Comparative Analysis and Expression of CLUL1, a Cone Photoreceptor-Specific Gene

Abstract

purpose. To characterize CLUL1, a cone photoreceptor-specific gene.

methods. A comparative genomics approach was used to analyze the gene organization and protein sequence of a retinal clusterin-like protein and to identify conserved elements between human and dog. Its expression was studied by Northern and Western analyses and its localization by in situ hybridization and immunocytochemistry.

results. The CLUL1 sequences of the human and dog share 85% and 73% identity, respectively, at the nucleotide and deduced amino acid level. The gene is organized into nine exons and shows strong homology, not only in exonic but also in some intronic sequences between the species. The polypeptide homology of CLUL1 to CLU, a molecular chaperone, indicates structural similarity of the two proteins. However, these data demonstrated that they present different expression profiles in the tissues, in retinal development, and in retinal diseases. Finally, CLUL1 was localized to retinal cone photoreceptor cells and a different immunolabeling in light- and dark-adapted retinas was shown.

conclusions. CLUL1 represents a potentially important gene and a candidate locus for retinal disease, particularly those diseases that affect cones.

With the sequencing of the human genome nearing completion and efforts to sequence the genomes of other species now in progress, comparative genomic analysis is becoming important to understand the function of orthologous and homologous genes, their regulation, and their role in inherited diseases. Using sequence information across species, and analysis of related gene families, we can study comparatively the gene organization and regulation and identify the functional elements conserved actively through evolution. The comparative approach is becoming indispensable in elucidating gene function, especially of novel genes, because functional elements are not obvious unless sequences are aligned across species to reveal the evolutionary conservation. In this regard, animal models such as the dog can make an important contribution to functional genomics and biomedical research, particularly in determining the molecular basis and pathogenesis of inherited disease. Recent examples include canine narcolepsy ¹ and canine X-linked retinal degeneration, ² an orthologue gene of the retinitis pigmentosa guanosine triphosphatase (GTPase) regulator (RPGR) that is mutated in human X-linked retinitis pigmentosa-3 (RP3). ^{3 4}  

Previously, we reported the cloning of a novel retinal specific clusterin (CLU)-like protein (CLUL1) cDNA as a downregulated transcript in some retinal diseases. ⁵ CLU, a secreted glycoprotein expressed ubiquitously in many tissues, is a highly conserved protein that is translated from a single-chain precursor, and the polypeptide is internally cleaved into two subunits linked by five disulfide bonds. ⁶ Recently, CLU has been suggested to function as a molecular chaperone. ⁷ A BLAST search revealed that the CLUL1 and CLU proteins have a 23% to 25% sequence identity. ⁵ The conserved residues, especially the 10 cysteine residues consisting of five disulfide bonds in CLU, may indicate the structural similarity of the two proteins. Although it has not yet been mapped to any retinal disease locus, CLUL1 is a strong candidate gene in the zero-recombination regions of two human inherited diseases: bipolar mood disorder ⁸ and familial high myopia. ⁹ Because CLUL1 is expressed in retina, ⁵ it is a strong retinal disease candidate in humans and other species. Information on the mapping, gene organization, and cellular localization is needed to examine the function and regulation of CLUL1. 

In this study, we identified a new CLUL1 splice variant in dog that has a length and sequence conservation similar to that of human CLUL1, cloned the human and dog genes, and analyzed their organization to identify conserved elements. We also mapped the canine CLUL1 locus and localized CLUL1 to cone photoreceptor cells. We demonstrated that CLU and CLUL1 are different in expression, and localization, regardless of the structural similarities they share. 

Materials and Methods

Animals

Dogs were housed in indoor kennels under cyclic illumination (12 hours of light, 12 hours of dark). Light-adapted retinas were collected from pentobarbital-anesthetized, light-entrained animals 3 to 4 hours after light onset or from dark-adapted animals 8 hours after light offset. Tissues were collected from dark-adapted dogs under dim red light. All research conducted was in full compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

BAC Library and Sequence

The RPCI81 canine bacterial artificial chromosome (BAC) library was used (http://www.chori.org/bacpac/mcanine81.htm/ provided in the public domain by the national Center of Excellence in Genomics at Children’s Hospital Oakland Research Center, Oakland, CA). Positive clones were retrieved by hybridization with a CLUL1 cDNA probe. Long-range PCR was used to amplify introns for cloning, and sequencing was performed at the core sequencing facility of Cornell University. 

Radiation Hybrid Mapping

The canine/hamster radiation hybrid (RH) 3000 panel was used for RH mapping. ¹⁰ A microsatellite marker, (CAAA)₁₀, associated with canine CLUL1 was isolated from a positive BAC clone. All other markers are from the Fred Hutchinson Cancer Research Center (Seattle, WA; http://www.fhcrc.org/science/dog_genome/markers/all600.html/).The data were analyzed with MultiMap software (http://compgen.rutgers.edu/multimap/ hosted in the public domain by Rutgers University, New Brunswick, NJ). ¹¹  

Hybridization Screening of Human Retinal cDNA Library

Human CLUL1 clones were identified by hybridizing a human retinal cDNA library (Stratagene, La Jolla, CA) with a canine CLUL1 cDNA probe. The inserts of positive clones were sequenced from both strands using vector primers first, followed by primer-walking, using gene specific primers designed from the verified sequence. 

RNA Extraction, 5′- and 3′-RACE

A rapid amplification of cDNA ends (RACE) cDNA amplification kit (SMART; Clontech, Palo Alto, CA) was used to amplify the canine alternative splice sequences. One microgram of total mRNA extracted ¹² from retinas of three normal light-adapted dogs (58–62 days) was used for first-strand cDNA synthesis for either 5′- or 3′-RACE libraries. A reverse primer (L2R5), based on the CLUL1 cDNA sequence conserved between human (AF395889) and dog (AF147784), was used with a universal primer mix (UPM) for 5′-RACE. A forward primer (L2F7) was used with UPM for 3′-RACE. Touchdown PCR used for RACE was performed as follows: 94°C for 30 seconds for initial denaturing, 94°C for 5 seconds and 72°C for 3 minutes for 5 cycles; 94°C for 5 seconds and 70°C for 3 minutes for 5 cycles; and 94°C for 5 seconds and 68°C for 3 minutes for 25 cycles. The PCR products were analyzed with 1.2% agarose gel and cloned into a vector (pCR2.1; Invitrogen, Carlsbad, CA). Positive clones were identified by hybridization screening using PCR probes made with primer pairs L2F1 and L2R1 (1131 bp) for 5′-RACE and L2F4 and L2R2 for 3′-RACE (300 bp; Table 1 ). 

Northern Blot Analysis

Northern blot analysis was performed with 20 μg total RNA in a 1% (wt/vol) agarose/0.66% M formaldehyde gel, blotting to a nylon membrane in 10× SSC, and hybridizing with RNA probes (10⁶ cpm/mL hybridization buffer). ¹² The blots contained multiple tissues and retinas from normal dogs at different stages of postnatal development and of dogs with different nonallelic forms (described later) of inherited retinal degeneration. It should be noted that the membranes used for the present studies previously had been used to examine CLUL1 expression using a cDNA probe. ⁵ For the present studies, the membranes were rehybridized with RNA probes of CLUL1, CLU, and CRX to examine the expression profiles of these three genes. Probes were prepared by cloning PCR products into a commercial vector (pCRII-TOPO; Invitrogen, Carlsbad, CA) that includes promoters for synthesis of single-strand sense and antisense RNA probes. Linearized plasmid vector DNA was used for in vitro transcription of a 483-bp α-³²P-UTP-labeled RNA probe with a kit (MaxiScript; Ambion, Austin, TX). The membranes were rehybridized with α-³²P-dCTP-labeled 18S cDNA probe to normalize RNA loading. RNA expression was examined by Northern blot analysis using the following light-adapted retinal samples: (1) 6-week-old normal subjects; (2) 6-week-old hemizygous XLPRA2 affected males. At this age, the XLPRA2 retina shows morphologically very mild disease with limited degeneration ² ; and (3) 5.3-week-old hereditary cone degeneration (cd) showing mild disease affecting only the cones. ^{13 14} These models were selected based on the relative early-onset of their respective diseases, and because of the expression pattern of CLUL1 in the retinal photoreceptors (described later). For analysis of the developmental pattern of CLUL1 expression, normal retinas between 6 and 48 days of age were used. 

In Situ Hybridization

CLUL1 RNA probes were synthesized with digoxigenin (DIG)-labeled UTP using a DIG RNA labeling kit (Roche Diagnostics, Indianapolis, IN). Frozen retinal sections (7 μm) fixed in 4% paraformaldehyde were hybridized with 100 ng DIG-labeled probe at 50°C for 16 hours in a humidified chamber. Slides were washed in 2× SSC twice and 1× SSC twice at 37°C and then digested with 10 μg/mL RNase A. Immunodetection was performed by incubation with alkaline-phosphatase-conjugated anti-DIG antibody (1:500 dilution) at room temperature for 1 hour, using a DIG nucleic acid detection kit (Roche Diagnostics), followed by incubation with 5-bromo-4-chloro-3-indoyl phosphate/ nitroblue tetrazolium (5-BCIP-NBT) chromogenic substrates (Roche Diagnostics). 

Antibody Production and Immunocytochemistry

The synthetic peptides used to prepare CLUL1-specific antibody were chosen from a region of strong homology between human and dog CLUL1 (sequence, Ac-QEFDQTFQSYFMSDTDL-amid; New England Peptide, Inc., Fitchburg, MA; see Fig. 1 ), but no homology to CLU protein. Rabbit antisera, without affinity-purification, were used for immunocytochemistry (ICC); negative controls included preimmune serum and unrelated total rabbit IgG. Seven-micrometer sections of optimal cutting temperature (OCT) embedded, paraformaldehyde (PF) fixed retinas (4% PF in 0.1 PBS for 3 hours; 2% for 24 hours; cryoprotected in 30% sucrose) were used with the following primary antibodies: anti-CLUL1 (1:250), monoclonal COS-1 (1:125) and OS-2 (1:30,000) antibodies directed against chicken cone pigments, ¹³ anti-human cone arrestin polyclonal antibody. ¹⁵ Control sections were treated in the same way with omission of primary antibody, or replacement with preimmune rabbit serum (for CLUL1 control). Secondary antibodies included goat anti-rabbit IgG conjugated to either of two chromophores (Alexa Fluor 488, green; Alexa Fluor 568, red; Molecular Probes, Eugene, OR), and goat anti-mouse IgG conjugated to the red fluorophore (Alexa Fluor 568; Molecular Probes). Sections were examined with a microscope using epifluorescence and DIC optics (Axioplan; Carl Zeiss Meditec, Thornwood, NY). Images were digitally captured (Spot 3.3, Diagnostic Instrument, Inc., Sterling Height, MI), and imported into a graphics program for image-analysis (Photoshop; Adobe Systems, Mountain View, CA). 

Western Blot Analysis

Canine tissues (retina, brain, kidney, liver, heart, and skeletal muscle) and mouse retina were homogenized on ice in buffer (50 mM Tris-HCl [pH 7.5], 100 mM NaCl, 5 mM EDTA, 1% Triton X-100, 0.05% SDS, 2.5% glycerol, and 1.0 mM phenylmethylsulfonyl fluoride). The homogenates were centrifuged at 100,000g for 30 minutes and then supernatants were taken. Protein concentration was determined using a protein assay reagent (catalog no. 500-0006; Bio-Rad, Hercules, CA). The soluble proteins (30 μg) were separated on 10% SDS-PAGE, and transferred onto a PVDF membrane. The blot was preblocked with 10% nonfat dry milk in Tris-based saline buffer containing 0.1% Tween 20 (TBST) at 4°C overnight before incubation with the CLUL1 primary antibody (not affinity purified) at a dilution of 1:100 for 24 hours at 4°C. Blots were then washed with TBST and incubated with a goat anti-rabbit secondary antibody at a dilution of 1:25,000 for 1 hour at room temperature. Chemiluminescent substrate (Super Signal; Pierce, Rockford, IL) was used for signal detection. 

Bioinformatics

The acquired nucleotide and amino acid sequences were aligned against the GenBank databases (nucleotide, expressed sequence tag [EST], and protein) at the National Center for Biotechnology Information (NCBI, Bethesda, MD) using BLAST to search for sequence matches. The alignments were performed with Clustal W program (http://www2.ebi.ac.uk/ClustalW/ provided in the public domain by the Bioinformatics Institute, European Molecular Biology Laboratory, Heidelberg, Germany). The potential functional sites for the protein were predicted using the ScanProsite software (http://expasy.hcuge.ch/tools/scnpsite.html/ provided in the public domain by the Swiss Institute of Bioinformatics, Geneva, Switzerland). For constructing the gene structure and establishing the exon-intron boundaries, human CLUL1 cDNA (GenBank accession No AF395889) was used for a BLAST search of the Celera Genomics (Rockville, MD) human genome database to identify a genomic contig containing CLUL1 (hCG38026). The canine genomic sequences were assembled from cDNA and gDNA sequences of BAC clones and long-range PCR products. Repeat sequences were masked with RepeatMasker (http://ftp.genome.washington.edu/cgi-bin/RepeatMasker/ provided in the public domain by the University of Washington Genome Center, Seattle, WA) program, and the entire CLUL1 genomic sequences of human and dog were compared with PipMaker (http://bio.cse.psu.edu/PipMaker/ provided in the public domain by Penn State University, State College, PA). ¹⁶ cDNA sequences have been deposited in GenBank with accession numbers AF241221 (dog), and AF395889 (human). 

Results

Mapping of Human and Canine CLUL1

Based on database searches of human EST and UniGene (Hs.26886), human CLUL1 was mapped in silico to a position close to the telomere of chromosome 18 (HSA 18p) between the CETN1 and YES1 genes. From a deposited contig (GenBank NT_011005), the genes residing in the region, from telomere to centromere, are CENT1, CLUL1, TYMS, rTS, and YES1, respectively. CLUL1 is approximately 35 kb telomeric to CENT1 and 18 kb centromeric to YES1, which is located on HSA 18p11.3. The canine orthologue was mapped to canine chromosome 7 (CFA 7) and was located between gene-associated markers of TYMS and RAB12 (see CFA7 at http://www.fhcrc.org/science/dog_genome/map/map3/mapgroups.html/Fred Hutchinson Cancer Research Center) by RH mapping (data not shown). With internal and end sequencing of the CLUL1 BAC, we identified CLUL1, rTS, and YES1 genes in the same clone, thus confirming the conservation of gene content in the region between human and dog. 

Cloning and Characterization of Human and Dog CLUL1 cDNA

We previously cloned a canine CLUL1 cDNA transcript that, in comparison to the human sequence, lacked approximately 128 amino acids from the N terminus. ⁵ In the present study, we isolated by 5′-RACE a new canine sequence that matches the human cDNA, and found no splicing variants by 3′-RACE. In addition, we obtained a full-length human cDNA clone from the retinal library by hybridization screening. The total number of nucleotides is 1764 and 1804 bp for human and dog cDNAs, respectively. The characterized regions of the CLUL1 cDNA include the 5′-untranslated region (UTR), coding sequence, and the 3′-UTR, which are 60 bp, 1410 bp, and 285 bp in human and 80 bp, 1398 bp, and 288 bp in dog cDNA, respectively. The CLUL1 cDNA sequences of the two species share 85% identity. There are three polyadenylation signal sites (AATAAA) in human (AF395889; 1529-1534, 154-1552, 172-1734) and two in dog (AF241221; 156-1570, 174-1748). Human and dog CLUL1 amino acid sequences share 73% identity (Fig. 1) . According to the prediction software (Psort sever, http://psort.nibb.ac.jp/form2.html/ provided in the public domain by the Human Genome Center, Institute for Medical Science, University of Tokyo, Japan), the putative protein is most likely a cytoplasmic glycoprotein, a prediction that is supported by its distinct localization in the cone photoreceptor cells in light and dark conditions (described later). 

Gene Organization of CLUL1

The CLUL1 gene consists of nine exons, spanning approximately 35 kb of genomic sequence in humans and 28 kb in the dog (Table 2 , Figs. 1 2 ). The exon-intron sizes and their boundaries, identified by alignment of cDNA and genomic sequences, are presented in Table 2 . Although the coding exons (exons 2–9) have the same length in the two species, the length of some introns in the dog generally is shorter than in humans. There is good agreement also with the gt/ag rule in the splice donor and acceptor sites, except for the splice donor site of intron 5 where the initial nucleotides in both human and dog are gc instead of gt. Most intronic sequences flanking the splice donor and acceptor sites are well conserved, suggesting their functional importance in splicing. The conserved gene sequences are illustrated in Figure 2 . The exonic sequences show an 83% identity on average between human and dog, and the general homology of intronic sequences range from 60% to 90% identity in some regions, especially for sequences related to splicing. In addition, some intronic sequences show much higher conservation in clusters, for example introns 2, 5, and 7. The sequences of intron 6 and 8 show the least conservation between the two species. 

Homology of CLU and CLUL1

Alignment of human and dog CLU and CLUL1 amino acid sequences shows approximately 23% identity and 41% to 43% similarity with conserved substitution (data not shown; see Ref. ⁵ ). There are 10 cysteine residues consisting of five disulfide bonds linking the α- and β-subunits of CLU. ⁶ These are conserved entirely in human CLUL1, and, with one substitution for tyrosine, in dog CLUL1. As well, the identical and conserved residues are distributed evenly throughout the whole sequence. These results suggest that CLUL1, like CLU, may adopt a similar folded conformation with the disulfide bonds and that they become divergent from a common ancestor during evolution, thus belonging to the same gene family. However, as noted below, they have different expression patterns although, at least in retina, their function may be similar. 

Expression of CLUL1 and CLU

For comparison, we examined the expression of CLU and CLUL1 in the same tissue blots (Fig. 3A) . Using a CLUL1 RNA probe in the same membranes used previously to examine the expression of this gene with a cDNA probe, ⁵ we confirmed that the transcription of CLUL1 is detectable only in the retina. In contrast, we found that CLU is expressed abundantly in liver, retina, and different regions of the brain. We also examined the developmental expression of CLU, CLUL1, and CRX ¹⁷ using, for each, RNA probes on the same retinal blot of samples taken from animals of different ages. CLU is expressed early in development and reaches a maximal level of expression by 10 days postnatal, but CLUL1 is developmentally regulated, and the expression of the transcripts parallels retinal differentiation (Fig. 3B) . In contrast, the CRX mRNA level is relatively constant between 6 and 48 days after birth. These results show the differences in expression and regulation for these three genes during postnatal retinal development in the dog. When comparing the expression of CLUL1 and CLU in two different retinal diseases, we found reduced CLUL1 levels in the hemizygous XLPRA2-affected retina, but not in retina affected with cone degeneration (Fig. 3C ; cd). In contrast, the CLU transcript level was increased in XLPRA2, a suggestion that the diseased retinal cells were undergoing apoptosis. ¹⁸ Western blot analysis using a CLUL1 polyclonal antibody demonstrated the presence of single band at 55 kDa in canine and mouse retina, which matches the predicted size (Fig. 3D) . We examined the protein expression in different tissues, including retina, brain, kidney, liver, heart, and skeletal muscle, and found only the 55-kDa band in the retina (data not shown). 

Localization of CLUL1 to Cone Photoreceptor Cells

By in situ hybridization, the CLUL1 transcript was identified in the inner segments of photoreceptor cells (Fig. 4A1-3) . Signals appear in cone inner segments, but, because of the section thickness, the signals in rods could not be conclusively identified. The labeling intensity gradually decreased from the central to the peripheral retina (Fig. 4A 4-6) . In the light-adapted retina, CLUL1 was localized to the outer segment (OS) of all cone photoreceptors. This was confirmed by colabeling with COS-1 and OS-2 Abs which, respectively, label red/green and blue cones ^{13 19} (Figs. 4B 4C) . However, in the dark-adapted retina, there was a redistribution of CLUL1 immunolabeling, with very weak to negative labeling of the cone outer segments, and intense labeling of the outer plexiform layer in the region of contact between the cone pedicles and second order neurons (Fig. 4D) . A similar change in immunolabeling pattern was observed with the cone arrestin antibody, as reported recently for the mouse retina. ²⁰ In the light, cone arrestin immunolabeling was most intense in the outer segments, with lower levels in the remainder of the cone cells. In contrast, the dark-adapted retina showed that cone-associated immunolabeling intensity shifted to the cone inner segment and synaptic terminals (Fig. 4E) . 

Discussion

In this study, we characterized human and dog CLUL1 cDNA and gene and examined CLUL1 retinal expression and cellular localization. From sequence analyses, the coding exons show higher identity (>80%) than noncoding regions (<75% in 3′) between human and dog. It is noteworthy that some intronic sequences show very high sequence homology. For example, intron 2 is almost completely conserved (approximately 90% identity) and intron 7 shows 50% identity, reflecting a functional constraint during evolution. As shown in Figure 2 (horizontal bars) and Table 2 , the high identity of flanking intronic sequences of most exons may represent a conserved function in splicing. However, the apparent higher sequence conservation of some (e.g., introns 2 to 5 and 7), but not all (e.g., introns 6 and 8), suggests a functional role other than splicing. 

Based on a BLAST conserved domains search, the deduced CLUL1 protein sequence has homology only to CLU (approximately 23% identity). The identical residues are spread throughout the whole CLUL1 and CLU sequences, thus implying relatively strong homology. It is generally true that the three-dimensional structure of proteins in a family are more conserved than their primary sequences during divergence. ²¹ Therefore, if similarity between two proteins is detectable at the amino acid level, structural similarity can usually be assumed. Based on the conservation of cysteine and other residues that are distributed evenly in the entire sequence, our data suggest that CLU and CLUL1 may share a closely folded structure and adopt a similar structure of disulfide-linked heterodimers, even though their function in cells may differ. It appears that CLU and CLUL1 have become evolutionarily divergent from a common ancestor. 

While pointing out the sequence similarity of the two genes, we also note several differences. First, CLU is ubiquitously expressed in many tissues, including retina, ²² whereas CLUL1 is expressed in retina, particularly in cone photoreceptor cells. Second, CLU appears regulated under stress or a cellular insult such as neuronal degeneration. This is in accord with its recently suggested function as a secreted mammalian chaperone that binds in vivo to many different biological ligands. ⁷ In contrast, CLUL1 is differentially regulated in the retina ⁵ and expressed in the latter stages of retinal development. Third, the clusterin element, a heat-shock–like transcription cis element identified in the CLU promoter, ²³ was not found in the CLUL1 putative promoter sequence (data not shown). Instead, we found the CRX-binding and PCE-1/Ret 1 elements in the putative promoter region of CLUL1 (data not shown). Last, the fact that CLUL1 transcript level was decreased, but CLU increased in XLPRA2 suggests that the two genes respond differently to the same retinal disease. In contrast, no difference in CLUL1 expression was found in the cd-affected retina; this is not surprising, because there is minimal to no disease in cones at the ages examined. ¹³ Although it is still premature to suggest a functional role of CLUL1 based on the current data, knockout or spontaneous animal models with retinal phenotypes would provide insights into the role of CLUL1 in the retina. 

With a polyclonal antibody generated against a region of strong sequence homology between the dog and human CLUL1, we detected a predominant band at 55 kDa in canine and mouse soluble retinal proteins. This is the predicted size of the CLUL1 protein in dog retina; this protein was not found in other canine tissues examined. Expression studies suggest that CLUL1 may have an important role in the retina. Northern blot analysis (Ref. ⁵ and present study) not only detected CLUL1 transcript in the retina, but showed that the transcript is developmentally regulated during photoreceptor differentiation and its expression decreased in an inherited disease affecting the rods and cones (XLPRA2). Furthermore, the CLUL1 transcript was localized to the inner segment of retinal photoreceptor cells using in situ hybridization, and the message was most intense and predominant in cone inner segments. In agreement with this observation, the expression level gradually declines from the central retina to the periphery. 

Using a polyclonal antibody against CLUL1, and antibodies that identified the different classes of cones, we localized the protein to the OS of all cone photoreceptor cells in the light-adapted retina. In the dark-adapted retina, however, there was a change in immunolabeling to the cone synaptic terminals in the outer plexiform layer. A similar change was observed in cone arrestin. This shift in immunolabeling during the light–dark cycle may reflect a CLUL1 function. The presence of intense labeling in the cone OS in the light, and conversely, the absence of staining in the dark, indicates a possible role for CLUL1 in cone phototransduction or a related function rather maintenance of cone cell structure. This hypothesis is supported by studies in rats and mice showing that cytoplasmic proteins involved in the activation or deactivation phases of phototransduction (e.g., cyclic GMP phosphodiesterase, rod transducin α and β, rod and cone arrestins, and phosducin) undergo a light-dependent translocation within different compartments of the photoreceptor cells. ^{20 24 25 26}  

²  Present address: Department of Cell Biology, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma.

 Supported by National Eye Institute Grants EY06855 and EY13132, The Foundation Fighting Blindness, the Morris Animal Foundation/The Seeing Eye Inc., and the Van Sloun Fund for Canine Genetic Research.

 Submitted for publication November 25, 2002; revised April 14, 2003; accepted May 27, 2003.

 Disclosure: Q. Zhang, None; W.A. Beltran, None; Z. Mao, None; K. Li, None; J.L. Johnson, None; G.M. Acland, None; G.D. Aguirre, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Sequence and Location of Primers Used for Mapping and RACE

Primer Sequence (5′ to 3′)	S/AS	Primer ID	Use	Location*
GGTATCTGATTCCACTGACTGC	S	L2F1	Probe	473–495
GATGTTCTCAGCTCCTGGGG	AS	L2R1	Probe	1585–1604
CCAATAGCTATGAAACTGCCATTC	AS	L2R2	Probe	1961–1984
TGCAGCACAGAGGACCGTATGG	S	L2F4	Probe	1306–1327
AGAGCAATAGGTTGTTCTGGATAGC	AS	L2R4	Mapping	169–193
TCCTCAATGAACTTTTCACCGACAGGG	AS	L2R5	5′-RACE	902–928
TGCCGTGAAGTTGTATGACCTTGC	S	L2F6	Mapping probe	5–28
GAGCAGTGTGACACAAAGACTTGG	AS	L2R6	Probe	758–781
ATTAGCTACATGCTGGAGAAAGCTGTGC	S	L2F7	3′-RACE	1749–1776
CATAGAGGACCAACTAGGTTGGC	AS	L2R7	Probe	480–502, **
AAGGTCCTTTCAGAAGCTGGTGG	S	L2F8	Probe	19–42, **

Figure 1.

 

Homology of deduced amino acid sequences of human and canine CLUL1 proteins. Gaps have been inserted to maximize the homology, and the identical amino acids are boxed. Note that the shaded portion of the sequence represents previously published data comparing the amino acid sequence of the previously reported shorter CLUL1 canine transcript with the human sequence. ⁵ The cleavable signal peptide, the ER membrane retention signal, and the bipartite nuclear localization signal are identified, as are the exon–intron boundaries (vertical lines). The peptide sequence (amino acids 207-223) used to generate the polyclonal antibody is underlined.

Table 2.

 

Exon/Intron Organization of the Human and Canine CLUL1 Genes

E/I No.	Exon Size (nt)	Intron Size (nt)	Splice Acceptor Site (3′)*	Splice Donor Site (5′)*
1	45	1221		AAAGTTAGTGgtaaatacactatattaaaaagttttgttt
	75	834		AAAATTAGTGgtaaatatatttcgttaaaaagtttcattt
2	119	1106	atttgatgcgggtttatttttcctttgcagTAACAGCGGG	AACCTGAAGAgtacgtttggtttcttacctgtgctgtgtc
	119	1391	atttggtacaggtttatttttcttttgcagTAACAGCAGG	AACCTGAAGGgtacgtttggtttctttatcttctgcacgg
3	149	5504	tcaaagaaaaaattgccacatgtcttttagGTTTTTCTGA	AGAAAAGCAGgtacagtcattgaaaataatgtctgttctt
	149	4977	tcaaaggaaaaaactgcaacatatcttcagGTTTTTCTGA	AGAAAACCAGgtacagtaattgagaataacacctgttctt
4	168	2073	aattggacttttgtttctacttttaactagGAGGCCCTGA	GAAAAATAAGgtaagagaaaaagagagctcaagatttcac
	168	2010	atagagtgtttgtgtctattttttaattagGAAGCCCTGA	GAAAAATACGgtaagggaaaaacagagctcaaggtttcag
5	433	5768	tccatttgtgtcccctactctactccacagATTGAACGGT	CAAGACAAAGgcaagtattaaaagattacttttacttaga
	433	4967	cttattccatttttatcctctaccctacagGTCGAACAGT	CAAGACAATGgcaagtattaaaagatgatctttacttaga
6	138	7867	actaacgtgtaaatgttatgttccctgtagCTCCTGACCA	CTATCTGAAGgtaaataattgctattttgttttttattct
	138	5587	attaatgtgtaaacactatgttacctgcagATTGTAACCA	CTGTGGGAAGgtaaataattattattttccagtgtctcaa
7	215	3368	ccttctccacattactttcttctctgctagATTGTCCTGA	TTCAATACAGgtaaaggagagacccaagagcagatacgga
	215	3051	cctttaccatgttgtcttcttctttgctagACTGTCCCGA	TTCCACAAAGgtaagggagagacccaagaaccgatatgga
8	188	4803	aacttctactgtataatccttcttctgtagGTAGTTCCAA	TTAAAACCTGgtaagcagagtgcctggttaggaatgcctt
	185	1172	aatcatttactgtgtaatcctttttcttagATGCTTCCAA	TTAAAACTTGgtaagcagagtcactgtttagatatgcctt
9	289		atcattgctgcctttttttttttttttaagGTAAGAAGAT
	292		attttggtttttgctgctttttttttaaagGTAAGACCTG

Figure 2.

 

Alignment of the human and dog CLUL1 genes. The figure was generated with PipMaker and depicted as a percent identity plot. The human sequence is presented along the horizontal axis with length and percent identity alignment between the species shown as horizontal bars, or dots. The exons and conserved elements in introns—for example, CpG islands, SINEs, or LINEs—are placed above the horizontal line.

Figure 3.

 

Expression of CLUL1 and CLU mRNAs in different tissues (A; normal male beagle, 8.3 weeks), during normal postnatal retinal development (B), and in two nonallelic inherited photoreceptor diseases (C). The expression pattern of CRX is included in the retinal development blot. Numbers to the left of panels indicate sizes of the transcripts hybridized to the RNA probes identified on the right side of the panels. The 18S cDNA probe was used to estimate consistency of loading of the RNA samples among different lanes. Note that the CLUL1 expression results obtained with an RNA probe are the same as those obtained with a cDNA probe. ⁵ (D) Western blot analysis. A single 55-kDa band is detected in normal dog and mouse retina with the CLUL1 polyclonal antibody.

Figure 4.

 

In situ hybridization (A) and immunolocalization (B–E) of CLUL1 in retinal tissues. (A) With the antisense probe, CLUL1 message was localized to the inner segment region, just external to the OLM (A1, arrow), and no signal was detected in control sections (A2, A3). The CLUL1 message was most abundant in the central retina and decreased in a central to peripheral gradient (A4–6). (B–E) Immunostaining of light- and dark-adapted retinas with CLUL1, COS-1, OS-2, and human cone arrestin antibodies. In the light-adapted retina, CLUL1 was localized in all red/green (B, COS-1) and blue (C, OS-2) cone outer segments. In the dark-adapted retina, there was a change of CLUL1 labeling from the cone outer segment to the cone synaptic terminals in the OPL (D, top). This is best visualized in the merged fluorescence and DIC images (D, bottom). A similar change in immunolabeling was observed with cone arrestin. In the light, cone arrestin labeling was most intense in the outer segments (E, right, arrows), but in the dark it translocated to the inner segments and cone synaptic terminals (E, center, arrowheads). OS, outer segment; IS, inner segment; OLM, outer limiting membrane; ONL, outer nuclear layer; OPL, outer plexiform layer. Scale bars: (A1–3, B–E) 25 μm; (A4–6) 50 μm.

The authors thank Keith Watamura for assistance with figures and graphics, Pam DiDia for technical assistance, Ágoston Szél and Steven Daiger for providing, respectively, the COS-1/OS-2 antibodies and human retinal cDNA library, and Cheryl Craft and Xuemie Zhu (the Mary D. Allen Laboratory for Vision Research, Doheny Eye Institute and the Keck School of Medicine, University of Southern California, Los Angeles, CA) for providing the human cone arrestin antibody.