Arrestin migrates in photoreceptors in response to light: a study of arrestinlocalization using an arrestin-GFP fusion protein in transgenic frogs

James J. Petersona,1, Beatrice M. Tamb,1, Orson L. Moritzb, Charles L. Shelamera, DonaldR. Duggera, J. Hugh McDowella, Paul A. Hargravea, David S. Papermasterb, W. Clay Smitha,*

aDepartment of Ophthalmology, University of Florida, 1600 SW Archer Road, D4-32, Gainesville, FL 32610-0284, USAbDepartment of Neuroscience, University of Connecticut, Farmington, CT, USA

Received 25 September 2002; accepted in revised form 31 January 2003

Abstract

Subcellular translocation of phototransduction proteins in response to light has previously been detected by immunocytochemistry. This

movement is consistent with the hypothesis that migration is part of a basic cellular mechanism regulating photoreceptor sensitivity. In order

to monitor the putative migration of arrestin in response to light, we expressed a functional fusion between the signal transduction protein

arrestin and green fluorescent protein (GFP) in rod photoreceptors of transgenic Xenopus laevis. In addition to confirming reports that arrestin

is translocated, this alternative approach generated unique observations, raising new questions regarding the nature and time scale of

migration. Confocal fluorescence microscopy was performed on fixed frozen retinal sections from tadpoles exposed to three different lighting

conditions. A consistent pattern of localization emerged in each case. During early light exposure, arrestin-GFP levels diminished in the inner

segments (ISs) and simultaneously increased in the outer segments (OSs), initially at the base and eventually at the distal tips as time

progressed. Arrestin-GFP reached the distal tips of the photoreceptors by 45–75 min at which time the ratio of arrestin-GFP fluorescence in

the OSs compared to the ISs was maximal. When dark-adaptation was initiated after 45 min of light exposure, arrestin-GFP rapidly re-

localized to the ISs and axoneme within 30 min. Curiously, prolonged periods of light exposure also resulted in re-localization of arrestin-

GFP. Between 150 and 240 min of light adaptation the arrestin-GFP in the ROS gradually declined until the pattern of arrestin-GFP

localization was indistinguishable from that of dark-adapted photoreceptors. This distribution pattern was observed over a wide range of

lighting intensity (25–2700 lux). Immunocytochemical analysis of arrestin in wild-type Xenopus retinas gave similar results.

q 2003 Elsevier Science Ltd. All rights reserved.

Keywords: arrestin; migration; fusion protein; green fluorescent protein; phototransduction; rhodopsin; photoreceptor; rod cell; confocal microscopy; Xenopus

laevis; transgenic; immunocytochemistry; light activation; cellular regulation

1. Introduction

The class of proteins known as arrestins is involved in the

inactivation of many G-protein coupled receptor cascade

systems (Wilson and Applebury, 1993; McDonald and

Lefkowitz, 2001). Visual arrestin is a cytoplasmic protein

found in photoreceptors and is known to quench

phototransduction by blocking the interaction of photo-

activated rhodopsin with its G-protein, transducin (Kuhn,

1978; Kuhn et al., 1984a,b). Thus, arrestin might be

expected to localize principally with rhodopsin in the outer

segment (OS) of the photoreceptor cell. Several independent

immunocytochemical investigations have shown, however,

that arrestin is immunologically detected primarily in the

inner segment (IS) in the dark and appears to migrate to the

OS in response to light (Broekhuyse et al., 1985; Mangini

and Pepperberg, 1987, 1988; Whelan and McGinnis, 1988;

Organisciak et al., 1991; McGinnis et al., 1992; Nir and

Ransom, 1993; Mangini et al., 1994; Loeffler and Mangini,

1995). Particularly intriguing is the fact that transducin

migrates in the opposite direction under the same lighting

conditions (Philp et al., 1987; Whelan and McGinnis, 1988;

Sokolov et al., 2002). It has been suggested that migration of

0014-4835/03/$ - see front matter q 2003 Elsevier Science Ltd. All rights reserved.

DOI:10.1016/S0014-4835(03)00032-0

Experimental Eye Research 76 (2003) 553–563

www.elsevier.com/locate/yexer

1 These authors contributed equally to the manuscript.

* Corresponding author. Dr W. Clay Smith, Department of

Ophthalmology, University of Florida, 1600 SW Archer Road, Room

D4-32, Gainesville, FL 32610-0284, USA.

E-mail address: csmith@eye1.eye.ufl.edu (W.C. Smith).

Abbreviations: FITC, fluorescein isothiocyanate; GFP, green

fluorescent protein; IS, inner segment; OS, outer segment; OCT, optimal

cutting temperature; PBS, phosphate buffered saline; WGA, wheat germ

agglutinin; Ar-GFP, Xenopus arrestin-GFP fusion protein.

signal transduction proteins may be part of a cellular

regulatory mechanism that is consistent with the role of rod

cells in supporting vision only under low lighting conditions

(McGinnis et al., 1991, 2002; Sokolov et al., 2002).

Most previous studies suggest that arrestin migration is

intersegmental, moving from the IS to the OS in response to

light (Broekhuyse et al., 1985; Whelan and McGinnis, 1988;

McGinnis et al., 1992; Mangini et al., 1994; Loeffler and

Mangini, 1995). However, other authors remain skeptical

(Nir and Ransom, 1993). Some of the technical limitations

of immunocytochemical imaging and subcellular fraction-

ation of signal transduction proteins have been outlined in a

discussion by Roof et al. (1988) (e.g. fixation artifacts,

epitope masking, covalent modification of proteins, inac-

curate fractionation of proteins from different subcellular

components, and non-linearity of densitometry of proteins

on SDS–PAGE). These and other criticisms have stimu-

lated renewed efforts in the search for experimental

alternatives for documenting light-activated migration of

signal transduction proteins (Sokolov et al., 2002).

This question can now be approached using green

fluorescent protein (GFP) fusion proteins (Barak et al.,

1997). GFP is an intrinsically fluorescent probe that requires

no additional cofactors, post-translational modifications, or

staining additives to visualize. Thus, GFP fusion proteins

are exempt from many of the technical limitations of an

immunocytochemical analysis (Chalfie and Kain, 1998). In

this study, we use fusions of GFP with arrestin to

conclusively demonstrate that arrestin moves from the IS

to the OS in response to light and we characterize the time

scale of both dark and light adaptation. In addition, we

observed a remarkable re-localization of arrestin-GFP

during prolonged light adaptation.

2. Materials and methods

2.1. Construction of arrestin-GFP fusions

For expression of arrestin-GFP in transgenic frogs we

utilized a modified form of the Clontech vector, eGFP-N1

(BD Biosciences Clontech, Palo Alto, CA, USA) containing

a 1·3 kb fragment of the Xenopus laevis opsin promoter,

XOP1·3eGFP-N1 (Tam et al., 2000). Xenopus arrestin

cDNA was isolated by reverse transcription-PCR from

Xenopus retinas (‘QuickPrep’ Micro mRNA purification,

Amersham; first strand cDNA synthesis kit, Amersham,

Piscataway, NJ, USA), introducing Xho I and Not I sites at

the 50 and 30 termini, respectively. An Nhe I site was also

incorporated immediately prior to the stop codon. The

amplicon was cloned into pPIC-Z (Invitrogen, San Diego,

CA, USA), and linearized with Nhe I (Promega, Madison,

WI, USA). An enhanced version of GFP (a kind gift from Dr

W. Hauswirth (Zolotukhin et al., 1996)) was inserted using

GFP to which Nhe I sites had been added by PCR. The

Xenopus arrestin-GFP (Ar-GFP) cDNA was then shuttled

into XOP1·3eGFP-N1 vector at the Xho I and Not I sites. The

resulting cDNA encodes a fusion protein consisting of GFP

fused to the C-terminus of Xenopus arrestin. Proper

construction of the recombinant plasmids was confirmed

by DNA sequencing.

2.2. Antibody production

Anti-GFP antisera was produced in rabbits using His(6)-

tagged GFP purified from E. coli BL21(DE3) cells. An anti-

arrestin mouse monoclonal antibody xAr1-6 was produced

using Xenopus arrestin expressed in E. coli BL21(DE3) cells

using previously established protocols (Adamus et al.,

1991). By peptide competition, this antibody was shown to

be specific for the N-terminal domain of arrestin.

2.3. In vitro assays of rhodopsin binding and measurement

of GFP fluorescence

Recombinant Ar-GFP was expressed in Pichia pastoris

(Invitrogen, San Diego, CA, USA), and purified as

previously described (McDowell et al., 1999). Rhodopsin,

and phosphorylated rhodopsin were prepared from either

bovine retinas or Rana catesbeiana (bullfrog) retinas

according to published procedures (McDowell, 1993).

Purified arrestin-GFP fusions were concentrated using either

a Centriprep-30, or a Centriprep-10 followed by a Centricon

Ultrafree-4 (50 kDa cut-off) membrane concentrator (Ami-

con, Beverly, MA, USA), snap frozen in liquid nitrogen and

stored at2808C before being tested for functionally specific

binding to rhodopsin using a centrifugation assay (Palc-

zewski et al., 1991). Fusion protein concentrations were

determined using the Bradford method. Absorbance and

fluorescence measurements were performed on Cary 50 UV

spectrophotometer and Photon Technologies Quantum

Master 1 instruments, respectively. Fluorescence properties

of the fusion proteins were then compared to those of His(6)-

tagged GFP alone. For this comparison, GFP (Zolotukhin

et al., 1996) containing an N-terminal His(6) tag was

expressed in E. coli BL21(DE3) cells and purified using a

Ni-NTA column (Qiagen, Valencia, CA, USA).

The extent of Ar-GFP binding to bullfrog rhodopsin was

estimated by scanning densitometry of Western blots

probed using anti-GFP polyclonal sera (500:1). Binding of

wild-type bovine arrestin to bovine rhodopsin was assessed

directly from gels stained with Coomassie brilliant blue.

Scion Image (version beta 4·0·2) was used to calculate area

and mean density of each band. Results were averaged

between blots or gels and normalized to the highest average

product of band area and density.

2.4. Generation of transgenic Xenopus laevis

Transgenic frogs were generated using a nuclear

transplantation approach (Kroll and Amaya, 1996) with

minor modifications (Moritz et al., 1999; Tam et al., 2000).

J.J. Peterson et al. / Experimental Eye Research 76 (2003) 553–563554

Xenopus sperm nuclei were incubated with diluted high-

speed egg extract, Sfo I restriction enzyme (NEB, Beverly,

MA or Promega, Madison, WI, USA) and 100–200 ng

linearized vector DNA. The final reaction mixture was

diluted to 0·3 nuclei/nL and 10 nL was injected per egg

using an injection pump (KD Scientific, model 100, New

Hope, PA, USA). Egg injections were performed at 18–

228C. The resulting embryos were kept at 188C and screened

for fluorescence as described previously (Moritz et al.,

1999). Transgenic adult males were mated with wild-type

females to produce F1 offspring, which were used for light/

dark-adaptation experiments. After 2–3 weeks, F1 tadpoles

were visually screened for green fluorescence and segre-

gated into four tanks corresponding to strong, medium,

weak, or no transgene expression.

2.5. Expression levels in transgenic tadpoles

In order to obtain an estimate of the expression levels of

the transgene relative to the native arrestin, eyes from

tadpoles ranked as ‘medium’ or ‘strong’ expressors were

excised and sonicated in 250 ml of reducing Laemmli

sample buffer (Laemmli, 1970). Aliquots of the eye extract

were separated on 10% SDS–PAGE gels, transferred to

PVDF membrane, and probed with xAR1-6, a monoclonal

antibody that recognizes both the native arrestin and the

arrestin-GFP fusion protein. The fusion protein was

quantified relative to the endogenous arrestin by scanning

densitometry. This experiment was repeated, using four

tadpoles in each category.

2.6. Experimental conditions for light and dark-adaptation

For light adaptation experiments, transgenic tadpoles

were initially dark-adapted overnight. The tadpoles were

then placed in transparent plastic bottles (Fisher cat. No.

097615, Pittsburgh, PA, USA) and light adapted for several

time points as shown in Fig. 2. At the end of light

adaptation, all tadpoles associated with a given time point

were netted and transferred as a group to fixative within

30 sec. Light adaptation was carried out under three

different lighting conditions: (1) over a fluorescent light

box (30 W American Medical Sciences model 201D) in a

well-lit laboratory (ca. 2700 lux); (2) in a well-lit laboratory

placed 1 m under a bank of two 32 W fluorescent bulbs (ca.

860 lux); (3) in a dimly lit, shaded area of the laboratory

receiving only indirect light (ca. 25–50 lux). Light levels

were measured using an International Light IL1700 radio-

meter. As a comparison, light that was measured outdoors

on a partly cloudy day under a light tree canopy at mid-day

ranged from 4000 to 80 000 lux. For dark-adaptation

experiments, tadpoles were initially dark-adapted overnight,

light adapted over the light box for 45 min, and then

dark-adapted for 5, 10, 15, 30 and 45 min. At the end of

dark-adaptation, the tadpoles were immediately placed in

fixative in the dark. Tadpoles used for these experiments

ranged from 2 to 6 weeks old (Stages 47–54). In most

experiments, light adaptation was initiated at ca. 8 AM.

However, to detect any possible circadian component of

migration we performed additional light adaptation exper-

iments at 860 lux beginning at 1 PM and also at 6 PM.

2.7. Tissue preparation, and immunohistochemistry

Tadpoles were fixed in chilled 73% methanol/3·7%

formaldehyde/23·3% H2O and incubated overnight at 08C.

Fixed tadpoles were re-hydrated at room temperature in a

series of methanol/phosphate buffered saline (PBS) mix-

tures containing 60, 40, 20, and 0%methanol for 15–30 min

per mixture at RT followed by overnight incubation in 30%

sucrose/PBS at 48C. Eyes were excised, embedded in

optimal cutting temperature (OCT) media (Sakura Finetek,

Torrance, CA, USA), cryosectioned at2208C (12–14 mm),

collected on slides, and stored over ice before use. Sections

from transgenic animals were rinsed with 1 £ PBS for

40 min to remove OCT, and immediately sealed in mowiol

solution (2·61 M glycerol, 12 mM n-propyl gallate,

96 mg ml21 mowiol (Aldrich, St Louis, MO, USA), 0·1 M

Tris pH 8·5) with a cover slip for quantitative imaging of

intrinsic green fluorescence. Sections from wild-type

animals were rinsed and treated with 0·1% NaBH4 in PBS

for 30 min, 1·0% Triton X-100 in PBS for 25 min, and

finally 6 M guanidinium HCl 50 mM Na3PO4 pH 7 for

20 min at RT. Sections were then washed 3 £ 5 min in

deionized water and blocked in 1% gamma globulin-free

horse serum (GGHS) (Life Technologies, Carlsbad, CA,

USA) for 2 hr. Primary antibody incubation with anti-

arrestin antibody xAr1-6 (50:1) was carried out in a flat,

sealed, humid chamber on a belly dancer (Stovall Life

Science Inc., Greensboro, NC, USA) set to 12–13 rpm in

the presence of 0·2% Triton X-100 and 0·2% GGHS in PBS

for 36–42 hr at RT. Secondary antibody incubation with

anti-mouse Texas red-X (100:1) (Molecular Probes,

Eugene, OR, USA) was performed under the same

conditions for 24 hr. Slides were subsequently washed

3 £ 30 min in PBS where the second wash also contained

350 nM Sytox green (Molecular Probes) and sealed in

mowiol solution as before.

2.8. Confocal microscopy

Confocal fluorescence microscopy was performed on a

Biorad 1024ES instrument using excitation lasers and

emission filters optimized for FITC and Texas Red and

20x, 60x-oil (NA 1·4), and 100x-oil (NA 1·35) objectives.

Due to the wide variety of expression levels, no

combination of settings of iris, gain, and laser power setting

was found that would provide suitable visualization of all

transgenic animals. Therefore, only animals with strong

J.J. Peterson et al. / Experimental Eye Research 76 (2003) 553–563 555

Ar-GFP expression were used in quantitative experiments.

Laser settings were chosen such that the presence of

autofluorescence could not be detected in wild-type retinas,

and the GFP signal was not saturated. These low gain

settings were rigorously fixed for each image during

quantification. For noise reduction, 5–10 Kalman noise

reduction algorithms were performed to produce each image

(Bartoli and Cerutti, 1983). Qualitative and quantitative

images were constructed from central retina using a Z-series

of 30–35 images spaced at 0·4 mm intervals that were

combined to create a 2D vertical projection (a sum of

the individual Z-slices, projected in two dimensions) using

Biorad Lasersharp version 3·0 software.

2.9. Quantitative analysis of confocal images

The FITC channel of each vertical projection was saved

as a TIFF file, and the grey scale image inverted in Adobe

Photoshop version 6·0 and imported into Scion Image

version 4·0·2. The image was screened for the presence of

Fig. 1. Bovine arrestin and Ar-GFP binding to rhodopsin. Wild-type bovine arrestin and Ar-GFP fusion were tested for specific binding to bovine and bullfrog

rhodopsin, respectively. (A) Coomassie blue stained SDS–PAGE gel of wild-type bovine arrestin bound to bovine rhodopsin in ROS membranes. Arrow

indicates arrestin (the band migrating below the 45 kDa marker is rhodopsin). (B) Western blot of SDS–PAGE gel of Xenopus Ar-GFP bound to bullfrog

rhodopsin in ROS membranes. Proteins were transferred to PVDF membrane and stained using anti-GFP polyclonal sera. Anti-rabbit alkaline phosphatase

conjugate served as the secondary antibody. Arrow indicates the Ar-GFP fusion protein. Fractions bound to the following forms of rhodopsin were separated by

SDS–PAGE: Non-phosphorylated rhodopsin in the dark (R); phosphorylated rhodopsin in the dark (RP); light-activated, non-phosphorylated rhodopsin (R*);

and light-activated, phosphorylated rhodopsin (R*P). Molecular mass standards are indicated to the left. Bound arrestin and Ar-GFP were quantified by

scanning densitometry and represented in the bar graphs, averaging 4 gels for arrestin (A) and 6 blots for Ar-GFP (B) (^SEM).

Fig. 2. Ar-GFP localization in retinal sections of transgenic tadpoles in response to various lighting intensities following overnight dark-adaptation. Each image

is representative of 5–11 retinas examined. (A) Light adaptation at 2700 lux. Green fluorescent confocal images of transgenic tadpoles expressing Ar-GFP that

were dark-adapted overnight and either fixed at time ¼ 0 (D.ON), or fixed following light-adaptation for t ¼ 15; 30, 45, 60, 90, or 150 min. Lower right-hand two

panels are controls, where a tadpole expressing unfused GFP was dark-adapted overnight or light adapted for 60 min. Arrow indicates the banding pattern (b)

observed in outer segments. Arrowhead indicates dense concentration of Ar-GFP in the area of the axoneme (a) adjoining the interconnecting cilium. Outer

segment (os); Inner segment (is); Nucleus (n); Synapse (s). (B) Light adaptation at 860 lux. Green fluorescent confocal images of transgenic tadpoles expressing

Ar-GFP that were dark-adapted overnight and light adapted for t ¼ 17; 30, 45, 67, 90, or 240 min. (C) Light adaptation at 25–50 lux. Green fluorescent confocal

images of transgenic tadpoles expressing Ar-GFP that were dark-adapted overnight and light adapted for t ¼ 17; 45, 76, 143, or 251 min. (D) Higher

magnification image from the retina of a tadpole light adapted for 45 min at 860 lux. The Ar-GFP fluorescence is distributed in a banded pattern in the OS region.

A, B, C Scale bar: 20 mm. D Scale bar: 10 mm.Q

J.J. Peterson et al. / Experimental Eye Research 76 (2003) 553–563556

J.J. Peterson et al. / Experimental Eye Research 76 (2003) 553–563 557

saturation and a group of 7–25 photoreceptor OSs from

each retina were encircled and measured for mean density

and area. These values were also quantified for the

corresponding ISs. The product of mean density and area

were calculated for IS and OS and the OS:IS ratios for these

products were plotted together with their standard error (5–

11 retinas per time point). Thus, each OS:IS ratio was

derived from a group of several adjacent photoreceptors

calculating the total green fluorescence from the OS divided

by that of the IS, averaging ratios from several sections from

a minimum of 5 retinas per time point. The mean values of

consecutive time points were compared using Student’s t-

test to establish statistical significance.

3. Results

3.1. Arrestin-GFP specific binding properties

A centrifugation assay was used to assess the binding

properties of Xenopus arrestin-GFP (Ar-GFP) to rhodopsin

(Fig. 1). In this assay, bovine arrestin was mixed with

bovine rhodopsin (both obtained from bovine retinas) in

disc membranes, and Ar-GFP (expressed in Pichia pastoris)

was mixed with bullfrog rhodopsin (obtained from Rana

catesbeiana retinas) in disc membranes. Like native

arrestin, Ar-GFP preferentially bound light-activated, phos-

phorylated rhodopsin (R*P). This result indicates that fusing

GFP to the C-terminus of arrestin does not interfere with the

binding selectivity of arrestin. In addition, fusion of these

two proteins did not interfere significantly with the

fluorescence properties of GFP (data not shown).

3.2. Expression of arrestin-GFP in transgenic

Xenopus laevis

Under the conditions used for our transgenesis,,20% of

the eggs injected resulted in properly cleaving embryos. Of

these, approximately 30–50% were positive for GFP

fluorescence, although expression levels were variable

from animal to animal and, occasionally, from cell to cell.

Expression levels of the transgene were determined in

tadpoles using western blots of extracts prepared from the

whole eye as described in the methods (data not shown).

Tadpoles that were visually categorized as ‘medium’

expressors had the Ar-GFP transgene expressed at

56·7 ^ 7·2% of the endogenous arrestin ðn ¼ 4Þ: ‘Strong’

expressors had the transgene expressed at 68·8 ^ 6·8% of

the native arrestin levels ðn ¼ 4Þ:

3.3. Light dependent migration of Ar-GFP from the IS

to the OS during early light adaptation and re-localization

to the IS after prolonged light exposure

Comparison of dark-adapted and light-adapted trans-

genic tadpoles revealed that Ar-GFP localization changed in

response to light. In order to observe the process of Ar-GFP

migration over time, transgenic tadpoles were dark-adapted

overnight and fixed after increasing periods of light

adaptation at 2700 lux (Fig. 2(A)), 860 lux (Fig. 2(B)),

and at 25–50 lux (Fig. 2(C)). Confocal images of frozen

retinal sections showed that green fluorescence from Ar-

GFP in dark-adapted retinas was predominantly associated

with the IS and axoneme (Fig. 2(A), D.ON ðt ¼ 0Þ,) and

intrinsic fluorescence levels were markedly lower in the OS.

Fig. 3. The localization of arrestin in wild-type tadpoles detected immunocytochemically is consistent with that of Ar-GFP in transgenic tadpoles. Sections

from wild-type Xenopus tadpole retinas were labelled using an anti-arrestin monoclonal antibody xAr1-6 (red channel). Immunocytochemically detected

arrestin is localized in the IS and axonemes of dark-adapted retinas (DA (ON)), in a banded, proximal-to-distal gradient in the OS during early light adaptation

LA (450), and in the IS and axonemes after extended light adaptation LA (2400) at 860 lux. Photoreceptors from retinal sections stained with secondary antibody

conjugate alone L. (45/NP) are not labelled. Nuclei are stained with Sytox green (blue channel). Outer segment (os); Inner segment (is); Nucleus (n); Synapse

(s): Axoneme (a); Bands (b). Scale bar: 20 mm.

J.J. Peterson et al. / Experimental Eye Research 76 (2003) 553–563558

Comparison of the dark-adapted time point ðt ¼ 0Þ of Fig.

2(A) with early light-adapted time points in Fig. 2(A)–(C)

clearly indicates that Ar-GFP moved from the IS, through

the proximal OS, then ultimately throughout the distal

portions of the OS. At early light adaptation time points,

regardless of light intensity, the majority of Ar-GFP signal

left the IS and was increased in the OS. Surprisingly, by

90 min of light exposure, the levels of Ar-GFP in the OS had

started to decline (Fig. 2(A)–(C)) and by 150–240 min

levels were essentially the same as in a dark-adapted retina

regardless of light intensity. This distribution of Ar-GFP

was different than that seen in animals expressing GFP

alone where no light-dark difference of GFP was observed

(Fig. 2(A), GFP controls). Moreover, the migration of Ar-

GFP does not appear to be modulated by a circadian cycle

since the migration pattern was identical in experiments

initiated at 8 AM, 1 PM, or 6 PM (data not shown).

3.4. Immunocytochemical analysis of arrestin migration in

wild-type tadpoles and adults is consistent with Ar-GFP

movement in transgenic tadpoles

One potential concern is that Ar-GFP in transgenic

tadpoles localizes differently than native arrestin in wild-

type tadpoles or adult frogs. This concern was addressed

immunocytochemically, by staining frozen sections of wild-

type tadpole and adult Xenopus retinas with an anti-arrestin

antibody. Our early experiments employing standard

immunocytochemical methods failed to detect arrestin in

the disc region of the OS. However, arrestin migration could

be assessed immunocytochemically after treatment with

borohydride, guanidinium hydrochloride, 1% Triton X-100,

and unusually long antibody incubation times (Fig. 3). After

these conditions were applied, arrestin immunolabelling

was intensely localized in the proximal OS, and concomi-

tantly diminished in the IS during early light adaptation.

After prolonged light adaptation, arrestin labelling was

diminished in the OS and increased in the IS, becoming

indistinguishable from that of dark-adapted animals. These

results were obtained from wild-type tadpoles (Fig. 3) and

adults (data not shown) and were essentially identical to

results obtained in transgenic tadpoles (Fig. 2(A)–(C)),

including the concentration of arrestin to the IS after

extended light adaptation, the formation of bands in the OS,

and the persistence of arrestin in the IS and synapses of

wild-type retinas after early light adaptation. It should be

emphasized that immunocytochemical results were highly

method dependent. Standard methods not employing

chaotropic agents and high detergent (1% Triton X-100)

resulted in poor staining of arrestin in the OS. In addition,

without primary antibody incubation times of 32 hr or more,

residual arrestin in the IS during early light adaption and

arrestin in the axonemes of dark-adapted photoreceptors

were poorly stained (data not shown). Control sections

(Fig. 3; L.(450/NP)) show that the harsh labelling conditions

and prolonged antibody incubation times did not lead to

non-specific background.

3.5. Ar-GFP and arrestin are distributed in a banded

pattern in the OS during early light adaptation

The Ar-GFP fusion was densely contained within the

cilia connecting the IS to the OS and in the axoneme in both

light- and dark-adapted tadpoles. In addition, both Ar-GFP

Fig. 4. Quantitative densitometric assessment of Ar-GFP localization

during light- and dark-adaptation. (A) Assessment of Ar-GFP localization

during light adaptation under three separate lighting conditions (2700 lux—

diamonds; 860 lux—squares; 25–50 lux—circles). Regardless of light

intensity the ratio of OS:IS fluorescence consistently increased during early

light adaptation, and gradually decreased during extended light adaptation.

The OS:IS fluorescence ratio was calculated for each time point using

densitometric analysis of confocal images from a minimum of 5 retinas per

time point. (B) Assessment of migration during dark-adaptation. Ar-GFP

migrates rapidly to the IS during dark-adaptation. Following 45 min of light

adaptation at 2700 lux, tadpoles were fixed following 5, 10, 15, 30 or

45 min in the dark. (See filled triangles at t ¼ 50; 55, 60, 75, or 90 min).

Light adaptation curves from Fig. 4(A) are included for reference. The first

data point (t ¼45 min) represents OS:IS ratios from tadpole retinas light

adapted for 45 min. The OS:IS fluorescence ratio was calculated for each

time point using densitometric analysis of confocal images from 7 to 8

retinas per time point.

J.J. Peterson et al. / Experimental Eye Research 76 (2003) 553–563 559

in transgenic tadpoles (Fig. 2) and immunocytochemically

labelled native arrestin in wild-type tadpoles (Fig. 3) formed

a banded pattern in the OS during early light adaptation.

This pattern was particularly evident at times when the

OS:IS ratio of Ar-GFP was maximal (Fig. 2(D)). This

banding was also seen in adult eyes from both transgenic

and wild-type animals (data not shown). The Ar-GFP fusion

was largely excluded from the photoreceptor nuclei,

probably due to the size restriction imposed by the nuclear

pores. In contrast, when GFP was expressed alone (i.e. not

in fusion with arrestin), GFP was present in the nucleus, did

not appear to preferentially localize in the connecting cilia

or axoneme, formed no evident bands in the light-adapted

OS, and did not appear to move between cell compartments

during light or dark-adaptation (control panels in Fig. 2(A)),

essentially as described by Knox et al. (1998) and Moritz

et al. (1999).

3.6. The effects of light intensity on the migration of Ar-GFP

To determine the effect of different lighting intensities

on the migration of Ar-GFP, transgenic tadpoles were light

adapted at 50, 860 and 2700 lux, analyzing migration both

qualitatively (Fig. 2) and quantitatively (Fig. 4). After 15–

17 min of light exposure, the basal third of the OS was

labelled with green fluorescence and appeared banded

regardless of light intensity. The fluorescence associated

with the ciliary body and axoneme appeared diminished,

but may simply have been obscured by increased

fluorescence in the OS as a whole. Under bright or normal

laboratory lighting (2700 or 860 lux, respectively), Ar-GFP

fluorescence reached the distal tips of most photoreceptors

by 45–60 min, at which point the OS:IS ratio of Ar-GFP

was maximal. Under dim lighting conditions (50 lux) the

measured OS:IS ratio of Ar-GFP was also nearly maximal

by 45 min, but the basal to apical flooding of the OS was

significantly slower and was not complete until after

76 min (Fig. 2(C)). Unexpectedly, the observed OS:IS ratio

of Ar-GFP for tadpoles that were light adapted for 45 min

was consistently higher at 860 or 50 lux than at 2700 lux.

After 60 min of light exposure at 860–2700 lux or after

80 min at 50 lux, Ar-GFP was distributed throughout the

photoreceptor, the OS was distinctly banded, and the OS:IS

ratio began to fall (Figs. 2 and 4(A)). The intrinsic

fluorescence in the cilium and axoneme was again intense,

comparable to that of dark-adapted tadpoles. Even at the

lower light intensities (Figs. 2(B),(C) and 4(A)), continued

light adaptation resulted in a gradual decline of the OS:IS

ratio of Ar-GFP as seen with the bright illumination (Figs.

2(A) and 4(A)). Under all three lighting conditions, after

approximately 4 hr of light adaptation both the appearance

of Ar-GFP distribution (Fig. 2) and the OS:IS ratio

(Fig. 4(A)) were indistinguishable from those of dark-

adapted retinas.

3.7. Time course and nature of Ar-GFP localization from

OS to IS during dark-adaptation

The distribution of Ar-GFP during dark-adaptation was

also monitored quantitatively (Fig. 4(B)) and qualitatively

(Fig. 5) immediately following 45 min of light adaptation at

2700 lux. After 5 min of dark-adaptation the magnitude of

the OS:IS ratio (Fig. 4(B)) was reduced (Fig. 5). Gradual

relative fading of Ar-GFP intensity in the OS compared to

that of the IS continued until by 30 min the OS:IS ratio had

fallen essentially to the level observed in retinas that were

dark-adapted overnight. In the fully dark-adapted tadpole,

Ar-GFP localized primarily to the IS, but also significantly

in the connecting cilia and axoneme (Fig. 2(A)). Note that

the rate of Ar-GFP depletion in the OS during dark-

adaptation appeared to be faster than occurred during

extended light adaptation (Fig. 4(A) and (B)). Intense

fluorescence was occasionally noted in the retinal pigment

epithelium in several dark- as well as light-adapted retinas

(see image of 15 min dark-adapted retina of Fig. 5), but this

was not consistently observed even though RPE was present

in the frozen section. This punctate fluorescence was not

stained in adjacent sections during subsequent immuno-

cytochemical treatment using GFP anti-sera, and has

broader emission than GFP, thus ruling out GFP as the

source.

4. Discussion

The intrinsic fluorescence of GFP enabled us to obtain

unequivocal evidence that arrestin is not simply localized

differentially as a result of synthesis, degradation, or even

epitope masking but actually moves to the OS in response

Fig. 5. Ar-GFP localization in retinal sections from Ar-GFP tadpoles during

dark-adaptation following 45 min of light adaptation at 2700 lux. Dark-

adaptation is characterized by a steady, uniform decrease of Ar-GFP in the

OS. Fluorescence confocal images were collected from tadpoles dark-

adapted for 5 min (D. 50), 10 min (D. 100), 15 min (D. 150), 30 min (D.300),

and 45 min (D. 450). Each image is representative of 7–8 retinas examined.

Outer segment (os); Inner segment (is); Axoneme (a); Nucleus (n); Synapse

(s). Scale bar: 20 mm.

J.J. Peterson et al. / Experimental Eye Research 76 (2003) 553–563560

to brief light exposure, in general agreement with

immunocytochemical studies in mammals (Philp et al.,

1987; Whelan and McGinnis, 1988; McGinnis et al., 1992;

Mirshahi et al., 1994). In addition, we contribute unex-

pected results that have implications for photoreceptor cell

function.

First and most remarkable, we found that arrestin and Ar-

GFP was concentrated in the IS and depleted in the OS

during prolonged light exposure. This result has not been

observed in previous studies; rather, arrestin has been

reported to remain in the OS even after several hours of light

exposure. Previous studies have used either mice or rats.

Based on this, one might speculate that our result in

Xenopus is a species-specific phenomenon. However,

recently McGinnis et al. (2002) also examined light-induced

migration of arrestin in Xenopus. In contrast to our results,

they find that after 240 min of exposure to standard

laboratory lighting, the majority of arrestin is localized to

the OS (with no observable banding). In the IS, only minor

amounts of arrestin remain in the form of puncta. McGinnis’

study used only immunohistochemistry to detect arrestin

localization, and results from such methods can vary due to

even subtle differences in technique (Whitehead et al.,

1999). For this reason we have used both antibody labelling

and detection of intrinsic GFP fluorescence in transgenic

animals. Our results from these two independent techniques

were consistent with each other. However, before we

implemented the use of harsh denaturants and prolonged

antibody incubations on our retinal sections, we obtained

inconsistent antibody labelling of arrestin in both paraffin

embedded and frozen sections (data not shown). Therefore,

differences in the exact method of tissue preparation and

antibody labelling might be the source of the discrepancy

between our results and McGinnis et al. Furthermore, we

believe that detection of intrinsic fluorescence from the GFP

fusion proteins provided us with a very reliable method of

assessing protein localization. Our results were not

complicated by circadian effects since they were unaltered

by the time of day that light adaptation was initiated. Also,

although most of our images were derived from tadpoles,

experiments repeated with adult frogs gave identical results.

The second observation unique to our investigation is the

presence of a distinct banding pattern formed by arrestin and

Ar-GFP in the OS during early light adaptation (Figs. 2

and 3). This effect is not due to structural barriers to

diffusion, since GFP alone does not form bands in the OS

during either light or dark-adaptation. In addition, the fact

that transgenic tadpoles expressing GFP were fixed and

embedded in the same way as tadpoles expressing Ar-GFP

is evidence that this pattern is not the result of simple

fixation or embedding artifacts. Moreover, strikingly similar

observations have been previously reported under a variety

of experimental methods. For example, when the integral

membrane protein rhodopsin is fused to GFP, a similar

banding pattern is observed, which has a regular periodicity

very similar to that of Ar-GFP (Moritz et al., 2001). Studies

of birefringence bands in frog OS reveal the same

approximate periodicity (Kaplan et al., 1982; Andrews

et al., 1984). This birefringence banded pattern was found to

coincide with the daily rhythm of disc synthesis, but the

precise nature of the bands has not been characterized. In

Xenopus, membrane disc assembly is stimulated by light in

the context of a 24-hr diurnal cycle, whereas opsin synthesis

is not (Hollyfield et al., 1982). Thus, it is possible that

variable rates of disc formation will result in periodic bands

of increased rhodopsin concentration in the ROS, and that

this is the cause of the periodic differences in birefringence

as well as the pattern of Ar-GFP distribution seen in the OS

after early light adaptation. The observed pattern suggests

that Ar-GFP is not cytoplasmic in the OS, but rather is

associating with disc membranes, presumably by binding to

light-activated, phosphorylated rhodopsin.

A third observation unique to our investigation concerns

the clearance of arrestin from the IS during early light

adaptation. Previous studies (Nir and Ransom, 1993;

McGinnis et al., 2002) showed nearly complete depletion

of immunologically detectable arrestin from the IS in light-

adapted animals and low labelling of the cilia. In the present

study some arrestin or Ar-GFP was always detectable in

the IS. There are several potential explanations for these

differences. First, it is likely that the intrinsic fluorescence of

the Ar-GFP is more sensitive than standard methods of

immunocytochemistry, particularly if epitope masking is

affecting antibody recognition. Agreement of our transgenic

and wild-type immunocytochemical results was obtained

only after extreme, denaturing conditions were applied

(compare Figs. 2 and 3). This agreement also suggests that

over-expression of Ar-GFP is not causing a change in

arrestin distribution. Alternatively, it is also possible that

since the magnitude of the OS:IS ratio of Ar-GFP is

dependent upon both the light intensity and the duration of

light exposure prior to fixation, other time points may exist

where the IS Ar-GFP levels are lower. For example,

tadpoles light adapted for 45 min at 50 lux have essentially

no Ar-GFP in the IS (Fig. 2(C)).

What are the underlying mechanisms of arrestin

migration under these various lighting conditions?

The simplest explanation for outward migration is that

mass action is drawing arrestin into the OS by passive

diffusion as a consequence of the affinity of arrestin for

bleached, phosphorylated rhodopsin as proposed by Man-

gini et al. (1994). Return to the IS would then be driven by

the release of arrestin as metarhodopsin decays and is

dephosphorylated, though this driving force would have to

be coupled with an additional mechanism to reduce arrestin

in the OS to the level observed in this study. The observation

that detectable amounts of GFP could be found in the OS as

well as the IS in animals expressing GFP (Moritz et al.,

1999) suggests that soluble proteins can be transported

throughout the photoreceptor cytoplasm through passive

diffusion. However, since rhodopsin is phosphorylated and

dephosphorylated on a time scale that is much faster than

J.J. Peterson et al. / Experimental Eye Research 76 (2003) 553–563 561

the rate of Ar-GFP migration, an alternative mechanism for

arrestin migration may be required. What this mechanism

would be is unknown at this point.

A possible mechanism for removal of arrestin from the

dark-adaptedOS is proteolysis associatedwith normal protein

turnover (Azarian et al., 1995). Another mechanism for

arrestin removal from the OS is active transport as previously

speculated (Whelan and McGinnis, 1988). Recently, it has

been demonstrated that the normal intracellular trafficking of

arrestin and rhodopsin can be disrupted by defects in kinesin-

II, a microtubule motor protein, without similar defects in the

trafficking of other cytoplasmic proteins (e.g. transducin)

(Marszalek et al., 2000).

Since arrestin functions to quench photoactivated and

phosphorylated rhodopsin, one would expect arrestin to be

persistently present in the ROS where rhodopsin is light

activated. Perhaps arrestin has an alternative function. It has

been suggested that the complementary translocations of

arrestin and transducinmay serve as amolecular basis for the

regulation of rod cell sensitivity in response to the lighting

environment (McGinnis et al., 1991; Williams and Mangini,

1991; Sokolov et al., 2002). In support of this hypothesis, it

has been shown that arrestinmigration in response to light (as

measured immunologically) is significantly slower in cones

relative to rods (Mirshahi et al., 1994). On the other hand, our

demonstration that arrestin leaves the OS after prolonged

light adaptation poses a challenge to this hypothesis. Perhaps

arrestin and transducin redistribution are only required

during early light adaptation. During prolonged light

adaptation it is possible that other redundant mechanisms

become sufficient to attenuate the rod photoreceptor

response, allowing arrestin localization to revert to that of

the dark-adapted state. For example, the time scale and

intercellular nature of retinoid processing inmammalian rods

(Palczewski et al., 1999 and Kang Derwent et al., 2002)

compared to that of cones (Mata et al., 2002) suggests rod

retinoid depletion as one such possibility. It is noteworthy

that light-driven shuttling of proteins as a method of

regulating photosensitivity is a common theme in invert-

ebrate photoreceptors as well (Bahner et al., 2002). We are

currently implementing experiments designed to elucidate

the basic mechanism and rationale for intersegmental

migration of signal transduction proteins.

Acknowledgements

The authors appreciate the generosity of several

individuals for materials and equipment, including Drs

W. Hauswirth, W. Dawson, M. Hope, B. Jordan and

L. Bloom. We also acknowledge the excellent assistance of

A. Dinculescu and Tricia Clark with many technical aspects

of this research. W.C.S. is the recipient of a Career

Development Award from the Research to Prevent Blind-

ness Foundation (RPB). P.A.H. is a Senior Scientific

Investigator of RPB. This research was supported by NIH

grants EY06225, EY06226, EY08571, and EY6891 from

the National Eye Institute, by an RPB grant to the University

of Florida Department of Ophthalmology, and by the

Foundation Fighting Blindness, and endowment of a chair

by John A. and Florence Mattern Solomon (D.S.P.).

References

Adamus, G., Zam, Z.S., Arendt, A., Palczewski, K., McDowell, J.H.,

Hargrave, P.A., 1991. Anti-rhodopsin monoclonal antibodies of defined

specificity: characterization and application. Vision Res. 31, 17–31.

Andrews, L.D., Mackenzie, J.M., Basinger, S.F., 1984. Birefringent

periodicities in amphibian rod outer segments. Curr. Eye Res. 3,

1265–1272.

Azarian, S.M., King, A.J., Hallett, M.A., Williams, D.S., 1995. Selective

proteolysis of arrestin by calpain. Molecular characteristics and its

effect on rhodopsin dephosphorylation. J. Biol. Chem. 270,

24375–24384.

Bahner, M., Frechter, S., Da Silva, N., Minke, B., Paulsen, R., Huber, A.,

2002. Light-regulated subcellular translocation of Drosophila TRPL

channels induces long-term adaptation and modifies the light-induced

current. Neuron 34, 83–93.

Barak, L.S., Ferguson, S.G., Zhang, J., Caron, M.G., 1997. A b-arrestin/

green fluorescent protein biosensor for detecting G protein-coupled

receptor activation. J. Biol. Chem. 272, 27497–27500.

Bartoli, F., Cerutti, S., 1983. An optimal linear filter for the reduciton of

noise superimposed to the EEG signal. J. Biomed. Engng 5, 274–280.

Broekhuyse, R.M., Tolhuizen, E.F.J., Janssen, A.P.M., Winkens, H.J.,

1985. Light induced shift and binding of S-antigen in retinal rods. Curr.

Eye Res. 4, 613–618.

Chalfie, M., Kain, S., 1998. Green Fluorescent Protein. Properties,

Applications, and Protocols, Wiley, New York.

Hollyfield, J.G., Rayborn, M.E., Verner, G.E., Maude, M.B., Anderson,

R.E., 1982. Membrane addition to rod photoreceptor outer segments:

light stimulates membrane assembly in the absence of increased

membrane biosynthesis. Invest. Ophthalmol. Vis. Sci. 22, 417–427.

Kang Derwent, J.L., Qtaishat, N.M., Pepperberg, D.R., 2002. Excitation

and desensitization of mouse rod photoreceptors in vivo following

bright adapting light. J. Physiol. 541.1, 201–218.

Kaplan, M.W., Robinson, D.A., Larsen, L.D., 1982. Rod outer segment

birefringence bands record daily disc membrane synthesis. Vision Res.

22, 1119–1121.

Knox, B.E., Schlueter, C., Sanger, B.M., Green, C.B., Besharse, J.C., 1998.

Transgene expression in Xenopus rods. FEBS Lett. 423, 117–121.

Kroll, K.L., Amaya, E., 1996. Transgenic Xenopus embryos from sperm

nuclear transplantations reveal FGF signalling requirements during

gastrulation. Development 122, 3173–3183.

Kuhn, H., 1978. Light regulated binding of rhodopsin kinase and other

proteins to cattle photoreceptor membranes. Biochemistry 17,

4389–4395.

Kuhn, H., 1984a. Interactions between photoexcited rhodopsin and light-

activated enzymes in rods. Prog. Ret. Res. 3, 123–156.

Kuhn, H., Hall, S.W., Wilden, U., 1984b. Light induced binding of 48-kDa

protein to photoreceptor membranes is highly enhanced by phosphoryl-

ation of rhodopsin. FEBS Lett. 176, 473–478.

Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly

of the head of bacteriophage T4. Nature 227, 680–685.

Loeffler, K.U., Mangini, N.J., 1995. Anti-arrestin immunoreactivity in the

human retina: difference between light- and dark-adaptation. Cur. Eye

Res. 14, 1165–1168.

Mangini, N.J., Garner, G.L., Okajima, T.L., Donoso, L.A., Pepperberg,

D.R., 1994. Effect of hydroxylamine on the subcellular distribution of

arrestin (S-Antigen) in rod photoreceptors. Vis. Neurosci. 11, 561–568.

J.J. Peterson et al. / Experimental Eye Research 76 (2003) 553–563562

Mangini, N.J., Pepperberg, D.R., 1987. Localization of retinal ‘48K’ (S-

Antigen) by electron microscopy. Jpn. J. Ophthalmol. 31, 207–217.

Mangini, N.J., Pepperberg, D.R., 1988. Immunolocalization of 48K in rod

photoreceptors. Invest. Ophthalmol. Vis. Sci. 29, 1221–1233.

Marszalek, J.R., Liu, X., Roberts, E.A., Chui, D., Marth, J.D., Williams,

D.S., Goldstein, S.B., 2000. Genetic evidence for selective transport of

opsin and arrestin by kinesin-II in mammalian photoreceptors. Cell 102,

175–187.

Mata, N.L., Radu, R.A., Clemmons, R.S., Travis, G.H., 2002. Isomerization

and oxidation of vitamin A in cone-dominant retinas: a novel pathway

for visual-pigment regeneration in daylight. Neuron 36, 69–80.

McDonald, P.H., Lefkowitz, R.J., 2001. BetaArrestins. New roles in

regulating heptahelical receptors’ functions. Cell Signal 13, 683–689.

McDowell, J.H., 1993. Preparing rod outer segment membranes, regener-

ating rhodopsin, and determining rhodopsin concentration. Meth.

Neurosci. 15, 123–130.

McDowell, J.H., Smith, W.C., Miller, R.L., Popp, M.P., Arendt, A.,

Abdulaeva, G., Hargrave, P.A., 1999. Sulfhydryl reactivity demon-

strates different conformational states for arrestin, arrestin activated by

a synthetic phosphopeptide, and constitutively active arrestin. Bio-

chemistry 38, 6119–6125.

McGinnis, J.F., Matsumoto, B., Whelan, J.P., Cao, W., 2002. Cytoskeleton

participation in subcellular trafficking of signal transduction proteins in

rod photoreceptor cells. J. Neurosci. Res. 67, 290–297.

McGinnis, J.F., Stepanik, P.L., Lerious, V., Whelan, J.P., 1991. Molecular

mechanisms regulating cellular and subcellular concentrations of gene

products in mouse visual cells. In: Hollyfield, J., Anderson, R.E.,

LaVail, M.M. (Eds.), Retinal Degenerations, CRC, Boca Raton, FL, pp.

503–516.

McGinnis, J.F., Whelan, J.P., Donoso, L.A., 1992. Transient, cyclic

changes in mouse visual cell gene products during the light-dark cycle.

J. Neurosci. Res. 31, 584–590.

Mirshahi, M., Thillaye, B., Tarraf, M., de Kozak, Y., Faure, J., 1994. Light-

induced changes in S-antigen (Arrestin) localization in retinal

photoreceptors: differences between rods and cones and defective

process in RCS rat retinal dystrophy. Eur. J. Cell Biol. 63, 61–67.

Moritz, O.L., Tam, B.M., Knox, B.E., Papermaster, D.S., 1999.

Fluorescent photoreceptors of transgenic Xenopus laevis imaged in

vivo by two microscopy techniques. Invest. Ophthalmol. Vis. Sci.

40, 3276–3280.

Moritz, O.L., Tam, B.M., Papermaster, D.S., Nakayama, T., 2001. A

functional rhodopsin-green fluorescent protein fusion protein localizes

correctly in transgenic Xenopus laevis retinal rods and is expressed in a

time-dependent pattern. J. Biol. Chem. 276, 28242–28251.

Nir, I., Ransom, N., 1993. Ultrastructural analysis of arrestin distribution in

mouse photoreceptors during dark/light cycle. Exp. Eye Res. 57,

307–318.

Organisciak, D.T., Xie, A., Wang, H.M., Jiang, Y.L., Darrow, R.M.,

Donoso, L.A., 1991. Adaptive changes in visual cell transduction

protein levels: effect of light. Exp. Eye Res. 53, 773–779.

Palczewski, K., Buczylco, J., Imami, N.R., McDowell, J.M., Hargrave,

P.A., 1991. Role of the carboxyl-terminal region of arrestin in binding

to phosphorylated rhodopsin. J. Biol. Chem. 266, 15334–15339.

Palczewski, K., Van Hooser, J.P., Garwin, G.G., 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, 12012–12019.

Philp, N.J., Chang, W., Long, K., 1987. Light stimulated protein movement

in rod photoreceptor cells of the rat retina. FEBS Lett. 225, 127–132.

Roof, D.J., Heth, C.A., Brann, M.R., 1988. Expression of transducin in

retinal rod photoreceptor outer segments. Science 241, 845–847.

Sokolov, M., Lyubarsky, A.L., Strissel, K.J., Govardovskii, V.I.,

Savchenko, A.B., Govardovskii, V.I., Pugh, E.N. Jr., Arshavsky,

V.Y., 2002. Massive light-driven translocation of transducin between

the two major compartments of rod cells: a novel mechanism of light

adaptation. Neuron 33, 95–106.

Tam, B.M., Moritz, O.L., Hurd, L.B., Papermaster, D.S., 2000. Identifi-

cation of an OS targeting signal in the COOH terminus of rhodopsin

using transgenic Xenopus laevis. J. Cell Biol. 151, 1369–1380.

Whelan, J.P., McGinnis, J.F., 1988. Light-dependent subcellular movement

of photoreceptor proteins. J. Neurosci. Res. 20, 263–270.

Whitehead, J.L., Wang, S.Y., Bost-Usinger, L., Hoang, E., Frazer, K.A.,

Burnside, B., 1999. Photoreceptor localization of the KIF3A and KIF3B

subunits of the heterotrimeric microtubule motor, kinesin II in

vertebrate retina. Exp. Eye Res. 69, 491–503.

Williams, M.A., Mangini, N.J., 1991. Immunolocalization of arrestin (S-

antigen) in rods of pearl mutant and wild-type mice. Curr. Eye Res. 10,

457–462.

Wilson, C.J., Applebury, M.L., 1993. Arresting G-protein coupled receptor

activity. Curr. Biol. 3, 683–686.

Zolotukhin, S., Potter, M., Hauswirth, W.W., Guy, J., Muzyczka, N., 1996.

A ‘Humanized’ green fluorescent protein cDNA adapted for high-level

expression in mammalian cells. J. Virol. 70, 4646–4654.

J.J. Peterson et al. / Experimental Eye Research 76 (2003) 553–563 563