits degeneration despite prevention of
rhodopsin/arr2 complex formation. (D) norpA/Arr2
-/-
mutant retinal section shows a rescue from
degeneration seen in norpA mutant due to removal of arrestin2. All flies were dark-reared and
exposed to 6 days of continuous room light. Eyes were fixed and 0.5 µm retinal sections were
obtained. Insets show higher magnification transmission electron micrographs of the same retinal
section shown. (Figure adapted from Alloway 2000)
1.4 Arrestins
1.4.1 Beta-Arrestin & Endocytosis
Moreover, it is now accepted that βarrestin bound to G-protein coupled receptors
can, through endocytosis, create signaling endosomes. The nature of βarrestin allows it to
act as a scaffolding protein for other signaling pathways (Castro-Obregón 2004).
Previous studies have applied this notion to visual arrestin and it seems to be a feasible
hypothesis as to how the stable rhodopsin/arrestin complex may cause retinal
degeneration. It seems that visual arrestin is acting in a similar fashion as βarrestin, in
which case it may generate a cell death signal through endocytosis. One process of
endocytosis requires the binding of adaptor proteins such as, AP-2, which initiates
23
clathrin-coated pit assembly (Fig. 1.9) (Laporte 2000). The AP-2 binding domain is
preserved across various forms of arrestins, including visual arrestin. Therefore, it seems
as though the K296E/arrestin complex is capable of undergoing endocytosis and it may
be this process that triggers cell death, as seen in similar drosophila models.
Fig. 1.9. Schematic model of GPCR internalization via β-arrestin complex formation. endocytic
vessel formation. Upon activation and subsequent phosphorylation of the β
2
-Adrenergic Receptor
(β
2
-AR) (1), β-arrestin translocates and binds to the receptor (2). The β
2
-AR/β-arrestin complex
can then be targeted to pre-existing clathrin coated pits (3a) or may recruit the binding of adaptor
protein AP-2 (3b). AP-2 can then initiate the formation of clathrin-coated pits for internalization
of the complex. β-arrestin has both a clathrin binding domain and an AP-2 binding domain.
(Figure adapted from Laporte 2000)
1.4.2 Endogenously Expressed Arrestin Splice Variant, p44
There are two forms of visual arrestin: the more abundant (~10x) common form,
full-length arrestin1 (p48) and a naturally occurring arrestin1 splice variant (p44), which
is truncated at residue 370 (Fig. 1.10 A) (Palczewski 1994). The truncation leaves p44
without the C-terminal AP-2 binding domain. Serendipitously, p44 is a perfect candidate
24
to examine whether endocytosis is responsible for retinal degeneration in K296E mice. It
has been shown that replacing arrestin1 with p44 does not have an adverse effect on
phototransduction. p44’s affinity for and binding to activated rhodopsin is similar to that
of arrestin1 and normal phototransduction deactivation occurs as demonstrated through
suction electrode recordings (Fig. 1.10 B) (Burns 2006). Thus, p44 can be expressed
confidently in K296E
+
/arrestin
-/-
mice, to investigate whether endocytosis of the stable
complex is necessary to induce retinal degeneration, since p44 lacks the AP-2 binding
domain.
Fig. 1.10. Amino acid sequence of the C-terminus of p44 (m44) and its functional properties. (A)
Alignment of the C-terminus amino acid sequences of full length arrestin and p44. (B) Rod
suction electrode recordings of the flash responses to increasing light intensities in WT, Arr
-/-
,
m44
arr-/-
, and p48
arr-/-
backgrounds. Expression of m44 in the arr
-/-
background exhibits normal
response kinetics. (Figure adapted from Palczewski 1994 and Burns 2006, respectively)
25
1.5 Dissertation Outline
Over the course of many years as a graduate student in Dr. Jeannie Chen’s Lab,
the techniques I have learned to investigate the molecular properties of phototransduction
and retinal degeneration run the gamut from cell culture assays to electrophysiology.
With these skills and the collaborative efforts with my fellow colleagues, I have
attempted to answer several questions throughout my PhD studies. The bulk of my time
has been spent on elucidating the molecular mechanism that leads to retinal degeneration
in Autosomal Dominant Retinitis Pigmentosa (ADRP). A couple of other projects I
worked on lead to some promising results surrounding the interlink between rhodopsin
phosphorylation, arrestin binding, and rhodopsin regeneration required for dark
adaptation.
In Chapter 2, I provide my investigation of a genetic mutation effecting
rhodopsin, K296E, which causes ADRP. The molecular mechanism responsible for
retinal degeneration seen in K296E mutations has been unclear. It has been postulated
that formation and endocytosis of stable K296E/Arrestin1 complexes may be triggering
cell death. Substituting arrestin1 with its splice variant, p44, which lacks the AP-2
binding domain necessary for endocytosis, should in theory prevent endocytosis-
mediated cell death. I employed morphological, electrophysiological, biochemical, and
biophysical techniques to test this theory.
In Chapter 3, focus is shifted onto the mechanisms of dark adaptation, the
molecular pathway that restores photosensitivity to rhodopsin after light exposure. The
two critical components of this recovery are the timely inactivation of photoactivated
26
rhodopsin and the regeneration of chromophore 11-cis-retinal. The inactivation of
rhodopsin is a two-step process involving phosphorylation by rhodopsin kinase followed
by arrestin binding. The regeneration of chromophore involves the various steps of the
visual cycle. Both processes are closely linked and very interdependent. I investigated the
effect on rhodopsin regeneration when arrestin is knocked out. I further investigated the
roles of specific phosphorylation sites in regulating phosphorylation kinetics and its
effect on response kinetics.
In Chapter 4, I present work performed on retinal degeneration 1 (rd1) mice,
which exhibit retinal degeneration induced by defective rod cGMP-phosphodiesterase
failing to hydrolyze cGMP. This defect increases cGMP levels in rod photoreceptors,
which results in detrimental amounts of Ca
2+
inside the cell due to prolonged opening of
cGMP-gated channels (CNG). I provide insight into treatments with variants of
Tetracaine, a drug that preferentially blocks CNG channels. I used morphological studies
to examine whether the drug is effective enough in blockade that retinal degeneration is
slowed or prevented.
27
Chapter 2
Visual Arrestin Interactions with Clathrin Adaptor AP2 Regulates
Photoreceptor Survival
2.1 Abstract
Arrestins bind ligand-activated, phosphorylated G protein-coupled receptors
(GPCRs) and terminate activation of G proteins. Additionally, non-visual arrestin/GPCR
complex can initiate G protein-independent intracellular signals through their ability to
act as scaffolds that bring other signaling molecules to the internalized GPCR. Like non-
visual arrestins, vertebrate visual arrestin (ARR1) terminates G-protein signaling from
light-activated, phosphorylated GPCR, rhodopsin. Unlike the other arrestins, however, its
role as a transducer of signaling from internalized rhodopsin has not been observed in the
retina. Formation of signaling complexes with arrestins often requires recruitment of the
endocytic adaptor protein, AP-2. We have previously shown that K296E, which is a
naturally occurring rhodopsin mutation in certain humans diagnosed with autosomal
dominant retinitis pigmentosa, causes toxicity through forming a stable complex with
ARR1. Here we investigated whether recruitment of AP-2 by the K296E/ARR1 complex
is required to generate the cell death signal in a transgenic mouse model of retinal
degeneration. We measured the binding affinity of ARR1 for AP-2 and found that,
although the affinity is much lower than that of the other arrestins, the unusually high
concentration of ARR1 in rods would favor this interaction. We further demonstrate that
p44, a splice variant of ARR1 lacking the AP-2 binding motif, rescues retinal
28
degeneration as well as visual function in K296E mice. These results reveal a novel role
of ARR1 in a G protein-independent signaling cascade leading to cell death.
2.2 Introduction
Autosomal dominant retinitis pigmentosa (ADRP) is a blinding disorder in humans
that is commonly caused by rhodopsin mutations (Malanson & Lem 2009). Some of
these mutations constrain the molecule in an active conformation, and it was previously
thought that the constitutive activation of transducin, the visual G-protein, underlies the
pathogenic mechanism. However, a study using transgenic mice that express such an
activating mutation, K296E, showed that is its deactivated by phosphorylation and
arrestin binding (Li 1995). How this class of rhodopsin mutations leads to death of
vertebrate rods remains unclear. In Drosophila, stable rhodopsin/arrestin complexes were
found to be cytotoxic (Alloway 2000; Kiselev 2000) through recruitment of AP-2 (Orem
2006), a key endocytic protein, by visual arrestin (DroArr2), followed by accumulation of
endocytosed rhodopsin in late endosomes (Chinchore 2009).
Like many other G-protein-coupled receptors, endocytosis of light activated
Drosophila rhodopsin is part of normal cell physiology but internalization has not been
observed for vertebrate rhodopsin in the intact retina. Drosophila phototransduction also
diverges substantially from that of vertebrates in terms of utilization of Gq, which
couples to phospholipase C, whereas vertebrates utilize transducin, which belongs to the
Gi class and activates a cGMP phosphodiesterase, PDE6 (Arshavsky 2002; Montell
2012). Furthermore, Drosophila phototransduction occurs in the rhabdomeres, which are
29
contiguous with the rest of the cell where rhodopsin/arrestin complex has direct access to
the endocytic machinery. In contrast, vertebrate rhodopsin is synthesized in the inner
segment and transported to the membranous outer segment compartment, which is
separated from the rest of the cell by a slender cilium. Once rhodopsin is incorporated
into outer segment discs, its concentration there remains constant (Young 1967),
indicating a lack of internalization. Based on the divergence between vertebrate and
invertebrate phototransduction, as well as the differences in the architecture of the
photoreceptor cell, a conserved pathogenic mechanism of rhodopsin/arrestin complex
cannot be assumed.
We used transgenic mice that express K296E to investigate whether AP-2
recruitment by rhodopsin/arrestin complex contributes to the cell death signaling pathway
in the vertebrate retina. We show that ARR1 lacks a high-affinity motif for binding the
clathrin-adaptor protein AP-2, but the high concentration of Arr1 in rods can still drive
this interaction. Importantly, we show that retinal morphology and function is preserved
when K296E is expressed with a naturally occurring ARR1 splice variant, p44, which
deactivates photolyzed rhodopsin normally but lacks the AP-2 binding element. This
rescueis long-lasting, in the timeframe of up to two years. Furthermore, we provide
immuncytochemical evidence that the K296E/ARR1 complex recruits AP-2 and clathrin.
Together, these results provide the first evidence for a role for vertebrate ARR1 in a
signaling pathway distinct from its known function in quenching phototransduction.
30
2.3 Results
We have previously shown that stable rhodopsin/arrestin complex is toxic to the
mammalian retina in a cell death pathway that is conserved from flies to mice (Chen et
al., 2006). To see whether endocytosis of K296E/ARR1 mediates an apoptotic signaling
pathway, we investigated the ability of ARR1 to bind AP-2, the adaptor protein central to
coated pit formation (Kirchhausen 1999; Owen 2004; Robinson, 2004; Traub 2003). The
AP2 complex is composed of two large subunits (α and β2) and two smaller subunits (µ2
and σ2). The appendage domains of α and β2 are found at the end of flexible linkers that
extend from the core of the adaptor complex (Traub 2003). These appendages recruit
interacting partners from the cytosol and thereby facilitate their concentration at sites of
coated pit formation (Schmid 2006). Non-visual arrestins and DroArr2 interact
specifically with the β2 appendage of AP2 (Laporte 2002; Orem 2006). The molecular
basis of this interaction can be seen in a crystal structure, where the contact sites were
identified to be DxxFxxFxxxR (Schmid 2006). A closer look at the amino acid alignment
of these residues shows that the binding motif is largely conserved in DroArr2 and
vertebrate non-visual arrestins (Fig. 2.1 A). However, a critical residue in AP-2
interaction, R384, is not conserved in mammalian ARR1 (Fig. 2.1 B). Previous
mutagenesis studies indicate that this is a key residue in the interaction between β-
arrestin and AP-2 (Kim 2002; Laporte 2000, 2002; Milano 2002). To quantify the effect
of the difference in amino acid at position 384 for AP-2 binding, we synthesized spin-
labeled peptides that contain the binding motif for AP-2 in the various arrestins and
measured their binding affinity for AP-2 using electron paramagnetic resonance. This
31
assay is based upon the change in line shape and the concomitant loss in signal amplitude
that occurs when the small peptides become ordered and exhibit reduced tumbling upon
binding to AP-2. The most pronounced loss in amplitude was observed for non-visual
arrestin (Fig. 2.1 C) indicating that this peptide exhibits the strongest binding.
Quantification of the spectral components of the bound and unbound states results in a
Kd of 9 µM. This low micromolar binding affinity is consistent with previous reports
(Schmid 2006). In contrast, the binding for ARR1 was weaker with a Kd of 306 µM. This
loss in binding affinity was almost entirely compensated by the N384R mutation, which
effectively converted the conserved binding sites of ARR1 to DroArr2 (Fig. 2.1 A).
These data show that a large difference in binding affinity exists between ARR1 and the
other arrestins for AP-2. However, since the concentration of ARR1 in rods is estimated
to be >2 µM (Song 2011; Strissel 2006) as compared to the ubiquitously expressed
arrestin-2 and -3 that are expressed at ≤ 0.2 µM (Gurevich 2004), binding between ARR1
and AP-2 could occur in vivo.
32
Fig. 2.1. The interacting domain of arrestins and binding to the β-appendage of AP-2. (A) The
residues on the arrestins involved in AP-2 binding. ARR1: visual arrestin; βarr-1 and βarr-2, two
forms of the ubiquitously expressed β-arrestin; DroArr2, the predominant form of Drosophila
visual arrestin; ARR3, cone arrestin. Human and mouse sequences have the h and m prefix,
respectively. (B) Side chain interactions of residues D374, F377, F380 and R384 from β-arrestin
with the β-appendage of AP-2 bound as shown in the crystal structure. (C) EPR spectra of spin
labeled peptides derived from β-arrestin, ARR1-N348R and ARR1 in the absence (black lines)
and presence (red lines) of the AP-2 beta domain. The line broadening causes the reduction in
amplitude upon the presence of AP-2, which is the consequence of spin label immobilization
upon binding. The individual spectral pairs in are normalized to the same number of spins.
Subtraction of the spectrum of the free peptide from that in the presence of AP-2 was used to
determine the percentage of the bound form. Concentrations in the first two spectral pairs (β-
arrestin and ARR1-N348R) were 20 µM peptide and 200 µM AP-2. Concentrations in ARR1
were 13 µM peptide and 267 µM AP-2.
33
2.3.1 Rescue of Retinal Degeneration in K296E
+
/p44
+
/ARR1
-/-
Retinas
Rod cells express two splice variants of ARR1: the abundant, full-length 48 kDa
form that is nearly stoichiometric with rhodopsin and a less abundant, truncated variant
called p44 in bovine rods (Palczewski & Smith 1996; Smith 1994). By expressing each
singly in the ARR1
-/-
rods, we have previously shown that both are equally efficient in
deactivating photolyzed, phosphorylated rhodopsin (R*-P) (Burns, 2006). Interestingly,
the p44 ARR1
1-370A
is truncated at the carboxyl-terminus and does not possess the AP-2
binding motif shown in Fig. 2.1 A. If AP-2 recruitment by ARR1 mediates the retinal
degeneration caused by K296E, then a prediction can be made that cell death would be
prevented in ARR1
-/-
rods that express p44. Furthermore, R* inactivation would proceed
normally in these rods. To test these predictions, K296E was introduced into the p44
ARR1-
/-
background and compared to various control mouse lines. Fig. 2.2 shows retinal
morphology from a 10–week old mouse with only the K296E
+
transgene. This mouse
showed thinning of the outer nuclear layer (ONL) when compared with a transgene-
negative littermate (compare Fig. 2.2 B to 2.2 A). Thinning of the ONL as well as marked
shortening of the outer segments was also observed when K296E was introduced into the
ARR1
-/-
background although the toxic rhodopsin/ARR1 complex would not be formed.
We attributed this degeneration to the unmasking of the catalytic activity of K296E when
ARR1 is not present to stop the signaling, creating a source of “equivalent light”. It is
known that such light exposure would lead to an adaptive behavior termed “photostasis”
(Penn & Williams 1986; Shi 2005), whereby the outer segment is shortened to decrease
photon catch. In support of this mechanism, retinal degeneration in K296E+ARR1-/-
34
mice was prevented in the rod transducin knockout (GNAT1
-/-
) background (Fig. 2.2 D,
see also Chen 2006). Also as predicted, the retinal morphology was noticeably improved
in littermates of K296E
ARR1-/-
that expressed p44: the outer segment length and ONL
thickness were similar to that of K296E-negative littermate (compare Fig. 2.2 E to 2.2 A).
The degree of retinal degeneration can be quantified by measuring the thickness of the
ONL across the entire span of the retina, which showed a ~30% reduction in the
thickness of K296E retinas when compared to that of transgene-negative littermates. This
reduction was statistically significant over the central region of the retina (Fig. 2.2 F
asterisks, p<0.01). Retinal degeneration was not observed in K296E
+
/p44
+
/ARR1
-/-
retinas; the ONL thickness was comparable to that of K296E transgene-negative control
retinas (Fig. 2.2 F). Thus, p44 appears to protect the K296E
+
retina from photostasis
without introducing the toxic effects of the rhodopsin/ARR1 complex.
35
Fig. 2.2. Retinal degeneration in K296E mice is rescued by the ARR1 splice variant, p44. Retinal
sections were prepared from 10-week old mice. (A) Normal retinal morphology from K296E
-
littermate showing organized outer segment structures and 10-12 layers of photoreceptor cell
nuclei at the outer nuclear layer. (B) The outer nuclear layer is thinned in the K296E
+
mice due to
photoreceptor cell death. (C) K296E
ARR1-/-
retina showed shortened outer segments and decreased
thickness of the outer nuclear layer. Retinal morphology was improved when K296E
ARR1-/-
mice
were further bred into the GNAT1
-/-
background (D) or when they were crossed with p44
transgenic mice to obtain K296E
+
p44
+
ARR1
-/-
mice (E). The photomicrographs in panels A-E
were taken at the same magnification. Scale bar = 25 mm. ONL, outer nuclear layer; IS, inner
segment; OS, outer segment. (F) Morphometric measurements of ONL thickness across the entire
span of the retina along the central superior-inferior axis. K296E
-
(□, N=3), K296E
+
(▲, N=3),
K296E
+
/p44
+
/ARR1
-/-
(●, N=6). Error bars represent standard error of the mean. The groups were
compared by ANOVA followed by Tukey’s HSD. * denotes a significant difference, p<0.01. S,
superior; I, inferior.
36
2.3.2 Recovery of Visual Function in K296E
+
/p44
+
/ARR1
-/-
Mice
A key distinction between rescue of K296E
ARR1-/-
retina by transgenic expression
of p44 (Fig. 2.2 E) and by GNAT
-/-
(Fig. 2.2 D) is that p44 can provide functional rescue,
whereas GNAT
-/-
cannot due to the absence of rod transducin. We used
electroretinography (ERG) on 10-week-old mice to investigate whether recovery of
visual function accompanied morphologic improvement in the K296E
+
/p44
+
/ARR1
-/-
retina. To do so, there should be sufficient quantities of p44 to bind K296E and light-
activated rhodopsin, R*. We have previously determined the amount of K296E protein to
be 5% of total rhodopsin pool, corresponding to ca. 25 pmol per retina (Shi 2005). The
amount of p44 was estimated to be 12% of endogenous ARR1, or ca. 50 pmol (Song
2011). This expression level, albeit low, is nevertheless expected to restore rapid
recovery of the light response (Song 2011). The ERG a-wave reflects the summed
responses from the photoreceptor cells whereas the b-wave corresponds to responses
from the bipolar cells. Consistent with the loss of rods, K296E
+
mice exhibited an
attenuated ERG and an increased light threshold when compared to their transgene-
negative littermates. In contrast, the K296E
+
/p44
+
/ARR1
-/-
mice showed similar
responses as the control mice (Fig. 2.3 A). Quantification of the amplitudes of the a- and
b-wave is shown in Fig. 2.3 B and C, respectively. Both were significantly reduced in the
K296E
+
mice (p < 0.01), but not in K296E
+
/p44
+
/ARR1
-/-
mice, which showed similar a-
wave responses and even increased b-wave responses when compared with the transgene
negative mice. Thus p44 restores visual function to K296E mice in addition to preventing
K296E-induced retinal degeneration. Because p44 is functionally equivalent to full length
37
ARR1 in binding R* but lacks the AP-2 binding sequences, this result strongly implicates
AP-2 recruitment by K296E/ARR1 complex as a mechanism for cell death.
Fig. 2.3. Visual function is restored in K296E mice expressing p44. (A) Electroretinogram
recordings show that K296E
+
mice are less sensitive and evoke smaller responses to light flashes
than their transgene-negative littermate. The expression of p44 restored similar sensitivity and
response waveform in K296E
+
/p44
+
/ARR1
-/-
similar to control mice. Comparison of the a-waves
(B) and b-waves (C) from control (∆, N=3), K296E
+
(■, N=4) and K296E
+
/p44
+
/ARR1
-/-
(●,
N=6) mice. Both showed responses from K296E
+
/p44
+
/ARR1
-/-
were similar to transgene-
negative mice. Flash strengths are in number of photons (f)•mm
-2
delivered.
38
2.3.3 Long Lasting Rescuing Effect of p44
Older animals were examined for the long-term effectiveness of p44 in preventing
retinal degeneration in K296E mice. In one-year old K296E mice, the ONL progressively
thinned to three to four cell layers; the outer segment shortened as well (Fig. 2.4 A).
Remarkably, the expression of p44 in K296E
+
transgenic mice was effective in
prolonging photoreceptor cell survival even after 18 and 24 months (Fig. 2.4 B and C,
respectively). ERGs were performed to test whether retinal function is correspondingly
retained. Fig. 2.4 D shows that this was indeed the case. As expected from the degree of
degeneration, responses from one-year old K296E showed a further decrease from 10
weeks. In contrast, a 2-year old K296E
+
/p44
+
/ARR1
-/-
mouse showed robust responses
indicating that visual function was maintained over long periods of time.
39
Fig. 2.4. The rescuing effect of p44 on K296E is long-lasting. (A) At one year of age the K296E
+
retina showed further decrease in outer segment length and outer nuclear layer thickness. In
contrast, the morphology of a K296E
+
/p44
+
/ARR1
-/-
mouse at 18 months (B) and 24 months (C)
was better preserved as evidenced by the increased length of the outer segment and thickness of
the outer nuclear layer. Scale bar = 25 mm. (D) Electroretinogram measured from mice shown in
(A) and (C) showing functional preservation up to 24 months. Flash strengths are in number of
photons (f)•mm
-2
delivered.
40
2.3.4 Increased Level of Endocytic Proteins in K296E Rod Outer Segments
Internalization of ligand-activated GPCRs into endocytic vesicles can be visualized
as puncta on the plasma membrane using immunofluorescent labeling of AP-2 or clathrin
(Burtey 2007; Scott 2002). We looked for co-localization of K296E with these endocytic
markers in retinal sections to support the genetic evidence of AP-2 recruitment by the
K296E/ARR1 complex. Staining of retinal sections from transgene-negative mice with an
antibody against AP-2 and clathrin showed puncta that are characteristic of endocytic
vesicles residing in the inner segment and cell body, but absent from the outer segment
(Fig. 2.5 A and B, WT. The boundary between outer and inner segment is demonstrated
by rhodopsin immunoreactivity (red) at the outer segment shown in the panel to the
right). This is consistent with previous observations for the primary sites of endocytosis
in rod cells (Cotter 1989; Hollyfield & Rayborn 1987). In contrast, AP-2 and clathrin
reactive puncta were seen to extend into the outer segment of the K296E
+
retina where
K296E is also localized with endogenous wildtype rhodopsin (Fig. 2.5 A and B,
K296E
+
). Similar to transgene-negative mice, AP-2 and clathrin reactivity in the outer
segment was absent in the K296E
+
/P44
+
/ARR1
-/-
retina (Fig. 2.5 A and B, K
+
p44
+
A
-/-
),
perhaps due to the absence of the AP-2 binding domain on p44. At the outer nuclear
layer, mislocalized rhodopsin in K296E
+
retinas is seen to occasionally co-localize with
AP-2 and clathrin-positive puncta (Fig. 2.5 C, arrows). Staining using antibody to
endophilin, another key endocytic vesicle marker, yielded similar results (data not
shown). Together, the data suggest that the AP-2 binding motif on full length ARR1
recruits endocytic proteins to K296E. When the AP-2 motif is not present, as in the case
41
of p44, K296E is not recruited to endosomes.
Fig. 2.5. AP-2 and clathrin are targeted to the outer segments of K296E
+
retina. Localization of
AP-2 and clathrin (green, A and B respectively) in the indicated retinal sections. The boundary
between outer (os) and inner (is) segment is distinguished by rhodopsin immunoreactivity in the
outer segment (red) shown in the panels at the right. In WT K296E
-
retinal sections, AP-2 (A) and
clathrin (B) immunoreactivity (green) is restricted to the IS and ONL whereas rhodopsin (red) is
predominantly localized to the OS. In the K296E
+
retina, AP-2 and clathrin positive puncta can be
seen to extend to the OS (middle panels, A and B, respectively) where K296E is also localized.
The altered distribution pattern of AP-2 and clathrin is reversed in the K296E
+
/p44
+
/ARR1
-/-
retina (A and B, right panels). (C) Localization of AP-2 and clathrin (green, left panels) and
rhodopsin (red, middle panels) in the ONL of K296E
+
retina. The number of rhodopsin-positive
puncta ranged from 0-8 per cell, the average being 4 ± 3 (Sdev, N=20). The sections were
counterstained with DAPI to visualize nuclei. These panels are representative of independent
experiments repeated ≥ 4 times. Scale bar = 25 mm (A, B). Scale bar = 10 mm (C)
Western blot was performed on isolated rod outer segments (ROS) from K296E
+
,
K296E
-
, control C57, and K296E
+
/p44
+
/ARR1
-/-
mice as an independent assay to
compare the level of endocytic proteins recruited to this cellular compartment (Fig. 2.6).
Gβ5 served as a loading control: Gβ5 has two splice variants; the long form is
42
specifically localized to the ROS whereas the short form is expressed in the inner
segment and in other retinal cells (Watson 1996). The level of Gβ5L therefore serves as a
loading control for the amount of ROS in each lane, whereas the signal from Gβ5S is
indicative of contamination from other cellular compartments or cell types. A
representative western blot is shown in Fig. 2.6 A, and quantified values of the
endophilin signal normalized to that of Gβ5L are shown in Fig. 2.6 B. Indeed, the amount
of endophilin appeared to be increased ten-fold in K296E
+
ROS when compared to the
control samples. This trend was reversed when K296E retinas expressed p44 instead of
ARR1. In summary, results from the immunocytochemistry and Western blot
experiments are consistent with a role of K296E/ARR1 in the recruitment of endocytic
machinery, and abrogation of this recruitment by replacing Arr1 with the mutant lacking
AP-2 binding site prevents cell death.
43
Fig. 2.6. Endophilin is more abundant in the outer segments of K296E
+
retina. (A) Western blot
of isolated outer segments from the indicated mice showed more endophilin in the K296E
+
sample. Gβ5L served as a loading control for the amount of ROS whereas the amount of Gβ5S is
indicative of contamination from non-ROS sources. The blot is representative of three
independent experiments. (B) The signals derived from endophilin and Gβ5L was quantified
using the LI-COR Odyssey software package. The value obtained for endophilin was normalized
to that of Gb5L. The bar graph shows mean ± SEM. The relative amount of K296E
+
sample is
significantly different than that of control, which represents combined K296E
-
and C57 samples
(2-tailed student’s t-test, p<0.0002).
2.4 Discussion
Rhodopsin consists of the opsin protein moiety and the 11-cis retinal
chromophore, which is attached to residue K296 through a protonated Schiff base linkage
(Bownds 1967). Photon absorption converts 11-cis to all-trans retinal and initiates a
series of changes in the protein leading to the catalytically active form, R*. Eventually,
the Schiff base linkage is hydrolyzed and all-trans retinal dissociates from opsin, and the
apoprotein molecule is constrained to an inactive conformation by a salt bridge that is
44
formed by K296 and E113. The K296E mutation destroys the retinal attachment site as
well as the salt bridge (Nathans 1990; Sakmar 1989; Zhukovsky & Oprian 1989).
Consequently, K296E does not bind 11-cis retinal, is insensitive to light, and is locked in
an active conformation (Robinson et al., 1992). Based on our results, we propose a model
that recruitment of AP-2 by K296E/ARR1 induces rod cell death (Fig. 2.7 A). The
identity of the signaling molecules recruited by this complex would be of great future
interest. When the K296E/ARR1 complex formation is prevented in the ARR1
-/-
background, K296E in turn activates transducin and induces cell death through a G-
protein dependent pathway (Fig. 2.7 B). This mechanism of cell death is circumvented in
the transducin knockout background (Fig. 2.2 D and Chen 2006). A full morphologic and
functional rescue occurs when p44 is expressed. In this scenario, K296E is silenced and
recovery of the light response is restored (Fig. 2.7 C). Interestingly, the b-wave of
K296E
+
/p44
+
/ARR1
-/-
retinas tends to be larger than in control mice, suggesting an
additional synaptic phenotype. The mechanism that underlies this phenotype will be the
subject of future investigations.
45
Fig. 2.7. Model for the role of ARR1 in K296E-induced rod photoreceptor cell death. (A) K296E
in an active conformation is highly phosphorylated by rhodopsin kinase and is bound to ARR1.
AP-2 is recruited to K296E/ARR1 complexes, which becomes a target for clathrin-coated
vesicles. Through recruitment of other signaling molecules to these clathrin-coated vesicles, a cell
death signal is generated. (B) In the ARR1 knockout background, K296E constitutively activated
transducin, which then leads to cell death. (C) The constitutive activity of K296E is quenched by
p44. Because p44 does not contain an AP-2 binding domain, endocytosis does not occur and the
cell death signal is not generated.
Each rod cell contains 3 mM rhodopsin and nearly equimolar of ARR1 (Song
2011; Strissel 2006). Under bright light exposure, millimolar concentration of R*/ARR1
complex would be formed. What may prevent this complex from causing cell death?
First, sequestration of these complexes to the outer segment and away from the endocytic
46
machinery in the cell body may have evolved to ensure that the potentially toxic product
of phototransduction does not interfere with normal endocytosis. Second, the low affinity
between ARR1 and AP-2 may have evolved as a protective mechanism against formation
of excessive amount of R*/arrestin complex in the inner segment for those newly
synthesized rhodopsin molecules that are en route to the outer segment. Third, residues
375-377 of ARR1, which overlap with the AP-2 binding element (FVF in bovine),
participate in a network of intramolecular interactions that are responsible for stabilizing
the basal conformation of ARR1 (Gurevich 2004); thus, the AP-2 interacting element is
masked while ARR1 exists in the latent, inactive conformation but exposed upon receptor
binding (Burtey 2007). In contrast to wildtype rhodopsin, which is effectively transported
to the outer segment discs, K296E is also mislocalized to the plasma membrane
surrounding the cell body (Li 1995). Although our data show that K296E recruited
endocytic proteins to the outer segment, we hypothesize that toxicity of K296E arises
primarily form the formation of stable K296E/ARR1 complex in the cell body, where it
has access to other signaling molecules for execution of cell death cascade. Alternatively,
the K296E/ARR1 complex may cause cell death by depleting endocytic proteins from the
cell body and disrupting normal endocytosis.
Rhodopsin mislocalization occurs in a different class of naturally occurring,
disease-causing mutations at its carboxyl-terminus (Concepcion & Chen 2010; Li 1996;
Sung 1994). These mutants form a visual pigment with 11-cis retinal and activate
transducin in a light-dependent manner (Kaushal & Khorana 1994; Sung 1993).
Therefore, they could potentially generate elevated levels of R*/ARR1 complex in the
47
IS/ONL compartments upon light exposure. We investigated this possibility in a
transgenic mouse line that expressed Q344ter. We found that Q344ter is a substrate for
light-dependent phosphorylation in the IS/ONL compartments, but in this case preventing
formation of rhodopsin/arrestin complex by raising the mice in darkness had little effect
on retinal degeneration (Concepcion & Chen 2010). Rhodopsin mislocalization also
occurs when its transport machinery is defective, such as in the case of functional loss of
kinesin-2, a major motor for anterograde transport (Marszalek 2000). In this disease
model neither rhodopsin/arrestin complex or light-activation were necessary for cell loss
(Lopes et al., 2010). What could be the explanation for the lack of effect by these
mislocalized rhodopsins? A major difference between K296E and these other models is
the stability of the K296E/ARR1 complex: since K296E does not bind 11-cis retinal and
exists in an active conformation independent of light, it can be phosphorylated and form a
stable complex with ARR1 as soon as it is synthesized. In the case of mislocalized
Q344ter or wildtype rhodopsin in kinesin-2 loss-of-function, ARR1 will dissociate from
opsin after all-trans retinal is hydrolyzed and released. Thus, despite mistrafficking of
these rhodopsin molecules, there may not be sufficient levels of R*/ARR1 complex
formed in the IS/ONL to recruit AP-2.
Like vertebrate rods, Drosophila photoreceptors express two visual arrestins: the
more abundant Arr2 and a less abundant, shorter Arr1 that is present at ~10% of the full
length arrestin (Dolph 1993; Hyde 1990; LeVine 1990; Smith 1990). It was previously
shown that Arr1 serves to scavenge phosphorylated rhodopsin, thus preventing
accumulation of toxic rhodopsin/Arr2 (Satoh & Ready 2005). Does p44 have a similar
48
pro-survival role in normal rods by preventing recruitment of AP-2 by ARR1? Evidence
in support of this idea would be a light-dependent decrease in cell survival in retinas that
express only the full length arrestin, such as that seen in the Drosophila. However, we
found no evidence of increase rod cell death when the full length arrestin was expressed
alone in the ARR1
-/-
mice (Burns 2006). In this regard, the function of the minority
arrestin species is likely to be different between invertebrate and vertebrate
photoreceptors. This is not be surprising given the compartmentalization of R*/ARR1
complex from the IS/ONL compartments where endocytosis takes place. We hypothesize
that the pathogenesis of K296E lies in the stability of the K296E/ARR1 complex and the
mislocalization of this complex to the cell body at high levels sufficient to recruit AP-2.
The transcript level of K296E in this transgenic mouse line was estimated to be 25% of
total rhodopsin (Li 1995), whereas the protein level was ascertained to be 5% (Shi 2005),
corresponding to 150 µM that can bind ARR1 and recruit AP-2. The lowered protein
level may reflect increased protein degradation of endocytosed K296E.
Could non-visual arrestins mediate endocytosis of K296E? They are capable of
binding R*P in vitro (Gurevich1995) and are ubiquitously expressed, albeit at much
lower concentrations (≤ 0.2 µM) (Gurevich 2004). Two pieces of evidence suggest that
this is unlikely. First, the observation that K296E is retained in the cell body through its
interaction with ARR1 but is released to the OS in the absence of ARR1 (Chen 2006)
suggest that K296E/ARR1 is the primary complex formed in the rods. Second, the rescue
of retinal degeneration in K296E by p44 arrestin also implicates K296E/ARR1 as the
primary toxic species. However, these results do not rule out a minor role of non-visual
49
arrestins in K296E toxicity.
Previous cell culture experiments using ADRP-associated rhodopsin R135G,
R135L, and R135W showed that these mutants were substrates for phosphorylation and
arrestin binding in vitro (Shi 1998), and co-localized with visual arrestin (ARR1) in
endocytic compartments when they were co-transfected into HEK cells (Chuang 2004).
Although these results suggest a role of endocytosis in the toxicity of the vertebrate
rhodopsin/arrestin complex, the in vitro system has several caveats. First, it was noted
that depending on expression level, R135 mutants could already appear in endosomes
when transfected alone; this behavior was not observed for K296E, which was primarily
targeted to the plasma membrane in HEK cells (Chuang 2004). Second, R135L was
phosphorylated by an endogenous kinase expressed by the cultured cells, which is not the
native GRK1 that phosphorylates rhodopsin in photoreceptors. This may be of
importance inasmuch as the sites and extent of rhodopsin phosphorylation affects arrestin
binding (Maeda 2003; Mendez 2000; Vishnivetskiy 2007). Third, whether ARR1
interacts with endocytic proteins will depend on their binding affinity and the
concentration of the reactants. Our results show that millimolar concentration of ARR1 is
required to drive this interaction whereas a much lower concentration of non-visual
arrestin would be needed. In this regard, it is possible that non-visual arrestins are
involved in the toxicity of R135 mutants expressed in cultured cells. Differences in the
pathogenic mechanisms of R135 mutants and K296E warrant further investigation.
Mutations yielding constitutively active GPCRs have been seen in many different
diseases (Schöneberg 2004; Seifert & Wenzel-Seifert 2002). The ability of p44 to restore
50
visual function and prevent cell loss caused by K296E has implications for therapy where
a combination of approaches to over-express p44 and knockdown K296E and ARR1 may
be more effective than any single approach alone.
2.5 Materials and Methods
2.5.1 Generation of Mouse Lines
All mice used in the experimental procedures were treated in compliance with
regulations set forth by the ARVO Statement for the Use of Animals in Ophthalmic and
Visual Research as well as with USC Institutional Care and Use Committee guidelines.
K296E
+
transgenic mice were obtained from Tiansen Li (Li 1995) and p44
+
mice (Burns
2006) were both bred with ARR1 knockout mice to obtain K296E
+
/ARR1
-/-
and
p44
+
/ARR1
-/-
mice, respectively. These mice were then bred together to obtain
K296E
+
/P44
+
/ARR1
-/-
mice and their complement negative controls. All mice were
maintained on a C57BL6/J background and dark-reared as to prevent any associated
light-dependent retinal degeneration.
2.5.2 Interaction Between AP-2 and Different Arrestin Peptides Monitored by Electron
Paramagnetic Resonance
Human AP-2 beta-appendage (AP-2 beta-ear, residues 701-937) expression and
purification were described elsewhere (Schmid 2006). Briefly, N-terminally His
6
-tagged
human AP-2 beta-ear was expressed in BL21 cells, lysed, purified with NiNTA, followed
by anion exchange. The protein was dialyzed into 20 mM Hepes pH7.4, 150 mM NaCl.
51
The following peptides were obtained from Biomer Technology (Pleasanton, CA):
ARR1 wild type (DENLVFEEFARQNLKDTGEC), ARR1-N384R
(DENLVFEEFARQRLKDTGEC), and arrestin-2 (DDDIVFEDFARQRLKGMKDC).
C-terminal cysteines of the peptides were labeled with MTSL spin label [(1-oxyl-2,2,5,5-
tetramethylpyrroline-3-methyl)-methanethiosulfonate] for 1 hour at RT. Samples were
mixed at 1:1, 1:20, and 1:40 molar ratio of arrestin peptide to AP-2 beta-ear protein.
Spectra were recorded with a scan width of 100 Gauss at 12 milliWatt incident
microwave power using ER4119 HS resonator and normalized to the same number of
spins. These normalized spectra were then used to subtract out any signal of unbound
peptide using the peptide only samples. This allowed quantification of the fraction of
bound and unbound peptide and, thereby, determination of binding constants.
2.5.3 Retinal Morphometry
The superior pole of the cornea was cauterized for orientation before enucleation of
the eyeballs. The cornea and lens were removed and the remaining eyecup was fixed
overnight at 4
o
C in ½ Karnovsky buffer (2.5% glutaraldehyde, 2.0% paraformaldehyde,
in 0.1M cacodylate buffer [pH 7.2]). The eyecups were washed with 0.1 M cacodylate
buffer before being treated with 1% osmium tetraoxide diluted in 0.1 M cacodylate
buffer. Eyecups were washed again with 0.1 M cacodylate buffer, dehydrated, embedded
in epoxy resin and cut into one µm thick sections and stained with Richardson stain (1%
toluidine blue, 1% azure II, 1% borax). Thickness of the outer nuclear layer (ONL) was
measured at 10 points equally spaced along the superior hemisphere and at 10 points
52
equally spaced along the inferior hemisphere. At each of the 20 total positions, three
measurements were taken and averaged. Measurements were made using a camera lucida
connected to a light microscope, a WACOM graphics tablet (Vancouver, WA) and
AxioVision LE Rel. 4.1. software (Zeiss Co., Goettingen, Germany). Prior to each
measurement session, the setup was calibrated using a stage micrometer (Klarmann
Rulings, Litchfield, NH)
2.5.4 Western Blot Analysis
Western blot analysis was performed using isolated rod outer segments (ROS) as
described (Tsang 1998). Ten retinas were pooled from ten-week old dark-reared K296E,
K296E
+
/p44
+
/Arr1
-/-
mice and their respective negative controls. The retinas were
suspended in Ringer's buffer (130 mM NaCl, 3.6 mM KCl, 2.4 mM MgCl
2
, 1.2 mM
CaCl
2
, 10mM HEPES, 0.02 mM EDTA, pH 7.4) containing 8% Opti-prep and vortexed
for two minutes. The supernatant was removed and collected, then 8% Opti-prep buffer
was added to resuspend the pellet and the process was repeated four more times. The total
supernatant collected was centrifuged on a discontinuous gradient containing 10% and
18% Opti-prep and centrifuged at 70K RPM (Beckman TLA100 rotor) for 1 hour. The
resultant ROS band was then removed and diluted in three volumes of Ringer’s buffer.
The sample was divided into three equal amounts and spun down at 45K for 20 mins. The
supernatant was then discarded and the remaining pellets were stored at -80
o
C. Western
blot was performed by vortexing the frozen ROS pellet in 100µl of Ringer's Buffer
containing 1% n-Dodecyl-ß-maltoside (EMD Biosciences Inc., La Jolla, CA) and
53
protease inhibitors (Roche Diagnostics, Indianapolis, IN). After 30 minutes of
incubation at room temperature, an equal amount of ROS homogenate from each sample
was loaded onto a 12% Bis-Tris SDS-PAGE gel (Invitrogen, Carlsbad, CA). Proteins
were transferred onto nitrocellulose membrane (VWR, Brisbane, CA) and incubated with
desired antibody: rabbit anti-endophilin-1 (Invitrogen, Carlsbad) at a 1:250 dilution and
rabbit anti-Gß5 at a 1:3000 dilution (CT215, M. Simon). A fluorescent-labeled secondary
antibody (IRDye-800 anti-rabbit IgG, 1:20000 dilution, LI-COR, Lincoln, NE) was used
to visualize the signals using an LI-COR Odyssey Infrared Imaging system.
2.5.5 Immunocytochemistry
Eyecups were prepared from dark-reared 10-week old Wildtype, K296E
+
,
K296E
+
/P44
+
/Arr1
-/-
mice and fixed in 4% paraformaldehyde and 0.5% glutaraldehyde in 0.1M
cacodylate buffer (pH 7.2). Tissues were washed and infiltrated overnight at 4
o
C with 30%
sucrose in 0.1M cacodylate buffer. Eyecups were cut in half, embedded in O.C.T. (Tissue Tek),
and later sectioned into 10 mm thick slices using a Leica Cryostat (CM3050S). The sections were
blocked in 5% BSA, 5% serum in PBS with 0.5% Triton-X 100 detergent for one hour and then
incubated with mouse anti-α-Adaptin (AP2) or mouse anti-clathrin heavy chain (Clathrin)
antibody (1:200 and 1:400 dilution, respectively, Sigma-Aldrich, St. Louis, MI) in dilution buffer
(5% BSA, 5% serum, in PBS) overnight at 4
o
C. The sections were then washed in dilution buffer
and incubated with the secondary antibody (Fluorescein anti-mouse IgG, 1:400 dilution, Vector
Laboratories, Burlingame, CA) for one hour. To visualize rhodopsin the same sections were
incubated again with blocking buffer for one hour, and then incubated with mouse anti-rhodopsin
C-terminus (1D4) antibody (1:800 dilution, Vector Laboratories, Burlingame, CA). The sections
54
were then washed in dilution buffer and incubated with the secondary antibody (Texas Red anti-
mouse IgG, 1:400 dilution, Vector Laboratories, Burlingame, CA). The sections were mounted
using Vectashield™ with DAPI (Vector Laboratories, Burlingame, CA) to visualize cell bodies.
Images were acquired using a spinning-disc confocol microscope (UltraVIEW VoX; Perkin
Elmer, Waltham, MA) under the same conditions.
2.5.6 Electroretinography
10-week old C57, K296E
+
, and K296E
+
/P44
+
/ARR1
-/-
mice are dark-adapted
overnight and processed under dim red light conditions. The mice were anesthetized by
intraperitoneal injection of a solution containing xylazine and ketamine (10 µg/g body
weight and 100 µg/g body weight, respectively) diluted in PBS. Pupils were dilated with
0.5% tropicamide and 2.5% phenylephrine hydrochloride solutions. To keep the eye from
drying out and to create an electrical contact between the eye and the corneal electrode, a
drop of hydroxyl propyl methylcellulose solution was placed on the eye. A reference
electrode was inserted subcutaneously near the eye. A family of flashes was delivered
using a narrow bandpass interference filter (500 nm), from dim to increasingly brighter
light intensities, which was controlled using neutral density filters. The ERG responses
acquired were then amplified using an AC/DC differential amplifier and filtered for
further analysis.
2.5.7 Statistics
For the ERG and morphometric measurements, ANOVA was performed on the
groups of data to detect differences, if any, between the means of each data group by
55
analyzing the variance of the individual data points within each group as well as between
each group using sum of squares. This was followed by Tukey HSD (Honestly
Significant Difference) test to parse out which groups were significantly different than
the others. For the bar graph that summarized the western data 2-tailed student’s t-test
was performed.
2.6 Acknowledgements
We thank Dr. Harvey McMahon (Medical Research Council, Cambridge, United
Kingdom) for providing the His
6
-tagged human AP-2 b-appendage and Dr. Robert S.
Molday (University of British Columbia, Vancouver, Canada) for the rhodopsin 1D4
antibody.
56
Chapter 3
Roles of Arrestin Binding and Phosphorylation in
Photoreceptor Sensitivity
3.1 Introduction
Our visual system has a remarkable ability to sense a vast range of light
intensities, due in large part to the sensitivity and temporal kinetics of the rods and cones
(Arshavsky & Burns 2012). The rods in particular have an extraordinarily high
sensitivity, which allows for reliable capture of single photons of light, most suitable for
dim light conditions. To extend the range into increasing levels of light, rods must adjust
their phototransduction kinetics to avoid extreme saturation. This process is known as
‘Light Adaptation’ and involves rapidly reducing sensitivity while speeding up
transduction events (Fain 2001). In contrast, recovering from exposure to bright or
prolonged light is a much slower process, known as ‘Dark Adaptation’. The restoration of
rhodopsin’s ‘dark’ photosensitivity involves the inactivation and decay of photoactivated
rhodopsin, the regeneration of new 11-cis-retinal chromophore via the retinoid cycle, and
reestablishing fresh rhodopsin (Lamb & Pugh 2004). These events work in concert to
regulate rod photoreceptor sensitivity and their exact contributions have been widely
studied, yet still remain unclear, both in regards to single photon responses as well as
under bright light conditions.
Upon photon absorption by rhodopsin, photoisomerization of 11-cis-retinal to all-
trans-retinal occurs, producing Meta II (MII), the active form of rhodopsin. The
formation of MII initiates the phototransduction cascade and remains active until its
57
catalytic activity is quenched when phosphorylated by rhodopsin kinase and subsequently
bound by arrestin. Prolonged or intense exposure to light results in slow recovery of rods
back into the ‘dark state’ due partially to the persistent phosphorylation of bleached
rhodopsin. Mouse rhodopsin has six available phosphorylation sites at its C-terminus,
consisting of three serine and three threonine residues. Previous studies have shown that
the three serine residues are predominantly phosphorylated during inactivation and are
necessary for desensitization (Mendez 2000; Kennedy 2001). However, phosphorylation
of all six residues seems to direct normal response recovery, as has been suggested by
single photo response (SPR) studies (Doan 2006). The physiological role of each
phosphorylation site in SPR kinetics has yet to be clarified.
Additionally, subsequent arrestin binding is highly dependent on the amount of
phosphorylated sites on rhodopsin and it is still unclear as to which combination is most
effective. It seems as though phosphorylation of two serine residues is sufficient to attract
arrestin, but having more greatly improves quenching of catalytic activity and in fact
stabilizes MII (Vishnivetskiy 2007). Interestingly, in conjunction with phosphorylation of
MII, neighboring unbleached rhodopsins may become phosphorylated by rhodopsin
kinase (Chen 2006). This phenomenon, termed transphosphorylation, aides in
desensitizing the rod during light adaptation and the amount of phosphorylation of
unbleached rhodopsin present is inversely proportional to the amount of bleached
rhodopsin (Shi 2005). Taken together, dark adaptation must include proper inactivation
of bleached rhodopsin as well as the dephosphorylation of both bleached and unbleached
species of rhodopsins. The presence of these phosphorylated species highlights the
58
importance of arrestin in regulating dark adaptation kinetics since it is effectively
involved in removing these opsin photoproducts. Moreover, arrestin plays a role in the
all-trans-retinal hydrolysis by all-trans-retinol dehydrogenase (RDH), thus influencing
the rate of 11-cis-retinal regeneration (Saari 1998). It may be the case that arrestin
binding has a significant role in regulating dark adaptation kinetics. This notion remains
to be determined and will be addressed later on in this study.
In addition to the roles of phosphorylation, arrestin binding, and
dephosphorylation required for dark adaptation, the release of all-trans-retinal and
regeneration of 11-cis-retinal chromophore is a critical component of dark adaptation
(Saari 2000; Palczewski 1999). The phosphorylation and arrestin binding of MII leads to
hydrolysis of all-trans-retinal by RDH to form all-trans-retinol. It has been previously
thought that all-trans-retinol is released from rhodopsin upon hydrolysis, but recent
studies suggest that release of the retinoid only occurs upon binding of new 11-cis-retinal
(Lamb and Pugh 2004). The dissociated all-trans-retinal is rapidly transported to the
retinal pigment epithelium (RPE) at which point all-trans-retinol is esterified into all-
trans-retinyl, then isomerized into 11-cis-retinol by the retinyl ester isomerohydrolase.
11-cis-retinol is oxidized into 11-cis-retinal, at which point the chromophore diffuses
back into the rod outer segment and regenerates the pigment by binding to opsin,
restoring photosensitive rhodopsin and completing dark adaptation (Lamb & Pugh 2004).
There is a lot of discrepancy between the attempts of previous studies to single
out the rate-limiting step of dark adaptation as well as the contributions of specific steps
in accounting for such slow recovery times. So far, the steps of the retinoid cycle seem to
59
contribute most to the slow kinetics of dark adaptation and it has been suggested that the
hydrolysis of all-trans-retinal is the rate-limiting step (Saari 1998). However, recent
studies have also demonstrated that the isomerization of retinyl esters is in fact the rate-
limiting step (Palczewski 1999). As such, there is an unquestionable need for elucidating
the exact process of dark adaptation and to parse out the contributions of each step.
Specifically, we have attempted to elucidate the role that arrestin binding may have on
the rate of rhodopsin regeneration with regards to the decay of MII. Moreover, we
provide insight into the distinct roles of specific phosphorylation sites in restoring
sensitivity to single photons in regards to phosphorylation rates.
3.2 Results
The slow recovery of dark adaptation has been attributed to rhodopsin
regeneration as well as the inactivation of photoactivated rhodopsin (Lamb and Pugh
2004). Arrestin has been largely set aside as having any major factor in the slow recovery
as it swiftly quenches phosphorylated rhodopsin. It leads one to assume that the rate of
phosphorylation is the most crucial while arrestin binding is just a secondary event.
However, previous studies and current knowledge of the retinoid cycle’s role in dark
adaptation may suggest otherwise and there may be a strong dependence on the activity
of arrestin and interlink of specific phosphorylation sites (McBee 2001). Phosphorylation
of serine and threonine residues on the C-terminus is the first step in reducing catalytic
activity (Lee 2010). Phosphorylation promotes subsequent arrestin binding, which fully
quenches the active rhodopsin. The specific presence and order of phosphorylation of the
60
three serines and three threonines highly regulate the response kinetics, specifically in
regards to arrestin binding and initial desensitization (Lee 2010; Vishnivetskiy 2007;
Arshavsky 2002; Kennedy 2001). Altering the arrangement of these residues has
provided a method to examine the significance of phosphorylation on the light response.
Previous studies on single photon responses have shown that response kinetics are
critically dependent on the amount of phosphorylation sites present on the C-terminus,
where three residues are sufficient for proper desensitization yet all six residues are in
fact necessary to provide reliable and reproducible responses (Doan 2006). Accordingly,
this data holds true for normal response conditions, which has set up previous work to
examine the order of phosphorylation events under constant illumination conditions.
However, there remains a void in this area of investigation as the exact roles of the
individual sites in determining the response kinetics remain unclear as well as their effect
on arrestin binding. In collaboration with Anthony Azevedo and Fred Rieke, we have
performed experiments on wildtype mice and two distinct transgenic mice that express
mutations on rhodopsin’s C-terminus to compare the specific roles of serines and
threonine phosphorylation sites, if any. The mutation effectively substitutes all three
serines or all three threonines with alanine, resulting in expression of threonines only
(Serine Triple Mutant, STM) or serines only (Threonine Triple Mutant, TTM),
respectively. Interestingly, TTM mice show near wildtype SPRs, while STM mice have a
dramatic variability in SPR recovery kinetics. In this study, we compared rhodopsin
levels in wildtype and arrestin-/- mice at different time points of dark recovery after
61
exposure to constant illumination as well as analyzed the phosphorylation properties in
TTM and STM mice.
3.2.1 Rhodopsin Regeneration Rate is Slower in Arrestin
-/-
Mice than in Wildtype Mice
To ascertain arrestin’s role in the rate of chromophore and rhodopsin
regeneration, wildtype and arrestin
-/-
mice were exposed to constant bright illumination
for 5 minutes and thereafter moved to complete darkness to initiate dark adaptation.
Retinas were extracted and analyzed for rhodopsin content after various recovery
durations. Preliminary results have provided some insight on the effect when arrestin is
removed. The rate of regeneration is actually quite normal in the absence of arrestin;
however the extent of regeneration appears to be delayed. Both wildtype and arrestin
knockout mice had somewhat similar increases of rhodopsin between 60, 90, and 120
minutes of dark recovery (Fig. 3.1). However, earlier time points after light exposure
suggests that rhodopsin concentrations remains somewhat steady until 30-60 minutes in
the arrestin-/- mice, unlike wildtype mice. Unexpectedly, the concentration of rhodopsin
did not regenerate back to normal levels after two hours of dark incubation time. Possible
explanations for this will be discussed in section 3.3.
62
Fig. 3.1 Slower rhodopsin regeneration rates in wildtype and Arrestin
-/-
mice. Mice were exposed
to 5 minutes of constant bright illumination and subsequently placed in the dark for recovery
incubation. Retinas were extracted after dark incubation times of 0, 15, 30, 45, 60, 90, and 120
minutes. The amount of rhodopsin concentration was quantified as percent of total and plotted
against time to observe regeneration rates. Arrestin
-/-
mice seem to regenerate rhodopsin at a
similar rate as Wildtype retinas, though it is markedly delayed as evidenced by a shift to the right.
To examine how phosphorylation is affected by the absence of arrestin, wildtype
and Arrestin
-/-
mice were exposed to 10 minutes of constant bright light then placed in the
dark for 15 minutes to facilitate dark adaptation. Comparing phosphorylated species of
rhodopsin directly after bleach to species found after 15 minutes of dark recovery,
revealed a persistent amount of phosphorylated rhodopsin in the Arrestin
-/-
mice (Fig.
3.2). This confirms that arrestin is an integral step in the removal of phosphorylated
rhodopsin and that the presence of increased amounts of these species may contribute to
the slowed regeneration rates.
63
Fig. 3.2 Comparison of phosphorylated species of rhodopsin in wildtype and Arrestin
-/-
mice.
Mice were exposed to constant bright illumination for 10 minutes. Retinas were extracted directly
after illumination or after subsequent dark recovery of 15 minutes. Isoelectric focusing of
phosphorylated species showed that Wildtype and Arrestin
-/-
mice exhibited strong
phosphorylation of all six sites directly after bleaching. However, the amount of phosphorylated
species was much higher in Arrestin
-/-
mice after 15 minutes of dark recovery, indicating the
persistence of phosphorylated rhodopsin in these mice.
3.2.2 Different Phosphorylation Aspects Between Serine-only and Threonine-only
Rhodopsin Mutations
The mutant mice studied here comprise of substitution of alanines for all three
threonines or all three serines in the rhodopsin C-terminus, TTM and STM mice,
respectively. The effect of these mutations on phosphorylation was investigated by
collecting retinas after various dark recovery times following a five minute exposure to
bright light, effectively bleaching rhodopsin. For wildtype mice, phosphorylation seemed
to function normally and quite rapidly. A full complement of residues were
Dark Recovery Time
64
phosphorylated by immediately and persisted still after 20 minutes in the dark (Fig. 3.3
A). For TTM mutants, all three serine residues were phosphorylated immediately and
persisted still after 20 minutes of dark incubation. However, over time the amount of
triply phosphorylated species seems to decrease, as did doubly phosphorylated species,
although at a slower rate in part due to previous triply phosphorylated species converting
into a double (Fig. 3.3 B). For STM mutants, it seems as though only weak mono- and di-
phosphorylated species were present, and only after until 20 minutes of dark recovery
time did all sites become phosphorylated (Fig. 3.3C). The phosphorylation rate for STM
mutants were much slower and weaker than TTM mutants.
Fig. 3.3 Comparison of phosphorylated species of rhodopsin in wildtype, TTM, and STM mutant
mice. Retinas were bleached under 5 minute constant bright illumination and extracted after
subsequent dark incubations times of 0, 2, 5, 10, and 20 minutes. While unbleached were
collected as well to serve as a control. (A) Wildtype mice exhibited strong phosphorylation of all
six sites and persisted until 20 minutes. (B) TTM mice exhibited strong phosphorylation at the
three available serine sites, while mono- and di-phosphorlayted species were most abundant. (C)
STM mutants also showed phosphorylation at all three available threonine sites, yet seemed
weaker and primarily contain mono- and di-phosphorylated species over this time course.
A B C
65
3.3 Discussion
Previous studies have shown that arrestin has an inhibitory effect on the
hydrolysis by RDH (Hofmann 1992). Moreover, arrestin has been implicated as a crucial
factor in the release of the retinoid depending on the number of phosphorylated sites
present on photoactivated rhodopsin, where the presence of five or more phosphates
result in faster release of the retinoid than when only four or less phosphates are present
(Vishnivetskiy 2007; Lamb & Pugh 2004). Given these findings, Saari et al. performed a
study to examine the effect of arrestin ablation on the rate of chromophore and rhodopsin
regeneration. In theory, the lack of arrestin will remove the imposed inhibition of RDH
and accelerate retinal hydrolysis and regeneration of chromophore. Analysis of visual
cycle retinoids was performed on wildtype and arrestin-/- retinas collected at various time
points during dark recovery incubation after exposure to a brief flash of light.
Interestingly, the absence of arrestin had little effect on the time course of rhodopsin
regeneration; 0.8%/min in arrestin-/- mice compared to 1%/min in wildtype mice. Yet, a
slight but insignificant change in 11-cis-retinal regeneration was observed as wildtype
mice regenerated chromophore 1.8 times faster than arrestin-/- mice (Saari 2000). These
results conflicted with the original hypothesis, most likely due to the experimental
procedure. An analysis of arrestin’s contribution must consider the relative amount of
arrestin present in the rod outer segment. Only 1-5% of arrestin resides in the ROS, while
the remaining population is located in the inner segment. Arrestin is known for its
massive translocation from the rod inner segment to outer segment upon exposure to
light. However, this feature is limited to bright and/or prolonged exposure to light. The
66
flash of light administered by Saari et al. may not have been strong enough or prolonged
enough to cause arrestin to enter the ROS. The results from our constant illumination
assay reveal delayed rhodopsin regeneration in arrestin knockout mice compared to
wildtype mice. This may serve as evidence supporting arrestin binding as a critical
component in efficient rhodopsin regeneration. However, the exact reason for slowing
needs to be investigated further. The notion that rhodopsin regeneration should have sped
up due to the lack of inhibition on RDH may have been wrongly assumed. The lack of
arrestin may have in fact released the inhibition on RDH and speed up hydrolysis of all-
trans-retinal, but the increase in hydrolysis may not necessarily translate into an increase
in chromophore production. It may be the case that increased hydrolysis might
overwhelm the downstream steps of the retinoid cycle and negatively impact
chromophore production or delivery. A possible analysis to consider is the amount of
accumulated retinyl esters produced by retinyl ester isomerohydrolase, which has been
identified previously as a possible rate-limiting step (Chen 2009). However, the focus on
RDH inhibition may only hold true under dim light conditions, while another mechanism
is affected in bright light conditions.
The amount of phosphorylation directly affects arrestin binding. In dim light, few
phosphates (~1-3) seem to be present which cause arrestin to be bound weakly and in fact
decrease the rate of hydrolysis. Whereas in bright conditions, four to six phosphates are
present which strongly bind arrestin and actually promote hydrolysis of all-trans-retinal.
Knocking out arrestin in both light conditions could essential have the same detrimental
effect on rhodopsin regeneration, albeit different mechanisms - improper downstream
67
retinoid steps or reducing the amount of retinoid hydrolysis and stability of MII,
respectively. How does specific phosphorylation sites determine the kinetics of rod
response given that arrestin binding and rhodopsin regeneration is critically dependent on
phosphorylation?
Altering the phosphorylation sites on rhodopsin’s C-terminus causes significant
changes in response kinetics. It has been long assumed that serine residues on the C-
terminus are the most important phosphorylation sites, since serines are rapidly
phosphorylated and are the most abundant phosphates after light exposure (Lee 2010;
Vishnivetskiy 2007; Kennedy 2001). However, solely having the three serines on the c-
terminus results in poor single photon response reproducibility and often have much
slower recovery times, which may be due to weak arrestin suppression (Brannock 1999).
In fact, arrestin binding seems to be dependent on the presence of at least three
phosphates while the full complement of six phosphates is ideal for reliable detect single
photons (Vishnivetskiy 2007; Doan 2006).
The TTM and STM mutant mice that Anthony Azevedo and Fred Rieke generated
allow for comparison of the role, if any, of the distinct residue. Given that serines seem to
be most the most prevalent phosphates, the TTM mutant would be expected to have
better response kinetics than STM mutant mice. However, Azevedo and Rieke collected
single photon responses from these mice and concluded that TTM mice have prolonged
responses and high variability, whereas STM mice have only slightly longer recovery
times but maintain near normal response kinetics (Fig. 3.4; Azevedo 2012 – manuscript
in prgoress). This seems to indicate that threonines are in fact critical in shutting off the
68
activity of rhodopsin as arrestin may have a higher affinity for them. This data examining
only three serine or three threonine residues corresponds well with previous data found in
vitro (Brannock 1999). The data as a whole counters the current ideology that rhodopsin
is shutoff due to preferential phosphorylation at serine residues, since STM mice are
capable of being phosphorylated at all three threonine sites and produce near wildtype
SPRs. The work presented here suggests there may be a preference for serines or
threonines by arrestin. In TTM mice, phosphorylation occurs at all three sites rapidly and
persists for as long as 20 minutes after 5 minutes of constant illumination. This
corresponds well with the notion that serines are phosphorylated rapidly and abundantly.
In contrast, STM mice seem to only be mono- and di-phosphorylated and at a much
slower rate.
69
Fig. 3.4 Response to single active wild-type and mutant rhodopsins. Panels A-C show individual
representative responses. Panels D-F collect responses from a representive cell, with identified
singles (~50) above and failures (~150) below: (A,D) wild-type; (B,E) SA; (C,F) TA.
(Figure adapted from Azevedo 2012 – manuscript in progress)
Previous studies have examined exact residue phosphorylation sequences, notably
with serine sites, but these have all been under dim, bright, or prolonged bleaches (Lee
2010; Kennedy 2001). There seems to be a lack of work presented on which exact
residues are critical during a single photon response. It has been reported that dim flashes
result in predominantly mono-phosphorylated rhodopsin species, so it might be the case
that under single photon conditions only one residue is phosphorylated. If an SPR lacks
70
multiple phosphorylated residues then arrestin may never actual bind, so why is the
presence of six sites necessary to produce reliable SPRs? And if in fact one phosphate is
produced, most likely Ser343 given that it is fastest to phosphorylate (Kennedy 2001),
this would weakly attract/bind arrestin so how does arrestin play a role during the shutoff
mechanism down at the single photon level? It has been shown that arrestin is necessary
for normal SPR kinetics, however it is important to note that arrestin knockout mice had
removed both full length arrestin (p48) and its splice variant (p44) (Burns 2006). It is
unclear whether the lack of p48 and/or p44 is responsible for the improper shutoff
kinetics; therefore it is important to consider p44 in this discussion as well.
The notion that arrestin may not even be involved in very dim conditions has been
proposed in a previous SPR study (Doan 2006) as well as a previous in vitro study
(Langlois 1996), yet an explanation of this idea was loose. I propose two possible
mechanisms that explain the process. First, given that the STM mice had only a slight
decrease in SPR kinetics, it may be possible that a single threonine phosphorylation is
enough to bind arrestin and quench activity. This may be possible since the analysis of
arrestin binding affinity for mono- and di-phosphorylated rhodopsin contained only
serines in the fraction (Vishnivetskiy 2007). Also, threonines are known to have a slow
rate of phosphorylation, which can explain the reduced recovery time of the STM SPR.
Second, a single photon response would most likely not induce translocation of arrestin
thus limiting the effect of an arrestin on the response if at all. Nevertheless, there exists 1-
5% of total arrestin in the ROS and arrestin must be considered. Given that, it may be
reasonable to assume that in fact the endogenous splice variant of arrestin, p44, is
71
responsible for shutoff at single photon levels. P44 may be the critical component for two
reasons. 1) All of p44 molecules reside in the outer segment, so they have rapid access to
a single photon response. 2) Most persuasively, p44 has a strong affinity for
unphosphorylated light activated rhodopsin species as well as phosphorylated
inactive/active species. P44 may regulate shutoff kinetics at single photon levels and may
only need one threonine phosphorylation to bind, thus making further phosphorylation
unnecessary. This may be what is happening in the STM data since mostly
monophosphorylated species are detected. A caveat to this idea is that if p44 does shutoff
unphosphorylated active rhodopsin species, TTM mutants shouldn’t have such slow
recovery times. However, this may indicate that p44 binding is dependent upon the exact
arrangement of C-terminus residues (specifically threonines) and not necessarily the
phosphorylation of them. Full length arrestin (p48) has been also implicated as being able
to shutoff SPRs when transgenically expressed in Arrestin
-/-
mice (Burns 2006).
However, in that study p48 expression was two-fold more than in wildtype outer
segments. This may indicate a law of mass action effect, which would increase the
probability of binding of arrestin to unphosphorylated active rhodopsin species due to
simply being more concentrated. It is interesting to note that responses to a family of dim
flashes seemed to have faster recover times in arrestin knockout mice which solely
expressed p44 (termed m44 in the study) compared to the sole expression of p48. This
aspect was not addressed by Burns et al. and may prove to be important since in the
transgenic p48 and p44 arrestin knockout mice, the ratio of p48 to p44 expression in the
ROS was 16:1 (~6%). Despite P44 being far less abundant in ROS, dim flashes responses
72
were faster to recover. This may further support the idea that p44 shuts off SPR activity
and unphosphorylated active rhodopsin species under dim light conditions. P44’s
inactivation ability broke down under bright light conditions.
This data and discussion provide the groundwork for future experiments to be
performed to gain further insight into the role of rhodopsin regeneration. For instance,
functional analysis via electroretinograms (ERGs) should be performed to observe the
physiological relevance of the abundance of phosphorylated species in arrestin knockout
mice. After rhodopsin regeneration, there should still be phosphorylated species of
inactive rhodopsin present. These species of rhodopsin have lower catalytic activity,
which would be detectable by ERG analysis. Moreover, it would be interesting to
examine the mutants that have the serines and threonines interchanged, as well as those
mutants that have a full complement of serines or threonines only. Lastly, the role of p48
and p44 in determining SPR kinetics should be further investigated via transgenic
expression of either protein in the arrestin knockout background, in conjunction with
rhodopsin C-terminus mutants.
3.4 Materials and Methods
3.4.1 Light Exposure
All mice were dark-adapted overnight. The light exposure used for the
phosphorylation and regeneration assays comparing wildtype and arrestin
-/-
mice were
performed by first anesthetizing the mice via intraperitoneal injection of a solution
containing xylazine and ketamine (10 µg/g body weight and 100 µg/g body weight,
73
respectively) diluted in PBS. Pupils were dilated with 0.5% tropicamide and 2.5%
phenylephrine hydrochloride solutions and then exposed to constant bright illumination
for 5-10 minutes. Afterwards, mice were transferred back into the dark and retinas were
extracted after desired recovery time (0 and 15 minutes for phosphorylation assays, and
0, 15, 30, 45, 60, 90, and 120 minutes for rhodopsin regeneration assays). Retinal
extraction was performed under infrared illumination on mice sacrificed by cervical
dislocation. Extracted retinas were subsequently placed in an eppendorf tube at which
point the process of either IEF or rhodopsin quantification was carried out. The light
exposure used for the phosphorylation assay comparing wildtype, TTM, and STM mutant
mice were performed on extracted retinas, after mice were sacrificed using cervical
dislocation and both retinas were subsequently removed. For each mouse, one retina was
used as a dark sample by immediately placing it in an eppendorf tube, frozen with dry ice
slurry, then wrapped in foil to prevent light exposure. The other retina was used as a
bleached sample by placing it in a dish filled with AMES medium and exposing for 5
minutes to a ~550nm-filtered light source. This exposure time necessary for a desired
fractional bleach (F) of the visual pigment was calculated using
€
F =1−e
(−I∗P∗t)
, where I
the intensity of the bleaching light, P is the photosensitivity of the visual pigment
(6x10
−9
), and t is the duration of exposure in seconds. The bleached retina was then
allowed to incubate in the dark for various recovery times (0, 2, 5, 10, and 20 minutes),
then immediately frozen with dry ice slurry and wrapped in foil. All retinal extracts were
stored at -80
o
C until thawed for IEF processing.
74
3.4.2 Rhodopsin Quantification
Retinas were dissected and collected under infrared light and placed into
eppendorf tubes containing 200µl of solubilization buffer (PBS, 1% DM, 1% Protease
Inhibitor Cocktail) and incubated for two hours. Samples were centrifuged at 4000rpm
for five minutes and 100µl of the supernatant was removed and placed into a cuvette for
UV-visible spectroscopy. Spectral analysis of samples was performed in the dark using a
wavelength scan spanning 300-600nm. Absorbance readings were recorded at 500nm
(‘dark’). Samples were then bleached under bright illumination for two minutes, after
which sample absorbance was recorded at 500nm (‘bleached’). Pigment levels were
quantified by taking the absorbance difference between ‘dark’ and ‘bleached’ states, and
dividing by the extinction coefficient, ε = 40,000 M
-1
cm
-1
.
3.4.3 Isoelectric Focusing (IEF)
Phosphorylation species detection was performed using Isoelectric Focusing
(IEF). All retinas were homogenized using a polytron for 15 seconds. The samples were
centrifuged at 12750rpm for 15 minutes, supernatants were discarded and pellets were
washed once in 10mM Hepes buffer. Pellets were resuspended in regeneration buffer
containing 9-cis retinal and nutated overnight in 4°C to allow for visual pigment
regeneration. Samples were centrifuged, supernatants were discarded, and pellets washed
as mentioned. Pellets were then solubilized overnight on a nutator in 4°C. The following
day, samples were isoelectrically focused on an acrylamide gel using a Pharmacia Flat
75
Bed Apparatus FBE300 with a cooling system set at 10°C. The gel was blotted with
mouse anti-rhodopsin (4D2) and phosphorylation was detected by chemiluminescence.
76
Chapter 4
Cyclic Nucleotide Channel Blockers Potential Photoreceptor Protection
4.1 Introduction
Retinitis Pigmentosa (RP) is a congenital disease leading to the progressive
degeneration of the retina. RP is hallmarked by rod photoreceptor cell death followed by
subsequent loss of cone photoreceptor cells. Many genetic mutations have been identified
as causes of RP and most notably these mutations affect specific proteins integral to the
phototransduction cascade, a G-protein coupled receptor signaling pathway that converts
photons of light into a chemical signal. Specifically, cyclic nucleotide-gated (CNG) ion
channels are responsible for this conversion by regulating the membrane potential of the
photoreceptor. The opening and closing of CNG channels in photoreceptors is dependent
on cytoplasmic cGMP concentration. Binding of cGMP opens the channel and allows an
influx of Ca
2+
and Na
+
ions. In the dark-adapted rod, a large concentration of cGMP is
present, which keeps the ion channels open and maintains a depolarized state resulting in
release of glutamate at the synaptic terminal. Upon light absorption by rhodopsin, the
phototransduction cascade is initiated, causing downstream activation of transducin and
the cGMP-phosphodiesterase (PDE). Activated PDE then hydrolyzes cGMP into GMP,
which lowers the intracellular concentration of cGMP. The decrease in cGMP
concentration results in closure of the Na
+
/Ca
2+
channels and a reduction of Na
+
/Ca
2+
ions
inside the cell. Thus the photoreceptor becomes hyperpolarized, leading to closure of
voltage-gated Ca
2+
channels and a decrease in glutamate release and transfer of the light
77
signal. Some forms of RP have been linked to genetic mutations affecting cGMP levels
inside the cell, which alters normal CNG channel activity and leads to cell death.
A readily used mouse model of RP is the retinal degeneration 1 (rd1) mouse, which
contains a mutation in the rod photoreceptor cGMP phosphodiesterase β-subunit
(PDE6b) (Farber 1995). A defect in PDE6b function results in elevated cGMP levels
inside the cell, roughly 10-fold more than normal (Paquet-Durand 2011; Sancho-Pelluz
2008). It has been proposed that the accumulation of cGMP causes an overactive CNG
channel leading to an excessive concentration of Ca
2+
that is toxic to the photoreceptor
cell. A rescuing effect has been achieved in rd1 mice when crossbred with mice deficient
in rod CNG channels (Cngb1
-/-
), thus eliminating the increase in Ca
2+
concentration in the
presence of elevated cGMP levels (Paquet-Durand 2011). Other studies have also shown
that a decrease in rod CNG channel activity in rd1 mice can significantly improve the
progression of RP (Vallazza-Deschamps 2005). This implies that clinically the rod CNG
channels may serve as a more effective target for therapy than a blanket blockade of Ca
2+
channels, which would disrupt many other signaling pathways.
Currently, there are several CNG channel antagonists such as l-cis-diltiazem and
tetracaine [2-(dimethylamino)ethyl 4-(butylamino)benzoate], yet their blocking of and
affinity for rod CNG channels are subpar as evidenced by patch clamp recordings. In
collaboration with Jeffrey Karpen, we investigated the effect of newly modified
tetracaines in rescuing rd1 and rd10 mouse models of RP through the analysis of retinal
morphology. These tetracaine derivatives have increased hydrolysis-resistance and a
higher affinity for and blockade of the CNG channels (Karpen 2011). Both in vitro and in
78
vivo studies were performed using retinal explants as well as intraperitoneal and
subretinal injections, respectively. We present a promising look into tetracaine
derivatives that exhibit a reduction in retinal degeneration in rd1 and rd10 mice.
4.2 Results
To curb retinal degeneration, an ideal rod CNG channel antagonist needs to
effectively block the channel as well as have a strong affinity for the channel. The most
commonly used antagonist, l-cis-diltiazem, has poor specificity for the rod CNG channel
and weakly blocks the channel. Another candidate for an effective CNG blocker is
Tetracaine, a local anesthetic that strongly blocks CNG channels yet exhibits weak
specificity for CNG channels, preferring the closed, inactive state. The ability to use
Tetracaine as a scaffold for creating new derivatives has paved the way for new
compounds that enhance the blocking effect and increase the affinity for the CNG
channels.
Using tetracaine as the framework for chemical composition, many derivatives
have been generated in an attempt to improve CNG channel block and affinity. Alkyl
substitutions followed by esterification or amidation led to Compound 8 (2-
(dimethylamino)ethyl 4-(butylamino)benzoate) and Compound 9 (4-(Octylamino)-N-(2-
(dimethylamino)ethyl)- benzothioamide). Compounds 8 and 9 were shown to have a
higher affinity for CNG channels than tetracaine itself, roughly an 8-fold increase. In
addition, these derivatives exhibited a strong resistance to hydrolysis by serum
cholinesterase, the most abundant esterase in ocular tissues. The substitution of thioamide
79
linkages in place of ester linakages vastly improved the ability for Compounds 8 and 9 to
block CNG channels (Karpen 2011). However, analysis of dissociation constants under
varied cGMP concentrations revealed that most compounds preferred the closed state of
the CNG channel, when cGMP levels are diminished. This property is suboptimal
because the retinal degeneration is largely due to open CNG channels. However, Karpen
et al. have constructed a tetracaine derivative on which a halogen substituent (2-Cl
modification) is placed the aromatic ring in the octyl tail. This compound, termed
Compound 2 in this study, has proved to be 40-50 fold more effective in binding CNG
channels, and may make up for the weaker affinity for open, active states of the channel
(Kirk 2011).
In this analysis, we compared four different compounds: tetracaine (Compound
1), 2-chloro octylcaine (Compound 2), thioamide tetracaine (Compound 8), and
thioamide octylcaine (Compound 9) and examined their effect on preventing retinal
degeneration through retinal morphological analysis. Three different assays were
performed; 1) in vitro rd1 retinal explants were bathed in media containing the drug, 2)
intraperitoneal injections of the drug were administered in vivo to rd10 mice, and 3)
subretinal injections locally delivered the drug in vivo to rd10 mice.
4.2.1 Varying Degrees of Rescuing Effect in Retinal Explants
The retinas of Wildtype (WT) and rd1 mice were explanted onto transwell plates
at age postnatal day12 (P12). Retinal explants were cultured for one week with media
containing one of four Tetracaine derivatives as well as a vehicle control. Explants were
80
then embedded in epoxy resin, sectioned at one micron thick, and examined for any
morphological changes. WT retinal explants exhibited normal morphology with outer
nuclear layer thickness (a measure of retinal health) consisting of ~8-9 rows of
photoreceptor cell bodies. As expected, retinal explants of rd1 mice with vehicle control
exhibited severe retinal degeneration with an ONL comprised of ~0-1 rows of
photoreceptor cell bodies. Compound 2 had a slight rescuing effect resulting in ~2-3 rows
of cells. Tetracaine also had a slight rescuing effect with ~3-4 rows of cells spanning the
ONL. Compound 8 seemed to have a slight rescuing effect and presence of pigment
granules with an ONL consisting of ~1-2 rows. Compound 9 had a noticeable rescuing
effect as well with ~4-5 rows of cells maintained throughout the ONL (Fig. 4.1).
Fig. 4.1 Retinal morphology of retinal explants treated with tetracaine derivatives. Complete loss of
photoreceptor cells and RPE layer in rd1 explants. Treatment with compound 8 preserved RPE cells and a
few rows of photoreceptor cells. Compound 2 was not as effective as seen by significant loss of
photoreceptor nuclei. Compound 9 revealed very strong rescue and improved upon the rescue seen in
tetracaine treated explants.
Tetracaine
81
4.2.2 Intraperitoneal Injections Less Effective
To examine whether the in vitro results are preserved in vivo, experiments were
performed on rd10 mice, in which the compounds were administered via intraperitoneal
injections. rd10 mice were injected with a vehicle control and the Compound 9
derivative, since it had the most significant rescuing effect in our retinal explant assay.
Retinas were analyzed for any morphological changes again using outer nuclear layer
thickness as a measure for retinal health. Retinas of rd10 mice treated with a vehicle
control had marked degeneration, consisting of ~0-1 rows of cells throughout most of the
ONL, while Compound 9 appeared to have a slight rescuing effect, ~1-5 row ONL
composition (Fig. 4.2). Our results were a bit inconsistent, which suggest that
intraperitoneal injections may be a poor delivery method of targeting the retina.
Fig. 4.2 Comparison of retinal morphology of rd10 mice injected with vehicle control versus
compound 9. Rd10 mice exhibited severe retinal degeneration without any drug treatment.
Compound 9 treated rd10 mice showed a rescue in the periphery but failed to preserve more
central photoreceptors. Insets show magnified view of retinal morphology.
rd10 + Vehicle Control rd10 + Compound 9
82
4.2.3 Subretinal Injections Exhibit Varying Degrees of Rescue
To improve targeting of the drug to the retina, we employed a subretinal injection
assay, in which the compound was delivered directly into the subretinal space. Retinas of
rd10 mice were examined after administering compounds 2 & 8 and compared to
uninjected rd10 retinas. Uninjected retinas of rd10 mice exhibited severe degeneration
with ~0-1 rows of cells along the ONL. Compound 8 showed a noticeable rescuing effect
showing ~2-5 rows of cells throughout the ONL. Compound 2 had only a slight rescuing
effect with about ~1-3 rows forming the ONL. These in vivo assays correlate well with
the in vitro experiments and proved to be a more effective delivery method for the drugs
during our assay analysis (Fig. 4.3).
Fig. 4.3 Retinal morphology of subretinal
injections of tetracaine derivatives at p12,
harvested after 5 days. Uninjected retinas
showed significant cell loss with 0-1 rows of
nuclei. Compound 2 showed a slight increase
in rows of nuclei (2-3). Compound 8 exhibited
a slight rescuing effecting, preserving 2-5
rows of nuclei. Insets show magnified view of
retinal morphology.
83
4.3 Discussion
Retinitis Pigmentosa (RP) is a progressive degenerative disease that causes loss of
rod photoreceptor cells. A well-documented mouse model of RP is the rd1 mouse that
contains a genetic mutation in the rod cGMP phosphodiesterase protein, which renders
the mouse incapable of hydrolyzing cGMP into GMP in the photoreceptor cell. This leads
to increased levels of cytosolic levels of cGMP and in turn keeps cGMP-gated ion
channels open. This activity causes in excessive accumulation of Ca
2+
inside the cell,
leading to cell death. The increase in free Ca
2+
can trigger an apoptotic pathway in rod
photoreceptors through activation of an endonuclease, promoting DNA fragmentation
and apoptosis. Increased Ca
2+
can also trigger cell death associated with calpain
activation and increased oxidative stress. Furthermore, increased cGMP concentrations
may induce one of many cGMP-dependent protein kinase (PKG) mediated cell death
pathways resulting from over-activation of PKG (Paquet-Durand 2010, Guo 2006,
Hofmann 2006). These known issues provide an opportunity to focus on therapeutic
methods to block CNG channels in the rod photoreceptor to prevent the toxic Ca
2+
concentration from developing. Common CNG channel blockers seem to fall short in
preventing cell death in large part due to low specificity and ineffective blockade.
Derivatives of a well-known CNG channel antagonist, Tetracaine, have been developed
that proved to have a higher affinity for CNG channels as well as improved channel
blockade. These derivatives may indeed lead to effective forms of therapeutic methods to
combat some forms of RP.
84
Here we have examined, both in vitro and in vivo, the efficacy of such derivatives
in preventing retinal degeneration. The results suggest that curbing retinal degeneration
can be achieved, albeit with varying effects. Our in vitro assay showed a marked
improvement in retinal morphology when Tetracaine derivatives were administered.
Markedly, Compounds 2 and 9 seem to have the greatest rescuing effect. Our in vivo
assays suggest that Compound 2 only had a slight effect, while Compound 8 had the
greatest effect. The discrepancy between our assays may be a result of our experimental
techniques as well as the exact concentration of the compounds effectively targeting the
CNG channels. A more concentrated delivery of Compound 8 has shown a more reliable
rescuing effect (data not shown).
It has yet to be determined whether these derivatives have specificity for rod CNG
channels over cone CNG channels. However, previous studies have shown that the
derivatives have a higher affinity for rod CNG channels than for cone CNG channels.
Tetracaine’s affinity for homomeric rod CNG channels (CNGA1 subunits) is
K
1/2
,tetracaine = 4.6 µM ± 0.70 µM, while its affinity for homomeric cone CNG channels
(CNGA3 subunits) is K
1/2
,tetracaine = 55.2 µM ± 5.3 µM, a roughly 10-fold difference
(Liu 2005). Although this seems like an advantage, the lack of direct specificity for the
rod CNG channel cannot be over looked. Also, these affinity measurements were
recorded using homomeric CNG channels, yet natural CNG channels are heteromeric in
rods and cones, 3CNGA1:1CNGB1 and 2CNGA3: 2CNGB3 stoichiometry, respectively
(Peng 2004; Weitz 2002; Zheng 2002; Zhong 2002). But presumably the affinity for the
85
rod channel may be maintained and thus useful to consider these tetracaine derivatives as
possible treatments.
While our in vitro and in vivo assays seem to conflict on which compounds seem to
be most effective, there is little question that an improved tetracaine derivative can
effectively rescue and slow down the retinal degeneration associated with RP. Karpen et
al. recently produced a tetracaine derivative that combines the halogen substitute with the
thioamide linkage, 2-chloro thioamide octylcaine. The combination of both
enhancements should prove to be more effective in rescuing degeneration, as the
hydrolysis resistance and blockade efficacy should be optimal. Our results have laid the
groundwork for further and more in depth in vivo studies looking at the viability of these
CNG blockers as an effective therapeutic strategy. A logically next step to take is to
examine the functional recovery after drug delivery, if any, using electroretinogram
analysis.
4.4 Materials and Methods
4.4.1 Retinal Explant Culture
Retinas were extracted from rd1 mice at age postnatal 11 and 12 and placed onto
Costar Transwell permeable membrane plates. Media comprised of DMEM high glucose
with 10% fetal bovine serum was added (1mL) to each well containing an explant.
Tetracaine derivatives dissolved in pure ethanol at a concentration of 100mM were
further dissolved into 0.9% NaCl (saline) to obtain a final concentration of 1mM. The
drugs were added to the medium as well to obtain final concentrations of 1.25µM,
86
2.5µM, or 5µM, while some wells contained the vehicle alone to serve as a control.
Retinal explants were cultured for one week, during which one media change after 2-3
days was performed. Retinal explants were then carefully removed from the wells for
retinal morphometry processing.
4.4.2 Intraperitoneal Injections
Intraperitoneal injections were performed on rd10 mice at postnatal day 18.
Tetracaine derivatives dissolved in pure ethanol at a concentration of 100mM were
further dissolved into 0.9% NaCl (saline) to obtain a final concentration of 1mM. Mice
were injected with 0.1ml of the vehicle control or drug daily for one week, after which,
mice were sacrificed by cervical dislocation and eyes were enucleated for retinal
morphometry processing.
4.4.3 Subretinal Injections
Subretinal injections were performed on rd1 mice. Tetracaine derivatives
dissolved in pure ethanol at a concentration of 100mM were further dissolved into 0.9%
NaCl (saline) to obtain a final concentration of 30µM. Mice were injected into the
subretinal space with 0.5µl of the vehicle control or drug to obtain a working
concentration of 5µM (assuming 3µl of vitreous volume). Injections were performed at
postnatal day 12 and allowed to take effect for 5 days, at which point mice were
sacrificed by cervical dislocation and eyes were enucleated for retinal morphometry
processing.
87
4.4.4 Retinal Morphometry
For the in vivo experiments, eyecups were collected as follows: the superior pole of
the cornea was cauterized for orientation before enucleation of the eyeballs. The cornea
and lens were removed, leaving an intact eyecup, which was placed in a glass vial used
throughout the preparation. For the retinal explants, fixation and dehydration steps were
performed in the cell culture dish and transferred to glass vials thereafter for the
remainder of the preparation. Eyecups and retinal explants were fixed overnight at 4
o
C in
½ Karnovsky buffer (2.5% glutaraldehyde, 2.0% paraformaldehyde, in 0.1M cacodylate
buffer [pH 7.2]). The eyecups and explants were washed with 0.1 M cacodylate buffer
before being treated with 1% osmium tetroxide diluted in 0.1 M cacodylate buffer.
Eyecups and explants were washed again with 0.1 M cacodylate buffer, dehydrated,
embedded in epoxy resin and cut into one mm thick sections and stained with Richardson
stain (1% toluidine blue, 1% azure II, 1% borax). For eyecups, thickness of the outer
nuclear layer (ONL) was analyzed, whereas for retinal explants, sections were grossly
observed for morphological changes.
4.4.5 Tetracaine Derivative Formation
The exact details of the composition of tetracaine derivative formation were as
described in Karpen et al.’s 2011 paper. General principles of the process are described
here. Using tetracaine as the framework for the derivatives, alkyl substituents were added
to the amino end to obtain alkylated benzoic acids with butyl (Compound 3) and octyl
(Compound 4) tails. Following activation at the carboxylic acid with 1,1’-
88
carbonyldiimidazole, compounds were subsequently esterified or amidated with 2-
(dimethylamino)ethanol (Compound 5) or N0,N0-dimethy-lethane-1,2-diamine
(Compounds 6 and 7), respectively. Compounds 6 and 7 were further treating with
Lawesson’s reagent to yield target thioamide tetracaine and thioamide octylcaine,
compounds 8 and 9, respectively. Compounds were later improved by halogen-
substitution on the aromatic ring of octyl tail derivatives to obtain 2-chloro octylcaine,
referred to here as Compound 2.
89
Chapter 5
Concluding Remarks
The overarching aim of this dissertation is to build upon previously laid
groundwork in the field of phototransduction, specifically addressing the molecular
interactions and properties underlying retinal impairments caused by genetic mutations.
Alterations in the molecular makeup of the phototransduction cascade have significant
effects. Presented here are insights into three such alterations: an examination of a
rhodopsin mutation that leads to autosomal dominant retinitis pigmentosa, a look into the
functional consequences of arrestin binding given rhodopsin C-terminal changes, and a
spotlight on a potential therapeutic strategy for cyclic nucleotide-gated channel mutation
that causes retinal degeneration. Though the phototransduction cascade has evolved to
highly streamlined and organized, this dissertation highlights the vulnerability of the
system as small perturbations have grand repercussions.
Autosomal dominant retinitis pigmentosa (ADRP) is a progressive retinal
degenerative disease hallmarked by rhodopsin genetic mutations. Many rhodopsin
mutations have been classified and their mechanisms of degeneration elucidated. Yet the
class of rhodopsin mutations leading to a constitutively active form has remained
shrouded in controversy. It has been suggested that over-activation of the cascade is the
leading cause of the degeneration. However, the information put forth in this dissertation
refutes this idea, supported by the analysis of the K296E rhodopsin mutation leading to
ADRP. Novel insight into the mechanism of photoreceptor cell death has been presented.
90
Specifically, the constitutively active rhodopsin species is hyperphosphorylated and
rapidly forms a tight complex with arrestin1. This complex then accumulates throughout
the rod photoreceptor and hypothesized to be endocytosed, disrupting normal endocytic
activity. This has been concluded given the data presented on the interaction with K296E
and p44, an arrestin1 splice variant that lacks the ability to become endocytosed due to a
lack of an AP-2 binding domain. The substitution of p44 for arrestin1 rescued the retina
from degeneration and restored normal visual function. This proved that the endocytosis
of the K296E/arrestin1 complex is the trigger for cell death. Despite this finding, there
still remains room for more elucidation regarding the exact steps of the cell death
mechanism, such as what downstream signaling proteins are involved. In addition, the
electroretinogram of the K296E/p44/Arr1
-/-
mice showed increased b-wave amplitude,
which implicates a change in bipolar cell activity. Therefore, a closer examination of this
phenotype should be performed using suction electrode analysis.
It has been a longstanding question as to how exactly rhodopsin is phosphorylated
and what the implications are for subsequent arrestin binding and rhodopsin regeneration.
Murine rhodopsin has six phosphorylation sites on its C-terminal tail comprised of three
serines and three threonines. It has been proposed that the number of phosphorylated sites
control the affinity for arrestin binding and the strength of the bond. It is believed that
three or four phosphorylated sites are necessary for arrestin binding and that all six sites
need to be available to produce reliable and reproducible single photon responses (SPR).
However, the notion that the number of sites is the most significant factor breaks down
when looking at the SPR kinetics of rhodopsin C-terminal mutant species that examine
91
the presence of only three serines (TTM) or only three threonines (STM). In collaboration
with Anthony Azevedo and Fred Rieke, we have shed some light on this matter. They
found that STM mice have a higher SPR variability, whereas TTM mice have near wild-
type SPRs. This finding suggests that the number of sites is not the defining component
of SPR kinetics, but rather that the identity of the individual phosphorylation site is most
important. This is especially significant given that it has long been held that serines are
phosphorylated first and rapidly, while threonine phosphorylation follows and is much
slower. However, it seems that the threonines are rather important for reliable SPRs. We
took this notion a step further and analyzed whether there is an issue with
phosphorylation when some sites are altered. The isoelectric focusing (IEF) results
presented here reveal that both the TTM and STM mutants can be fully phosphorylated.
This suggests that there may be preferential arrestin binding. It remains to be clarified
whether the individual amino acid residues or their specific position along the C-terminus
are significant in binding arrestin. Specifically, an examination of all six sites being
serine or threonine only, or switching the positions of the serines and threonines in both a
full complement of sites present as well as in the scenario of just three sites present.
Previous in vitro studies by Brannock et al. have exactly done this and their conclusion
suggest that the individual site is important and not the position. The in vivo work
remains to be performed, such as SPR analysis, ERG analysis, and IEF analysis.
Furthermore, it would be interesting to see the effect of p44 on these mutant rhodopsins
as it has been suggested in this dissertation that in fact, p44 is important for dim light
92
quenching of rhodopsin activity and the composition of the phosphorylation sites on the
C-terminus may play an important role in its binding affinity.
Arrestin binding is also very important in the rate of rhodopsin regeneration in
dark adaptation. Previous studies have attempted to examine the effect of removing
arrestin, yet they failed to discern a difference in rhodopsin regeneration. A glaring issue
with those studies is the use of dim light conditions to assay the effect. Arrestin is
primarily located in the inner segment and undergoes massive translocation to the outer
segment (where rhodopsin is located) after bright light exposure. So the lack of a
phenotype may be a result of the light conditions, since at dim light conditions it is
suggested that p44 binding to rhodopsin is the main desensitization step. In this
dissertation, the effect of removing arrestin was examined under bright light conditions.
The results suggest that the rate of regeneration remains the same, yet there is a delay in
the initiation of the process when compared to wildtype mice. Therefore, under bright
light conditions arrestin seems to be necessary for normal rhodopsin regeneration. Future
work remains to be performed to study rhodopsin regeneration between wildtype and
arrestin
-/-
mice under controlled dim light conditions.
The final genetic mutation analysis in this dissertation involves the rd1 mouse
model of retinal degeneration, which contains a mutation in the cGMP
phosphodiesterase. The effect of the mutation is an inability to hydrolyze cGMP into
GMP inside the rod photoreceptor cell, which causes an accumulation of cGMP thus
leaving the cyclic nucleotide-gated (CNG) channels open. The prolonged opening of
CNG channels causes an increase of intracellular concentration of Ca
2+
, which is toxic to
93
the photoreceptor and leads to retinal degeneration. Previous attempts to combat this
event utilized CNG channel blockers to curb the influx of Ca
2+
and prevent toxicity. The
most common CNG blockers used have been l-cis-diltiazem and tetracaine, yet they
exhibit poor affinity and blockade of rod CNG channels. In collaboration with Karpen et
al., we set out to test the use of novel CNG blockers derived from tetracaine, shown to
have a better affinity for and blockade of rod CNG channels in vitro. Here, we have
successfully shown that in vivo administration of the CNG blocker proves to rescue
retinal degeneration, albeit not in a dramatic fashion. Several tetracaine derivatives were
analyzed, such as having thioamide linkages as well as alkylated substitutes. The
thioamide linkage derivatives proved to be most effective in rescuing retinal degeneration
as the retinal morphology was markedly improved. The sole specificity for rod CNG
channels remains to be isolated and the introduction of halogen substitutes on the
aromatic ring show promise in resisting hydrolysis. These improvements need to be
tested in future studies, yet we have presented the feasibility for a potential therapeutic
strategy to counter the effects of this genetic mutation.
Generally speaking, this work has expanded the knowledge base of the
phototransduction cascade and sheds light onto the interactions between the molecules
involved in retinal degeneration caused by genetic mutations. There are many mutations
that lead to similar degenerative phenotypes addressed in this dissertation and the insight
presented may prove to be applicable across other signaling cascades. For instance, G-
protein coupled receptors have very similar properties yet are functional very different.
The defining difference may involve the C-terminus, specifically in reference to the
94
phosphorylation sites and activity. The phosphorylation sites may act as an identifier
amongst each other and thus handled accordingly by the rhodopsin kinase family, which
is far less varied than the GPCRs. Nevertheless, I believe great strides will be made in the
following years, both in future experiments that will follow the ones presented here as
well as in the scientific community as a whole. More and more information is being
gathered and at a faster rate than ever before and soon the focus will be shifted to direct
therapeutic strategies. Gene therapies may render many of these diseases a problem of the
past. I am proud of the work that I have been able to take part in and the findings that I
have helped to discover. I hope that my work here can keep the torch lit in the efforts to
treat retinal disease.
95
Bibliography
Alloway, P. G., Howard, L., & Dolph, P. J. (2000). The formation of stable rhodopsin-
arrestin complexes induces apoptosis and photoreceptor cell degeneration. Neuron,
28(1), 129–138.
Arshavsky, V. Y. & Burns, M. E. (2012). Photoreceptor Signaling: Supporting Vision
across a Wide Range of Light Intensities. Journal of Biological Chemistry, 287(3),
1620–1626.
Arshavsky, V. Y. (2002). Rhodopsin phosphorylation: from terminating single photon
responses to photoreceptor dark adaptation. Trends in Neuroscience, 25(3), 124–126.
Bownds, D. (1967). Site of attachment of retinal in rhodopsin. Nature, 216(5121), 1178–
1181.
Bovolenta, P. & Cisneros, E. (2009). Retinitis pigmentosa: cone photoreceptors starving
to death. Nature Neuroscience 12(1), 5-6.
Brannock, M. T., Weng, K., & Robinson, P. R. (1999). Rhodopsin’s Carboxl-Terminal
Threonines are Required for Wild-Type Arrestin-Mediated Quench of Transducin
Activation In Vitro. Biochemistry, 38(12), 3770-3777.
Burns, M. E. (2006). Deactivation of Phosphorylated and Nonphosphorylated Rhodopsin
by Arrestin Splice Variants. Journal of Neuroscience, 26(3), 1036–1044.
Burtey, A., Schmid, E. M., Ford, M. G. J., Rappoport, J. Z., Scott, M. G. H., Marullo, S.,
Simon, S. M., McMahon, H. T., & Benmerah, A. (2007). The Conserved Isoleucine-
Valine-Phenylalanine Motif Couples Activation State and Endocytic Functions of β-
Arrestins. Traffic, 8(7), 914–931.
Castro-Obregón, S., Rao, R. V., del Rio, G., Chen, S. F., Poksay, K. S., Rabizadeh, S.,
Vesce, S., Zhang, X., Swanson, R. A., & Bredesen, D. E. (2004). Alternative, non-
apoptotic programmed cell death: mediation by arrestin2, ERK2, and Nur77. Journal
of Biological Chemistry 279(17), 17543-17553.
Chen, C. K., Burns, M. E., Spencer, M., Niemi, G. A., Chen, J., Hurley, J. B., Baylor, D.
A., Simon, M. I. (1999) Abnormal photoresponses and light-induced apoptosis in
rods lacking rhodopsin kinase. Proceedings of the National Academy of Sciences of
the USA 96(7), 3718-3722.
96
Chen, J., Shi, G., Concepcion, F. A., Xie, G., Oprian, D., & Chen, J. (2006). Stable
Rhodopsin/Arrestin Complex Leads to Retinal Degeneration in a Transgenic Mouse
Model of Autosomal Dominant Retinitis Pigmentosa. Journal of Neuroscience,
26(46), 11929–11937.
Chinchore, Y., Mitra, A., & Dolph, P. J. (2009). Accumulation of Rhodopsin in Late
Endosomes Triggers Photoreceptor Cell Degeneration. (N. Perrimon, Ed.) PLoS
Genetics, 5(2), e1000377.
Chuang, J. Z. (2004). Structural and functional impairment of endocytic pathways by
retinitis pigmentosa mutant rhodopsin-arrestin complexes. Journal of Clinical
Investigation, 114(1), 131–140.
Concepcion, F., & Chen, J. (2010). Q344ter Mutation Causes Mislocalization of
Rhodopsin Molecules That Are Catalytically Active: A Mouse Model of Q344ter-
Induced Retinal Degeneration. (A. Lewin, Ed.) PLoS ONE, 5(6), e10904.
Cotter, J. R. (1989). Endocytosis of cationized ferritin by rat photoreceptors.
Neuroscience Letters, 106(1-2), 65–70.
Doan, T., Mendez, A., Detwiler, P. B., Chen, J., & Rieke, F. (2006). Multiple
Phosphorylation Sites Confer Reproducibility of the Rod’s Single-Photon Responses.
Science, 330(5786), 530-533.
Dolph, P. J., Ranganathan, R., Colley, N. J., Hardy, R. W., Socolich, M., & Zuker, C. S.
(1993). Arrestin function in inactivation of G protein-coupled receptor rhodopsin in
vivo. Science, 260(5116), 1910–1916.
Fain, G. L., Matthews, H. R., Corn wall, M. C., & Koutalos, Y. (2001). Adaptation in
Vertebrate Photoreceptors. Physiological Reviews, 81(1), 117-151.
Farber, D.B. (1995). From mice to men: the cyclic GMP phosphodiesterase gene in
vision and disease. The Proctor Lecture. Investigative Ophthalmology & Visual
Science 36(2), 263-275.
Fu, Y. & Yau, K. (2007). Phototransduction in mouserods and cones. Pflügers Archiv –
European Journal of Physiology 454(5), 805-819.
Guo, L. & Ruoho, A. E. (2006). Inhibition of cGMP-dependent protein kinase by the
regulatory sununit of photoreceptor cGmp phosphodiesterase. Neuroscience Letters
401(3), 252-255.
97
Gurevich, E. V., Benovic, J. L., & Gurevich, V. V. (2004). Arrestin2 expression
selectively increases during neural differentiation. Journal of Neurochemistry, 91(6),
1404–1416.
Gurevich, V. (2004). The molecular acrobatics of arrestin activation. Trends in
Pharmacological Sciences, 25(2), 105–111.
Gurevich, V. V., Dion, S. B., Onorato, J. J., Ptasienski, J., Kim, C. M., Sterne-Marr, R.,
Hosey, M. M., & Benovic, J. L. (1995). Arrestin interactions with G protein-coupled
receptors. The Journal of biological chemistry, 270(2), 720. ASBMB.
Harwood, S. M., Yaqoob, M. M., & Allen, D. A. (2005). Caspase and calpain function in
cell death: bridging the gap between apoptosis and necrosis. Annals of Clinical
Biochemistry 42(6), 415-431.
Hofmann, F., Feil, R., Kleppisch, T., & Schlossmann, J. (2006). Function of cGMP
-dependent protein kinases as revelaed by gene deletion. Physiological Reviews
86(1), 1-23.
Hofmann, K. P., Pulvermüller, A., Buczylko, J., Van Hooser, P., & Palczewski, K.
(1992). The Role of Arrestin and Retinoids in the Regeneration Pathway of
Rhodopsin. The Journal of Biological Chemistry, 267(22), 15701-15706.
Hollyfield, J. G., & Rayborn, M. E. (1987). Endocytosis in the inner segment of rod
photoreceptors: analysis of Xenopus laevis retinas using horseradish peroxidase.
Experimental Eye Research, 45(5), 703–719.
Hyde, D. R., Mecklenburg, K. L., Pollock, J. A., Vihtelic, T. S., & Benzer, S. (1990).
Twenty Drosophila visual system cDNA clones: one is a homolog of human arrestin.
Proceedings of the National Academy of Sciences of the USA, 87(3), 1008–1012.
Kaushal, S., & Khorana, H. G. (1994). Structure and function in rhodopsin. 7. Point
mutations associated with autosomal dominant retinitis pigmentosa. Biochemistry,
33(20), 6121–6128.
Kennedy, M. J., Lee, K. A., Niemi, G. A., Craven, K. B., Garwin, G. C., Saari, J. C., &
Hurley, J. B. (2001). Multiple Phosphorylation of Rhodopsin and the In Vivo
Chemistry Underlying Rod Photoreceptor Dark Adaptation. Neuron, 31(1), 87-101.
Kim, Y. M. (2002). Differential Roles of Arrestin-2 Interaction with Clathrin and
Adaptor Protein 2 in G Protein-coupled Receptor Trafficking. Journal of Biological
Chemistry, 277(34), 30760–30768.
98
Kirchhausen, T. (1999). Adaptors for clathrin-mediated traffic. Annual review of cell and
developmental biology, 15, 705–732.
Kiselev, A., Socolich, M., Vinós, J., Hardy, R. W., Zuker, C. S., & Ranganathan, R.
(2000). A molecular pathway for light-dependent photoreceptor apoptosis in
Drosophila. Neuron, 28(1), 139–152.
Lamb, T.D. & Pugh, E. N. (2004). Dark Adaptation and the Retinoid Cycle of Vision.
Progress in Retinal and Eye Research, 23(3), 307-380.
Langlois, G., Chen, C. K., Palczewski, K., Hurley, J. B., & Vuong, T. M. (1996).
Responses of the phototransduction cascade to dim light. Proceedings of the National
Academy of Sciences of the USA 93(10), 4677-4682.
Laporte, S. A. (2000). The Interaction of beta -Arrestin with the AP-2 Adaptor Is
Required for the Clustering of beta 2-Adrenergic Receptor into Clathrin-coated Pits.
Journal of Biological Chemistry, 275(30), 23120–23126.
Laporte, S. A. (2002). beta -Arrestin/AP-2 Interaction in G Protein-coupled Receptor
Internalization. Journal of Biological Chemistry, 277(11), 9247–9254.
Lee, K. A., Nawrot, M., Garwin, G. C., Saari, J. C., & Hurley, J. B. (2010). Relationships
among Visual Cycle Retinoids, Rhodopsin Phosphorylation, and Phototransduction
in Mouse Eyes during Light and Dark Adaptation. Biochemistry, 49(11), 2454-2463.
LeVine, H., Smith, D. P., Whitney, M., Malicki, D. M., Dolph, P. J., Smith, G. F.,
Burkhart, W., & Zuker, C. S. (1990). Isolation of a novel visual-system-specific
arrestin: an in vivo substrate for light-dependent phosphorylation. Mechanisms of
development, 33(1), 19–25.
Li, T., Franson, W. K., Gordon, J. W., Berson, E. L., & Dryja, T. P. (1995). Constitutive
activation of phototransduction by K296E opsin is not a cause of photoreceptor
degeneration. Proceedings of the National Academy of Sciences of the USA, 92(8),
3551-3555.
Li, T., Snyder, W. K., Olsson, J. E., & Dryja, T. P. (1996). Transgenic mice carrying the
dominant rhodopsin mutation P347S: evidence for defective vectorial transport of
rhodopsin to the outer segments. Proceedings of the National Academy of Sciences of
the USA, 93(24), 14176–14181.
Liu, C. & Varnum, M. D. (2005). Functional consequences of progressive cone
dystrophy-associated mutations in the human cone photoreceptor cyclic nucleotide-
gated channel CNGA3 subunit. American Jounal of Physiology: Cell Physiology
289(1), C187-C198.
99
Lohr, H. R., Kuntchithapautham, K., Sharma, A.K., & Rohrer, B. (2006). Multiple,
parallel cellular suicide mechanisms participate in photoreceptor cell death.
Experimental Eye Research 83(2), 380-389.
Lopes, V. S., Jimeno, D., Khanobdee, K., Song, X., Chen, B., Nusinowitz, S., &
Williams, D. S. (2010). Dysfunction of heterotrimeric kinesin-2 in rod photoreceptor
cells and the role of opsin mislocalization in rapid cell death. Molecular Biology of
the Cell, 21(23), 4076–4088.
Maeda, T. (2003). Rhodopsin phosphorylation: 30 years later. Progress in Retinal and
Eye Research, 22(4), 417–434.
Malanson, K. M., & Lem, J. (2009). Chapter 1 - Rhodopsin-Mediated Retinitis
Pigmentosa. Molecular Biology and Translational Science: G Protein-Coupled
Receptors in Health and Disease, Part A (1st ed. Vol. 88, pp. 1–31). Elsevier Inc.
Marszalek, J. R., Liu, X., Roberts, E. A., Chui, D., Marth, J. D., Williams, D. S., &
Goldstein, L. S. (2000). Genetic evidence for selective transport of opsin and arrestin
by kinesin-II in mammalian photoreceptors. Cell, 102(2), 175–187.
McBee, J. K., Palczewski, K., Baehr, W., & Pepperberg, D. R. (2001). Confronting
Complexity: the Interlink of Phototransduction and Retinoid Metabolism in the
Vertebrate Retina. Progress in Retinal and Eye Research, 20(4), 469-529.
Mendez, A., Burns, M. E., Roca, A., Lem, J., Wu, L. W., Simon, M. I., Baylor, D. A., &
Chen, J. (2000). Rapid and reproducible deactivation of rhodopsin requires multiple
phosphorylation sites. Neuron, 28(1), 153–164.
Milano, S. K., Pace, H. C., Kim, Y.-M., Brenner, C., & Benovic, J. L. (2002). Scaffolding
Functions of Arrestin-2 Revealed by Crystal Structure and Mutagenesis †,‡.
Biochemistry, 41(10), 3321–3328.
Nathans, J. (1990). Determinants of visual pigment absorbance: identification of the
retinylidene Schiff's base counterion in bovine rhodopsin. Biochemistry, 29(41),
9746–9752.
Orem, N. R. (2006). An essential role for endocytosis of rhodopsin through interaction of
visual arrestin with the AP-2 adaptor. Journal of Cell Science, 119(15), 3141–3148.
Owen, D. J., Collins, B. M., & Evans, P. R. (2004). ADAPTORS FOR CLATHRIN
COATS: Structure and Function. Annual review of cell and developmental biology,
20(1), 153–191.
100
Palczewski, K., Van Hooser, J. P., Garwin, G. C., Chen, J., Liou, G. I., & Saari, J. C.
(1999). Kinetics of Visual Pigment Regeneration in Excised Mouse Eyes and in Mice
with a Targeted Disruption of the Gene Encoding Interphotoreceptor Retinoid-
Binding Protein or Arrestin. Biochemistry, 38(37), 12012-12019.
Palczewski, K., & Smith, W. C. (1996). Splice variants of arrestins. Experimental Eye
Research, 63(5), 599–602.
Palczewski, K., Buczlko, J., Ohguro, H., Annan, R. S., Carr, S. A., Crabb, J. W., Kaplan,
M. W., Johnson, R. S., & Walsh, K. A. (1994). Characterization of a truncated form
of arrestin isolated from bovine rod outer segments. Protein Science 3(2) 314-324.
Paquet-Durand, F., Beck, S., Michalakis, S., Goldmann, T., Huber, G., Muhlfriedel, R.,
Trifunovic, D., Fischer, M. D., Fahl, E., Duetsch, G., Becirovic, E., Wolfrum, U., van
Veen, T., Biel, M., Tanimoto, N., & Seeliger, W. (2011). A key role for cyclic
nucleotide gated (CNG) channels in cGMP-related retinitis pigmentosa. Journal of
Neurochemistry 108(3), 796-810.
Paquet-Durand, F., Hauck, S. M., van Veen, T., Ueffing, M., & Ekstrom, P. (2009). PKG
activity causes photoreceptor cell death in two retinitis pigmentosa models. Human
Molecular Genetics 20(5), 941-947.
Peng, C., Rich, E. D., & Varnum, M.D. (2004). Subunit configuration of heteromeric
cone cyclic nucleotide-gated channels. Neuron 42(3), 401-410.
Penn, J. S., & Williams, T. P. (1986). Photostasis: regulation of daily photon-catch by rat
retinas in response to various cyclic illuminances. Experimental Eye Research, 43(6),
915–928.
Robinson, M. S. (2004). Adaptable adaptors for coated vesicles. Trends in Cell Biology,
14(4), 167–174.
Robinson, P. R., Cohen, G. B., Zhukovsky, E. A., & Oprian, D. D. (1992). Constitutively
active mutants of rhodopsin. Neuron, 9(4), 719–725.
Saari, J. C. (2000). Biochemistry of Visual Pigment Regeneration: the Friedenwald
Lecture. Investigative Ophthalmology & Visual Science, 41(2), 337-348
Saari. J. C., Garwin, G. C., Van Hooser, J. P., & Palczewski, K. (1998). Reduction of all-
trans-retinal Limits Regeneration of Visual Pigment in mice. Vision Research,
38(10), 1325-1333.
101
Sakmar, T. P., Franke, R. R., & Khorana, H. G. (1989). Glutamic acid-113 serves as the
retinylidene Schiff base counterion in bovine rhodopsin. Proceedings of the National
Academy of Sciences of the USA, 86(21), 8309–8313.
Sancho-Pelluz, J., Arango-Gonzalez, B., Kustermann, S., Romero, F. J., van Veen, T.,
Zrenner, E., Ekström, P., & Paquet-Drand, F. (2008). Photoreceptor cell death
mechanisms in inherited retinal degeneration. Molecular Neurobiology 38(3), 253-
269.
Satoh, A. K., & Ready, D. F. (2005). Arrestin1 Mediates Light-Dependent Rhodopsin
Endocytosis and Cell Survival. Current Biology, 15(19), 1722–1733.
Schmid, E. M., Ford, M. G. J., Burtey, A., Praefcke, G. J. K., Peak-Chew, S.-Y., Mills, I.
G., Benmerah, A., & McMahon, H. T. (2006). Role of the AP2 β-Appendage Hub in
Recruiting Partners for Clathrin-Coated Vesicle Assembly. PLoS Biology, 4(9), e262.
Scott, M. G. H. (2001). Recruitment of Activated G Protein-coupled Receptors to Pre-
existing Clathrin-coated Pits in Living Cells. Journal of Biological Chemistry,
277(5), 3552–3559.
Shi, G. W., Chen, J., Concepcion, F., Motamedchaboki, K., Marjoram, P., Langen, R., &
Chen, J. (2005). Light causes phosphorylation of nonactivated visual pigments in
intact mouse rod photoreceptor cells. The Journal of biological chemistry, 280(50),
41184–41191.
Shi, W., Sports, C. D., Raman, D., Shirakawa, S., Osawa, S., & Weiss, E. R. (1998).
Rhodopsin arginine-135 mutants are phosphorylated by rhodopsin kinase and bind
arrestin in the absence of 11-cis-retinal. Biochemistry, 37(14), 4869–4874.
Smith, D. P., Shieh, B. H., & Zuker, C. S. (1990). Isolation and structure of an arrestin
gene from Drosophila. Proceedings of the National Academy of Sciences of the USA,
87(3), 1003–1007.
Smith, W. C., Milam, A. H., Dugger, D., Arendt, A., Hargrave, P. A., & Palczewski, K.
(1994). A splice variant of arrestin. Molecular cloning and localization in bovine
retina. The Journal of biological chemistry, 269(22), 15407–15410.
Song, X., Vishnivetskiy, S. A., Seo, J., Chen, J., Gurevich, E. V., & Gurevich, V. V.
(2011). Arrestin-1 expression level in rods: balancing functional performance and
photoreceptor health. NSC, 174(C), 37–49.
Strissel, K. J. (2006). Arrestin Translocation Is Induced at a Critical Threshold of Visual
Signaling and Is Superstoichiometric to Bleached Rhodopsin. Journal of
Neuroscience, 26(4), 1146–1153.
102
Sung, C. H., Davenport, C. M., & Nathans, J. (1993). Rhodopsin mutations responsible
for autosomal dominant retinitis pigmentosa. Clustering of functional classes along
the polypeptide chain. The Journal of biological chemistry, 268(35), 26645–26649.
Sung, C. H., Makino, C., Baylor, D., & Nathans, J. (1994). A rhodopsin gene mutation
responsible for autosomal dominant retinitis pigmentosa results in a protein that is
defective in localization to the photoreceptor outer segment. The Journal of
neuroscience : the official journal of the Society for Neuroscience, 14(10), 5818–
5833.
Traub, L. M. (2003). Sorting it out: AP-2 and alternate clathrin adaptors in endocytic
cargo selection. The Journal of Cell Biology, 163(2), 203–208.
Vallazza-Deschamps, G., Cia, D., Gong, J., Jellali, A., Duboc, A., Forster, V., Sahel, J.
A., Tessier, L. H., & Picaud, S. (2005). Excessive activation of cyclic nucleotide-
gated channels contributes to neuronal degeneration of photoreceptors. European
Journal of Neuroscience, 22(5), 1013–1022.
Vishnivetskiy, S. A., Raman, D., Wei, J., Kennedy, M. J., Hurley, J. B., & Gurevich, V.
V. (2007). Regulation of arrestin binding by rhodopsin phosphorylation level. The
Journal of biological chemistry, 282(44), 32075–32083.
Wang, T. and Montell, C. (2007). Phototransduction and retinal degeneration in
Drosophila. Pflügers Archiv – European Journal of Physiology, 454(5), 821-847.
Watson, A. J., Aragay, A. M., Slepak, V. Z., & Simon, M. I. (1996). A novel form of the
G protein beta subunit Gbeta5 is specifically expressed in the vertebrate retina. The
Journal of biological chemistry, 271(45), 28154–28160.
Weitz, D., Ficek, N., Kremmer, E., Bauer, P. J., & Kaupp, U. B. (2002). Subunit
stoichiometry of the CNG channel of rod photoreceptors. Neuron 36(5), 881-889.
Wenzel, A., Grimm, C., Samardzija, M., & Reme, C. E. (2005). Molecular mechanisms
of light-induced photoreceptor apoptosis and neuroprotection for retinal
degeneration. Progreess in Retinal and Eye Research, 24(2), 275-306.
Wright, A. F., Chakarova, C. F., Abd El-Aziz, M. M., & Bhattacharya, S. S. (2010).
Photoreceptor degeneration: genetic and mechanistic dissection of a complex trait.
Nature Reviews Genetics, 11(4) 273-284.
Young, R. W. (1967). The renewal of photoreceptor cell outer segments. The Journal of
Cell Biology, 33(1), 61–72.
103
Zheng, J., Trudeau, M. C., & Zagotta, W. N. (2002). Rod cyclic nucleotide-gated
channels have a stoichiometry of three CNGA1 subunits and one CNGB1 subunit.
Neuron, 36(5), 891-896.
Zhong, H., Molday, L. L., Moday, R. S., & Yau, K. W. (2002) The heteromeric cyclic
nucleotide-gated channel adopts a 3A:1B stoichiometry. Nature, 420 (6912), 193-
198.
Zhukovsky, E. A., & Oprian, D. D. (1989). Effect of carboxylic acid side chains on the
absorption maximum of visual pigments. Science, 246(4932), 928–930.
Abstract (if available)
Abstract
Autosomal dominant retinitis pigmentosa (ADRP) is a blinding disorder whose most frequent causes are rhodopsin mutations. Of the more than 100 rhodopsin mutations that have been found, some disrupt the intramolecular interactions that constrain the molecule in an active conformation. This leads to a detrimental effect on the phototransduction cascade, which ultimately leads to retinal degeneration. One such mutation is Lys296Glu (K296E), in which case a lysine residue is replaced with a glutamic acid residue. It has been previously shown that opsin mutations in Drosophila form stable opsin/arrestin complexes that causes retinal degeneration via photoreceptor cell death. Similarly, mice expressing the K296E rhodopsin mutation also show progressive degeneration due to formation of a stable K296E/arrestin1 complex that is toxic to mammalian photoreceptors. The degeneration in Drosophila has been attributed to endocytosis of the rhodopsin/arrestin complexes via AP-2/clathrin binding. By effectively altering the genetics of K296E mice, I investigated whether the underlying causes and mechanisms that give rise to retinal degeneration associated with the K296E mutation are comparable to those found in Drosophila. This was achieved by substituting arrestin1 with a naturally occurring arrestin1 splice variant, p44, in which case the C-terminus is truncated at residue 370. This truncation eliminates the AP-2 binding domain and in turn prevents endocytosis. Expression of p44 in arrestin1 knockout mice reveals normal binding to rhodopsin and phototransduction activity. Accordingly, expressing p44 in K296E⁺/Arr⁻/⁻ mice should indicate whether endocytosis is mediating retinal degeneration and support our hypothesis that cell death is resultant of AP-2-mediated endocytosis of the K296E/Arrestin1 complex. ❧ Furthermore, I provide insight into another form of retinal degeneration, Oguchi disease, which is hallmarked by an inability to effectively dark-adapt from bright light conditions. The underlying cause of the disease has been identified as a mutation in arrestin and/or rhodopsin kinase, where a defect in either protein would prolong recovery from a light response. A further investigation of the roles of arrestin and rhodopsin kinase is still needed. I show that arrestin knockout mice exhibit a reduced rate of rhodopsin regeneration, thus reducing rod sensitivity. I go on to investigate whether specific phosphorylation residues on the C-terminus of rhodopsin regulate proper recovery kinetics and present their possible roles in dark adaptation. ❧ Lastly, a mutation affecting the rod cGMP-phosphodiesterase causes a form of retinitis pigmentosa. The rd1 mouse model of this disease exhibits an accumulation of cGMP concentrations inside the rod photoreceptor, which causes cell death due to the increase in cytosolic Ca2+ influx through cGMP-gated ion channels. I investigated the possible therapeutic treatment of CNG channel blockers in the rd1 mouse model. Preliminary treatments show promising results as retinal degeneration was slowed in several cases.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Multimodal investigation of rod photoreceptor structure and function in healthy and diseased states
PDF
New understanding of rhodopsin in retinal degeneration and high gain phosphorylation
PDF
Exploring alternative roles of visual arrestin 1 in photoreceptor synaptic regulation and deciphering the molecular pathway of retinal degeneration using mouse knockout technology
PDF
Cone arrestin 4 contributes to vision, cone health, and desensitization of the dopamine receptor D4
PDF
Spatial anomalies of visual processing in retinitis pigmentosa and a potential treatment
Asset Metadata
Creator
Moaven, Hormoz
(author)
Core Title
Visual arrestin interactions with clathrin adaptor AP-2 regulate photoreceptor survival
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Neuroscience
Publication Date
07/25/2012
Defense Date
03/13/2012
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
arrestin,CNG channels,k296e,OAI-PMH Harvest,p44,phosphorylation,photoreceptor,phototransduction,retina,Retinal degeneration,retinitis pigmentosa,rhodopsin,rhodopsin kinase,rod
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Chen, Jeannie (
committee chair
), Hinton, David R. (
committee member
), Sampath, Alapakkam P. (
committee member
)
Creator Email
hmoaven@gmail.com,moaven@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-64255
Unique identifier
UC11289088
Identifier
usctheses-c3-64255 (legacy record id)
Legacy Identifier
etd-MoavenHorm-993.pdf
Dmrecord
64255
Document Type
Dissertation
Rights
Moaven, Hormoz
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
arrestin
CNG channels
k296e
p44
phosphorylation
photoreceptor
phototransduction
retina
retinitis pigmentosa
rhodopsin
rhodopsin kinase
rod