Nature Genetics
20, 344 - 351 (1998)
doi:10.1038/3813
Rpe65 is necessary for production of 11-cis-vitamin A in
the retinal visual cycleT. Michael Redmond.1, Shirley Yu1, Eric Lee2, Dean Bok3, Duco Hamasaki4, Ning Chen5, Patrice Goletz5, Jian-Xing Ma5, Rosalie K. Crouch5
& Karl Pfeifer21 Laboratory of Retinal Cell and Molecular Biology, National
Eye Institute, Bethesda, Maryland 20892, USA. 2 Laboratory of Mammalian Genes and Development, National
Institute of Child Health and Human Development, National Institutes of Health,
Bethesda, Maryland 20892, USA. 3 Jules Stein Eye Institute, Brain Research Institute
and Department of Neurobiology, UCLA School of Medicine, Los Angeles,
California 90095, USA. 4 Bascom-Palmer Eye Institute, University of Miami School
of Medicine, Miami, Florida 33101, USA. 5 Storm Eye Institute, Medical University of South Carolina,
Charleston, South Carolina 29425, USA.
Correspondence should be addressed to T. Michael Redmond. redmond@helix.nih.govMutation of RPE65 can cause severe blindness from birth or early
childhood, and RPE65 protein is associated with retinal pigment epithelium
(RPE) vitamin A metabolism. Here, we show that Rpe65-deficient mice
exhibit changes in retinal physiology and biochemistry. Outer segment discs
of rod photoreceptors in Rpe65−/− mice are
disorganized compared with those of Rpe65+/+ and
Rpe65+/− mice. Rod function, as measured by electroretinography,
is abolished in Rpe65−/− mice, although
cone function remains. Rpe65−/− mice lack
rhodopsin, but not opsin apoprotein. Furthermore, all-trans-retinyl
esters over-accumulate in the RPE of Rpe65−/−
mice, whereas 11-cis-retinyl esters are absent. Disruption of the RPE-based
metabolism of all-trans-retinyl esters to 11-cis-retinal thus
appears to underlie the Rpe65-/- phenotype, although
cone pigment regeneration may be dependent on a separate pathway.Introduction
RPE65, a protein preferentially and abundantly expressed in the RPE (1), is essential for the maintenance of normal vision.
Mutations in RPE65 result in severe forms of early-onset retinal dystrophy,
including Leber congenital amaurosis (LCA; OMIM 180069 and 204100; http://www.ncbi.nlm.nih.gov/Omim) and autosomal recessive, childhood-onset severe retinal dystrophy2,
3,
4,
5 (arCSRD). Mutations in this gene may account for 15% of
cases of LCA, a genetically heterogeneous disorder, in North America5. RPE65 has been localized to human chromosome 1p31 (Refs 6,7) and its mouse homologue
(Rpe65) to distal mouse chromosome 3 (6).
By linkage analysis, certain arCSRDs were localized to the interval occupied
by RPE65 and were shown to be associated with mutations in RPE65
(3). In addition, candidate gene analysis
of RPE65 identified mutations in patients with LCA (4,5). Common features
of these patients include loss of vision from birth or early childhood associated
with complete night-blindness, extinguished rod electroretinography and severely
compromised cone responses, all suggestive of a role for RPE65 in retinal
function.
Although the precise function of RPE65 is not known, a role in vitamin
A metabolism is suspected2,
8,
9. Evidence for this is provided
by its biochemical association with retinol-binding protein9
(RBP) and 11-cis-retinol dehydrogenase10. RPE65, also
known as p63 (9), has been proposed to be
the RPE RBP receptor9. Association with 11-cis-retinol
dehydrogenase suggests that RPE65 is part of the visual cycle pathway, the
process by which the 11-cis-retinal chromophore of visual pigments
is photo-isomerized to the all-trans-isomer, which is then regenerated
in the dark11,
12. The chromophore of mammalian opsins, 11-
cis-retinal, is restricted to the eye. In the vertebrate eye, the major
site of production of 11-cis-retinoids is the RPE (Refs 13,14). The molecular identity of the
retinoid isomerase or isomerohydrolase, despite years of study, is still unknown15.
In view of the potentially important role of RPE65 in the physiology of
vision and perhaps the visual cycle, we generated a targeted disruption of
mouse Rpe65, as a natural mouse mutation is not known. In parallel
with the human diseases described above, mutant mice showed severe changes
in visual function. This may be related to supply of the 11-cis-retinal
chromophore, as rhodopsin was absent from Rpe65-deficient mice, although
opsin was present. These changes were restricted to the rod photoreceptors,
leaving cones relatively unaffected.
Results
Disruption of mouse Rpe65
We designed a gene
targeting protocol for the disruption of mouse Rpe65 that replaced
a 5´ region containing 1.1 kb of 5´ flanking region, exons 1,
2 and 3, intervening introns and 0.5 kb of intron c with the PGK-neo gene
(Fig. 1a), along with loss of a HindIII
site. Rpe65+/+ mice showed only the 4.9-kb HindIII
fragment on Southern-blot analysis and Rpe65−/−
only the mutant 5.9-kb HindIII fragment, whereas Rpe65+/−
progeny had a single copy of each (Fig. 1b
). A normal mendelian distribution of genotypes was observed, indicating
no effect of the mutation on viability. To confirm that transcription of
Rpe65 was disrupted in Rpe65−/− mice,
RT-PCR was performed on total RNA isolated from whole mouse eyes. No Rpe65
RT-PCR product was seen in Rpe65−/−
mice, although the Gapd control transcript was amplified (
Fig. 1c). As the Rpe65 transcript region amplified was
downstream of the disruption (exons 4-6), supernumerary transcription of a 5´-truncated
mutant mRNA could be discounted.
 | | Figure 1. Targeted disruption of the Rpe65 locus. | | |  | a, Strategy: (i), 5' end of wild-type Rpe65; (ii), targeting
vector; (iii), targeted locus; (iv), sizes of HindIII fragments from
wild-type and mutant loci, respectively, hybridized to probe indicated in
(iii). H, HindIII sites; E, EcoRI restriction sites. b,
Southern-blot analysis of genomic DNA from progeny of a heterozygous cross
digested with HindIII, using a flanking probe showing wild-type (4.9
kb) and/or mutant (5.9 kb) HindIII restriction fragments (a,
iv). c, RT-PCR analysis of Rpe65 mRNA expression in progeny
of a heterozygous cross. The left set of lanes are results of Rpe65-specific
RT-PCR on cDNA reverse transcribed from total RNA of whole eyes from Rpe65
+/-, Rpe65−/− and Rpe65
+/+ mice, whereas the right set are results of Gapd-specific
RT-PCR on the same cDNA. The Rpe65 mRNA-specific, 366-bp band is absent
from Rpe65−/−, whereas the Gapd-specific
452-bp band is present; S, 100-bp ReadyLoad DNA Ladder.
 Full Figure and legend (15K) |
| |
Subtle photoreceptor changes in Rpe65−/−
mice
By light microscopy, changes in the RPE of 7-week-old and
15-week-old Rpe65−/− mice were minimal compared
with Rpe65+/+ or Rpe65+/−
littermates (Fig. 2), although in Rpe65−/−
animals there were subtle and slowly progressing changes in the
photoreceptor cells. No differences were seen between Rpe65+/+
and Rpe65+/− littermates. At seven weeks of age,
the outer segment (OS) region of Rpe65−/−
photoreceptors stained less densely with toluidine blue than did those of
Rpe65+/+ or Rpe65+/− littermates
(Fig. 2a-c). This could be interpreted
as loss of, or less dense packing of, OS discs. This was more evident by 15
weeks (Fig. 2d,e), by which time the length
of the outer segments was significantly decreased in Rpe65−/−
mice and more voids were seen, consistent with loss of OS material.
A slow degeneration was also seen with loss of photoreceptor nuclei. In 7-week-old
and 15-week-old Rpe65+/+ and Rpe65+/−
mouse retinae, there were 10-11 layers of nuclei (
Fig. 2a,b,d), comparable with retina from 7-week-old
Rpe65−/− retina (Fig. 2c
). By 15 weeks of age, the number of nuclear layers had decreased
to 8-9 (Fig. 2e) in Rpe65−/−
retina, and by 28 weeks to 7 (data not shown).
| | |
By electron microscopy (EM), the OS of rod photoreceptors appear to be
slightly altered at seven weeks of age compared with those of wild type (Fig. 3a,b). Rod photoreceptor OS discs in
Rpe65−/− mice are not as tightly packed as
in Rpe65+/+ and Rpe65+/−
littermates, resulting in more apparent voids in the sections. This feature
contributes to the more disorganized appearance of these Rpe65−/−
OS (Fig. 3a,b). The Rpe65
−/− OS were also shorter in length than those of wild
type. At 15 weeks of age, the OS of Rpe65−/−
mice were more disorganized than at 7 weeks (Fig. 2e,
and data not shown). OS morphogenesis thus appears to be subtly affected initially
by Rpe65-deficiency in the RPE; OS discs are elaborated and deployed,
as in wild type, although they are not as regularly packed. In older animals
the level of disorganization increases. Alterations in OS phagocytosis by
the RPE were not evident. In the RPE of 15-week-old Rpe65−/−
mice, numerous lipid inclusions were also present (
Fig. 8d).
| | |
| | |
| | |
| | |
No gross changes were seen in tissues or organs outside the eye. To address
the possibility that a generalized systemic effect on vitamin A-requiring
organs occurred, other organs containing retinol-dependent cells, including
kidney and testis, were analysed for histological changes. No differences
were observed in Rpe65−/− mice when compared
with Rpe65+/+ or Rpe65+/−
littermates (data not shown).
Rod photoreceptor ERG is abolished in Rpe65−/−
mice
The electroretinograms (ERGs) recorded from dark-adapted
Rpe65+/+ and Rpe65+/− mice
were typical of ERG responses recorded from other strains of normal mice with
normal a-, b- and c-waves. The dark-adapted ERG responses of the Rpe65
−/− mice on the other hand were very different; the
a-wave was very small, the c-wave was not evident and the b-wave was small
and had fast rising and falling limbs. The mean and standard error of the
mean of the b-wave amplitude for Rpe65+/+, Rpe65
+/− and Rpe65+/− mice are shown
(Fig. 4). The full intensity stimulus was reduced by
use of neutral density (ND) filters. For both Rpe65+/+
and Rpe65+/− mice, a small (20-30  V) b-wave
was elicited by a stimulus of ND=6.0 (1.59 10-3 cd/m
2). Further increases in the stimulus intensity increased the b-wave
amplitude until it saturated at a stimulus intensity of ND=3.0 (1.59 cd/m
2). Additional increases in the stimulus intensity did not change
the amplitude until the stimulus intensity was ND=2.0 (1.5910 cd/m
2) when the photopic segment of the b-wave was evoked. The mean of
the maximum b-wave amplitude was 960.0 V for Rpe65+/+
and 937.6 V for Rpe65+/− mice. The mean intensity:response
curves for these mice did not differ significantly and resembled the curves
reported previously for mice and other animals. The mean intensity:response
curve for Rpe65−/− mice was very different;
the curve was displaced to higher intensities by approximately 4.5 log units,
and the maximum b-wave amplitude was 256.0 V. Thus, there is a very reduced
dark-adapted ERG response in Rpe65-deficient mice; although there is
a response to high intensity light. In the mouse, as in other animals, the
ERG response in the dark-adapted state is mostly due to rod activity, however,
under bright light conditions, rod activity is suppressed and only the cone
pathway contributes to the ERG (16).
Next, ERG responses were recorded under light-adapted or photopic conditions
by increasing stimulus intensities in the dark, and then with increasing levels
of steady background illumination (Fig. 5a).
In 14-week-old Rpe65+/+ mice in the dark (No Bkgd),
the amplitude of the b-wave increased (Fig. 4). With
an ND=6.0 steady background illumination, the intensity:response curve was
shifted to the right (higher intensities) by approximately 0.8 log units,
and the b-wave amplitude increased. Further increases in the background intensity
shifted the curve further to the right. Similar data were obtained from heterozygous
mice (data not shown). In 14-week-old Rpe65−/−
animals, however, the intensity:response curves for the eye in the dark, and
with background intensities of ND=4.0, 3.0 and 2.0, were approximately the
same. These findings suggest that the ERG responses of Rpe65−/−
mice are cone-driven; a plot of the mean intensity:response curve
for the Rpe65+/+ mouse with a ND=2.0 background intensity
was not significantly different from that of the Rpe65−/−
mouse obtained in the dark (data not shown). Further demonstration
that the ERGs from Rpe65−/− mice are cone-driven
comes from the shapes of the ERGs. The shape of the ERG responses of the
Rpe65−/− mouse elicited by the full intensity
stimulus in the dark closely resembles the photopic ERG responses of the
Rpe65+/+ mouse with a -2.0 background (
Fig. 5b, inset) and photopic ERGs previously reported for mice.
Additional evidence that the ERG responses of the Rpe65-deficient mouse
were cone-driven came from flicker testing. Two Rpe65−/−
mice responded only to high frequency flicker (in the cone response
range) and only at higher stimulus intensities, whereas Rpe65+/+
and Rpe65+/− mice responded to low
frequency flicker (rod range) only at low stimulus intensity and to all frequencies
at high stimulus intensity, as expected.
Rhodopsin is absent from Rpe65−/−
mouse retina
To better understand the total abrogation of rod
ERG response, we analysed the rhodopsin status of dark-adapted Rpe65
−/− mice compared with Rpe65+/+
or Rpe65+/− littermates (Fig. 6
a). Total dark-adapted retina rhodopsin was detergent extracted
and the unregenerated difference spectrum measured to assess endogenous rhodopsin.
A typical rhodopsin absorption maximum was observed at 500 nm. Normal amounts
of total rhodopsin were detergent extracted from Rpe65+/+
mouse retina of different ages (Fig. 6a, trace
A for a representative spectrum). Less rhodopsin was extractable from
Rpe65+/− littermates (Fig. 6a,
trace B for a representative spectrum), however, no rhodopsin was extractable
from dark-adapted Rpe65−/− mice retina of
different ages, as shown by the flat difference spectrum (
Fig. 6a, trace C for a representative spectrum). When exogenous
11-cis-retinal chromophore was added to rhodopsin extracts, some regeneration
of rhodopsin was seen in Rpe65−/− mouse
extracts, as shown by the slight increase in the difference spectrum (Fig. 6b, trace C), indicating that at least a fraction
of the opsin was structurally intact. Conversely, the limited increase suggests
that much of the opsin is not able to be regenerated. Considering its relatively
normal appearance in 7-week-old animals, morphogenesis of the Rpe65
−/− OS does not appear to be affected by lack of rhodopsin.
As a further measure of photoreceptor function, apoprotein opsin was measured
in Rpe65−/−, Rpe65+/+
and Rpe65+/− littermates. Opsin was detected by
immunoblot in all three genotypes (Fig. 7a, top;
Rpe65+/+>Rpe65+/−>Rpe65
−/−). The integrity of the photoreceptors was shown
by immunoblot using an antibody to the  -subunit of transducin, G
 t (Fig. 7a, bottom). Opsin was located
in the outer segments of the rod photoreceptors, the physiologically appropriate
compartment, in Rpe65+/− (Fig.
7b) and in Rpe65−/− (Fig. 7c) mice, as shown by opsin immunoelectron microscopy.
All-trans-retinyl esters over-accumulate in Rpe65
−/− RPE
To determine whether the evident
lack of 11-cis-retinal in the rod photoreceptors was due to lack of
uptake of vitamin A by the RPE, uptake of vitamin A but lack of transport
from the RPE of 11-cis-retinal (resulting in accumulation of 11-
cis-retinyl esters), or uptake of vitamin A but lack of synthesis of 11-
cis-retinoids, we performed HPLC analysis of RPE retinoids extracted from
Rpe65+/+ (Fig. 8a) and
Rpe65−/− (Fig. 8b)
mice. There was an over-accumulation of total retinyl esters in 15-week-old
Rpe65−/− mice. Accumulated esters in the
Rpe65−/− RPE were found to be all-trans
-retinyl ester by ester saponification (Fig. 8b,
inset). No peak was detected at the mobility of authentic 11-cis-retinol
standard (Fig. 8b, denoted by solid arrow). As
expected, a much smaller all-trans-retinol peak was seen in Rpe65
+/+ samples (Fig. 8a, inset), but
no 11-cis-retinol peak was detectable under the conditions used. As
11-cis-retinyl ester has been previously found to comprise only 5%
of total retinyl ester in normal 13-week-old mouse RPE (17), lack of detection of 11-cis-retinol in saponified
Rpe65+/+ samples was expected given the small amount of
retinyl ester present. Similarly, given the increased amount of retinyl ester
in the Rpe65−/− RPE (Fig.
8c), if 11-cis-retinyl ester were present in the same
relative 5% abundance, it should be detectable. On the basis of these data,
however, we cannot exclude the possibility that very low levels of 11-
cis-retinoids are present in the RPE of Rpe65-deficient mice. A
minor amount of the 13-cis-isomer (Fig. 8b,
peak 1), derived from sample handling, was detected. This selective over-accumulation
of all-trans-retinyl esters may be associated with morphological changes
evident in the RPE of Rpe65−/− mice. A feature
of the Rpe65−/− RPE at later ages, such
as 15 weeks, is the presence of numerous inclusions visible at the electron
microscope level (Fig. 8d) that appear to be
lipid droplets. They range in form from electron-lucent to somewhat electron-opaque.
This suggests that they contain lipid which has been partially or completely
extracted by histological processing. In addition, 15-week-old Rpe65
−/− RPE contained other unidentified structures with
electron-translucent centers (Fig. 8d).
Discussion
The results presented here confirm the role of RPE65 in the vitamin A metabolism
of the retina with the finding that the Rpe65-deficient mouse lacks
rhodopsin, although it appropriately expresses the opsin apoprotein in the
rod OS. This evident lack of 11-cis-retinal chromophore occurs alongside
the over-accumulation of all-trans-retinyl esters in the RPE and implicates
a block in the RPE visual cycle. At the same time, it appears that no 11-
cis-retinyl ester accumulates. Although isomerohydrolase activity can
be somewhat inhibited by 11-cis-retinol14, storage of
11-cis-retinyl ester does occur in mouse RPE (17), given an active isomerization step. This concurrent lack of
rhodopsin and over-accumulation of all-trans-retinyl ester (but apparently
not of 11-cis-retinyl ester) in the Rpe65-deficient mouse RPE
argues for a block in isomerization. Either one of these situations alone,
unaccompanied by the other, would implicate a block in transport of retinoids,
either into or out of the RPE into the photoreceptors. On the other hand,
11-cis-retinoids are, apparently, not completely absent from Rpe65
-deficient mouse retina, as cone photoreceptors are still functional,
although the availability of this chromophore pool appears to be restricted
to cones.
At the same time, our data argue against the hypothesis that RPE65 (9) is the RPE RBP receptor, because its absence does
not prevent uptake of vitamin A by the RPE, although an RBP receptor has been
identified in a variety of tissues, including the RPE, by binding studies18,
19,
20. Alternatively, cellular uptake of retinol by the RPE may
be governed by the level of intracellular binding proteins (high in the RPE; 21) and the rate of metabolism of vitamin A in target
cells22. In the RPE of Rpe65-deficient mice, all-
trans-retinol uptake is unimpaired, whereas over-accumulation of all
-trans-retinyl esters may result in the RPE lipid inclusions in 15-week-old
Rpe65-deficient mice, similar to those seen in the RPE of mice injected
with all-trans-retinyl esters23. Retinyl esters also
accumulate in the lipid droplets of frog RPE (24).
Although many components of the mammalian visual cycle have been characterized
at the molecular level, including rhodopsin25, interphotoreceptor
retinoid-binding protein26 (IRBP), ROS all-trans-retinol
dehydrogenase27, lecithin:retinol acyltransferase28
(LRAT; which has been cloned; D.B., unpublished data), 11-cis-retinol
dehydrogenase10 and 11-cis specific cellular retinaldehyde-binding
protein29,
30 (CRALBP), the all-trans: 11-cis-retinol
isomerase or isomerohydrolase13,
14,
15 remains incompletely
characterized. Association of RPE65 with the putative 11-cis-retinoid-generating
enzyme complex of the RPE is evident from its association with 11-cis-retinol
dehydrogenase9. In addition, the first appearance of RPE65 (1) coincides with the first appearance of 11-cis
-retinoids and retinol isomerase activity in the developing mouse and
rat retina17,
31. Our data indicate that metabolism of all-
trans-retinyl ester to 11-cis-retinoid is blocked in the RPE in
the absence of RPE65.
Rpe65-deficient mice are, in effect, functionally rhodopsin deficient.
Although both Rpe65−/− and Rho
−/− mice32 lack rhodopsin, the morphological
phenotypes expressed are distinct. In Rpe65−/−
mice, rhodopsin holoprotein is not required for OS morphogenesis, whereas
in Rho−/− mice, synthesis of opsin apoprotein
is essential. Despite the presence of appropriately localized opsin in
Rpe65−/− mouse retina, our results suggest
that it may not be capable of regeneration. Although it may not be necessary
for the initial synthesis of opsin, 11-cis-retinal may be required
for its continued stability. Thus, chronic lack of chromophore may affect
the ability of opsin to be regenerated in the OS. Further study of this possibility
is required. Additionally, whereas both mutant mouse rod photoreceptors are
non-functional, the OS-lacking Rho−/−photoreceptors
degenerate far faster than OS-bearing Rpe65−/−
photoreceptors32.
A central feature of the Rpe65-deficient phenotype is the abolition
of rod ERG responses. The OS of Rpe65−/−
are well-preserved compared with the OS of Rbp3−/−
mice, which have a rod ERG response, implying movement of retinoids despite
lack of the transport protein33. Thus even when transport is
impaired, some chromophore can still reach the photoreceptors. Although lowered
levels of rhodopsin (and opsin) are seen in Rpe65+/−
animals compared with Rpe65+/+ animals, their OS morphology
is identical and their ERG responses are normal, suggesting that rod photoreceptor
cells have a wide range of tolerance for rhodopsin level. This is in agreement
with similar findings in Rho+/− mice32.
In contrast, cone ERG function in the Rpe65-deficient mice is preserved.
This finding corroborates the conclusion that cone and rod pathways of regeneration
may be different or separate34 . These results suggest that
two separate visual cycles may operate in the mammalian retina: one centred
in the RPE supplying the rod photoreceptors and another supplying the cones.
Cultured chicken Muller retinal glial cells have been shown to secrete 11-
cis-retinol into media35 and may play a role. Furthermore,
mammalian cones, unlike rods, can resensitize with added 11-cis-retinol36, and amphibian cones34,
37 and cone-like pigments38 separated from the RPE can be regenerated. The recent identification
of RPE65-like mRNA expression in cone, but not rod, photoreceptors
of salamander is supportive of this39.
The Rpe65-deficient mouse is the first in vivo evidence that
a distinct mammalian cone-associated visual cycle, or mechanism of regeneration,
may exist. This raises differences between the Rpe65−/−
mouse and the human LCA phenotypes, as in the latter, both rod
and cone function are abolished or severely compromised3,
4,
5.
This may be due to the relative age at which ERGs are measured in mice and
human patients, or from species-specific differences in retinal anatomy and
cone distribution. Mice have two types of cone cells scattered throughout
the retina, one a middle-wave (510 nm) cone and the other a short-wave (370
nm)-sensitive cone that may be homologous to the human blue cone40.
In the human retina, red- and green-sensitive cone cells are concentrated
in the central area of high cone density known as the fovea, whereas blue-sensitive
cones are scattered throughout the retina41. Perhaps this very
dense concentration of highly active cones is not amenable to supply by the
associated Muller cells and is therefore dependent on the RPE for regeneration.
These Rpe65-deficient mice provide a model system to study potential
therapies for human diseases caused by RPE65 defects. RPE cells take
up both adenovirus42,
43 and adeno-associated virus44,
allowing the possibility of rescue experiments. As a therapeutic strategy,
RPE transplantation is easier than that of photoreceptors, and in addition,
the ultimate biochemical outcome of Rpe65 deficiency, lack of 11-
cis retinoids, may be amenable to pharmacological intervention.
Blockage of the metabolism of all-trans-retinyl ester to 11-
cis-retinol underlies the Rpe65-/- phenotype, indicating
that RPE65 may be the effector enzyme in the isomerization step yielding 11-
cis-retinol. Alternatively, RPE65 may be involved in a previously unknown
step required for isomerization. RPE65 is an obligate component of the isomerization
reaction, essential to both vision and the visual cycle. Continued function
of cones in the Rpe65-deficient mouse, however, provides new insight
into cone pigment regeneration.
Methods
Generation of Rpe65-disrupted mice and genotype analysis.
A P1 clone containing the entire mouse Rpe65 was isolated
from a 129SV mouse genomic library (Genome Systems). From this, a 7.7-kb
EcoRI fragment containing exons 1, 2 and 3, and 2.8 kb of 5´ flanking
region, was subcloned into Bluescript II SK (-) (T.M.R., unpublished data).
A 1.7-kb EcoRI/AflII fragment from the 5´ flanking region
was subcloned into pLPG9 and linearized as an EcoRI/BamHI fragment.
A downstream 2.2-kb Ecl136II/EcoRI fragment from intron c was
inserted into pLPG9 and linearized as an XhoI/NotI fragment.
These linearized segments were inserted into X-pPNT to generate pPNT-Rpe65.
pPNT-Rpe65 was linearized with NotI and electroporated into R1 ES cells45. Two ES clones with a targeted disruption were identified out of
31 clones isolated by the positive-negative G418/gancyclovir selection method.
Targeted clones were identified by Southern-blot analysis after HindIII
digestion of genomic DNA, using a flanking probe 3´ to the replacement.
The wild-type allele was 4.9 kb, the targeted allele 5.9 kb. Clone ES-29 was
microinjected into C57Bl/6 blastocysts and five chimaeric mice were produced.
These chimaeras were bred with C57Bl/6 mice to generate progeny heterozygous
for the mutation. Heterozygous mice were crossed to produce Rpe65
+/+, Rpe65+/− and Rpe65
−/− mice. Germline transmission of the mutation was determined
by Southern-blot analysis of tail DNA, as described above. All animal experiments
were carried out under protocols approved by institutional Animal Care and
Use Committees and followed the NIH "Using Animals in Intramural Research"
guidelines.
Rpe65 mRNA Assay.
Total RNA was isolated from
whole mouse eyes using an RNeasy kit (Qiagen) and reverse transcribed using
Moloney mouse leukaemia virus reverse transcriptase (Retroscript, Ambion).
cDNA was PCR-amplified using mouse Rpe65 exon 4 forward primer (5´-ATGATCGAGAAGAGGATTGTC-3´)
and exon 6 reverse primer (5´-CTGCTTTCAGTGGAGGGATC-3´. Primers
specific for mouse Gapd mRNA were: forward (starting at nt 586), 5´-ACCACAGTCCATGCCATCAC-3´
and reverse (starting at nt 1037), 5´-TCCACCACCCTGTTGCTGTA-3´.
Cycle conditions used were: hot start at 94 °C, followed by 35 cycles
(94 °C, 30 s; 55 °C, 30 s; 72 °C, 30 s). Product sizes were 366
and 452 bp for Rpe65 and Gapd, respectively.
Light and electron microscopy.
Tissues for light and
electron microscopy were processed as described46. Thin (0.5 m)
sections were cut with glass knives and stained with toluidine blue for light
microscopy. Ultrathin sections were cut with a diamond knife and stained with
uranium and lead salts for electron microscopy. For immunocytochemistry, mouse
eyes were processed as previously described46. Ultrathin sections
were incubated for 1 h on affinity-purified anti-bovine opsin polyclonal antibodies
diluted in 0.1 M sodium phosphate buffer (pH 7.2), containing 1% bovine serum
albumin (BSA-phosphate buffer). The grids with their sections were washed
in buffer and floated for 1 h on 1 part 18 nm colloidal gold adsorbed to affinity-purified
goat antirabbit IgG (Jackson ImmunoResearch) and 15 parts BSA-phosphate buffer.
Tissue contrast was enhanced by floating the specimens on 5% aqueous uranyl
acetate for 10 min and Sato's lead stain for 3 min.
Electroretinography.
Before measuring, ERG mice were
anaesthetized with a mixture of ketamine:xylazine:urethane, and pupils were
dilated with phenylephrine and atropine. ERGs were recorded with a wick-Ag:AgCl
electrode placed on the cornea, a subcutaneous hypodermic needle (30 gauge)
on the head as reference electrode and a ground placed subcutaneously in the
neck region, all fed to a Tektronix A39 preamplifier with the half-amplitude
bandpass set at DC to 3 kHz and displayed on an oscilloscope, and also fed
to a computer running a Biopac M100 signal averaging program. The stimulus
light source was either a quartz halogen lamp bulb (maximum stimulus intensity=1.5910
3 cd/m2) or a PS22 Grass stroboscopic flash unit (maximum
stimulus intensity=5.52102 cd/m2 set
at 8 and 40 Hz). The filament of the halogen lamp bulb was focused in the
plane of a Uniblitz shutter (Vincent Assoc.), and another lens focused the
filament on the tip of one arm of a bifurcated, randomized fiber optic bundle.
The Grass stroboscopic bulb was focused on the tip of the other arm. The common
arm of the bundle was brought into the Faraday cage and the tip was placed
1-2 mm from the cornea. The stimulus intensity was measured with a photometer
(UDT Instruments) with the detector placed at the position of the cornea.
Neutral density (ND) filters were used to reduce the full-intensity stimulus.
The stimulus intensity was increased in 0.5 log unit steps, and two responses
were averaged at the lower stimulus intensities (ND=6.0-3.5). Only one response
was recorded at the higher stimulus intensities (ND=3.0-0). For flicker testing,
responses were recorded to 2.5, 5, 7.5,10, 15 and 20 flashes/s at increasing
stimulus intensities.
Rhodopsin assay.
All procedures were carried out under
red lights. Mice were killed, their eyes removed and retinae dissected and
stored at -70 °C until required. Retinae were homogenized in 0.5 ml of
Tris-HCl (10 mM), EDTA buffer (1.0 mM, pH 7.5) containing PMSF (1.0 mM) and
DNase (10 g) by a series of syringe triturations using progressively smaller
needle sizes. (18G, 20G, then 26G). Syringes and tubes were rinsed with an
additional buffer (0.5 ml) and pooled with the homogenate. This was centrifuged
at 88,200g at r max for 15 min. For regeneration experiments,
the pellet is first resuspended in PBS and a fivefold excess of 11-cis
-retinal added (~4 nM), incubated for 1 h at 4 °C on a rotator, centrifuged
as above and the recovered pellet resuspended in 5% BSA in PBS and incubated
for 30 min on a rotator at 4 °C. The regenerated sample was centrifuged
as above to recover the pellet. Otherwise, after a PBS wash, the pellet was
resuspended in 100 l of 1% dodecyl maltoside in PBS and solubilized overnight
at 4 °C on a rotator. Unsolubilized material was then removed by centrifugation
at 109,000g at r max for 15 min. Supernatant was analysed by
spectroscopy on the WPI SpectroPette microspectrophotometer (World Precision
Instruments). A total of five spectra were summed for each sample (both dark
and bleached). For the bleached spectrum, light exposure was limited to 30
s due to extremely small sample volume (60 nanoliters per mm light path length).
Data were exported into Kaleidagraph for difference spectra calculations.
Western
-blot analysis. The membrane fractions
of single retinae were dissolved in 100 l PBS/1% dodecyl maltoside. For
anti-opsin immunoblot analysis, these extracts were diluted 1:40 and 1 l
aliquots were resolved with SDS-PAGE (8-16% acrylamide gel) along with COS-1
cell membrane extract (50 ng) and purified bovine rhodopsin (5 ng) and electrotransferred
onto Hybond-ECL nitrocellulose membrane (Amersham; 25 V, 2 h). For anti-transducin -subunit
immunoblot analysis, aliquots (10 l) of undiluted retina membrane fraction
extracts were used. The membranes were blocked by incubation with 5% (w/v)
BLOTTO/TBST (20 mM Tris-Cl, pH 7.6, 137 mM NaCl, 0.1% Tween 20) for 2 h, and
incubated with anti-rhodopsin monoclonal antibody 1D4 (22.5 g/ml in BLOTTO/TBST; 47) or anti-transducin -subunit (Calbiochem;
1:200 in BLOTTO/TBST) overnight at 4 °C. Membranes were washed three times
(15 min each at 25 °C) with BLOTTO/TBST and incubated in secondary antibody
(horseradish peroxidase-conjugated rabbit anti-mouse IgG (Amersham) diluted
(1:10,000) in BLOTTO/TBST for anti-opsin, or horseradish peroxidase-conjugated
goat anti-rabbit IgG (Amersham) diluted (1:2,000) in BLOTTO/TBST for anti-transducin
subunit) for 3 h at 25 °C. The membranes were washed four times in TBST
followed by 1 min incubation in the ECL (Amersham) developer substrate. The
membranes were drained, covered with SaranWrap and exposed to a Kodak X-Omat
film.
Retinoid and retinyl ester HPLC.
Retinyl esters were
extracted with methanol/hexane from dark-adapted whole mouse eyes or dissected
mouse eye tissue homogenized in phosphate buffer17 (0.1 M, pH
7.0). Retinyl esters were separated on a Lichrosorb SI-60 normal phase HPLC
column (Alltech) in an Applied Biosystems or Waters HPLC system, using hexane/dioxane48 or hexane/methyl t-butyl ether17 mobile phase elution
with detection at 320 nm. Experimental data were compared with 11-cis-retinal
and 11-cis-retinol (synthesized by RKC), all-trans-retinol and
all-trans-retinyl palmitate (Sigma) standards run on the same column.
Saponification was done by incubating the dried ester HPLC fractions with
0.06 N ethanolic KOH (1 ml) at 55 °C for 30 min in the dark (J.C. Saari,
pers. comm.). A minor amount of isomerization (<3%) occurs with this method.
Hexane extracts were analysed by HPLC as above using 11.2% ethyl acetate/2%
dioxane/1.4% 1-octanol in hexane as mobile phase49.
Received 27 July 1998; Accepted 30 October 1998
REFERENCES
- Hamel, C.P. et al. A developmentally regulated microsomal protein specific for the pigment epithelium of vertebrate retina. J. Neurosci. Res. 34, 414-425 (1993). | PubMed | ISI | ChemPort |
- Wright, A.F. A searchlight through the fog. Nature Genet. 17, 132-134 (1997). | Article | PubMed | ISI | ChemPort |
- Gu, S. et al. Mutations in RPE65 cause autosomal recessive childhood-onset severe retinal dystrophy. Nature Genet. 17, 194-197 (1997). | Article | PubMed | ISI | ChemPort |
- Marlhens, F. et al. Mutations in RPE65 cause Leber's congenital amaurosis. Nature Genet. 17, 139-141 (1997). | Article | PubMed | ISI | ChemPort |
- Morimura, H. et al. Mutations in the RPE65 gene in patients with autosomal retinitis pigmentosa or Leber congenital amaurosis. Proc. Natl Acad. Sci. USA 95, 3088-3093 (1998). | Article | PubMed | ChemPort |
- Hamel, C.P., Jenkins, N.A., Gilbert, D.J., Copeland, N.G. & Redmond, T.M. The gene for the retinal pigment epithelium-specific protein RPE65 is localized to human 1p31 and mouse 3. Genomics 20, 509-512 (1994). | Article | PubMed | ISI | ChemPort |
- Nicoletti, A. et al. Molecular characterization of the gene encoding an abundant protein specific to the retinal pigment epithelium. Hum. Mol. Genet. 4, 641-649 (1995). | PubMed | ISI | ChemPort |
- Hamel, C.P. et al. Molecular cloning and expression of RPE65, a novel retinal pigment epithelium-specific microsomal protein that is post-transcriptionally regulated in vitro. J. Biol. Chem. 268, 15751-15757 (1993). | PubMed | ISI | ChemPort |
- Båvik, C.-O., Lévy, F., Hellman, U., Wernstedt, C. & Eriksson, U. The retinal pigment epithelial membrane receptor for plasma retinol-binding protein. J. Biol. Chem. 268, 20540-20546 (1993). | PubMed | ISI | ChemPort |
- Simon, A., Hellman, U., Wernstedt, C. & Eriksson, U. The retinal pigment epithelial-specific 11-cis retinol dehydrogenase belongs to the family of short chain dehydrogenases J. Biol. Chem. 270, 1107-1112 (1995). | Article | PubMed | ISI | ChemPort |
- Wald, G. Molecular basis of visual excitation. Science 162, 230-239 (1968). | PubMed | ISI | ChemPort |
- Saari, J.C. Retinoids in photosensitive systems. in The Retinoids: Biology, Chemistry and Medicine (eds Sporn, M.B., Roberts, A.B. & Goodman, D.S.) 351-385 (Raven, New York, 1994). | ChemPort |
- Bernstein, P.S., Law, W.C. & Rando, R.R. Biochemical characterization of the retinoid isomerase system of the eye. J. Biol. Chem. 262, 16848-16857 (1987). | PubMed | ISI | ChemPort |
- Winston, A. & Rando, R.R. Regulation of isomerohydrolase in the visual cycle. Biochemistry 37, 2044-2050 (1998). | Article | PubMed | ISI | ChemPort |
- Barry, R.J., Canada, F.J. & Rando, R.R. Solubilization and partial purification of retinyl ester synthetase and retinoid isomerase from bovine ocular pigment epithelium. J. Biol. Chem. 264, 9231-9238 (1989). | PubMed | ISI | ChemPort |
- Peachey, N.S., Goto, Y., al-Ubaidi, M.R. & Naash, M.I. Properties of the mouse cone-mediated electroretinogram during light adaptation. Neurosci. Lett. 162, 9-11 (1993). | Article | PubMed | ISI | ChemPort |
- Carter-Dawson, L. et al. Rhodopsin, 11-cis Vitamin A, and interstitial retinol-binding protein (IRBP) during retinal development in normal and rd mutant mice. Dev. Biol. 116, 431-438 (1986). | Article | PubMed | ChemPort |
- Smeland, S. et al. Tissue distribution of the receptor for plasma retinol-binding protein. Biochem. J. 305, 419-425 (1995). | PubMed | ISI | ChemPort |
- Heller, J. & Bok, D. A specific receptor for retinol-binding protein as detected by the binding of human and bovine retinol-binding protein to pigment epithelial cells. Am. J. Ophthalmol. 81, 93-97 (1976). | PubMed | ISI | ChemPort |
- Sundaram, M., Sivaprasadarao, A., DeSousa, M.M. & Findlay, J.B.C. The transfer of retinol from serum retinol-binding protein to cellular retinol-binding protein is mediated by a membrane receptor. J. Biol. Chem. 273, 3336-3342 (1998). | Article | PubMed | ISI | ChemPort |
- Wiggert, B., Bergsma, D.R., Lewis, M., Abe, T. & Chader, G.J. Vitamin A receptors. II. Characteristics of retinol binding in chick retina and pigment epithelium. Biochim. Biophys. Acta 498, 366-374 (1977). | Article | PubMed | ISI | ChemPort |
- Noy, N. & Blaner, W.S. Interactions of retinol with binding proteins: Studies with rat cellular retinol-binding protein and rat retinol-binding protein. Biochemistry 30, 6380-6386 (1991). | PubMed | ISI | ChemPort |
- Robison, W.G. & Kuwabara, T. Vitamin A storage and peroxisomes in retinal pigment epithelium and liver. Invest. Ophthalmol. Vis. Sci. 16, 1110-111
|
