Molecular genetic tools have been used to dissect the evolutionary relationships of the dog-like carnivores, revealing their place in the order Carnivora, the relationships of species within the family Canidae, and the genetic exchange that occurs among conspecific populations. High rates of gene flow among populations within some species, such as the coyote and gray wolf, have suppressed genetic divergence, and where these species hybridize, large hybrid zones have been formed. In fact, the phenotype of the endangered American red wolf may be strongly influenced by hybridization with coyotes and gray wolves. Hybridization and habitat fragmentation greatly complicate plans to conserve the genetic diversity of wild canids.

The dog family, Canidae, is a diverse group of 34 species ranging in size and proportion from squat, dachshund-like bushdogs to the long-legged maned wolf, a species sometimes called;a fox-on-stilts (Fig. 1). This morphological diversity is matched by the diversity of their natural history: canids inhabit temperate and tropical forests, savanna, tundra and deserts throughout the world (Table 1). Moreover, canids have a broader appetite than is commonly realized; most include a substantial proportion of vegetable and insect matter in their diet¹. In the past, the evolutionary relationships of canids have been studied by morphological approaches, but parallel changes in several evolutionary lineages can make inferences uncertain. The use of molecular and biochemical techniques to examine genetic differences among species provides an alternative way to investigate phylogenetic relationships. Such methods also have inherent biases, but systematists can search for groupings that are supported by a number of different approaches, and are thus more likely to be genuine phylogenetic relationships. Molecular genetic approaches have provided information about evolutionary divergence at a number of different levels, ranging from the relationships of the Canidae to other carnivore families, to relationships among populations within a single species. This information is discussed in this review.

Geographic distribution of selected canids

Relationships of canids to other carnivore families
The order Carnivora includes the cat, hyena, bear, weasel, seal, mongoose, civet and dog families (Fig. 2). All have ancient origins some 40-60 million years ago and thus their relationships can be studied by comparing the sequences at single-copy genes that have only a modest rate of sequence evolution^2-4. The degree to which two single-copy DNA sequences have diverged can be estimated by the DT_m which is the difference between the melting temperature (the point where 50% of DNA is double stranded) for a homologous duplex (i.e. both strands from the same species) and a heterologous duplex (with constituent strands from different species). the value is normalized for the final percentage of hybridization and designated DT_mR (Ref. 4). A clustering phylogeny based on the DT_mR between carnivore species shpows that extant species are closely related to each other (DT_mR <4*C) but are only distantly related to species in other carnivore families[5] (DT_mR>18*C). Assuming a constant rate of sequence evolution, the Canidae diverged from other carnivore families approximately 50-60 million years ago, near the time when canids first appeared in the fossil records^6,7. Clearly, the Canidae diverged early in the evolution of carnivores, and one should be cautious about attempting to draw conclusions about carnivore gene structure and function from studies on canids alone.

Relationships of canids to each other
Patterns of evolution within the Canidae have been elucidated by use of protein electrophoresis to study allozyme variants and by comparison of G-banded metaphase chromosomes^8-10 (Fig. 3). The differences between allele frequencies for a large number of loci are first used to calculate the genetic distance between pairs of species; from these genetic distances, clusters of species can be discerned^8,11. Comparative analysis of chromosomes has also proved very useful because canids have a rich diversity of chromosome morphology ranging from species such as the red fox, which has a low diploid number of chromosomes (2n = 36) and all metacentric autosomes, to the gray wolf, which has a high diploid number (2n = 78) and all acrocentric autosomes (Table 1). The primitive canid karyotype has been reshuffled in different lineages, in a way that reveals the phylogenetic history of the group^8-10. The evolutionary sequence of chromosomal rearrangements is deduced by differentially staining chromosomes and matching segments of similar banding patterns in different species^9,10. The results of allozyme and chromosome analyses suggest several phylogenetic divisions within the Canidae (Fig. 3): (1) the wolf-like canids, including domestic dogs, gray wolves, coyotes, and jackals; (2) the South American canids, including species of diverse morphology but common recent ancestry; (3) the red-fox-like canids of the Old and New World, including red foxes and kit foxes; and (4) monotypic genera -- species such as the bat-eared fox and raccoon dog -- that have a long, separate evolutionary history (Table 1). The fossil record and genetic distances indicate that these divisions began about 7 -- 10 million years ago.

Table 1. Canid species, their distribution and chromosome number

Diploid chromosome number.
^b variable number of B-chromosomes present.
 

FIG 2 Relationship tree of carnivores based on differences in single-copy DNA sequences⁵ (DT_mR).

Relationships of the wolf-like canids
The wolf-like canids are a closely related group of large carnivores whose chromosomes are stable in morphology and number (2n = 78). Because of the recent common ancestry of the members of this group, genes that have high rates of sequence substitution, such as those found in the vertebrate mitochondrial genome, can be used to resolve their phylogenetic relationships¹². A phylogenetic analysis of 736 bp of the mitochondrial cytochrome b gene revealed a close kinship of gray wolves, dogs, coyotes and Simien jackals^13-16 (Fig 4). As a group, these three taxa were distinct from the African wild dog and from the golden, side-striped and black-backed jackals. The gray wolf and coyote may have had a recent common North American ancestor about two million years ago¹⁷ whereas the Simien jackal, found only in a small area of the Ethiopian highlands, is possibly an evolutionary relic of a past African invasion of gray wolf-like ancestors. The Simien jackal is the most endangered canid¹⁸ and should be called a wolf rather than a jackal to reflect its evolutionary heritage.

An unexpected result of this research was the high sequence divergence (about 8%) that was found between two black-backed jackals in the same popuation, or a segment of the mitochondrial cytochrome b gene¹⁵ (Fig. 4). This was the largest divergence in mitochondrial DNA (mtDNA) then recorded within a single population that was Freely interbreeding. (as indicated by analysis of morphology and nuclear genes)¹⁹. The mtDNA sequences of these two genotypes evolved at significantly different rates and probably diverged before the speciation event giving rise to black-backed jackals. These findings emphasize the need for caution in the interpretation of phylogenies based on mtDNA; such gene trees are not necessarily species trees and may not accurately reflect phylogenetic affiliations or divergence time²⁰.

The evolution of the domestic dog
The earliest remains of the domestic dog date from 10 to15 thousand years ago²¹; the diversity of these remains suggests multiple domestication events at different times and places. Dogs may be derived from several different ancestral gray wolf populations, and many dog breeds and wild wolf populations must be analysed in order to tease apart the genetic sources of the domestic dog gene pool. A limited mtDNA restriction fragment analysis of seven dog breeds and 26 gray wolf populations from different locations around the world has shown that the genotypes of dogs and wolves are either identical or differ by the loss or gain of only one or two restriction sites²². The domestic dog is an extremely close relative of the gray wolf, differing from it by at most 0.2% of mtDNA sequence^15,22,23.

In comparrison, the gray wolf differs from its closest wild relative, the coyote, by about 4% of mitochondrial DNA sequence¹⁴ (Fig. 4). Therefore, the molecular genetic evidence does not support theories that domestic dogs arose from jackal ancestors²⁴. Dogs are gray wolves, despite their diversity in size and proportion; the wide variation in their adult morphology probably results from simple changes in developmental rate and timing²⁵.

FIG 3 Consensus relationship tree of the dog family based on allozyme genetic distance and chromosome morphology^5,8-10. 

FIG 4 Left: A phylogenetic tree of wolf-like canids generated from maximum parsimony analysis of 736 bp of cytochrome b sequence. The tree was rooted by sequence data from the kit fox. The scale represents percentage sequence divergence. Right: A phylogenetic tree of gray wolves, coyotes, and captive and museum specimen red wolves, generated from maximum parsimony analysis of 398 bp of cytochrome b sequence". The tree was rooted by sequence data from the golden jackal. 

FIG 5 A most parsimonious phylogenetic tree of gray wolf (Wl -- W12) and select coyote (Cl -- C24) genotypes based on data from restriction analyses¹⁴. The tree was rooted at the midpoint. Genotypes with asterisks indicate gray wolf genotypes that are identical, or very similar, to those of coyotes. The scale represents percentage sequence divergence. 

Relationships of populations within species of wolf-like canids
Wolf-like canids can travel great distances and overcome sizeable topographic obstacles. Gray wolves, for example, have been observed to disperse over a thousand kilometers during their lifetimes²⁶. Because dispersing wolves may establish territories and reproduce, gene flow can occur over much larger distances than is usual for terrestrial vertebrates²⁷. A number of different subspecies of the gray wolf and the coyote have been described²⁸; do molecular genetic analyses support the existence of these subspecies, and if so, how are subspecies related? Because the mitochondrial genome evolves so rapidly, its analysis has been an important source of clues about the differentiation of populations within species. Analysis of mtDNA variation in several hundred coyotes and gray wolves has shown little geographic subdivision of mtDNA genotypes^22,29. Within each species, the same genotypes were present at widely spaced locations. There was no significant genetic difference among populations of coyotes, whereas wolves showed only a hint of genetic divergence between Alaskan and southern Canadian populations. Allozyme studies also showed low levels of differentiation among gray wolf populations³⁰.

The phylogenetic tree of mtDNA genotypes can also reveal evidence of geographic subdivision (Fig. 5). In small vertebrates that have poor dispersal ability, the phylogenetic relationships of mitochondrial DNA genotypes from different populations often correspond to the physical distance between the populations or to the presence of geographic barriers^31,32. The greater the geographic distance, the larger the genetic divergence. In gray wolves and coyotes, the relationship between genotypes did not reflect the geographic distance between localities. Closely-related coyote genotypes were found in regions as distant as California and Florida (for example, Cl and C14, Fig. 5) and distantly related genotypes were found at a single locality in southern California (for example, Cl and C7). This result supports the idea that gene flow is a force that homogenizes genetic variation, perhaps across large parts of the continent, but these findings also cast doubt on the validity of the dozen or more subspecies described for both species. The subspecies differences, which are based on pelage or skeletal morphology, may reflect inadequate sampling, rapid evolution of specific ecotypes through selection, or differences in food supply³³. The molecular genetic evidence suggests that these phenotypic differences do not signify a long history of genetic isolation.

The population structure of Old World wolves differs from that of their relatives in North America. In crowded Europe, wolf populations are highly fragmented and small in size. Analysis of mtDNA in European wolves showed that, with one exception, each population had a single genotype not found elsewhere²². The genetic differences among the seven observed genotypes were small: just one or two restriction sites among the 95 that were sampled. However, the structured distribution of these genotypes suggested geographic subdivision and thus led to the concern that each population should be conserved and bred separately²². Hundreds of years ago, gray wolves ranged throughout Europe, as they do now in northern Canada, and probably showed little geographic subdivision. As available habitats for wolves decreased and populations became small, genotypes were fixed at random in the remaining populations, leaving a fractured genetic landscape. Because this landscape reflects the recent activities of humans, preserving each population separately through captive breeding amounts to a continuation of artificial selection on a grand scale and is not justified.

Gene flow within other canid species
Do other wolf-like canids show more geographic structure in their distribution of genotypes than wolves and coyotes? The African wild dog, a large wolf-like canid found in subsaharan Africa, is a good candidate, since the Rift Valley lakes may effectively interrupt gene flow between the eastern and southern populations^16,18. Indeed, there seems to be no gene flow across this barrier, since eastern and southern African wild dogs do not share any mtDNA genotypes¹⁶. Moreover, the sequence divergence between the genotypes is substantial: about 1% of the sequence of the mitochondrial cytochrome b gene differs between the two genotype groups, a figure that is nearly an order of magnitude greater than the divergence between the most different genotypes within a population. Because the difference between populations was so much greater than that within each population, it was recommended that to preserve genetic diversity, east and south African wild dogs should not be interbred in captivity¹⁶.

Do the genotypes of small, less mobile canids have a geographic structure more like other small vertebrates, such as rodents, than that of their larger canid brethren? The diminutive kit fox, a species that lives in the arid lands of the American west, has a distribution that encircles the Rocky Mountains. Analysis of the mtDNA of this species showed two distinct genetic gradients. One was precipitous, and had developed between populations on either side of the Rocky Mountains³⁴; the difference between these populations was nearly as great as between either population and the arctic fox, a species classified in a separate genus. The other gradient was among populations on the same side of the Rockies, and was more gradual. Neighbouring populations shared a greater number of genotypes, and these were more similar to each other than to those of distantly separated populations. Thus, the kit fox showed the two common patterns characteristic of smaller, genetically well-partitioned vertebrates: isolation by topographic barriers, and genetic differentiation increasing with distance.

Interspecific hybridization and the origin of the red wolf
Species, such as wolves and coyotes, that are highly mobile and can interbreed under some conditions, may form large hybrid zones. Several hundred years ago, coyotes were numerous only in the southern United States and wolves were common toward the north. Where wolves are abundant, they will exclude the much smaller coyote from their territories³⁵. After the arrival of European settlers, agriculture and predator control programs caused wolf populations to dwindle, while the coyote, a remarkably flexible and opportunistic species, expanded its geographic range to areas north and east¹⁷. Today the coyote is found throughout most of North America. In eastern Canada, an area invaded b coyotes in the last 100 years, several genotypes identical or very similar to those found in coyotes were discovered in individuals phenotypically identified as gray wolves¹⁴ (genotypes with asterisks in Fig. 5). Wolves with these "coyote" genotypes increased in frequency toward the east, from 50% in Minnesota to 100% in Quebec (Fig. 6). The hypothesis advanced to explain this pattern was that coyotes and wolves had hybridized in areas of eastern Canada where wolves were rare and coyotes common. The interspecific transfer of mtDNA was asymmetric; none of the coyotes sampled had wolf-like genotypes although coyote genotypes were common in gray wolves. Because mtDNA is maternally inherited without recombination, this result reflects a mating asymmetry: male wolves mate with female coyotes, and their offspring backcross to wolves. Either the reverse cross is rare, or the offspring of such backcrosses to coyotes do not reproduce. This mating asymmetry may indicate that the smaller male coyotes cannot inspire the larger female gray wolves to mate with them.

Theory predicts that older hybrid zones between wolves and coyotes may be much larger than that in eastern Canada, and may be up to several thousand kilometers in width^15,36. Accordingly, attention has
been focused on a potentially older and more extensive hybrid zone in the southern United States. The zone includes populations of three wolf-like canids: the red wolf, gray wolf and coyote (Fig. 6). The red wolf is intermediate in size between coyote and gray wolves and can potentially hybridize with both species. It is also an endangered species that became extinct in the wild about 1975, and descendants of the last populations were used to found a successful captive breeding and reintroduction program. If the red wolf were a distinct species ancestral to wolves and coyotes³⁷, there should be unique mtDNA genotypes that define a separate species clade¹⁵, a pattern previously found in wolf-like canids^13-16 (Fig. 4).

However, captive red wolves had a genotype that was indistinguishable by restriction site analysis from those found in coyotes from Louisiana. Because hybridization was thought to occur between the two species as the red wolf became rare, the presence of the coyote-derived genotypes in captive red wolves could represent an accident of sampling and not be representative of the ancestral population. Subsequently, an additional mtDNA analysis of 77 samples obtained in about 1975 from areas inhabited by the last wild red wolves showed that all had either a coyote or gray wolf genotype¹⁵.

Conceivably, hybridization between gray wolves and coyotes alone could explain the intermediate morphology of red wolves. To test this hypothesis, DNA was isolated from six museum skins of red wolves obtained from Five states in about 1910, a time before hybridization of red wolves and coyotes was thought to be common. Phylogenetic analysis of 398 bp of the cytochrome b gene showed that red wolves at that time did not have a distinct genotype; all six had genotypes classified with gray wolves or coyotes, a result consistent with a hybrid origin for the species¹⁵ (Fig. 4). Although more research needs to be done, the implication of this result is troubling for the US Endangered Species Act because a policy on hybrids has not been formulated. In some situations we may wish to protect hybrids, such as the red wolf, because they are unique. Elsewhere, in Minnesota for example, hybridization may be undesirable because it jeopardizes the genetic integrity of the gray wolf, a threatened species. Similarly, in Italy, hybridization with domestic dogs may be changing the character of gray wolves that enter small towns to feed because their natural prey has been depleted. Even the highly endangered Simien jackal is threatened with hybridization by feral domestic dogs. Molecular genetic analyses offer a powerful means to determine if hybridization is changing the composition of these endangered populations.

Future research on the population genetics of canids should focus on the analysis of polymorphic nuclear genes to complement the mtDNA data. However, nuclear DNA domains that evolve as fast as highly variable mtDNA regions have yet to be identified, and may not exist. Hypervariable simple sequence repeat loci³⁸ may prove useful; these loci are abundant in the nuclear genome and evolve through loss or gain of repeat units rather than sequence substitutions. Analysis of simple sequence repeats will not provide the detailed picture of the succession of historical changes revealed by sequence data but may furnish estimates of gene flow and hybridization among closely related canid populations.
Examine the DNA bands carefully in the picture below.They represent,from left to right,strands from dog,coyote,wolf,wolfdog,wolfcoyote and wolf

 

FIG 6 Geographic range, sampling localities and hybrid zones of gray wolves, coyotes, and red wolves^14,15. The darkly shaded area indicates the present North American geographic range of the gray wolf. The stippled areas in eastern North America and the south-central US are hybrid zones between wolf-like canids, as assessed by analysis of mtDNA. Question marks indicate areas from which no samples have been obtained. For identification of the localities, see Refs 14 and 15. The past geographic range of the red wolf around the year 1700 is also shown. Mitochondrial DNA control region sequences were analyzed from 162 wolves at 27 localities worldwide and from 140 domestic dogs representing 67 breeds. Sequences from both dogs and wolves showed considerable diversity and supported the hypothesis that wolves were the ancestors of dogs. Most dog sequences belonged to a divergent monophyletic clade sharing no sequences with wolves. The sequence divergence within this clade suggested that dogs originated more than 100,000 years before the present. Associations of dog haplotypes with other wolf lineages indicated episodes of admixture between wolves and dogs. Repeated genetic exchange between dog and wolf populations may have been an important source of variation for artificial selection.

The archaeological record cannot resolve whether domestic dogs originated from a single wolf population or arose from multiple populations at different times (1, 2). However, circumstantial evidence suggests that dogs may have diverse origins (3). During most of the late Pleistocene, humans and wolves coexisted over a wide geographic area (1), providing ample opportunity for independent domestication events and continued genetic exchange between wolves and dogs. The extreme phenotypic diversity of dogs, even during the early stages of domestication (1, 3, 4), also suggests a varied genetic heritage. Consequently, the genetic diversity of dogs may have been enriched by multiple founding events, possibly followed by occasional interbreeding with wild wolf populations.

We sequenced portions of the mitochondrial DNA of wolves and domestic dogs. Initially, 261 base pairs (bp) of the left domain of the mitochondrial control region (5) were sequenced from 140 dogs representing 67 breeds and five cross-breeds and 162 wolves representing 27 populations from throughout Europe, Asia, and North America (Fig. 1) (6). Because all wild species of the genus Canis can interbreed (7) and thus are potential ancestors of the domestic dog, five coyotes (Canis latrans) and two golden, two black-backed, and eight Simien jackals (C. aureus, C. mesomelas, and C. simensis, respectively) were also sequenced.

Fig. 1. Substitutions and deletions ( -- ) observed in 261 bp of control region sequence from wolves (W) and dogs (D). The dog sequence D13 had the same sequence as D4 except for an insertion of a 67-bp tandem repeat. The numerals I, II, III, and IV indicate assignments to the four clades of dog sequences. Wolf localities: Bulgaria (n = 1, W7); Croatia (n = 5, W2); Estonia (n = 1, W10); France (n = 2, W4); Finland (n = 2, W10); Greece (n = 7; W2, W5, W8, and W9); Italy (n = 12, W4); Poland (n = 1, W3); Portugal (n = 19; W1 and W2); Romania (n = 4; W5 and W6); Russia (n = 3; W6, W10, and W26); Spain (n = 46; W1 and W3); Sweden (n = 2; W2 and W10); Afghanistan (n = 3, W18); China (n = 3; W14, W19, and W27); India (n = 1, W12); Iran (n = 6; W16 and W17); Israel (n = 16, W11); Saudi Arabia (n = 7; W7, W12, W13, W14, and W15]; Turkey (n = 2, W2); Alaska (n = 3, W20); Alberta (n = 1, W22); Labrador (n = 3, W22); Mexico (n = 5, W25); Montana (n = 1, W22); Northwest Territories (n = 3, W22); and Yukon (n = 3; W21, W23, and W24). Dog breeds: basenji (n = 1, D2); basset (n = 1, D6); boxer (n = 1, D4); Norwegian buhund (n = 1, D1); bulldog (n = 1, D6); Chinese crested (n = 2; D2 and D25); chow chow (n = 3; D1, 02, and D3); collie (n = 1, D1); border collie (n = 3; D1 and D5); wirehaired dachshund (n = 3; D5 and D10); Australian dingo (n = 4, D18); grey Norwegian elkhound (n = 9; D3 and D8); Eskimo dog (n = 1, D23); German shepherd (n = 8; D4, D5, D6, D7, and D19); greyhound (n = 1, D9); groenendael (n = 1, D6); Mexican hairless (n = 6; D3, D6, D21, and D26); hamilton-stovare (n = 1, D5); Afghanistan hound (n = 3, D6); Alaskan husky (n = 2; D4 and D7); Siberian husky (n = 3; D3, D7, and D18); jamthund (n = 3; D7 and D8); keeshond (n = 1, D5); kuvasz (n = 1, D4); Leonberger (n = 2; Dl and D4); Norwegian lundehund (n = 1, D16); Mareema (n = 1, D=6); Pyrenean mastiff (n = 1, D11); Newfoundland (n = 1, D4); otter hound (n = 1, D6); papillon (n = 2; D3 and 04); poodle (n = 1, D3); toy poodle (n = 1, D6); pug (n = 1, D26); Chesapeake Bay retriever (n = 1, D13); flat-coated retriever (n = 3; D4 and D10); golden retriever (n = 6; D4, D6, D15, and D24); labrador retriever (n = 6; D4 and D12); Rhodesian ridgeback (n = 1, D26); rottweiler (n = 2, D3); Samoyede (n = 3; D1, D4, and D5); St. Bernard (n = 1, D9); schipperke (n = 1, D4); giant schnauzer (n = 3; D4 and D7); miniature schnauzer (n = 1, D9); English setter (n = 4; D3 and D5); Irish setter (n = 3; D1 and D9); New Guinea singing dog (n = 2, D18); shar (n = 1, D26); Icelandic sheepdog (n = 1, D3); Old English sheepdog (n = 1, D5); shiba inu (n = 1, D20); Cavalier King Charles spaniel (n = 1, D17); Irish water spaniel (n = 1, D6); springer spaniel (n = 1, D3); Tibetan spaniel (n = 1, D6); spitz (n = 1, D22); Japanese spitz (n = 1, D3); airedale terrier (n = 1, D7); border terrier (n = 2, D3); fox terrier (n = 2; D3 and D14); Norfolk terrier (n = 2, D4); West Highland terrier (n = 2, D7); Tibetan terner (n = 2; D2 and  D9); wachtelhund (n = 1, 05); whippet (n = 1, D3); Irish wolfhound (n = 2, D11); and crossbreeds (n = 5; D1, D3, D4, D5, and D18).

The control region of wolves and dogs was highly polymorphic (Fig. 1). We found 27 wolf haplotypes that differed on average by 5.31 +/- 0.11 (+/- SE) substitutions (2.10 +/- 0.04%), with a maximum of 10 substitutions (3.95%). The distribution of wolf haplotypes demonstrated geographic specificity, with most localities containing haplotypes unique to a particular region (Fig. 1). Four haplotypes (W2, W7, W14, and W22) had a widespread distribution. In dogs, 26 haplotypes were found. Only haplotype D6 also occurred in some gray wolves from western Russia and Romania (W6). Sequence divergence among dogs was similar to that found among wolves. Dog haplotypes differed by an average of 5.30 +/- 0.17 substitutions

(2.06 +/- 0.07%), with a maximum divergence of 12 substitutions (4.67%). Mitochondrial haplotype diversity in dogs could not be partitioned according to breeds. For example, in eight German shepherds examined, five distinct sequences were found, and in six golden retrievers, four sequences were detected.
Moreover, many breeds shared sequences with other breeds. For instance, dog haplotypes D4, D3, D5, and Dl were found in 14, 14, 9, and 7 breeds, respectively. No dog sequence differed from any wolf sequence by more than 12 substitutions, whereas dogs differed from coyotes and jackals by at least 20 substitutions and two insertions. These results clearly support wolf ancestry for dogs. However, because mitochondrial DNA is maternally inherited, interbreeding between female dogs and male coyotes or jackals would not be detected. More limited studies of nuclear markers support the conclusion that the wolf was the ancestor of the domestic dog (8).

Several methods of phylogenetic analysis, including maximum likelihood (9), maximum parsimony (10), minimum spanning networks (1 I), and statistical parsimony (12), were used to investigate relationships among sequences. All analyses supported a grouping of dog haplotypes into four distinct clades, although the topology within and among clades differed among trees (13). As exemplified by the neighbor-joining analysis (Fig. 2A), three of the four monophyletic clades defined a larger clade containing all but three dog haplotypes and a subset of wolf haplotypes (W4 and W5). Clade I included 19 of the 26 dog haplotypes. This group contained representatives of many common breeds as well as ancient breeds such as the dingo, New Guinea singing Jog, African basenji, and greyhound (14). Clade II included dog haplotype D8, from two Scandinavian breeds (elkhound and jamthund), and was closely related to two wolf haplotypes found in Italy, France, Romania, and Greece (W4 and W5). Clade III contained three dog haplotypes (D7, D19, and D21) found in a variety of breeds such as the German shepherd, Siberian husky, and Mexican hairless. Finally, clade IV contained three haplotypes (D6, D10, and D24) that were identical or very similar to a wolf haplotype (W6) found in Romania and western Russia, which suggests recent hybridization between dogs and wolves. Many breeds contained representatives of more than one dog haplotype grouping (Fig. 1).

Because the overall bootstrap support for many of the internodes in Fig. 2A was low, 1030 bp of the control region were sequenced for 24 canids, including representatives of the four dog clades (15). Although the association of clades was different, the analyses of the longer sequences provided stronger support for the tour monophyletic groupings of dog haplotypes (Fig. 2B) (13). A Wilcoxon signed-rank test was used to assess the monophyly of dog clades (16). Monophyly of all dog haplotypes can be rejected, and monophyly of clades I, II, and III is marginally rejected (P = 0.0004 and P = 0.053, respectively). In both trees, dog haplotype clakes II and IV are most closely related to wolf sequences from eastern Europe (Greece, Italy, Romania, and western Russia).

The coyote and wolf have a sequence divergence of 0.075 +/- 0.002 (17) and diverged about one million years ago, as estimated from the fossil record (18). Consequently, because the sequence divergence between the most different genotypes in clade I (the most diverse group of dog sequences) is no more than 0.010, this implies that dogs could have originated as much as 135,000 years ago (19). Although such estimates may be inflated by unobserved multiple substitutions at hypervariable sites (20), the sequence divergence within clade I clearly implies an origin more ancient than the 14,000 years before the present suggested by the archaeological record (21). Nevertheless, bones of wolves have been found in association with those of hominids from as early as the middle Pleistocene, up to 400,000 years ago (1, 22). The  ancient dates for domestication based on the control region sequences cannot be explained by the retention of ancestral wolf lineages, because clade I is exclusively monophyletic with respect to dog sequences and thus the separation between dogs and wolves has been long enough for coalescence to have occurred. To explain the discrepancy in dates, we hypothesize that early domestic dogs may not have been morphologically distinct from their wild relatives. Conceivably, the change around 10,000 to 15,000 years ago from nomadic hunter-gatherer societies to more sedentary agricultural population centers may have imposed new selective regimes on dogs that resulted in marked phenotypic divergence from wild wolves (23).

Although individual breeds show uniformity with respect to behavior and morphology, most breeds show evidence of a genetically diverse heritage because they contain different haplotypes. Moreover, dog sequences cluster with different groups of wolf haplotypes. Therefore, after the origin of dogs from a wolf ancestor, dogs and wolves may have continued to exchange genes. Backcrossing events could have provided part of the raw material for artificial selection and for the extraordinary degree of phenotypic diversity in the domestic dog. Domestic species of plants and animals whose wild progenitors are extinct cannot be enriched through periodic interbreeding, and change under artificial selection may be more limited. Consequently, the preservation of wild progenitors may be a critical issue in the continued evolution of domestic plants and animals.

Extensive interbreeding occurred among multiple matriarchal ancestors during the domestication of dogs: Evidence from inter- and intraspecies polymorphisms in the D-loop region of mitochondrial DNA between dogs and wolves
To test the hypothesis that the domestic dogs are derived from several different ancestral gray wolf populations, we compared the sequence of the displacement (D)-loop region of the mitochondrial DNA (mtDNA) from 24 breeds of domestic dog (34 individual dogs] and 3 subspecies of gray wolf {Canis lupus lupus, C.l. pallipes and C. l. chanco; 19 individuals). The intraspecific sequence variations within domestic dogs (0.00 – 3.19%) and within wolves (0.00 – 2.88%) were comparable to the interspecific variations between domestic dogs and wolves (0.30 – 3.35%). A repetitive sequence with repeat units (TACACGTA/GCG) that causes the size variation in the D-loop region was also found in both dogs and wolves. However, no nucleotide substitutions or repetitive arrays were specific for domestic dogs or for wolves. These results showed that there is a close genetic relationship between dogs and wolves. Two major clades appeared in the phylogenetic trees constructed by neighbor-joining and by the maximum parsimony method; one clade containing Chinese wolf (C. l. chanco) showed extensive variations while the other showed only slight variation. This showed that there were two major genetic components both in domestic dogs and in wolves. However, neither clades nor haplotypes specific for any dog breed were observed, whereas subspecies-specific clades were found in Asiatic wolves. These results suggested that the extant breeds of domestic dogs have maintained a large degree of mtDNA polymorphisms introduced from their ancestral wolf populations, and that extensive interbreedings had occurred among multiple matriarchal origins.

INTRODUCTION

The dog, Canis familiaris, is the only member of the family Canidae, and is the "oldest" animal that can be fully domesticated in the world, since the historical evidence showing its strong connection with humans can be traced back to the far preagricultural age (Turnbell and Reed, 1974). During the history of dogs' domestication, more than 400 breeds with high morphological variations have been established by crosses between or within their ancestral stacks and also by artificial selection. Despite such a large number of extant breeds, only a few reports have been published about the genetic backgrounds or genetic variations among the breeds. For example, Tanabe et al. (1991) analyzed 25 blood protein loci, using approximately 3,000 individual dogs of 40 breeds, including Japanese, Asian, and European dog breeds, and showed that the dogs maintained low genetic variations among the populations. Those authors also suggested that most modern Japanese dogs have been affected by two major genetic components which were brought to Japan from Southeast Asia about 10,000 years ago and from the northwest through the Korean peninsula about 1,700 -- 2,300 years ago (Tanabe, 1990, 1991; Tanabe et al., 1991; Fujise et al, 1997). However, the low heterozygosity in the polymorphic biochemical loci did not allow further detailed investigations.

Although a high degree of morphological variation is present among the 400-plus breeds of dogs, the most widespread and accepted view, based mainly on morphological studies, places only one species, the wolf (Canis lupus), as the wild progenitor of domestic dogs (Lorenz, 1975; Zimen, 1981) However, the existence of more than 5 subspecies of wolves and the absence of detailed information on the genetic diversity among these subspecies underlie a fundamental question: From which of the wolf subspecies are domesticated dogs descended?

In the past two decades, polymorphisms of mitochondrial DNA (mtDNA) have been successfully applied in examination of the genetic relationships among populations of the same animal species and between closely related species (Avise, 1991; Wayne, 1993), since mtDNA evolves 5 – 10 times as much as do single-copy nuclear genes (Brown et al., 1979).The strictness of the maternal inheritance of mtDNA (Kaneda et al., 1995) and the lack of genetic re-combination in mtDNA also make mtDNA polymorphisms, in particular nucleotide substitutions, a useful tool for disclosing the evolutionary events in matriarchal Lineages of animal species. Since domesticated animals have generally been maintained through matriarchal lineages with the introduction of patarnal genes which are useful for the improvement of a breed, the mtDNA polymorphisms can be used to determine which wild species was the matriarchal ancestor of a domesticated animal of interest.

Using this molecular genetical analysis, Wayne et al. (1991, 1997) reported that the domestic dog is an extremely close relative of the gray wolf, differing from it by at most 0.2% of the DNA sequence. Okumura et al. (1996) made an extensive survey of the mtDNA D-loop region from 73 individual dogs of 24 breeds and found extensive polymorphisms in the mtDNA D-loop region even within breeds. Those authors also mentioned that they found no distinct correlations between the European and Japanese breeds of dogs, demonstrating that different aspects of the genetic relationships between dog breeds compared with allozyme analysis. However, they did not investigate the relationship between these dog breeds and wolves as their putative ancestor.

To address this issue, we performed a sequence comparison of the mtDNA D-loop region among 24 breeds of domestic dogs (34 individual dogs) and three subspecies of wolf (C. l. lupus. C. l. pallipes, and C. l. chanco), and found that the dogs were not genetically differentiated from the wolves. This result is the first evidence suggesting the dog has multiple domestication centers from wild wolves, and that extensive interbreedings occurred among the multiple matriarchal origins.

MATERIALS AND METHODS

DNA samples. Whole-blood samples were collected from 24 breeds of domestic dog (33 individuals) and 3 subspecies (19 individuals) of wolf European wolves (C. l. lupus) and Indian wolves (C. l. pellipes) were captured in Yugoslavia and Afghanistan, respectively. All wild individuals of both subspecies were temporarily bred at Kiel University. Wild individual Chinese wolves (C. l. chanco) were captured near Ulan Bator in Mongolia. DNA samples from four foxes and a raccoon dog were also used as outgroups of the dogs when phylogenetic trees were constructed. All DNA samples used in this study are listed in Table l.

DNA sequencing and sequence alignment. Before the analysis of the DNA samples mentioned above, we determined the complete sequence of the D-loop region of mtDNA, which was purified from the livers of two mongrel dogs by the method described by Yonekawa et al. (1981). Genomic DNAs were then extracted from the blood samples of all other dogs as described by Maniatis et at. (1982). Based on the sequence, four primers for direct sequencing were synthesized. The names and sequences of the primers were: 5'-tgtaaaacgacggccagtgctcttgctccaccatca-3' (DL14: a region within the tRNA{Pro}), 5'-caggaaacagctatggccccccttgatttttatgcg-3' (DH6: a region within the D-loop), 5'-tgtaaaacgacggccagtatcactcatctacgaccg-3' (DL10; a region within the D-loop) and 5'-caggaaacagctatggcccgtgcgactcatcttggc-3' (DH7; a region within the tRNA{Phe}).

Table I. List of source of DNA samples used in this study
 

* none of the individual wolves had been bred.
 

The sequence of the mtDNA D-loop regions were determined by direct sequencing with semi-nested polymerace chain reaction (PCR) (Kikkawa et al., 1997) using the four primers, followed by an automated method. We then aligned the nucleotide sequences of the D-loop region among the dogs, wolves, foxes, and raccoon dog. Phylogenetic Analysis. The sequence data were analyzed using distances corrected for multiple hits by the two-parameter method of Kimura (1980) with the DNADIST of PHYLIP program (Ver. 3.5c; Felsenstein, 1993). Phylogenetic trees were constructed by the neighbor-joining (NJ) method (Saitou and Nei, 1987) and by the UPGMA (Sneath and Sokal, 1973)incorporated into the NEIGHBOR program. The sequences of the foxes and raccoon dog were used for outgroups to construct phylogenetic trees. The phylogenetic analysis of character-state matrices using the parsimony method was performed with the DNAPARS, with a majority-rule consensus tree produced by the CONSENSE; a majority-rule consensus tree based on 100 or 1,000 boot-strap replicates was produced using the SEQBOOT.

RESULTS

Nucleotide substitutions and length variations in the D-loop region. We determined the complete sequences of the mtDNA D-loop region for domestic dogs, wolves, foxes and a raccoon dog. By aligning these sequences, we found that the D-loop region was divided into three parts; the 5'-highly polymorphic segment (672-4 bp), the repetitive sequence (from 200 to 300 bp), and the 3'- conserved segment (294 bp). Of the 52 nucleotide substitutions found in the D-loop region, 51 were restricted in the 5'-side segment to the repetitive sequence (674 bp), and only one substitution and 3 deletions/insertions could he observed in the 3'-side segment of the repetitive sequence (Fig. 1).

The dogs and wolves both had the repetitive sequence with two repetitive units of TACACGT([A/G])CG. The foxes and raccoon dog also had sequences with the TACACATACG, TACACACA([C/T])ACG and the TAC([A/G])CACG units, respectively(Fig. 1). The sequence of the units found in the dogs and wolves differed from three of the other two species, and thus the repetitive sequences were species specific. Among the dogs and wolves, the repetitive sequence caused extensive length variations and intra-individual heteroplasmy in the D-loop. We found no specific arrays of the units nor specific length variations among the breeds of domestic dogs or subspecies of wolf (data not shown).

No significance in the sequence divergence values among dogs and wolves. Using a pairwise method, we compared the nucleotide sequences in the 5'- highly polymorphic segments among dogs and wolves and calculated the intra- and inter-species sequence divergences. The values of sequence divergence within domestic dogs varied from 0% to 3.19%, and those within wolf subspecies varied from 0% to 2.88%. The sequence divergence was calculated to be 0.30 -- 3.35% between dogs and wolves (Table 2). These results showed that there were no significant differences among the intraspecific or interspecific divergence values. These results also showed that the domestic dogs tested maintain a large degree of mtDNA polymorphisms, suggesting that the polymorphisms had been introduced by the wolves. It is thus concluded that the wolf is the strongest candidate for the matriarchal ancestor of the dog.

We also calculated the sequence divergences between domestic dogs and foxes, between domestic dogs and the raccoon dog, and between foxes and the raccoon dog to be from 18.37% to 21.35%(the average value, 19.71%), 19.81% to 21.88% (the average value, 21.01%), and from 19.81% to 20.43% (the average value, 20.28%), respectively (Table 2).

Phylogenetic analysis and nucleotide substitution patterns. We constructed phylogenetic trees by the NJ method using the sequence divergence values. Two distinct clades. Clade A and Clade B, appeared in the N J trees, but no clades specific for any particular dog breeds were observed, consistent with the results reported by Okumura et al. (1986). It should be noted that five breeds of Asiatic native dog, i.e., Shiba (Mino), Kai, Atita, Korean Chejudo, and Mongolian native dog shared Clade A and Clade B, although the sample size was very small; namely, we examined only two individuals in each of these breeds (Fig. 2). These results suggest that most breeds of Asiatic native dog possess two major genetic components in their matriarchal lineages. The Asiatic wolves showed clade specificity among the subspecies; namely, Chinese wolves (C. l. chanco) were limited to Clade A, whereas Indian wolves (C.l. pallipes) were limited to Clade B. European wolves (C. l. lupus) shared both clades. We then tested the significance of tree topology by bootstrapping using the wolves, foxes, and raccoon dog. Domestic dogs were excluded from this test, since we considered that the exclusion would eliminate any bias caused by artificial intra- and interbreeding in the dog populations. The existence of two clades was confirmed to be significant by the bootstrapping values (Fig. 3).

After construction of the NJ tree, we compared the nucleotide substitutions within and between the clades. As mentioned above, 53 substitutions were restricted in the 5'- highly polymorphic segment. We found 6 transversions, 43 transitions, 2 transitions/transversions and 2 insertions/deletions in the segment. Five substitutions were specific for the Clade B which contains Indian wolves (Fig. 1). In particular, one substitution, which is located at the 27th nucleotide position from the beginning of the D-loop (in the 5' segment) and changed from A to G, occurred specifically in the Indian wolves, being a good diagnostic marker for the subspecies. However, no substitutions were specific for the Clade A.

DISCUSSION

The close relationships between the domestic dog and wolves. Based on the molecular analysis of mtDNA between domestic dogs and wolves, we present here direct molecular evidence that the ancestor of the domestic dog is the wolf. We compared their nucleotide substitutions in the mtDNA D-loop region. and found no differences in the specificity for the substitution between the two species (Fig. 1 and Table 2). Moreover, the clades of the individual domestic dogs were completely overlapped with those of the individual wolves, showing that the domestic dog was descended from the wolf.

From the morphological point of view, domesticated dogs were suggested to be descended from wolves (Wayne, 1986). This was supported by the results from the studies of behavior (Zimen. 1981), vocalization (Zimen, 1981) and molecular biology (Wayne et el., 1987a, b). Wayne et al (1991) also showed in a mtDNA restriction fragment length polymorphisms (RFLP) analysis that the domestic dog is an extremely close relative of the gray wolf, differing from it by at most 0.2% of the mtDNA sequence. Our results indicate that the mtDNA polymorphisms in the domestic dog are not distinguishable from those in the wolf. This slight difference may be caused by the difference of collection localities between our study and that of Wayne et al. (1991). We collected the wolf samples from an Old continent, whereas Wayne et al. (1991) had collected from a New continent. There may be a slight difference in mtDNA polymorphisms between the wolves from the two continents.

Wayne (1993) stated that dogs are gray wolves, despite their diversity in size and population, and that the wide variation in their adult morphology probably the result of simple changes in the developmental rats and timing. Recently, Wayne and his colleagues carried out extensive survey of DNA sequence polymorphisms in the mtDNA D-loop region among domestic dogs and wild wolves (Vila et al., 1997). They concluded that 1) wolves were the ancestors of dogs, and 2) dogs had multiple origins, based on extensive polymorphisms shown in dog populations. We also obtained the same conclusion by extensive survey of mtDNA D-loop sequences mainly among Japanese dogs and Asitatic wild wolves. The lack of difference between our conclusion and theirs suggests two possibilities on  the domestication processes of dogs: 1) Asia is the secondary domestication center for the dog and the ancestors of Asiatic dogs had maintained the extensive polymorphisms from the original ancestors of dogs, and 2) human migrations frequently occurred during the domestication of dogs and consequently modern dogs exhibit the extensive polymorphisms. Wayne and his colleagues (Vila et al., 1997) also suggests that dogs originated more than 100,000 years before the present, based on the experimental evidence that most dog sequences belonged to a divergent monophyletic clade sharing no sequences with wolves. However, we could not confirm this evidence because our phylogenetic tree shows no clades belonged to only dogs (Fig. 2).

Intraspecific mtDNA polymorphisms in the domestic dog. When we analyzed the phylogenetic relationships among domestic dog breeds based on the sequence polymorphisms of the mtDNA D-loop region, we found extensive polymorphisms We also observed two major clades in the breeds, showing that the domestic dogs possess two major genetic components. However, we found no clades or haplotypes specific for certain dog breeds, consistent with the results of Okumura et al. [1996]. Tanabe et al. (1991) suggested that Japanese dogs were derived from influxes from two distinct routes; one from Southwest Asia, and the other from the northwest through the Korean peninsula. Our present results are not consistent with this suggestion, because Japanese native dog breeds could not be delimited as distinct breeds. This discrepancy may be elucidated by the difference of the inheritance manner between nuclear-coded genes and mtDNA-coded genes. All of the genes that were investigated by Tanabe et al. (1991) are encoded by the nuclear genome, and they are biparentally transmitted. In contrast, mtDNA is maternally transmitted in mammals, and the maternal transmission is quite strict (Kaneda et al., 1995). When the genetic status of domesticated animals is to he improved by breeding, the most common method is that useful genes are introduced from males. Since domestic dogs have been improved in this way the results reported by Tanabe et al. (1991) may reflect the male effect. It is thus necessary to investigate male-specific genes much as unique genes encoded on the Y-chromosome.

Intraspecific mtDNA polymorphisms in the wolf. We obtained experimental evidence that Asiatic wolves have subspecies-specific clads in their phylogenetic trees. This suggests that the Asiatic subspecies of wolf are genetically differentiated from each other. Our phylogenetic analysis showed that the European subspecies of wolf shared two clades: one clade contains the Indian subspecies, and the other contains the Chinese subspecies. Two possibilities underlying this result can be posited. One possibility is that the European subspecies is the ancestor of the Asiatic subspecies; the other is that the haplotypes found in the Asiatic subspecies were introduced in the European subspecies. From the archeological point of view, historical remains of wolf bones have been found exclusively in the New Continent. This suggests that the ancestor of the wolf had arisen there and moved to the Old Continent. If this is true, the migration routes should be from the Old continent to the European Continent through the Asian Continent. Thus, the latter possibility might be more likely.

Wayne (1993) reported that European wolves possessed, with one exception, a single haplotype specific for each population. He suggested that the mtDNA genotypes were randomly fixed to be monomorphic because of the decreasing and fracturing of available habits for wolves and the decreases in their populations. Our results support his idea. We found only two haplotypes in the European subspecies, and the same was true for the Indian subspecies (Figs. 2 and 3). Not only in Europe but also in India, the wolves' habitats may have become smaller and fractured. In sharp contrast to these subspecies, the Chinese subspecies showed, as did the domestic dogs, extensive polymorphisms in their mtDNA. This suggests that the Chinese subspecies still maintain populations large enough to keep such polymorphisms, and that these wolves can move freely through their territories.

Estimation of divergence time. The average sequence divergence values were 19.71% (dogs and wolves - foxes), 20.28% (foxes - raccoon dog) and 21.01% (dogs and wolves - raccoon dog). The rate of sequence divergence of mtDNA in mammals is estimated to be 2-4% per million years (Brown et el., 1979). Based on this value, the divergence of the 3  genera occurred approximately 6 -- 10 million years ago. Wayne et al. (1987a.b) reported that the fossil record and genetic distances indicate that the division of canines began about 7 -- 10 million years ago, which is consistent with our results.

Did extensive interbreeding occur in the domestic dogs? As mentioned above, the Asiatic subspecies of the wolf showed subspecies-specific mtDNA polymorphisms. Though the European subspecies showed a much smaller degree of mtDNA polymorphisms, these wolves showed the existence of two genetic components in their mtDNA. In contrast, the domestic dogs exhibited extensive polymorphisms in their mtDNA, and these polymorphisms were not specific for any breed or for a geographical distribution (Fig. 2 and Table 4). Okumura et el. (1996) investigated 28 individuals of one Japanese native breed, the Shiba, and they found that at least 7 mtDNA haplotypes exist in the Shiba breed. The haplotypes are also found in the Ryukyu, an old Japanese native breed. Many haplotypes are shared among different dog breeds, not only Japanese native breeds but also non- Japanese breeds. These results suggest that extensive interbreeding occurred in the ancestral stocks of domestic dogs. We inferred, therefore, that the domestication from wolves to dogs occurred in several (at least two) places, and that the domestic dog may be derived from two wolf populations separated geographically (north and south).

There are several hypotheses regarding the domestication of dogs. The most widespread and accepted view regards the wolf as the ancestor of the domestic dog (Zeuner, 1963; Scott and Fuller, 1965; Lorenz, 1975; Herre and Rohrs, 1977). At present, a leading hypothesis is that the ancestor of dogs is the Arabian wolf [C. l. arabs) (Clutton-Brock, 1995). However, the report on skeletal anatomy, the dingo, C. familialis dingo, closely resembles the Indian wolf and the pariah dogs of Southeast Asia. It is probable that the dingo is a direct descendant of dogs that were originally domesticated from tamed Indian wolves (C. l. pallipes) (Corbett, 1985). We suggest that findings and ours indicate plural origins of the domestic dog. It is possible that the dogs were domesticated in a single place, mated with wolves along a migration route of humans. We are not able to conclude whether the domestication occurred in a single place or several places.

Repetitive sequences. We found that the repetitive sequences appeared in the identical positions near the 3' end of the mtDNA D-loop region among dogs, wolves, foxes, and a raccoon dog. We also found such repetitive sequences in the dhole (Cuon alpinus), bush dog (Speothos vanaticus), and marten (Martes melampus) (data not shown). Repetitive sequences in the mtDNA D-loop region have also been reported in other mammalian species, i.e., the white sturgeon (Buroker et al., 1990), evening bat (Wilkinson and Chapman, 1991), rabbit (Biju-Duval et al., 1990; Mignotte et al, 1990), harbor seal (Arnason and Johnsson, 1992), elephant seals (Hoelzel et al., 1993a), pig (Ghivizzani et al., 1993), and equine (Ishida et al., 1994). Our present result indicated the following characteristics of the repetitive sequence; the sequence of the repeat units is species specific, although the array and the length of the repetitive units vary among individuals in the same species. However, one exception occurs between dogs and wolves. In conclusion, our finding contain strong experimental evidence that dogs and wolves are members of the same species.