Journal of Heredity Advance Access originally published online on June 16, 2009
Journal of Heredity 2009 100(Supplement 1):S54-S65; doi:10.1093/jhered/esp031
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
This article appears in the following Journal of Heredity issue: Symposium Issue: Fourth International Conference on Advances in Canine and Feline Genomics and Inherited Diseases, Saint Malo, Brittany, France, 21-24 May 2008. [View the issue table of contents]
Original Articles |
Highly Endangered African Wild Dogs (Lycaon pictus) Lack Variation at the Major Histocompatibility Complex
Division of Ecology and Evolutionary Biology, Graham Kerr Building, University of Glasgow, Glasgow, G12 8QQ, UK (Marsden, Mable, and Cleaveland); Centre for Integrated Genomic Medical Research, University of Manchester, Manchester, UK (Kennedy); Institute of Zoology, Regent's Park, London, UK (Woodroffe); Wildlife Conservation Research Unit, Department of Zoology, University of Oxford, Tubney, Oxon, UK (Rasmussen); Painted Dog Research, Natural History Museum, Bulawayo, Zimbabwe (Rasmussen); Botswana Predator Conservation Trust, Maun, Botswana (McNutt); Tanzanian Wildlife Research Institute, Arusha, Tanzania (Emmanuel); and Royal Zoological Society of Scotland, Edinburgh zoo, Corstorphine Road, Edinburgh, EH12 6TS, UK (Marsden and Thomas)
Address correspondence to Clare Marsden at the address above, or e-mail: c.marsden.1{at}research.gla.ac.uk.
The major histocompatibility complex (MHC) is a set of highly polymorphic genes involved in the immune response. Extensive research on the canid MHC has found moderate-to-high levels of diversity at the DLA-DRB1, DLA-DRA, DLA-DQA1, and DLA-DQB1 class II loci with frequent transspecific polymorphism among Canis species. In this study, we assessed MHC variation in the more distantly related and highly endangered African wild dog (Lycaon pictus). We screened 168 African wild dogs from Eastern and Southern Africa as well as 200 samples from the European captive population for variation at MHC class II loci. As for all other canids screened to date, we found a single allele at DLA-DRA, which was the same as that found in Canis species. In contrast, we found 17 DLA-DRB1 alleles, one DLA-DQA1 allele, and two DLA-DQB1 alleles, all of which were unique to African wild dogs. At DLA-DRB1, African wild dogs were found to have comparable numbers of alleles but less overall amino acid variation than other canids. However, the low numbers of alleles at DLA-DQA1 and DLA-DQB1 are surprising, given that in other canids, these loci are also highly variable. Overall, our data suggest that African wild dogs are genetically depauperate at the MHC relative to other canids. These data are indicative of a loss of genetic variation, possibly as a result of population bottlenecks and declines experienced by this species.
Key Words: adaptive variation • DLA • Lycaon pictus • MHC • population bottleneck
The major histocompatibility complex (MHC) is a highly diverse set of vertebrate genes that code for molecules involved in the recognition of intra- and extracellular antigens and, therefore, form a fundamental component of immune responses (Eggert et al. 1998; Hedrick 2003; Piertney and Oliver 2006). MHC genes are renowned for their high allelic diversity and heterozygosity, which is thought to be the result of pathogen-driven balancing selection (Van Den Bussche et al. 1999). Diversity at the MHC is adaptively significant in disease resistance; high diversity has been shown to allow response to a wider range of parasites and pathogens than low diversity (Hedrick et al. 2001, 2003; Sommer et al. 2002). Given the importance of adaptive genetic variation for evolutionary change and rising concerns about infectious diseases in the conservation of endangered species (Daszak et al. 2000), assessments of MHC variation are increasingly incorporated into endangered species research (e.g., giant panda Ailuropoda melanoleuca (Wan et al. 2006), crested ibis Nipponia nippon (Zhang et al. 2006), and Mexican wolf Canis lupus baileyi (Hedrick et al. 2000)).
Considerable research has been conducted on the canid MHC (known as the dog leukocyte antigen, DLA) in the domestic dog Canis familiaris and more recently, wild Canis species: Gray wolf (Canis lupus) (Seddon and Ellegren 2002; Kennedy et al. 2007), Coyote (Canis latrans) (Seddon and Ellegren 2002), Ethiopian wolf (Canis simensis) (Kennedy LJ, unpublished data), and Mexican wolf (Hedrick et al. 2000). Research has focused on variation at 3 MHC class II loci: DLA-DRB1, DQA1, and DQB1, which are physically tightly linked and inherited as a haplotype (Kennedy et al. 2007). MHC class II loci are involved in the recognition of antigens of extracellular pathogens and parasites. However, strong linkage disequilbrium has been found between MHC class I and II loci in humans (Sanchez-Mazas et al. 2000), domestic dogs, and many other species studied, suggesting that variation at MHC class II loci can also reflect variation at MHC class I loci, which are involved in the recognition of intracellular pathogens such as viruses (Piertney and Oliver 2006). To date, 134 DLA-DRB1 alleles, 26 DLA-DQA1 alleles, and 68 DLA-DQB1 alleles have been assigned official names by the DLA Nomenclature Committee (Kennedy LJ, unpublished data). These genes have been shown to be polymorphic across the Canis genus, with particularly high levels of polymorphism in both the domestic dog and Gray wolf (Seddon and Ellegren 2002; Kennedy 2007; Kennedy et al. 2007). Transspecific polymorphism (allele sharing) has been found to be a recurring feature among Canis species at all 3 loci. A fourth locus, DLA-DRA, appears to be monomorphic for allele DLA-DRA*0101 in all canids screened to date (Kennedy LJ, unpublished data). Given the focus of research on the genus Canis, it is not currently known if these patterns of MHC polymorphism are specific to these species or a characteristic of canids in general.
African wild dogs are the sole member of the Lycaon genus and a distantly related member of the wolf-like canid clade, to which the genus Canis belongs (Girman et al. 1993). This highly endangered social species has suffered extensive declines in the wild to <6000 individuals distributed across a few remaining small and fragmented populations (Figure 1) (Woodroffe and Ginsberg 1997; Sillero-Zubiri et al. 2004). Disease is argued to represent a significant threat to African wild dogs, which share susceptibility to diseases of common sympatric canids such as jackals and domestic dogs (Alexander et al., forthcoming), outbreaks of which have resulted in both pack and population extinctions in the past (reviewed in Woodroffe et al. 2004). Consequently, knowledge of the MHC is particularly pertinent to African wild dog conservation.
|
In this study, we have characterized MHC class II DLA-DRB1, DRA, DQA1, and DQB1 variation in African wild dogs to extend knowledge of the canid MHC to more distantly related canid species (Lindblad-Toh et al. 2005). Specifically, we assessed levels of polymorphism at MHC class II loci and looked for evidence of allele sharing between African wild dogs and species in the genus Canis.
 |
Methods |
|---|
 Top Methods ResultsDiscussionSupplementary MaterialFundingReferences |
|---|
Blood, tissue, hair, and serum samples were provided from free-ranging study populations in Eastern and Southern Africa (Figure 1): Laikipia, Central Kenya (n = 56 from 13 packs; study population size
300); Serengeti, Northern Tanzania (n = 14 from 4 packs; study population size 160); Okavango, Northern Botswana (n = 53 from 8 packs; study population size 200); and Hwange, Western Zimbabwe (n = 15 from 7 packs; study population size 250). The sampled Serengeti population, hereafter referred to as New Serengeti, represents a population that is thought to have naturally re-established in the early 2000s, rather than the Serengeti population assessed in previous genetic studies (Girman et al. 1997, 2001), which was extirpated with the last pack disappearing in 1991 (Woodroffe et al. 1997). South African samples were derived from a set of animals artificially reintroduced and translocated between game reserves in South Africa and included some captive animals of South African origin (A. Bastos, n = 43). This South African sample set is considered as a managed group of animals rather than a free-ranging population. A further 6 samples from this managed group were collected from a set of 16 wild dogs that were translocated from Pilanesberg Game Reserve, South Africa, to Hwange National Park, Zimbabwe, in 2006. These animals were analyzed as part of the South African sample set rather than the Hwange sample set. The 15 Hwange samples do not include any animals recently translocated from South Africa or their offspring. Ear or muscle samples were also provided from carcasses collected in Kajiado district in Southern Kenya (R. Woodroffe, n = 1), Ghanzi district in Western Botswana (M. Swarner, n = 1), Northern Sofala province in Central Mozambique (J.-M. André, n = 3 from one pack), and Mangetti district in North Western Namibia (F. Stander, n = 1). We sampled 200 captive African wild dogs (75% of the total population) from the European Endangered Species Program, which are derived from founders from Southern Africa. This sample set was analyzed together and is hereafter referred to as EU zoos. Details of the 31 contributing institutions are given in Supplementary table 1. DNA was extracted from samples using DNeasy extraction kits (Qiagen, Crawley, UK) according to the manufacturer's instructions, with the following modifications: Tissue samples were lysed for 18 h rather than 3; blood spots and hair samples were lysed for 3 h rather than one. A negative control was conducted with all extractions to detect contamination.
Sequence-based typing was conducted on exon 2 of the DLA-DRB1, DLA-DQA1, and DLA-DQB1 loci using locus-specific intronic domestic dog primers that gave products of 303 bp (DLA-DRB1), 345 bp (DLA-DQA1), and 300 bp (DLA-DQB1). Primers were as follows (M13 and T7 tails are underlined): DRBln1: ccg tcc cca cag cac att tc (Wagner et al. 1996); DRBln2M13r: cag gaa aca gct atg acc tgt gtc aca cac ctc agc acc a (Wagner et al. 1996); DQAln1: taa ggt tct ttt ctc cct ct (Wagner et al. 1996); DQAIn2: gga cag att cag tga aga ga (Wagner et al. 1996). DQB1BT7 taa tac gac tca cta tag gg ctc act ggc ccg gct gtc tc (Wagner et al. 1996); DQBR2: cac ctc gcc gct gca acg tg (Kennedy, Barnes, Happ, Quinnell, Bennett, et al. 2002). A fourth MHC class II locus DLA-DRA, which has been shown to be monomorphic in all other canids tested to date (Kennedy LJ, unpublished data), was examined using locus-specific exonic primers: DRAF: gag cac gta atc atc cag gc; DRAR: ggt gtg gtt gga gcg cgc ttt a (Wagner JL, personal communication) and gave products of approximately 261 bp.
Polymerase chain reactions (PCR) were performed in 25 µl reactions containing 1 x Q solution (Qiagen), 1 x PCR buffer containing 15 mM MgCl2 (Qiagen), 1 mM MgCl (Qiagen), 0.4 mM of each DNTP (Invitrogen, San Diego, CA), 0.04 uM of each primer, 0.1 µg/µl BSA (Promega), 1 unit of Hot Startaq (Qiagen), and approximately 25 ng of template DNA. To detect contamination, each PCR was run with both the DNA extraction negative and a PCR-negative control containing no template DNA. Reactions were run on PTC-200 DNA engine machines (MJ Research Inc.). PCR amplifications were conducted with a touchdown protocol: 15 min at 95 °C; 14 touchdown cycles of 95 °C for 30 s; followed by 1 min annealing, starting at 62 °C (DLA-DRB1), 62 °C (DLA-DRA), 52 °C (DLA-DQA1), 68 °C (DLA-DQB1), and reducing at 0.5 °C per cycle; and 72 °C for 1 min. This was followed by 20 cycles of 95 °C for 30 s, 60 °C (DLA-DRB1), 55 °C (DLA-DRA), 50 °C (DLA-DQA1), 65 °C (DLA-DQB1) for 1 min, and 72 °C for 1 min. The protocol ended with a final extension of 72 °C for 10 min. The number of amplifications in the second stage of the PCR protocol was increased from 20 to 30 cycles for DNA derived from hair, blotting paper, and serum samples, which typically yielded lower quantities of DNA.
PCR products were cleaned using ExoSAP-IT (USB) according to the manufacturer's instructions and sequenced on an ABI 3730 sequencer. Sequencing was conducted in both directions for DLA-DRB1, using primers DRBln1 and M13r (cag gaa aca gct atg acc). To reduce costs, unidirectional sequencing was used for DLA-DQA1 and DLA-DQB1, using primers DQAln1 and DQB1BT7, respectively. Sequence data were analyzed using Match Tools and Match Tools Navigator (Applied Biosystems), as described in Kennedy, Barnes, Happ, Quinell, Courtenay, et al. (2002). This method relies on an allele library built from homozygotes. We had 6 heterozygous individuals (South Africa = 3, EU zoos = 3) that did not match any pair of known alleles. Therefore, we cloned these 6 individuals using the TOPO TA cloning system (Invitrogen) and identified a single new allele DRB1*90301. This allele was subsequently found in a further 12 heterozygotes. There were 3 pairs of alleles, which could not be distinguished using the Match tools analytical method because some allele pairs gave the same heterozygous sequence (DRB1*90601/90202 and DRB1*90602/90201; DRB1*90101/90201 and DRB1*90102/90202; DRB1*90101/90601 and DRB1*90102/90602). These ambiguous combinations were resolved using a combination of reference strand-mediated conformation analysis (RSCA) and pedigree information from the zoo populations. RSCA is a genotyping method that separates allelic variants based on conformation-dependent mobility through a gel (Kennedy et al. 2005) and was used to distinguish between ambiguous DLA-DRB1 heterozygous sequences by running ambiguous samples alongside a set of candidate alleles in homozygous form. For the EU zoo samples, individuals with ambiguous allele combinations could be resolved using pedigree information to examine the alleles of siblings, parents, and offspring. For example, individual #P20791 was found to be heterozygous for either 1) DRB1*90601/DRB1*90202 or 2) DRB1*90602/DRB1*90201. Five of its siblings were found to have the following 4 alleles DRB1*90101, DRB1*90201, DRB1*90301, and DRB1*90602, which means that #P20791 must be heterozygous for DRB1*90602/DRB1*90201. Pedigree data were also used to examine segregation of DLA-DRB1 alleles and lineages within families. Chi-square goodness-of-fit tests were used to compare observed segregation patterns to expected genotype combinations under random segregation at a single locus. Pedigree information for the EU zoo samples were provided by H. Verberkmoes. Pedigrees were drawn using SmartDraw 2009.
Preliminary sequencing of 30 individuals for DLA-DQA1, DLA-DQB1, and DLA-DRA revealed just 1, 2, and 1 alleles, respectively. Consequently, we used RSCA (together with sequenced samples as controls) to screen for further variation at these loci. For DLA-DQA1, DLA-DQB1, and DLA-DRA, RSCA analysis was conducted on samples from EU zoos (n = 92), Laikipia (n = 56), New Serengeti (n = 9), Okavango (n = 53), Hwange (n = 13), South Africa (n = 6), and the 6 carcass samples. DNA from 5 New Serengeti, 2 Hwange, and 43 South Africa samples were not available in time for RSCA analysis; however, sequence-based typing detected no new DLA-DRB1 alleles in these samples. DLA-DQB1 typing was conducted on an additional 25 Okavango samples that were not successfully typed at the DLA-DRB1 due to low-quality DNA and RSCA failures. Because RSCA was used to screen for new variants and the EU zoos included large family groups, we did not type offspring if we typed both parents and screened a maximum of 3 animals per litter. In total, we typed 92 individuals representative of 36 sibling groups from the 200 captive samples.
The new alleles identified in this study were submitted to the DLA nomenclature committee. Those that met the appropriate criteria were recognized and assigned official names by the committee. Prior to this study, preliminary data (Kennedy LJ, Bacon H, Radford A, unpublished data) based on 4 African wild dog museum samples provided by the National Museums of Scotland (A. Kitchener) had identified 3 DLA-DRB1 alleles (DLA-DRB1*90101, 90102, and 90201), 1 DLA-DQA1 allele (DLA-DQA1*01901), and 2 DLA-DQB1 alleles (DLA-DQB1*90101 and 90201). One allele did not fulfill the naming criteria and is referred to by its local name "fmut."
Sample sizes varied from 14–56 for nonmanaged populations. Therefore, we used rarefaction to compensate for sampling disparity between study populations by standardizing to a population size of 10 using the program HP-Rare v.4.1 (Kalinowski 2005). We calculated nucleotide diversity in populations as the average number of segregating sites 
and pairwise diversity 
, in DnaSP 4.20 (Rozas and Rozas 1995), using a Jukes–Cantor model of substitutions and standard errors calculated with 5000 bootstrap replications. We tested for an excess of heterozygosity relative to Hardy–Weinberg proportions, which is indicative of selection on the current generation, using the U test in Genepop 4.0 (Raymond and Rousset 1995). Synonymous and nonsynonymous genetic distances were calculated separately for putative peptide-binding region (PBR) sites and non-PBR sites using the Nei–Gojobori method with a Jukes–Cantor model of substitutions in Mega 4.0 (Tamura et al. 2007). Putative PBR sites were based on the human HLA-DRB1 (Brown et al. 1993). Due to the recombining nature of MHC genes, phylogenetic trees are not strictly appropriate for analysis of the MHC and there is too much variation to allow a network approach. However, MHC allele trees are a useful tool for displaying relationships among alleles. Phylogenetic trees were constructed using African wild dog sequences alongside 105 alleles from Canis species made available by LJ Kennedy, who collates these data on behalf of the DLA nomenclature committee (Kennedy et al. 2001). We also tested alternative phylogenetic models but these did not affect the resolved relationships within the tree. Therefore, we have only shown neighbor joining trees with Kimura's 2-parameter model as implemented in Mega 4.0 to demonstrate relationships. Following Seddon and Ellegren (2002), a human HLA sequence with 80% similarity to dog DLA-DRB1 alleles was used as an out-group (HLA-DRB1*03011, accession number AF352294). Bootstrapping was conducted with 5000 replicates.
|
Results |
|---|
|
TopMethodsResultsDiscussionSupplementary MaterialFundingReferences |
|---|
African wild dogs were found to have 17 DLA-DRB1 alleles (n = 368), 1 DLA-DQA1 allele (n = 234), 2 DLA-DQB1 alleles (n = 234), and 1 DLA-DRA allele (n = 234). Fewer samples were analyzed at DLA-DQB1, DLA-DQA1, and DLA-DRA because of the lack of variation found. However, we did type representative individuals for all DLA-DRB1 alleles. This is important because in domestic dogs and other Canis species, there is strong linkage between MHC class II loci. Therefore, new DLA-DQB1, DLA-DQA1, and DLA-DRA variants would be most likely found in individuals with new DLA-DRB1 alleles. There was no evidence of pseudogenes (stop codons or frameshift mutations), indicating that functional genes were being amplified. All DLA-DRB1, DQA1, and DQB1 alleles detected in African wild dogs were new and have not been identified in any other canid species to date; accession numbers DQA1 (AM182470 [GenBank] ), DQB1 (FJ648575, FJ648576), and DRB1 (FJ648559-FJ648574). As with all other surveyed wolf-like canids (Kennedy LJ, unpublished data), African wild dogs were monomorphic at DLA-DRA for allele DRA*00101, which was originally identified in domestic dogs (Wagner et al. 1995).
African wild dog DLA-DRB1 alleles varied at 31 polymorphic sites across 95 codons, with 14 substitutions at the first codon position, 10 at the second codon position, and 7 at the third codon position. These changes corresponded to 17 amino acid differences among alleles (Figure 2). This included unique amino acid residues at 2 codons not seen in other canids and 1 new polymorphic site at a putative PBR residue which is monomorphic in all other canids. All DLA-DRB1*907011 alleles differed from each other at the amino acid level, except for DLA-DRB1*907011 and DRB1*907012, which indicates a high level of nonsynonymous substitutions. The majority of nucleotide (22/31) and amino acid (14/17) differences between DLA-DRB1 alleles were found to occur within the 3 hypervariable regions (HVR) (Figure 2) (Kennedy et al. 2007). Nine of the 22 functionally important putative PBR sites of DLA-DRB1 based on human HLA-DRB1 were variable in African wild dogs. The ratio of nonsynonymous to synonymous substitutions at the putative PBR sites was greater than 1.0 and larger than in non-PBR, but it was not found to be significant (PBR: dN = 0.2, dS = 0.117, dN/dS = 1.709, P = 0.073; non-PBR: dN = 0.031, dS = 0.022, dN/dS = 1.409, P = 0.307).
|
DLA-DRB1 alleles consisted of two highly divergent allelic lineages, which we have called A (7 alleles) and B (10 alleles). Alleles within lineages were relatively similar, whereas alleles from different lineages were highly divergent (Figure 2). Lineage A alleles have identical HVR1 and HVR2 sequences. Lineage B alleles have the same HVR1 sequence (which is different from that in lineage A) and 1 of 2 very similar HVR2 sequences which differed by just one amino acid. At HVR3, there were 5 different sequences, 3 of which were shared among lineages, and 2 of which were specific to Lineage B. Overall, the average numbers of nucleotide differences within alleles of the same lineage were 6.0 (lineage A) and 6.8 (lineage B), compared with an average of 22.9 nucleotide differences between alleles from different lineages. Because RSCA analysis, DNA cloning, and sequencing did not detect more than 2 alleles in any individual and less than half of the individuals sampled (46%) had alleles from both lineages, we are confident that these 2 allelic lineages are derived from a single locus. Furthermore, pedigree data clearly show cosegregation of the 2 allelic lineages within families (Figure 3).
|
Phylogenetic analyzes on African wild dog DLA-DRB1 alleles were conducted alongside alleles from Canis species. The highly polymorphic nature of these genes resulted in insufficient resolution to determine specific relationships between groups of alleles; however, they were used to indicate the positioning of African wild dog alleles relative to the alleles of other canids (Supplementary Figure 1). African wild dog DLA-DRB1 alleles were clearly shown to cluster into 2 distinct and separate monophyletic branches rather than being scattered across branches, as found with Gray wolf and Ethiopian wolf alleles. Furthermore, African wild dog alleles were clearly positioned within, rather than peripheral to, the canid DLA-DRB1 allele tree, indicating similarity to other canid alleles. In particular, comparison of amino acid sequences highlight that certain African wild dog DLA-DRB1 lineage B alleles and certain Ethiopian wolf alleles differ by just one amino acid at HVR1 and are identical at HVR2 (data not shown).
DLA-DRB1 alleles from both A and B lineages were found in all populations with more than 3 samples. Four of 7 lineage A and 5 of 9 lineage B DLA-DRB1 alleles were detected in two or more sampling areas, which were often separated by large geographic distances (see Table 1). For example, DLA-DRB1*90202 was found in countries across Eastern (Laikipia, Kenya; New Serengeti, Tanzania) and Southern Africa (Hwange, Zimbabwe; Okavango, Botswana; NW Namibia; and South Africa).
|
African wild dog populations were found to differ from each other in DLA-DRB1 allelic composition, allelic diversity, and heterozygosity. For nonmanaged populations, the number of alleles per population varied between 3 and 9 and average observed heterozygosity varied from 53.6% to 92.9% (see Table 1). Despite being the most thoroughly sampled population, Laikipia had the smallest number of alleles (3 alleles, n = 56) and, correspondingly, also had the lowest observed heterozygosity (53.6%). However, nucleotide diversity was actually highest in this population (
= 0.0758, = 0.0716), suggesting that the 3 alleles are highly divergent; there were 29 variable sites among these 3 alleles. In contrast, nucleotide diversity was lower in the 3 other populations, which had between 7 and 9 alleles ( = 0.0509, 0.0595, 0.0613; = 0.0435, 0.0435, 0.0484). Rarefaction was used to standardize population sample sizes to n = 10 and showed Hwange to be most diverse in terms of numbers of alleles expected with that sample size (7.8 alleles), although New Serengeti and Okavango had only slightly lower levels of diversity (6.1 and 5.8 alleles, respectively). All 3 of these populations had at least 50% more diversity than Laikipia (2.9 alleles). Although levels of observed heterozygosity were generally high, there was not an excess of heterozygosity relative to Hardy–Weinberg expectations in any nonmanaged population. The South African sample set consisted almost entirely of heterozygotes (46/49). However, this is a managed group of animals derived from multiple sources rather than a natural population. Together, the EU zoos were found to have 12 of the 14 DLA-DRB1 alleles detected in Southern African populations and levels of heterozygosity (82%) comparable to nonmanaged wild populations (53.6–92.9%). One allele (DRB1*90101) was found at high frequency among the zoo samples (33.5%). The 2 DLA-DQB1 alleles differed at 8 sites within HVR2, resulting in 5 amino acid differences. This included 1 new polymorphic amino acid site that is monomorphic in other canids tested to date and 4 unique amino acid residues. DLA-DQB1*90101 was considerably more frequent (87.5%) than DLA-DQB1*90201 (12.5%) resulting in a predominance of DLA-DQB1*90101 homozygotes (81%). In fact, we found just 6 DLA-DQB1*90201 homozygotes in 234 samples. Both DLA-DQB1 alleles were found across Eastern and Southern Africa (Table 2); however, DLA-DQB1*90201 was noticeably absent from Hwange, Zimbabwe. Previous research has shown strong linkage disequilibrium between the canid DLA-DRB1, DQA1, and DQB1 loci (Kennedy et al. 2007). There was insufficient variation at the DLA-DQA1 and DQB1 loci for haplotype designation in African wild dogs. However, we did detect an association between DLA-DQB1*90201 and DLA-DRB1 lineage A alleles. Six of 7 individuals homozygous for DLA-DQB1*90201 had only lineage A DLA-DRB1 alleles (DRB1*90101, *90201, *90202, or *90204). Furthermore, all DLA-DQB1*90201 heterozygotes had at least one DLA-DRB1 lineage A allele, most commonly DRB1*90101, DRB1*90201, or DRB1* 90202.
|
|
Discussion |
|---|
|
TopMethodsResultsDiscussionSupplementary MaterialFundingReferences |
|---|
Research on MHC class II loci in Canis species has shown moderate-to-high levels of diversity at the DLA-DRB1, DLA-DQA1, and DLA-DQB1 class II loci with frequent transspecific polymorphism (allele sharing) among Canis species. In this study, we conducted a geographically widespread survey of MHC class II variation in the highly endangered African wild dog to extend knowledge of the canid MHC to more distantly related canid species. African wild dogs belong to a monotypic genus that is phylogenetically and morphologically divergent from Canis species (Wayne et al. 1997; Bardeleben et al. 2005). In total, we found 17 alleles at the DLA-DRB1 locus, 1 allele at the DLA-DQA1 locus and 2 alleles at the DLA-DQB1 locus, all of which are currently unique to African wild dogs. At DLA-DRA, African wild dogs were monomorphic for the same allele found in other canids.
Balancing selection is a key mechanism in the maintenance of variation at MHC loci (reviewed in Garrigan and Hedrick 2003) and is indicated by an increased ratio of nonsynonymous (dN) to synonymous (dS) substitutions at the amino acid residues of the functionally important PBR (Seddon and Ellegren 2002). Although dN/dS was elevated at putative PBR sites of DLA-DRB1 alleles in African wild dogs, there was not a significant excess of nonsynonymous substitutions (P = 0.073). This is not typical of canid DLA-DRB1 alleles; there was a significant excess of dN/dS at PBR sites in Gray wolves, Coyotes, and domestic dogs (Seddon and Ellegren 2002). Whereas dN/dS ratios provide information on historical selection, excess heterozygosity can provide an indication of current selection at a locus (Garrigan and Hedrick 2003; Aguilar et al. 2004). Despite the high heterozygous frequencies found in nonmanaged free-ranging populations, the observed heterozygosity did not exceed Hardy–Weinberg expectations. This is not atypical for MHC studies (Garrigan and Hedrick 2003).
The distribution of alleles from polymorphic loci under balancing selection are predicted to show very different distributions from that of neutral loci. In particular, they are expected to show lower levels of differentiation in allele composition between populations (Schierup et al. 2000). Neutral genetic markers show strong structuring and differentiation between African wild dogs populations, in particular between Eastern and Southern Africa (Girman et al. 2001). At the MHC, we found 17 DLA-DRB1 alleles, which clustered into 2 highly distinct lineages. These 2 lineages showed no evidence of geographic structuring; all areas where more than 3 animals were sampled had alleles from both lineages. Similarly, individual DLA-DRB1 alleles were not geographically restricted, with many alleles detected in populations spanning Eastern and Southern Africa. The discordance between patterns of MHC and neutral variation could indicate that selective forces are shaping patterns of MHC diversity across African wild dog populations; for example, selection for alleles which confer resistance to diseases common to most populations.
Two DLA-DQB1 alleles were detected in African wild dogs. However, allele DLA-DQB1*90201 was considerably rarer (12.5%). This rare allele was found across Eastern and Southern Africa but was absent from Hwange. This may be the result of the low frequency of DLA-DRB1 lineage A alleles in Hwange (20%), which appear to be associated with DLA-DQB1*90201. The stark differences in frequency of the 2 DLA-DQB1 alleles may be indicative of selection on adaptive differences between these alleles or haplotypes.
High MHC allelic diversity in a population and high heterozygosity in individuals is thought to be important because it theoretically expands the range of pathogens to which a population or individual can respond (Doherty and Zinkernagel 1975; Sommer et al. 2002). We found that the number of DLA-DRB1 alleles and levels of heterozygosity varied between populations (Table 1), even after population sample sizes were standardized using rarefaction. This may reflect differences in demographic history and connectivity. The highest allelic diversity in nonmanaged populations was found in Hwange (9 alleles, n = 15), which is a long-standing stable population located within an admixture zone (Girman et al. 2001). In contrast, the lowest number of alleles was found in Laikipia (3 alleles, n = 56), a recently recolonized population, which is also relatively isolated (Woodroffe et al. 2007). Clearly, however, recolonization does not always result in low numbers of alleles because the recently recolonized New Serengeti population was considerably more diverse than Laikipia. However, the New Serengeti is linked to a number of other African wild dog populations and therefore may have been recolonized by a mixture of founders from multiple source populations. Despite lower allele numbers, nucleotide diversity among alleles was higher in the 2 recently recolonized populations (Laikipia and New Serengeti) than in two long-standing populations (Hwange and Okavango) (Table 1). The lower nucleotide diversity measures in Hwange and Okavango likely reflect the presence of closely related or similar alleles and may indicate that in these populations, new diversity has been accumulating, whereas in the recently recolonized populations, there has been insufficient time for the evolution of new variants. More research is required to explore whether differences in MHC diversity between populations reflect differences in disease characteristics of populations or neutral processes such as size of historical bottlenecks.
The use of fitness-related genes, such as the MHC, in endangered species management remains a contentious issue. Nonetheless, it is valuable to evaluate the impact of management actions, such as translocations and captive breeding, on adaptive genes. In 2006, 16 African wild dogs were translocated from South Africa to Hwange, Zimbabwe. Sampling of 6 of these South African translocated animals detected 1 allele (DLA-DRB1*90301) not present in the 15 resident Hwange samples. This may indicate that the translocation has introduced new MHC diversity into the Hwange population. Our results show that 12 of the 14 DLA-DRB1 alleles found in Southern African populations and both DLA-DQB1 alleles are represented in the European zoo African wild dog population. Nonetheless, allele DLA-DRB1*90101 clearly dominates this population (33.8%). High frequency of this allele does not appear typical to Southern Africa where the EU zoo founders originated; it has less than 10% representation in Hwange, Okavango, and South Africa. Mapping DLA-DRB1 alleles onto the EU zoo pedigree (data not shown) shows that overrepresentation of this allele is the result of an extreme bias in founder contributions and is a major cause of homozygosity in the EU zoos (21/35 homozygotes were homozygous for DRB1*90101). Management of this population is now focusing on equalizing founder representation.
The patterns of MHC variation detected in African wild dogs are best interpreted through comparison with other canids. Extensive research on the MHC in Canis species shows frequent transspecific polymorphism at DLA-DRB1, DQA1, and DQB1 loci (Kennedy et al. 2001; Seddon and Ellegren 2002; Kennedy et al. 2007). By contrast, all alleles characterized at these 3 loci in African wild dogs were unique to this species and not yet found in any species of Canis. Furthermore, phylogenetic analyses of African wild dog DLA-DRB1 alleles showed clustering into 2 distinct branches (species-specific allelic clustering) rather than a scattered distribution throughout the DLA-DRB1 tree indicative of transspecific polymorphism (as seen in Gray wolves and Ethiopian wolves). Such a distribution may suggest that the canid DLA-DRB1, DLA-DQA1, and DLA-DQB1 allele lineages diverged prior to speciation within the genus Canis 1–2 Ma (Seddon and Ellegren 2002) but after the divergence of the Lycaon and Canis genera approximately 4–5 Ma (Wayne et al. 1997). However, given that allele sharing is most common among species of Canis at DLA-DQA1 and DLA-DQB1 loci (Seddon and Ellegren 2002; Kennedy et al. 2007), whereas there are 1 and 2 alleles, respectively, in African wild dogs, it is also possible that shared alleles have been lost.
Allelic diversity at DLA-DQA1 and DLA-DQB1 was much lower in African wild dogs than expected based on the pattern found in other canids (Table 3, Hedrick et al. 2000; Seddon and Ellegren 2002; Kennedy, Barnes, Happ, Quinnell, Courtenay, et al. 2002; Kennedy 2007; Kennedy et al. 2007). This cannot be explained by the endangered status of African wild dogs or differences in sampling because they had lower levels of DLA-DQA1 and DLA-DQB1 variation than Ethiopian and Mexican wolves; 2 other endangered canids sampled from single populations (Table 3). It is particularly striking that 5 DLA-DQA1 alleles were found in fewer than 7 Mexican wolves sampled from a single population, whereas in this study, we found just a single allele in 234 African wild dogs sampled across Eastern and Southern Africa. The lack of variation at these loci does not appear to be the result of nonmatching primers as all samples amplified successfully; if a mutation had occurred in the primer site, homozygous individuals for these alleles should fail to amplify. African wild dogs showed the most variation at the DLA-DRB1 loci, where they had the same number of DLA-DRB1 alleles to Gray wolves sampled across a similar geographic range in both European and North American regions (Table 3) and slightly higher numbers of DLA-DRB1 alleles in single populations than other endangered canids. However, because African wild dog DLA-DRB1 alleles are derived from just 2 allelic lineages, amino acid diversity among alleles was considerably lower than for other canids: 17 variable amino acid sites across 17 DLA-DRB1 alleles in African wild dogs compared with 26 variable amino acid sites across 17 alleles in the North American Gray wolf (Kennedy et al. 2007) and 22 variable amino acids sites amongst just 4 alleles in a single Ethiopian wolf population (Kennedy LJ, unpublished data). Furthermore, there was less variation at the putative PBR site residues, which are thought to be primarily responsible for functional differences between alleles (Sommer 2005), in African wild dogs compared with Ethiopian wolves (Kennedy LJ, unpublished data) and North American Gray wolves (Kennedy et al. 2007): total number of variable PBR sites, 9, 11, and 15, respectively; average number of residues/PBR site, 1.5, 1.7, and 2.2, respectively. Consequently, one might speculate that although African wild dogs have a large number of DLA-DRB1 alleles, they may have little functional diversity. Overall, our data suggest that African wild dogs are genetically depauperate at the MHC relative to other canids. They have uncharacteristically low amino acid diversity at the DLA-DRB1 locus and low numbers of alleles at the DLA-DQA1 and DQB1 loci, for a canid, even for an endangered one.
|
African wild dogs may have lost allelic diversity across all MHC class II genes due to historical bottlenecks, with strong disease pressures subsequently maintaining or generating MHC variation at the least conserved region, in this case the DLA-DRB1 locus. The presence of just 2 highly divergent monophyletic allelic lineages for both DLA-DRB1 and DLA-DQB1 is consistent with the hypothesis that this species suffered severe bottlenecks, resulting in the loss of alleles and subsequent evolution of new diversity (Van Den Bussche et al. 1999). However, both DLA-DQB1 alleles and both DLA-DRB1 lineages were represented across African wild dog populations. A range-wide bottleneck would be unlikely to produce such a consistent pattern of diversity loss across populations because this would result in the random loss of variation. It is more likely that African wild dogs suffered local population extinctions across most of the African wild dog range with remnant populations retaining both allelic lineages and subsequently expanding to recolonize their former range.
It is clear from our study that African wild dogs are atypical in their patterns of MHC diversity among the canids, which have been studied to date. However, all canids previously studied at the MHC (domestic dogs, Gray wolves, Coyotes, Ethiopian wolves, Red wolves, and Mexican wolves) have been closely related species of Canis, which have a long history of hybridization (Lehman et al. 1991; Gottelli et al. 1994; Garcia-Moreno et al. 1996; Roy et al. 1996; Vila et al. 1997; Verginelli et al. 2005). Consequently, we cannot distinguish whether African wild dogs show different patterns of MHC polymorphism to Canis species because of factors related to African wild dog demographic history, rather than their distant phylogenetic relationship to the Canis genus, or the fact that they lack extensive hybridization in their recent evolutionary history. Future work is planned on other nonhybridizing species of the wolf-like clade to investigate these alternative hypotheses.
Adaptive genetic variation is of primary interest in conservation genetics; therefore, MHC data have particular application to endangered species programs (Aguilar et al. 2004). In this study, we have shown that the highly endangered African wild dog has a reduced level of MHC variation compared with other canids, perhaps as a result of historical bottlenecks. Among African wild dog populations, levels of MHC diversity were found to vary, but more research is required to investigate the significance of this in relation to differences in disease incidence and exposure. Our data have shown that the distribution of MHC variation does not match the pattern of neutral genetic variation highlighted in previous studies (Girman et al. 2001), indicating that conservation plans based on neutral genetic data alone may not adequately conserve adaptive genetic variation. It is promising, however, that such a high proportion of MHC diversity from free-ranging populations has been successfully conserved within the European captive breeding programme. This high diversity is likely the result of the diverse origin of individuals that founded the European captive population.
|
Supplementary Material |
|---|
|
TopMethodsResultsDiscussionSupplementary MaterialFundingReferences |
|---|
Supplementary material can be found at http://www.jhered.oxfordjournals.org/.
|
Funding |
|---|
|
TopMethodsResultsDiscussionSupplementary MaterialFundingReferences |
|---|
Joint grant between Natural Environment Research Council and the Royal Zoological Society of the Scotland-Edinburgh Zoo (NER/S/A/2006/14139); The Royal Zoological Society of Scotland.
|
Acknowledgments |
|---|
We would like to thank the following African wild dog field projects for contributing samples: African Wild Dog Conservation Project Mozambique, Botswana Wild Dog Project, Painted Dog Research Project Zimbabwe, Samburu-Laikipia Wild Dog Project Kenya, Serengeti Wild Dog Conservation Project Tanzania. In addition, we thank Matt Swarner, Jean-Marc Andre, Tshepo Matjila, Amanda Bastos, Hanru Strydom, Laurence Frank, Flip Stander, and Andrew Kitchener for assistance with sample provision. We are very grateful to all of the European zoos that contributed samples to this research and would like to thank Hanny and Wim Verberkmoes for coordinating captive sampling and providing assistance with analysis of stud book data.
|
Footnotes |
|---|
Corresponding Editor: Francis Galibert
Received November 22, 2008
Revised January 26, 2009
Accepted April 22, 2009
|
References |
|---|
|
TopMethodsResultsDiscussionSupplementary MaterialFundingReferences |
|---|
-
Aguilar A, Roemer G, Debenham S, Binns M, Garcelon D, Wayne RK. High MHC diversity maintained by balancing selection in an otherwise genetically monomorphic mammal. Proc Natl Acad Sci USA (2004) 101:3490–3494.
Alexander KA, McNutt JW, Briggs MB, Standers PE, Funston P, Hemson G, Keet D, van Vuuren M, Forthcoming. Multi-host pathogens and carnivore management in Southern Africa. Comp Immunol Microbiol Infect Dis (2008) doi:10.1016/j.cimid.2008.10.005.
Bardeleben C, Moore RL, Wayne RK. A molecular phylogeny of the Canidae based on six nuclear loci. Mol Phylogenet Evol (2005) 37:815–831.[CrossRef][Web of Science][Medline]
Brown JH, Jardetzky TS, Gorga JC, Stern LJ, Urban RG, Strominger JL, Wiley DC. Three-dimensional structure of the human class II histocompatibility antigen HLA-DR1. Nature (1993) 364:33–39.[CrossRef][Medline]
Daszak P, Cunningham AA, Hyatt AD. Emerging infectious diseases of wildlife; threats to biodiversity and human health. Science (2000) 287:443–449.
Doherty PC, Zinkernagel RM. A biological role for the major histocompatibility antigens. Lancet (1975) 305:1406–1409.[CrossRef]
Eggert F, Muller-rucholtz W, Fertsi R. Olfactory cues associated with the major histocompatibility complex. Genetica (1998) 104:191–197.[CrossRef][Web of Science][Medline]
Garcia-Moreno J, Matocq MD, Roy MS, Geffen E, Wayne RK. Relationships and genetic purity of the endangered Mexican wolf based on analysis of microsatellite loci. Conserv Biol (1996) 10:376–389.[CrossRef][Web of Science]
Garrigan D, Hedrick PW. Perspective: detecting adaptive molecular polymorphism: lessons from the MHC. Evolution (2003) 57:1707–1722.[CrossRef][Web of Science][Medline]
Girman DJ, Kat PW, Mills MGL, Ginsberg JR, Borner M, Wilson V, Fanshawe JH, Fitzgibbon C, Lau LM, Wayne RK. Molecular genetic and morphological analyses of the African wild dog (Lycaon pictus). J Hered (1993) 84:450–459.
Girman DJ, Mills MGL, Geffen E, Wayne RK. A molecular genetic analysis of social structure, dispersal, and interpack relationships of the African wild dog (Lycaon pictus). Behav Ecol Sociobiol (1997) 40:187–198.[CrossRef][Web of Science]
Girman DJ, Vila C, Geffen E, Creel S, Mills MGL, McNutt JW, Ginsberg J, Kat PW, Mamiya KH, Wayne RK. Patterns of population subdivision, gene flow and genetic variability in the African wild dog (Lycaon pictus). Mol Ecol (2001) 10:1703–1723.[Medline]
Gottelli D, Sillero-Zubiri C, Applebaum GD, Roy MS, Girman DJ, Garcia-Moreno J, Ostrander EA, Wayne RK. Molecular genetics of the most endangered canid: the Ethiopian wolf Canis simesis. Mol Ecol (1994) 3:301–312.[Medline]
Hedrick PW. The major histocompatibility complex (MHC) in declining populations: an example of adaptive variation. In: Reproduction science and integrated conservation—Holt WV, Pickard AR, Rodger JC, Wildt DE, eds. (2003) Cambridge (UK): Cambridge University Press. 97–113.
Hedrick PW, Kim TJ, Parker KM. Parasite resistance and genetic variation in the endangered Gila topminnow. Anim Conserv (2001) 4:103–109.[CrossRef]
Hedrick PW, Lee RN, Buchanan C. Canine parvovirus enteritis, canine distemper, and major histocompatibility complex genetic variation in Mexican wolves. J Wildl Dis (2003) 39:909–913.[Abstract]
Hedrick PW, Lee RN, Parker KM. Major histocompatibility complex (MHC) variation in the endangered Mexican wolf and related canids. Heredity (2000) 85:617–624.[CrossRef][Web of Science][Medline]
Kalinowski ST. HP-Rare 1.0: a computer program for performing rarefaction on measures of allelic richness. Mol Ecol Notes (2005) 5:187–189.[CrossRef][Web of Science]
Kennedy LJ. 14th International HLA and Immunogenetics Workshop: report on joint study on canine DLA diversity. Tissue Antigens (2007) 69:269–271.[CrossRef][Web of Science][Medline]
Kennedy LJ, Angles JM, Barnes A, Carmichael LE, Radford AD, Ollier WER, Happ GM. DLA-DRB1, DQA1, and DQB1 alleles and haplotypes in North American Gray Wolves. J Hered (2007) 98:491–499.
Kennedy LJ, Angles JM, Barnes A, Carter SD, Francino O, Gerlach JA, Happ GM, Ollier WER, Thomson W, Wagner JL. Nomenclature for factors of the dog major histocompatibility system (DLA), 2000: second report of the ISAG DLA Nomenclature Committee. Anim Genet (2001) 32:193–199.[CrossRef][Web of Science][Medline]
Kennedy LJ, Barnes A, Happ GM, Quinnell RJ, Bennett D, Angles JM, Day MJ, Carmichael N, Innes JF, Isherwood D, et al. Extensive interbreed, but minimal intrabreed, variation of DLA class II alleles and haplotypes in dogs. Tissue Antigens (2002) 59:194–204.[CrossRef][Web of Science][Medline]
Kennedy LJ, Barnes A, Happ GM, Quinnell RJ, Courtenay O, Carter SD, Ollier WER, Thomson W. Evidence for extensive DLA polymorphism in different dog populations. Tissue Antigens (2002) 60:43–52.[CrossRef][Web of Science][Medline]
Kennedy LJ, Quarmby S, Fretwell N, Martin AJ, Jones PG, Jones CA, Ollier WER. High-resolution characterization of the canine DLA-DRB1 locus using reference strand-mediated conformational analysis. J Hered (2005) 96:836–842.
Lehman N, Eisenhawer A, Hansen K, Mech LD, Peterson RO, Gogan PJP, Wayne RK. Introgression of coyote mitochondrial-DNA into sympatric North-American gray wolf populations. Evolution (1991) 45:104–119.[CrossRef][Web of Science]
Lindblad-Toh K, Wade CM, Mikkelsen TS, Karlsson EK, Jaffe DB, Kamal M, Clamp M, Chang JL, Kulbokas EJ, Zody MC, et al. Genome sequence, comparative analysis and haplotype structure of the domestic dog. Nature (2005) 438:803–819.[CrossRef][Medline]
McNutt JW, Mills MGL, McCreery K, Rasmussen G, Robbins R, Woodroffe R. Lycaon pictus. In: IUCN 2008. 2008 IUCN Red List of Threatened Species (2008) Available from: http://www.lucnredlist.org.
Mech LD, Boitani L. Canis lupus. In: IUCN 2008. 2008 IUCN Red List of Threatened Species (2008) Available from: http://www.iucnredlist.org.
Piertney SB, Oliver MK. The evolutionary ecology of the major histocompatibility complex. Heredity (2006) 96:7–21.[Web of Science][Medline]
Raymond M, Rousset F. GENEPOP (Version 1.2): population genetics software for exact tests and ecumenicism. J Hered (1995) 86:248–249.
Roy MS, Geffen E, Smith D, Wayne RK. Molecular genetics of pre-1940 red wolves. Conserv Biol (1996) 10:1413–1424.[CrossRef][Web of Science]
Rozas J, Rozas R. DnaSP, DNA sequence polymorphism: an interactive program for estimating population genetics parameters from DNA sequence data. Comput Appl Biosci (1995) 11:621–625.
Sanchez-Mazas A, Djoulah S, Busson M, de Gouville IL, Poirier JC, Dehay C, Charron D, Excoffier L, Schneider S, Langaney A, et al. A linkage disequilibrium map of the MHC region based on the analysis of 14 loci haplotypes in 50 French families. Eur J Hum Genet (2000) 8:33–41.[CrossRef][Web of Science][Medline]
Schierup MH, Vekemans X, Charlesworth D. The effect of subdivision on variation at multi-allelic loci under balancing selection. Genet Res (2000) 76:51–62.[CrossRef][Web of Science][Medline]
Seddon JM, Ellegren H. MHC class II genes in European wolves: a comparison with dogs. Immunogenetics (2002) 54:490–500.[CrossRef][Web of Science][Medline]
Sillero-Zubiri C, Hoffman M, Macdonald DW. Canids: foxes, wolves, jackals and dogs. Status survey and action plan (2004) Gland (Switzerland): IUCN.
Sommer S. The importance of immune gene variability (MHC) in evolutionary ecology and conservation. Front Zool (2005) 2:16.[CrossRef][Medline]
Sommer S, Schwab D, Ganzhorn JU. MHC diversity of endemic Malagasy rodents in relation to geographic range and social system. Behav Ecol Sociobiol (2002) 51:214–221.[CrossRef][Web of Science]
Tamura K, Dudley J, Nei M, Kumar S. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol Biol Evol (2007) 24:1596–1599.
Van Den Bussche RA, Hoofer SR, Lochmiller RL. Characterization of Mhc-DRB allelic diversity in white-tailed deer (Odocoileus virginianus) provides insight into Mhc-DRB allelic evolution within Cervidae. Immunogenetics (1999) 49:429–437.[CrossRef][Web of Science][Medline]
Verginelli F, Capelli C, Coia V, Musiani M, Falchetti M, Ottini L, Palmirotta R, Tagliacozzo A, De Grossi Mazzorin I, Mariani-Costantini R. Mitochondrial DNA from prehistoric canids highlights relationships between dogs and South-East European wolves. Mol Biol Evol (2005) 22:2541–2551.
Vila C, Savolainen P, Maldonado JE. Multiple and ancient origins of the domestic dog. Science (1997) 276:1687–1689.[CrossRef][Web of Science][Medline]
Wagner JL, Burnett RC, DeRose SA, Storb R. Molecular analysis and polymorphism of the DLA-DQA gene. Tissue Antigens (1996) 48:199–204.[Web of Science][Medline]
Wagner JL, Burnett RC, Works JD, Storb R. Molecular analysis of DLA-DRBB1 polymorphism. Tissue Antigens (1996) 48:554–561.[Web of Science][Medline]
Wagner JL, Derose SA, Burnett RC, Storb R. Nucleotide-sequence and polymorphism analysis of canine DRA cDNA clones. Tissue Antigens (1995) 45:284–287.[Web of Science][Medline]
Wan Q-H, Zhu L, Wu HUA, Fang S-G. Major histocompatibility complex class II variation in the giant panda (Ailuropoda melanoleuca). Mol Ecol (2006) 15:2441–2450.[Medline]
Wayne RK, Geffen E, Girman DJ, Koepfli KP, Lau LM, Marshall CR. Molecular systematics of the Canidae. Syst Biol (1997) 46:622–653.
Woodroffe R, Cleaveland S, Courtenay O, Laurenson MK, Artois M. Infections disease in the management and conservation of wild canids. In: Biology and conservation of wild canids—Macdonald DW, Sillero-Zubiri C, eds. (2004) Oxford (UK): Oxford University Press. 123–142.
Woodroffe R, Ginsberg J. Past and future causes of wild dogs’ population decline. In: The African wild dog: status survey and action plan—Woodroffe R, Ginsberg J, Macdonald DW, eds. (1997) Gland (Switzerland): IUCN. 58–73.
Woodroffe R, Ginsberg J, Macdonald DW. The African wild dog: status survey and conservation action plan (1997) Gland (Switzerland): World Conservation Union.
Woodroffe R, Lindsey PA, Romanach SS, Ranah S. African wild dogs (Lycaon pictus) can subsist on small prey: implications for conservation. J Mammalogy (2007) 88:181–193.[CrossRef]
Zhang B, Fang SG, Xi YM. Major histocompatibility complex variation in the endangered crested ibis Nipponia nippon and implications for reintroduction. Biochem Genet (2006) 44:113–123.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||