A Comparative Gene Map of the Horse (Equus caballus)
- Alexandre R. Caetano1,2,6,
- Yow-Ling Shiue1,2,6,
- Leslie A. Lyons3,
- Steven J. O'Brien3,
- Thomas F. Laughlin4,
- Ann T. Bowling1,5, and
- James D. Murray2,5,7
- 1Veterinary Genetics Laboratory, University of California Davis, Davis, California 95616-8744 USA; 2Department of Animal Science, University of California Davis, Davis, California 95616 USA; 3Laboratory of Genomic Diversity, National Cancer Institute, Frederick Cancer Research and Development Center, Frederick, Maryland 21702-1201 USA; 4Department of Biological Sciences, East Tennessee State University, Johnson City, Tennesee 37614 USA; 5Department of Population Health and Reproduction, University of California Davis, Davis, California 95616 USA
Abstract
A comparative gene map of the horse genome composed of 127 loci was assembled based on the new assignment of 68 equine type I loci and on data published previously. PCR primers based on consensus gene sequences conserved across mammalian species were used to amplify markers for assigning 68 equine type I loci to 27 horse synteny groups established previously with a horse-mouse somatic cell hybrid panel (SCHP, UC Davis). This increased the number of coding genes mapped to the horse genome by over 2-fold and allowed refinements of the comparative mapping data available for this species. In conjunction with 57 previous assignments of type I loci to the horse genome map, these data have allowed us to confirm the assignment of 24 equine synteny groups to their respective chromosomes, to provisionally assign nine synteny groups to chromosomes, and to further refine the genetic composition established with Zoo-FISH of two horse chromosomes. The equine type I markers developed in this study provide an important resource for the future development of the horse linkage and physical genome maps.
The rapid progress being made in the development of genetic maps for humans and mice (Hudson et al. 1995; Stewart et al. 1997; Rhodes et al. 1998) has led to a recent boom in the construction of genome maps for a number of domesticated mammalian species of economic importance (Bishop et al. 1994; Archibald et al. 1995;O'Brien et al. 1997a; de Gortari et al. 1998). Many recent technological advances have contributed to the generation of these data which will greatly increase our ability to find and isolate genes that lead to genetic diseases and/or have an effect on economically important production traits of livestock (Cockett et al. 1994;Georges and Andersson 1996; Grobet et al. 1997).
Molecular markers based on repetitive or anonymous DNA sequences (Type II markers, O'Brien et al. 1993) such as microsatellites, Random Amplified Polymorphic DNA (RAPDs), and Amplified Fragment Length Polymorphisms (AFLPs) have been used extensively to saturate the genetic maps of various species because of their technical advantages and high degree of polymorphism (O'Brien et al. 1993; Georges and Andersson 1996; Andersson et al. 1996). Type II markers, however, are seldom informative across mammalian orders and therefore do not allow for the comparison of genetic maps from different species (O'Brien et al. 1993; Georges and Andersson 1996; Andersson et al. 1996). Conversely, molecular markers for functional genes conserved across species (Type I anchor loci, O'Brien et al. 1993) can be used for this purpose.
In conjunction with chromosome painting (Zoo-FISH) studies, mapping studies with type I loci have revealed that genomes from related species have a high degree of synteny conservation (O'Brien et al. 1993; Andersson et al. 1996; Wakefield and Graves 1996). This information forms the basis for comparative genome mapping, a discipline that allows the prospect of using information from highly characterized genomes to study genetic phenomena in map-poor species (O'Brien et al. 1993; Georges and Andersson 1996; Andersson et al. 1996; Wakefield and Graves 1996).
Although the horse genome is not as highly characterized as the genomes of other domestic animals, much progress has been made recently. Somatic cell hybrid (SCH) panels have been used to make synteny assignments of 240 type II markers (Bailey et al. 1995; Shiue et al. 1999), many of which were physically assigned to horse chromosomes by FISH (Sakagami et al. 1995; Tozaki et al. 1995; Breen et al. 1997;Godard et al. 1997, 1998). In addition, the first low-resolution microsatellite-based linkage maps of the horse have been published (Lindgren et al. 1998; Guérin et al. 1999). In the horse, a framework for comparative mapping has been established with chromosome painting studies (Raudsepp et al. 1996, 1997; Rettenberger et al. 1996), FISH assignments of type I loci (Lear et al. 1998a,b,c; Godard et al. 1998), and by synteny assignments of type I loci with SCH panels (Williams et al. 1993; Bailey et al. 1995; Caetano et al. 1999a,b). Using human chromosome-specific probes, Zoo-FISH was used to determine that a minimum of 21 chromosomal breaks, followed by the appropriate rearrangements, are necessary to reconstruct the horse karyotype relative to the human karyotype (Raudsepp et al. 1996; Rettenberger et al. 1996; Raudsepp et al. 1997; Chaudhary et al. 1998). However, in addition to these advancements in the equine genome map, a larger group of type I anchor loci needs to be mapped to allow for thorough comparisons with maps from model species and the study of the fine intrachromosomal gene structure of the horse genome.
Despite their value for comparative mapping, the incorporation of type I anchor loci to genome maps of multiple species has not been used extensively because of technical reasons. Large collections of PCR primer sets for type I loci [i.e. universal mammalian–sequence-tagged sites (UM-STS), comparative anchor tagged sequences (CATS) have been designed to amplify mammalian sequences by taking advantage of public sequence databases (Venta et al. 1996; Lyons et al. 1997). In this report we present the synteny assignment of 68 equine type I loci by analysis of a horse-mouse SCH panel with PCR-based markers generated with universal primers for mammalian genes.
RESULTS
Development of PCR-Based Markers for Equine Type I loci
In addition to amplifying gene-specific fragments, a key feature necessary for the effective use of PCR-based markers in synteny mapping with interspecific SCH panels is the ability to differentiate between fragments amplified from the donor species (i.e. horse) and the murine background in each hybrid cell line. Most of the primers used in this study were designed to anneal to conserved short exon sequences flanking introns, which should therefore result in the amplification of fragments of different sizes and/or sequences in the two species.
A total of 289 previously published primers for mammalian type I loci (Venta et al. 1996; Lyons et al. 1997) were tested for their ability to amplify horse-specific fragments useful for synteny mapping with a SCH panel. PCR conditions were optimized to amplify markers for 42 equine type I loci (Table 1). Digestions of PCR products from mouse and horse using restriction enzymes were used to obtain gene-specific equine markers with an additional 26 primer sets. In one instance (RB1), an equine-specific primer was designed (Table 1) based on the sequence obtained from a fragment amplified from horse with primers published previously for this gene (Venta et al. 1996). The remaining primers amplified no fragments (5 primers), multiple equine fragments (180 primers), or fragments from horse and mouse that could not be differentiated with the techniques used (36 primers).
PCR Conditions for Amplifying 68 Equine Type I Loci with Universal Primers
Sequencing PCR Products
To verify that the PCR products used for making synteny assignments correspond to the expected equine gene homologs, a total of 50 isolated PCR fragments amplified from a thoroughbred horse were cloned and sequenced. The remaining 18 horse PCR products were gel-isolated and partially sequenced by direct sequencing (Table1). All of the nucleotide sequences obtained were used to search GenBank with the BLAST routine and confirm the identity of 66 of the genes mapped. In two instances (KRAS2, SOD2), the cloned equine PCR fragment lacked the expected intron sequence, sug gesting that the PCR products mapped are pseudogenes.
Many of the primers that amplified horse-specific fragments were designed to target specific genes from large paralogous gene families. Several of the isolated PCR products cloned were found to contain multiple fragments of highly homologous sequences, most likely amplified from different syntenic isoforms of paralogous genes (Table1). For example, the MYH6 primers (Lyons et al. 1997) amplified two equine fragments that are 97% identical at the expected exon regions. These equine fragments are highly homologous to the human myosin heavy polypeptide alpha and beta genes (MYH6 andMYH7, respectively), which are closely linked and show 93% sequence identity. Similarly, the primers used to amplify two highly homologous equine fragments from the equine IL1B gene (92% identity) were originally designed to amplify IL1A (Lyons et al. 1997), two genes that are closely linked in human Chromosome 2. Therefore, although these markers were useful for determining the equine synteny assignment of the respective loci, further characterization will be necessary for determining which gene sequence was amplified from these paralogous gene families in the horse.
In one instance (SST), the equine PCR product contained fragments with the expected sequence in addition to fragments with completely unrelated sequence, except for the primer binding sites. Primers specific to the anonymous equine fragment were designed (F-5′-TTTCCATGGACTTATTTCCC-3′/R-5′-TCCCTTGTTACCTGGAGTATG-3′). The PCR product amplified with these primers was only found in the SCH clonal lines where the SST product was found, suggesting this unrelated sequence is syntenic with the equine SST gene.
SCH Panel Analysis with Equine PCR-based Type I Loci
The genome of the horse consists of 31 autosomes plus the X and Y sex-chromosomes. Previous work with the UC Davis horse-mouse SCH panel established 33 equine synteny groups. A total of 25 synteny groups were assigned to their respective horse chromosomes by correlation with microsatellites and type I loci mapped previously by FISH (Shiue et al. 1999; Caetano et al. 1999a), with microsatellites assigned by DNA typing of trisomic horses (Bowling et al. 1997) and with sex-chromosome type I markers (Shiue 1999). Chromosome assignments of synteny groups were made based on a range of one to six physically assigned loci.
The 68 markers for equine type I loci we characterized were mapped to 27 different synteny groups (Table 2). Forty-five of these markers were mapped to 18 synteny groups that had been assigned previously to specific horse chromosomes. The remaining 23 markers were mapped to 9 synteny groups which are not presently assigned to equine chromosomes.
Correlation Values of 68 Equine Type I Loci with Markers Previously Assigned to UCD Davis Syntenic Groups and/or Horse Chromosomes
DISCUSSION
Sixty-eight makers for equine type I loci were characterized and mapped by SCH panel analysis. These data, combined with 26 type I loci previously assigned with the UC Davis SCH panel (Caetano et al. 1999a,b) and 32 type I loci mapped by other groups working on the characterization of the horse gene map, provide the most comprehensive comparative mapping data currently available for the horse genome (Table 3), when considered in conjunction with chromosome painting studies of the horse karyotype (Raudsepp et al 1996, 1997; Rettenberger et al. 1996; Chaudhary et al. 1998). These data have allowed us to confirm the assignment of 24 equine synteny groups to their respective chromosomes, to provisionally assign nine synteny groups, and to further refine the genetic composition of two horse chromosomes established previously with Zoo-FISH.
Human–Horse Comparative Gene Map Using Synteny, Linkage and in Situ Hybridization (127 Genes)
Provisional Assignment of Equine Synteny Groups by Comparative Mapping
The lack of markers physically mapped to certain horse chromosomes has not allowed for the assignment of all 33 established equine synteny groups to their respective chromosomes. Nonetheless, the consideration of the synteny assignments of the equine type I loci we mapped and the other markers mapped with the UC Davis SCH panel, in conjunction with the data obtained by Zoo-FISH studies of the horse karyotype have allowed the tentative assignment of nine equine synteny groups.
A total of four equine type I loci were mapped to synteny group UCD5 (AT3, GBA, LAMC1, NGFB). The corresponding human homologs of these genes have been mapped to human Chromosome 1 (Table 1). Raudsepp et al. (1996) reported that painting probes from Hsa1, hybridize to three horse chromosomes (ECA2, 5, and 30). The previous assignment of UCD synteny groups 2 and 30 to ECA2 and ECA30, respectively (Shiue et al. 1999), therefore suggests that UCD5 is located on horse chromosome 5.
Four equine genes were assigned to UCD synteny group D (Table 2). The human homologs of equine FN1 and CHRNG have been mapped to Hsa2, and the homologs of PFKM andIFNG have been mapped to Hsa12. Furthermore, we have recently assigned PAX3 (Hsa2q35) and PMEL17(Hsa12q13) to this synteny group and physically mappedPMEL17 to ECA6 by FISH (A.T. Bowling et al., unpubl. observations). The q arm of horse Chromosome 6 was shown to hybridize to painting probes from Hsa12, whereas the p arm did not hybridize to probes from any human chromosome (Raudsepp et al. 1996). Therefore, the assignment of UCD-D to horse Chromosome 6 is in agreement with ZooFISH studies of the horse karyotype and suggests that the p arm of ECA6 contains material orthologous toHsa2.
We had previously assigned UCD6 to ECA6 (Shiue et al. 1999) based on the assignment of a microsatellite (ASB14) by FISH (Breen et al. 1997), but this assignment has now been withdrawn (M. Breen, pers. comm.). Two horse genes, TCF1 (Hsa12) and TS(Hsa18), were assigned to UCD6 (Table 2). Therefore, the consideration that ECA8 is the only horse chromosome to hybridize to painting probes from human Chromosomes 12 and 18 (Raudsepp et al. 1996), and that no synteny groups have been assigned to this horse chromosome, suggests that UCD6 should be tentatively assigned toECA8.
Painting probes from human Chromosomes 3, 9, 10, and 13 were shown to hybridize exclusively to horse Chromosomes 16 and 19, 23 and 25, 1 and 29, and 17, respectively (Raudsepp et al. 1996). Equine genes with homologs located on these human chromosomes were assigned to UCD synteny groups 16 and 19, 23 and 25, 29, and 17, respectively (Table2). The consideration that synteny groups UCD1, UCD19, and UCD23 were assigned respectively to ECA1, ECA19, andECA23 (Shiue et al. 1999), suggests the provisional assignment of UCD synteny groups 16, 17, 25, and 29 to ECA16,ECA17, ECA25, and ECA29, respectively. The assignment of equine ADRB2 (Hsa5) to synteny group UCD14 is in agreement with the tentative assignment of this synteny group to horse Chromosome 14 as proposed by Caetano et al. (1999b).
Equine gene homologs mapping to human Chromosomes 12 and 22 were assigned to UCD synteny group C. Three horse chromosomes (ECA1, ECA8, and ECA28) have been shown to have blocks orthologous to Hsa12 and Hsa22 (Raudsepp et al. 1996). As mentioned previously, UCD1 has been assigned toECA1 (Shiue et al. 1999) and we have tentatively assigned UCD6 to ECA8. Therefore, the most consistent explanation of the current data is the provisional assignment of UCDC to horse Chromosome 28.
In view of the described data, ECA27 is the only horse chromosome without an assigned or provisionally assigned synteny group. UCDA is the only established synteny group that has not been assigned to a horse chromosome. Therefore, the tentative assignment of UCDA to horse Chromosome 27 is reasonable, given the current data.
In addition, the assignments of type I loci to horse synteny groups have allowed two refinements of the human-horse comparative data produced with in situ hybridization experiments. The q arm ofECA13 was shown to be orthologous to Hsa16, whereas the orthology of the p arm of this horse chromosome was not determined (Raudsepp et al. 1996). The human homologs of two equine genes (ELN and GUSB) assigned to ECA13 have been mapped to Hsa7, which therefore suggests that the p arm of this horse chromosome contains material orthologous to human Chromosome 7. Similarly, the orthology of the p arm of ECA6 could not be determined (Raudsepp et al. 1996) but the assignment of equine genes with homologs located on Hsa2 to ECA6 (UCDD), suggests that the p arm of this horse chromosome contains material orthologous to human Chromosome 2. Physical mapping of these equine genes mapped by synteny will be necessary to confirm these inferences.
Application of Universal Primers for Type I Loci
Although the published universal primers used for generating type I equine markers were highly useful for producing a first generation comparative map of the horse genome with the techniques used, additional markers could be readily developed from a few of the remaining primers tested. A total of 36 of the primers tested amplified candidate fragments for equine type I loci which could not be amplified in the somatic cell hybrids, while mouse-specific bands were amplified instead, and therefore could not be used for synteny mapping. We speculate that small deviations of the equine gene sequence from the consensus at the primer binding sites caused the observed preferential amplification of mouse products. Further characterization by cloning and sequencing of horse-specific PCR products from these primers, followed by redesigning of horse-specific primers, could result in additional markers for equine type I loci useful for synteny mapping.
It is likely that the observed amplification of no fragments, multiple fragments, and anonymous equine sequences by a number of the primers used in this study may have been caused by deviations of the equine gene sequences from the consensus sequences used for designing the primers. In addition, expansion of intron sequences within targeted gene regions to sizes beyond the amplification range of Taq DNA polymerase may also have caused the observed results. In these instances, markers for equine type I loci may be developed by redesigning new universal primers to target other intron regions of the respective genes using the reported strategies (Venta et al. 1996;Lyons et al. 1997).
General Considerations
Because synteny mapping by SCH panel analysis does not allow for the precise determination of the relative order and distance between syntenic markers, further work is needed to integrate the markers from this study into the International Equine Linkage Map (Guérin et al. 1999). Intraspecific sequence variation at the intron sequences amplified in most of the markers we mapped can be used to produce genetic markers useful for linkage mapping studies with pedigreed families (Lyons et al. 1997), and therefore order these genes relative to each other and to other markers, in the horse linkage map. In addition, dinucleotide repeat sequences found in the intron regions of a few of the equine PCR products we sequenced are currently being characterized (Caetano 1999) and may be useful in linkage studies as markers for the respective loci. The integration of type I loci into the horse linkage map will be useful for further refining the comparative map of the horse and revealing existing intrachromosomal rearrangements which may have occurred during the evolution of the horse genome.
The assignment of synteny groups to horse chromosomes has been highly dependent on the physical mapping of type I and II markers by FISH. Additional physical mapping data is needed to confirm proposed assignments, ratify provisional assignments, and moreover, to proceed with the establishment of a basic framework map of the horse genome. The small average size of the PCR-based type I markers we characterized precludes their direct use as probes in in situ hybridizations to metaphase chromosome spreads. The technical requirements of current FISH procedures demand larger DNA probes, which can be isolated from available horse BAC libraries (Godard et al. 1998) with the equine markers we characterized. In addition to their use in physical mapping, the isolated large-insert genomic clones may also be used for isolating microsatellite sequences associated with the respective equine genes, which could then be used in linkage studies.
Although type I gene homologs from all human chromosomes have been assigned to the horse genome map (Table 3), no equine type I loci have been mapped to horse Chromosomes 27 and 30. Zoo-FISH was used to show that ECA30 hybridizes to DNA fragments from human Chromosome 1 (Raudsepp et al. 1996) but the homeology of ECA27 could not be determined with this technique. A continued effort to develop and map markers for equine type I loci will be necessary to fill the gaps in the horse-human comparative map and to continue its refinement.
METHODS
Optimization of PCR Amplification Parameters
The methods utilized for optimizing PCR conditions to amplify horse-specific fragments with each primer set for use in synteny mapping with a horse-mouse SCH panel have been described previously (Caetano et al. 1999a). Optimal conditions for amplifying 68 equine type I loci are indicated in Table 1. Horse and mouse PCR products of the same length were subjected to a panel of 15 restriction enzymes (not shown), in the buffer and temperature conditions recommended by the suppliers, prior to electrophoresis. Restriction enzymes selected to distinguish these PCR products from the two species are indicated in Table 1.
Sequencing
Plugs containing DNA fragments amplified from a thoroughbred horse with each of the primer sets selected for mapping were taken from agarose gels with glass Pasteur pipettes and incubated at room temperature in 75 μl of low TE buffer (10 mm Tris-HCl, 0.1 mm EDTA, pH 8.0) overnight. This solution was used as template for a PCR reaction under the conditions used to amplify the original product. An aliquot from this PCR reaction was quantified in an agarose gel and the remaining was used for cloning (TA Cloning Kit, Invitrogen). A minimum of 3 clones from each fragment was sequenced with ABI Prism sequencing kits and sequencing products were analyzed with an ABI 377 automated sequencer. Large PCR fragments that could not be readily cloned with the techniques used were isolated from agarose gels with the QIAEXII kit (Qiagen) and sequenced by direct sequencing. A few of the type I markers characterized were not completely sequenced because of the presence of extensive intron sequences which could not be used for sequence comparisons with other species. Each sequence obtained was subjected to BLAST searches of GenBank at the National Center for Biotechnology Information internet server (http://www.ncbi.nlm.nih.gov/). Sequences from 68 equine genes were submitted to GenBank.
Somatic Cell Hybrid Panel Analysis
The establishment of the UC Davis SCH panel has been described (Shiue et al. 1999). DNA from the same 108 horse-mouse heterohybridoma cell lines used in previous studies (Shiue et al. 1999; Caetano et al. 1999a,b) were used in this study. DNA from each cell line was amplified with each of the 68 primer sets (Table 1) and scored for the presence or absence of horse-specific fragments after electrophoresis. Amplification products obtained from each hybrid cell line with 26 primer sets were digested with the respective restriction enzymes prior to electrophoresis (Table 1). Correlation coefficients were calculated between all of the markers in the UC Davis SCH panel database and each of the 68 loci studied. A correlation value ≥0.70 (Table 2) was accepted as evidence for synteny between two markers (Chevalet and Corpet 1986).
Acknowledgments
A.R.C. was supported by a fellowship from Conselho Nacional de Desenvolvimento Cientifico e Tecnológico (CNPq), Brasília, Brazil.
