The Dog Gene Map

Matthew Binns, Ph.D., Nigel Holmes, Ph.D., and Matthew Breen, Ph.D., are with the Centre for Preventive Medicine, Animal Health Trust, Lanwades Park, Kentford, Newmarket, Suffolk CB8 7UU, United Kingdom. Matthew Binns, Ph.D., Nigel Holmes, Ph.D., and Matthew Breen, Ph.D., are with the Centre for Preventive Medicine, Animal Health Trust, Lanwades Park, Kentford, Newmarket, Suffolk CB8 7UU, United Kingdom. Matthew Binns, Ph.D., Nigel Holmes, Ph.D., and Matthew Breen, Ph.D., are with the Centre for Preventive Medicine, Animal Health Trust, Lanwades Park, Kentford, Newmarket, Suffolk CB8 7UU, United Kingdom. It can be argued that the domestic dog ( Canis familiaris ) demonstrates the power of selective breeding more than any other domesticated species. Indeed, within the 300 or so breeds of pedigree dogs, there exists a greater range of morphological types than is seen in any other mammalian species. With regard to size and weight for example, there is at least a 30-fold difference between the Chihuahua and the Saint Bernard. The vast majority of pedigree dog breeds have been selectively bred over the last 200 years, although the origin of many of the breeds is uncertain. The first Kennel Club was established in the United Kingdom in 1873; before this time dogs were bred mainly for a range of practical functions, many of which are still evident in the names of present breeds such as retriever, deerhound, and shepherd. Over the last 100 years, the increasing popularity of dog shows has altered the pattern of breeding such that the majority of dogs are now bred largely for their appearance. The dog family Canidae is thought to have diverged from other carnivore families 50 to 60 million years ago. The family, which now comprises 34 extant species, shows a wide range of chromosome morphologies, with the diploid chromosome number varying from 2n=36 (with mainly metacentric autosomes) in the red fox ( Vulpes vulpes ) to 2n:78 (with all autosomes being acrocentric) in the domestic dog and also a number of wolf-like canids such as the gray wolf ( Canis lupus ). The chromosomal rearrangements observed in the different species have been used to deduce the phylogenetic history of the group ( Wayne and others 1987a , b ). Mitochondrial DNA sequences have also been used to examine the evolution of the Canidae and the origins of the domestic dog ( Wayne 1993 ). The results demonstrate that the domestic dog is an extremely close relative of the gray wolf, with as little as 0.2% variation in mitochondrial DNA sequence between the 2 species. This contrasts with 4% variation in mitochondrial sequences between gray wolves and their nearest wild relative, the coyote ( Canis latrans ). The timing of the divergence of the dog from the gray wolf is controversial, with a discrepancy between the archaeological record and recent molecular studies ( Vila and others 1997 ). Many historical sources depict the type of dogs used by peoples such as the ancient Greeks and Romans. It is clear that there were already different basic “types” of dog several thousand years ago, and it is likely that their domestication occurred independently in several places followed by selection for particular functions. The ancient types of dog include examples closely resembling the modern day greyhound, mastiff, Pekingese, and spitz breeds. Genetic mapping in the dog will produce results of veterinary importance and, through comparative genetics, will provide data of medical and biological interest. Different dog breeds developed from small founder populations lollowed by carefully controlled breeding. They became valuable genetic resources in the same way that isolated human populations such as the Finnish and Icelandic people are extensively used for mapping genetic traits. When the genetic basis for an interesting disorder has been established, it is relatively easy to generate large pedigrees segregating the disease due to the large litter size and short generation intervals of the dog. Much recent interest in dog genetics has resulted from a desire on the part of veterinary scientists to reduce the problem of inherited diseases in pedigree dogs. Of the 350 or more identified inherited disorders, the majority that have been well characterized are inherited as simple recessive traits. This situation reflects the high level of inbreeding that has been practiced, as well as the small number of founder animals. Many of the disorders are breed-specific; and even in conditions such as retinal dysplasias, which occur in several breeds, cross-breeding experiments have revealed that at least 3 different genes ( rcd1, rcd2, and erd ) are involved in the different breeds ( Acland and others 1989 ). In contrast, it is expected that the mutations causing a particular disease within a breed will be identical by descent. Many of the inherited disorders in dogs are thought to be homologues of human inherited diseases. For example, progressive retinal atrophy (PRA 1 ) is equivalent to human retinitis pigmentosa (RP 1 ). It is possible that in the future, the identity of some human RP genes may become known from the identification of dog PRA genes mapping within regions of conserved synteny that contain human RP mutations. Human gene therapy failures using therapies developed in mice may be due in part to differences in physical size and longevity between mouse and human. Dogs therefore have potential as animal models for gene therapy experiments, and although dogs have some disadvantages as experimental animals, they may be suitable intermediate-sized models with their greater lifespan allowing longer term studies than are possible in mice. At the time of this writing, very few of the inherited diseases in dogs have been characterized at the molecular level. The mutation for PRA in Irish setters has recently been identified within the β-subunit of a retinal cGMP phosphodiesterase gene ( Suber and others 1993 )--the same gene that is mutated in the rd mouse ( Pittier and Baehr 1991 ) and in humans with RP ( McLaughlin and others 1993 ). All affected Irish setters tested to date possess the same mutation (G to A transition at position 2420), which truncates the β-subunit by 49 amino acid residues ( Ray and others 1994 ). Genetic screening tests are now being used by Irish setter breeders to identity PRA carriers and to exclude them from breeding programs. Recently, markers linked to canine progressive rod-cone degeneration ( prcd ) have been mapped to a region of dog chromosome 9 showing conserved synteny with human chromosome 17q, the mapped location of retinitis pigmentosa RP17 ( Acland and others 1998 ). Single strand conformation polymorphism (SSCP 1 ) studies have indicated that the mutation in β-cGMP phosphodiesterase is probably not responsible for PRA in other breeds with this disease. Dogs will also be a valuable species lot mapping a number of complex genetic diseases including heart disease, hip dysplasia, narcolepsy, atopy, and behavioral traits. Refinement of the dog map will facilitate the identification of candidate genes for these complex disorders in human and other species through comparative mapping. The dog also presents a special opportunity for studying the genetic basis of morphological and behavioral traits. The canine genetic map is in its infancy, although rapid progress is now being made. Two recent papers have reported extensive genetic linkage studies in the dog ( Lingaas and others 1997 ; Mellersh and others 1997 ). Lingaas and others (1997 ) mapped 94 markers onto 2-generation reference families comprising purebred German shepherds and beagles. Sixteen linkage groups of 2 or more markers were identified, and 2 were assigned to defined chromosomes L13 to CFA20 and L 16 to CFA 18. In contrast, Mellersh and others (1997 ) mapped 150 microsatellite markers onto large 3-generation cross-bred reference families to generate a framework map, and they identified 30 linkage groups comprising 2 or more markers. The latter map is estimated to cover 2073 cM. At the time of this writing, no accurate estimates of the genetic length of the dog genome exist, although the physical sizes of most canine chromosomes have been estimated by comparison of their flow karyotype peaks with human chromosome 4 ( Langford and others 1996 ). Several genes have been physically mapped by fluorescence in situ hybridization (FISH 1 ) analysis and are shown in Table 1 . In addition, a limited number of microsatellites isolated from cosmid libraries have been assigned to chromosomes by FISH mapping (for example, Fischer and others 1996 ; Dolf and others 1997 ). Genes physically mapped in the dog Genes physically mapped in the dog The markers used in the construction of the maps are mainly microsatellites. Several hundred polymorphic dinucleotide microsatellites have been characterized ( Ostrander and others 1995 ). Recently it has been reported that tetranucleotide microsatellites are highly polymorphic in dogs ( Francisco and others 1996 ), with the (GAAA) n motif particularly polymorphic. Because tetranucleotide repeats have definite advantages over dinucleotide repeats for scoring alleles (generally showing a simpler signal on analysis and being more polymorphic), their use is becoming increasingly popular. Several resources have been developed to complement the basic genetic linkage mapping of microsatellite markers that has been carried out on the reference families described above. These include a somatic cell hybrid panel ( Langston and others 1997 ), which has been used to establish 31 syntenic groups containing both microsatellite and type I markers. The commercially available mapping panel comprises 43 microcell hybrid clones containing 1 to 7 canine chromosomes, and 3 whole cell hybrid clones, with each one including 10 to 20 canine chromosomes. The majority of the established synteny groups are correlated with linkage groups so that as more of the linkage groups become fixed to chromosomes, gross comparative gene organization in the dog will rapidly become defined. Cytogenetic studies in the dog have been constrained by the complex karyotype that comprises 38 pairs of acrocentric autosomes. A standard karyotype for chromosomes 1 through 21 has recently been established ( Switonski and others 1996 ). Chromosome paints representing all chromosomes of the canine karyotype have been generated by bivariate flow sorting ( Langford and others 1996 ), thereby providing reagents for nonambiguous chromosome identification. The paints are being used to aid the identification of characteristic features for autosomes 22 through 38. Together with canine cosmids (containing microsatellites), the paints are also being used in 2-color FISH experiments to ensure that at least 1 physically anchored microsatellite marker is available for every canine chromosome. A canine bacterial artificial chromosome (BAC 1 ) library of approximately 150,000 clones has recently been constructed (the Internet address of Roswell Park Canine BAC Library is provided below). With an average insert size of 155 kb representing an 8-fold genome coverage, this library will be an important resource in the future mapping of the mutations responsible for inherited diseases. Bedlington terriers suffer from copper toxicosis, in which dietary copper accumulates in the liver to toxic levels, a condition similar to Wilson's disease in humans. A microsatellite marker linked to the disease locus has recently been characterized, enabling identification of affected and carrier animals in pedigrees containing at least I member with confirmed Copper toxicosis ( Yuzbasiyan-Gurkan and others 1996 ). This screening test should result in the improved health of the breed and lead ultimately to the identification of the disease gene. The canine X-linked severe combined immunodeficiency (SCID 1 ) locus has been mapped to proximal Xq, and a mutation within the g chain of the IL-2 receptor gene has been identified, establishing that the canine disease is a homologue of human X-linked SCID ( Henthorn and others, 1994 ). The mutation responsible for X-linked hereditary nephritis (HN 1 ) in a family of Samoyed dogs has recently been identified within the a5 chain of collagen type IV and should provide an excellent model for HN in humans, for whom mutations in this gene are common ( Zheng and others 1994 ). The availability of dog chromosome paints also enables reciprocal ZOO-FISH (cross-species) experiments to be undertaken with human chromosome paints and metaphase spreads to define the comparative chromosome organization in the 2 species. The generation of a radiation hybrid panel for the dog (L. McCarthy, University of Cambridge, personal communication, 1997) should facilitate high-resolution mapping in the dog and enable maps containing both type I and II markers to be generated. Results from such work will be particularly useful in identifying positional candidate genes once markers linked to disease traits have been located. Dog chromosome paints will also be useful in investigating the extensive karyotype evolution that has taken place during the evolution of the Canidae. Many of the microsatellites derived from the domestic dog are polymorphic in other canids, and indeed they have been used to look at wild canid populations. For example, microsatellites derived from the domestic dog were used to analyze hybridization between the Ethiopian wolf (the world's most endangered canid) and the domestic dog. Results indicated that hybridization had already occurred in 1 population of wolves and that the variability within and between populations was very low, indicating that captive breeding may be necessary to preserve genetic variability ( Gottelli and others 1994 ). Behavioral attributes are important characteristics of each dog breed and have been subject to strong selection pressure since the domestication of the dog. Because the instinctive behaviors inherited from wild ancestors have been selected to varying degrees in different breeds, certain behavior patterns are now strongly associated with particular breeds. Different dog breeds therefore present unique opportunities for behavioral genetic studies. The availability of a large number of markers will allow the evolutionary relationships between the breeds to be investigated in more detail and should allow breed histories to be established on a more scientific basis than is currently possible. Genetic mapping in the dog is a rapidly developing science. Readers are directed to the following available dog genetic resources on the Internet: Dogmap http://ubeclu.unibe.ch/itz/dogmap.html Fred Hutchinson Cancer Research Center (FHCRC) Dog Genome Project http://tiberius.fhcrc.org/home/dog.html University of California Berkeley Dog Genome Project http://mendel.berkeley.edu/dog.html University of Michigan Microsatellite http://www.cvm.msu.edu/main/res/microsat.html University of Michigan Type I Markers http://www.cwn.msu.edMnain/res/anchor.html Roswell Park Canine BAC Library http://bacpac.med.buffalo.edu/canine-bac.html On-line Mendelian Inheritance in Animals (OMIA) http://probe.nalusda.gov:8300/animal/omia.html The canine genome project is entering an exciting phase in which the majority of tools necessary to map traits of interest have been established, and an increasing number of linkages to important diseases are being reported. The dog offers many opportunities for the mapping of complex traits that are important for veterinary medicine and for the development of animal models of human diseases. Work on the dog genome at the Animal Health Trust is generously supported by the Guide Dogs for the Blind Association.