Linkage Mapping of Canine Rod Cone Dysplasia Type 2 (rcd2) to CFA7, the Canine Orthologue of Human 1q32

Abstract

purpose. To map the canine rcd2 retinal degeneration locus. Rod–cone dysplasia type 2 (rcd2), an early-onset autosomal recessive form of progressive retinal atrophy (PRA), is phenotypically similar to early-onset forms of retinitis pigmentosa collectively termed Leber congenital amaurosis and segregates naturally in the collie breed of dog. Multiple genes have previously been evaluated as candidates for rcd2, but all have been excluded.

methods. A set of informative experimental pedigrees segregating the rcd2 phenotype was produced. A genome-wide scan of these pedigrees using a set of 241 markers was undertaken. To refine the localized homology between canine and human maps, an RH map of the identified rcd2 region was built using a 3000 cR panel. A positional candidate gene strategy was then undertaken to begin to evaluate potentially causative genes.

results. A locus responsible for the rcd2 phenotype was mapped to CFA7 in a region corresponding to human chromosome 1, region q32.1-q32.2. Maximum linkage was observed between rcd2 and marker FH3972 (θ = 0.02; lod = 25.53), and the critical region was flanked by markers FH2226 and FH3972. As CRB1 is the closest gene on human chromosome 1q known to cause retinal degeneration, canine gene–specific markers for CRB1 were developed, and this gene was excluded as a positional candidate for rcd2.

conclusions. The rcd2 locus represents a novel retinal degeneration gene. It is anticipated that when identified, this gene will offer new insights into retinal developmental and degenerative processes, and new opportunities for exploring experimental therapies.

The dog, Canis familiaris, is a diverse species that is divided into more than 300 genetically closed subpopulations, termed breeds. Each is distinguished by a set of morphologic and behavioral phenotypes that are rigorously maintained through tightly regulated closed breeding practices. An artificial selection focused on retaining these traits has led to the accumulation of disease-causing mutations within many breeds. Several of these represent natural models of human disorders of interest. ^{1 2 3 4}  

The arsenal of modern canine genetics has expanded rapidly in the past 5 years. Public domain resources now include a 7.5× high quality draft sequence of the dog ⁵ (see http://www.ncbi.nlm.nih.gov/genome/guide/dog/), a large collection of canine-specific SNPs (http://www.broad.mit.edu/mammals/dog/snp/, provided in the public domain by the Massachusetts Institute of Technology, Cambridge, MA), and a canine integrated map featuring more than 4200 markers. ^{6 7} In aggregate, these provide extensive support for linkage and association studies in dogs. Canine pedigrees thus have enormous genetic mapping potential, especially when well-characterized canine models are used for the analysis and treatment of corresponding human disorders. ^{8 9 10}  

Rod cone dysplasia type 2 (rcd2), an early-onset form of progressive retinal atrophy (PRA), is phenotypically similar to early-onset forms of human retinitis pigmentosa (RP), and segregates naturally in the collie breed of dog. ¹¹ The disease has been thoroughly characterized electrophysiologically, morphologically, and biochemically. ^{12 13 14 15} As in RP, night blindness is the earliest clinical sign of rcd2, detectable in 6-week-old affected dogs. By 6 to 8 months of age, rcd2 dogs become functionally blind. Ophthalmoscopic abnormalities can be detected at 3.5 to 4 months of age, including tapetal hyperreflectivity, retinal vascular attenuation, and optic nerve pallor. Retinal dysfunction can be detected by electroretinogram (ERG) as early as 16 days of age. Both rods and cones in the affected retina fail to develop normal outer segments. ¹² At 6 weeks of age, when the photoreceptors of normal dogs are fully developed, only a few underdeveloped outer segments are visible in rcd2 dogs. By 2 to 2.5 months age, the outer segments completely disappear in the affected retina. Both types of photoreceptors subsequently degenerate—cones more slowly than rods. 

Biochemically, rcd2 disease is characterized by a 10-fold increase in retinal cyclic guanosine monophosphate (cGMP) content, and a corresponding deficiency in cGMP-phosphodiesterase activity. ¹³ This hallmark of disease is also seen in murine rd and in canine rcd1, ¹⁶ both of which represent mutations in the gene (PDE6B) for the beta subunit of rod cGMP-phosphodiesterase. PDE6B has been excluded, however, as the rcd2 locus by experimental breeding ¹⁴ and molecular studies. ¹⁷  

Inheritance of the rcd2 phenotype as a monogenic autosomal recessive trait has been observed in multiple natural collie pedigrees and described in a large body of literature. ^{14 18 19 20} An extensive candidate gene approach has been used in attempts to identify the rcd2 gene. Ten genes that participate in the visual phototransduction cascade (PDE6A, PDE6B, PDE6G, PDE6D, opsin, arrestin, GNAT1, GNBT1, GNGT1, RDS/peripherin) as well as CRX and ROM-1 have been tested by linkage analysis and excluded as rcd2 candidates (Kukekova AV, et al. IOVS 2003;44:ARVO E-Abstract 2325). ^{17 19 20 21 22 23 24 25} Other known canine PRA loci (erd, prcd) have also been excluded (Acland GM, unpublished data, 2005). ¹⁴  

In this study, we report the mapping of rcd2 to CFA7, the canine orthologue of HSA1q32, and the testing and exclusion of CRB1, a positional candidate gene located in this region. As none of the other genes in this interval are obvious candidates, our results indicate that rcd2 represents a novel retinal-degeneration locus. 

Methods

Animals

A colony of rcd2 dogs is maintained at the Retinal Disease Studies Facility (RDSF; Kennett Square, PA), as part of a National Institutes of Health (NIH)–funded resource project (R01-EY06855, Models of Hereditary Retinal Degeneration) to develop, maintain, and exploit canine models of retinitis pigmentosa. The rcd2 strain of dogs was founded with several distantly related collie dogs affected with rcd2. These rcd2-affected collies were initially bred to unrelated mixed-breed dogs to maximize heterozygosity in F1 dogs. One intercross (F2) and 13 backcross, rcd2-informative, three-generation pedigrees from this colony, comprising altogether 148 individuals, were used in the current project. Rigorous diagnostic standards including ophthalmoscopic, ERG, and morphologic studies have been developed and implemented for assignment of the rcd2 phenotype. ¹⁴ Animals from experimental pedigrees were observed ophthalmoscopically and evaluated by ERG at approximately 8 weeks of age. Canine retinal tissues were collected after death and examined by light microscopy. All procedures involving animals were performed in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Institutional Animal Care and Use Committees (IACUCs) at Cornell University and the University of Pennsylvania. 

DNA Samples

Tissue samples (blood, spleen) were collected from each animal, and DNA was extracted by standard methods. ²⁷ The quality of DNA samples was evaluated by agarose gel electrophoresis and estimated by the A₂₆₀:A₂₈₀ ratio. 

Markers

A genome-wide scan of the rcd2 experimental pedigrees with the Marshfield canine screening set of 241 markers was undertaken by Mammalian Genotyping Service (Marshfield, WI). This marker set represents a selection of canine microsatellite markers distributed relatively evenly across the canine genome. 

Power–Simulation Study

To calculate the power of the proposed rcd2 pedigrees, we performed two-point linkage analyses of these pedigrees, using simulated marker data and MultiMap. ²⁷ Because markers from the standard canine screening set are approximately equally distributed across the canine genome at an average distance of approximately 10 cM, we tested simulated markers under two assumptions: (1) a recombination fraction (Θ) of 0.0 between rcd2 and the simulated marker; and (2) Θ = 0.1. For the best-case scenario (Θ = 0) and assuming marker polymorphism information content (PIC) of 0.37 (heterozygosity, 0.49), our simulation study yielded a lod score = 28.297; for Θ = 0.1 a lod score of 14.282 was obtained. 

Analysis of Genome-Wide Screen Data

Marker genotypes were analyzed using the MultiMap software package, ²⁷ as described previously. ^{28 29 30} Genotypes were checked for Mendelian segregation, using the Prepare option of MultiMap. Linkage between the rcd2 locus and each marker was determined using the MultiMap Best-Twopoint function. The marker order on the chromosome of interest was determined by multipoint analysis. Markers were assigned to linkage groups (using the Find All Linkage Groups function of MultiMap) if linked to at least one other marker in the group with a Θ ≤ 0.4 and a lod score of at least 3.0 (equivalent to odds of 1000:1 in favor of linkage). A sex-averaged framework map was then constructed beginning with the pair of markers with the highest joint PIC value and for which a recombination fraction of 0.05 to 0.4 was supported with a lod score ≥3.0. To order disease locus and genetic markers along the chromosome, further markers were added to a linkage group in decreasing order of informativeness until no further markers could be localized to a unique interval with a lod score ≥3.0. The FLIPS function of MultiMap was used to ensure that the odds in favor of the final order of each marker pair were at least 1000:1 over alternative orders. 

Development of CRB1-Associated Polymorphic Markers

The human CRB1 gene sequence was obtained from http://genome.ucsc.edu/ (provided in the public domain by UCSC Genome Bioinformatics, University of California at Santa Cruz, Santa Cruz, CA) and analyzed by using BLAST against the canine genome sequence present in the Trace archive database (http://www.ncbi.nih.gov/Traces/trace.cgi?, National Center for Biotechnology Information, Bethesda, MD). Canine sequences identified through the blast search were combined in a contig and aligned against the human CRB1 gene using the Cornell University server: http://ser-loopp.tc.cornell.edu/cbsu/align2genome.htm. Simple repeat elements were identified in canine CRB1 introns 3, 5, and 7. Oligonucleotide primers were designed to amplify three gene-specific microsatellites. One of the repeats within intron 5 was found to be polymorphic in rcd2-informative pedigrees. The sequence of the PCR product was consistent with the original canine sequence. Primers CRB1MF4 (5′-ATATGAGTTAAGAAGTCCTGGCT-3′) and CRB1MR4 (5′-GCATGATAACCTTGGAAACCACT-3′) were used for genotyping. The CRB1 marker was amplified from 50 ng of genomic DNA by using the following conditions: 96°C for 2 minutes; 30 cycles of 96°C (20 seconds), 58°C (20 seconds), and 72°C (20 seconds); and a final extension at 72°C for 5 minutes. PCR products were analyzed by electrophoresis through 10% nondenaturing polyacrylamide gels. 

Radiation Hybrid Mapping of the CFA7 rcd2 Region

A commercially available 3000-rad canine radiation hybrid panel (RH3000) was used (Research Genetics, Huntsville, AL). Microsatellite markers that showed linkage to rcd2 and selected markers previously mapped to the same region of canine CFA7 using a different 5000-rad panel (RH5000) ³¹ were mapped with the RH3000 panel (Fig. 1) . A CRB1 gene-specific marker was designed based on the sequence of canine CRB1 intron 5 retrieved from the Trace archive database. The same pair of CRB1 primers was used for RH mapping and genotyping. Primers RHSYT14F3 (5′-GCTAACTGGAAACAGTGACCAGA-3′) and RHSYT14R2 (5′-GTCACTTGGAACATCTTCTTCGT-3′) were also designed and used for RH mapping of synaptotagmin XIV (SYT14), another gene located on HSA1q32. Primers for previously published markers used in the present study (FH2226, CD34, CPH20, REN314O07, FH1031, FH3972, REN300M13, REN319J07, and CENPF) are available from the Web site (http://www.research.nhgri.nih.gov/dog_genome/guyon2003/guyonmarkers_data/CFA07MarkerTable.html). Each marker was amplified on the RH panel under the following conditions: 96°C for 2 minutes; 30 cycles of 96°C (20 seconds), 58°C (20 seconds), and 72°C (20 seconds); and a final extension at 72°C for 5 minutes. All PCR reactions were performed using 25 ng of genomic DNA from each cell line in a final volume 15 μL, and the products were separated on a 1.8% agarose gel. PCR products were visualized by ethidium bromide staining. All markers were scored, and the complete set was analyzed with MultiMap. ²⁷ Multipoint analysis was used to order markers and determine intermarker distances. Distances (D) were calculated as D = −ln(1 − θ), where θ is the frequency of breakage. Distances are expressed in centirays3000 (cR3000). 

Results

Experimental rcd2 Informative Pedigrees

Because the rcd2 strain of dogs was developed from several distantly related rcd2-affected founders, allelism of disease among these founders was confirmed by test breeding. Segregation of the rcd2 phenotype, evaluated in a subset of 12 of the pedigrees used in the present study, was compatible with Mendelian expectation for a single-locus autosomal recessive trait (χ² = 0.504; P = 0.478). Dominant transmission was ruled out by multiple examples of affected pups produced from nonaffected-to-nonaffected breeding (data not shown; but see, for example, individual 13, Fig. 2 ). X-linked inheritance was ruled out by multiple examples of affected female pups produced from nonaffected-to-nonaffected breeding (data not shown). 

Linkage Analyses of the Genome-Wide Scan Data for rcd2

The genome-wide scan data comprised genotypes of 148 dogs for the 241 microsatellite markers of the Marshfield canine screening set. Linkage between rcd2 and each marker was evaluated using the MultiMap Best-Twopoint function. ²⁷ The rcd2 locus mapped to telomeric CFA7. The highest linkage was observed between rcd2 and markers FH3972 (Θ = 0.02; lod = 25.535), VIASD10 (Θ = 0.07; lod = 17.08), and FH2226 (Θ = 0.072; lod = 14.128; Table 1 ). The marker order on the chromosome of interest was determined using multipoint analysis (Fig. 1) . Haplotypes for CFA7 markers were assembled (Fig. 2)and the rcd2 interval was identified as positioned between markers FH2226 and FH3972, a distance of 61.4 cR. In several dogs, non-Mendelian inheritance of marker alleles was observed (see, for example, individual 6, allele 203, Figure 2 ). This phenomenon is not unexpected in microsatellite genotyping, especially for a marker like FH2226, which is a complex (CCTT)_n(CTTT)_n repeat. In this data set, four new alleles for FH2226 were observed in 172 meioses—a rate of 0.023. Fortunately, haplotype analysis allowed each of these to be resolved. 

Assignment of rcd2 Interval and Comparison with HSA1

The telomeric region of CFA7 corresponds to HSA1q, based on the integrated map of the canine genome. To refine the homology between the canine and human maps in the rcd2 region, we built an RH map of the region using a 3000-cR RH panel. Microsatellite markers linked to rcd2, and markers located in the same region of CFA7 on the canine 5000 cR RH map were placed on this RH3000 map. Additional gene-specific markers were developed and RH mapped to establish the level of conservation of gene order between the canine and human orthologous regions (Fig. 1) . 

Exclusion of CRB1 as a Positional Candidate for rcd2

Canine sequences corresponding to the human CRB1 gene were obtained from the Trace archive database and aligned in a contig using the Cornell server. The structure of the canine CRB1 gene was similar to that of human CRB1. A polymorphic microsatellite was identified in CRB1 intron 5 and used as a marker to test for cosegregation between CRB1 and rcd2. Four rcd2 informative pedigrees were tested, and three recombinant animals were identified. An identified canine CRB1 microsatellite was placed on RH map 3000 to confirm the location of CRB1 on CFA7 close to the rcd2 flanking marker FH2226, but outside of the critical rcd2 interval. CRB1 was also observed to be outside the rcd2 interval in the assembled sequence of the canine genome, and outside the paralogous human genome interval. Based on linkage and RH data, CRB1 was excluded as a candidate gene for rcd2 (Figs. 1 2) . 

Discussion

We report the identification of a new locus causing retinal degeneration in dogs. Canine rcd2 is an early-onset form of PRA that is phenotypically similar to severe forms of early-onset RP or Leber congenital amaurosis in humans and segregates naturally in the collie breed of dog. A genome-wide scan of rcd2-informative pedigrees mapped the rcd2 locus to CFA7 in a region corresponding to human chromosome 1q at region 32. The CRB1 gene, which maps to HSA1q32, is known to cause retinal disease (RP12) in humans. ^{32 33} RP12 is an early-onset form of retinal degeneration that causes severe visual impairment in humans by 20 years of age. The CRB1 gene encodes an extracellular protein with 19 epidermal growth factor (EGF)-like domains, three laminin A G-like domains, and a C-type lectin domain. The function of CRB1 in the retina is not clear, but the protein is most likely involved in neuronal development of the retina and may have a role in the organization or polarity of retinal cells. ³² We tested CRB1 as a positional candidate for rcd2 and excluded it by both linkage and RH mapping. We will proceed to apply fine mapping and LD mapping approaches to find the rcd2 causative gene. 

From linkage analyses, rcd2 is located between markers FH2226 and FH3972. The closest canine gene-specific markers to rcd2, located by RH mapping, are CD34 and SYT14. From the first assembly of the canine genome, the location of microsatellites FH2226 and FH3972 can be identified as CFA7:8,730,713-8,730,961 and CFA7:12,595,597-12,595,819 respectively, making the zero recombination interval 3,864,634 nucleotides (nts). Refseq genes FCAMR (CFA7:8,748,646-8,752,169) and SLC30A1 (CFA7:12,560,688-12,564,152) approximately delimit this interval in the canine genome sequence, and allow precise identification of the paralogous human interval to HSA1:203,519,748-208,140,494, a distance of 4,620,747 nts, assuming complete conservation of synteny and gene order. Cytogenetically, this corresponds to the human interval HSA1q32.2-q32.3. No genes causing retinal degeneration in humans have been identified in the region corresponding to canine rcd2, although conservation of synteny suggests a possible overlap of the canine rcd2 and murine rd3 map intervals. ³⁴ We expect that rcd2 may account for some of the cases of early-onset RP or Leber congenital amaurosis for which no causative gene has yet been identified. 

Biochemical studies have demonstrated that rcd2 is associated with a deficiency in cGMP-phosphodiesterase activity. ^{13 15} Identification of the rcd2-causative gene is thus expected to provide new insight into the mechanisms of both the phototransduction cascade and retinal degeneration. 

Canine breed-specific forms of PRA provide critical models for the study and treatment of corresponding human diseases. ^{1 2 9 35} Once the rcd2 gene and mutation responsible are identified, it will be feasible to test this gene for candidacy in potentially homologous human disorders. Progress toward therapies for human retinal degenerations is clearly accelerating, ^{9 36 37 38 39 40 41} aided by a variety of approaches and model systems. It is evident that different therapies work in different disease models and that no single therapy is generally applicable. ^{35 42 43 44 45} This has led to a growing appreciation of the importance of a rich set of animal models to identify specifically promising and appropriate therapies and move them from the laboratory into clinical use. In this regard, the rcd2 dog, as a new large animal model of retinal degeneration, encoding a clearly novel gene, is a worthy addition to the armamentarium. 

 Supported by National Eye Institute Grants MH069688, EY06855, EY13729 and EY13132; The Foundation Fighting Blindness; the Morris Animal Foundation; Seeing Eye Inc., and the Van Sloun Fund for Canine Genetic Research.

 Submitted for publication July 6, 2005; revised October 7, 2005; accepted December 27, 2005.

 Disclosure: A.V. Kukekova, None; J. Nelson, None; R.W. Kuchtey, None; J.K. Lowe, None; J.L. Johnson, None; E.A. Ostrander, None; G.D. Aguirre, None; G.M. Acland, 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.

 

Figure 1.

 

Comparative maps of the rcd2 interval. (A) Canine meiotic linkage map of telomeric CFA7; (B) RH3000 map of the corresponding canine genomic interval; (C) map of the assembled canine genome sequence for the same region of CFA7; (D) the corresponding region of the human chromosome 1 sequence. Meiotic linkage places rcd2 in an 8-cM interval flanked by FH2226 and FH3972. The distance between these markers measures 61.4 cR on the RH3000 map, and 3,864,637 nts in the current assembly of the canine genome sequence. CRB1 maps outside of this interval on the meiotic map, the RH3000 map, and the canine genome sequence and is well outside the paralogous region (FCAMR to SLC30A1) of the human genome sequence.

Figure 2.

 

Representative experimental pedigree segregating the rcd2 phenotype. Markers (CRB1, FH2226, FH3972, and VIASD10) are shown in the order established by multipoint linkage analysis. Haplotype shading indicates parental origin. The rcd2 locus is represented by + for the wild-type allele and − for the mutant. The rcd2 locus is positioned between FH2226 and FH3972 in recombinant individuals 8, 9, and 13. The position of rcd2 in individuals 10, 12, and 14 is consistent with this, assuming haplotypes (as shown) that minimize parental recombinations. Allele 203 for FH2226 in individual 6 is an example of non-Mendelian inheritance.

Table 1.

 

Results of Best Two-Point Linkage Analysis between the rcd2 Locus and Markers from CFA7

Loci	FH2226	rcd2	FH3972	VIASD10	REN162C04	FH2174	REN149P06	FH2201
CRB1
θ	0	0.143	0.128	0.257	0	0.454	0.5	0.271
Z	8.429	2.28	3.813	1.527	0.903	0.02	0	0.943
FH2226
θ		0.072	0.096	0.144	0.105	0.262	0.458	0.344
Z		14.128	15.111	11.621	5.585	2.154	0.089	1.1
rcd2
θ			0.02	0.072	0.217	0.204	0.328	0.337
Z			25.535	17.082	3.387	4.401	2.051	1.869
FH3972
θ				0.086	0.2	0.223	0.315	0.352
Z				23.8	3.767	4.62	3.452	2.108
VIASD10
θ					0.25	0.179	0.202	0.302
Z					2.045	5.796	9.645	3.521
REN162C04
θ						0.031	0.219	0.261
Z						7.7	2.332	2.381
FH2174
θ							0.13	0.203
Z							6.112	4.819
REN149P06
θ								0.101
Z								16.238

 Data in bold indicate loci linked to rcd2 with suprathreshold lod score (i.e., >3.5). Markers are ordered according to results of multipoint analysis in the current study. Note that recombinants are observed (θ = 0.143) between CRB1 and rcd2, excluding CRB1 as a positional candidate. θ, recombination fraction; Z, LOD score.

The authors thank Amanda Nickle, Gerri Antonini, and the staff of the Retinal Disease Studies Facility for technical research assistance; Marshfield Laboratories (Marshfield, WI) for genotyping service; Leonid Kruglyak for helpful discussion, Sue Pearce-Kelling, Julie Jordan, and Pam Hammond for excellent technical assistance; and Keith Watamura for preparation of the figures.