Mitochondrial diversity of Bulgarian native dogs suggests dual phylogenetic origin

The dog has been the first domesticated animal to have a central role in human society from ancient times to present day. Although there have been numerous investigations of dog phylogeny and origin, genetic data of dogs in the region of the Balkan Peninsula (South-Eastern Europe) are still scarce. Therefore, the aim of the present study was to perform phylogenetic analysis of three native Bulgarian dog breeds. A total of 130 samples were analyzed at HVR1 (hypervariable region, D-loop region). The samples were taken from two hunting dog breeds (Bulgarian Hound Dog: Barak, n = 34; Bulgarian Scenthound Dog: Gonche, n = 45) as well as from a Bulgarian Shepherd Dog (n = 51). The first two breeds are reared in a flat region of the country (the Northern part of Bulgaria, the Danubian Plain), while the last breed is a typical representative of the mountainous part of the country. The results have shown the presence of almost all main clades—A, B, C and D—in the three dog breeds taken together, except clades E and F, as expected. With regard to haplogroups distribution, there are clear differences among investigated breeds. While hunting breeds exhibit a prevalence of clade C, the mountainous Shepherd dog shows presence of the D2 haplogroup but absence of the C clade. In conclusion, the present study has been the first to investigate the mitochondrial DNA diversity of native dog breeds in Bulgaria. The results have revealed a clear difference of haplogroups dissemination in native hunting and shepherd dogs, which suggests a dual independent phylogenetic origin, without hybridization events between these dogs.


INTRODUCTION
The origin and evolution of the domestic dog has remained for a long time a controversial question for the scientific community with regard to basic aspects such as the place of origin (Wang et al., 2016). It is known that two main processes have occurred in dog (Canis lupus familiaris) history: first, primitive dogs were domesticated from their wild predecessor, the gray wolves (Canis lupus lupus) before the beginning of the last ice age (about 30,000 BP). Second, the primitive forms were further selected to form many dog breeds with specialized The hunting dogs are some of the oldest hunting dogs in the Balkans. Presumably, they originated back in the Thracian period (around 2500 BC), based on pictures from this epoch. In general, on the territory of the Balkan Peninsula, hounds are divided into smooth-and rough-haired.
In Bulgaria, one of the most widely distributed smooth-haired hunting dog breeds is the Bulgarian Scenthound dog (BSD, Gonche). The breed is also known as ''Bicolor hound'' or ''Ludogorsko gonche'' due to the area where it is largest populations are found-in the Ludogorie region in Northeastern Bulgaria. Dogs were used for hunting of large and small hairy game and predators. Some of the nearest members of this breed are the Hungarian hound (Transylvanian Scenthound), the Serbian Hound (Tricolor hound) and the Greek Harehound.
The Barak (Bulgarian Barak Hound, BBH, Barak) is a rough-haired Bulgarian dog. The Barak belongs to the group of hounds chasing game over a freshly traced trail. This breed is prevalent in the Central, North and North-Western (the Danubian Plain) parts of the country. Some most closely related member of this breed is the Slovak Rough-haired Pointer.
The Bulgarian Shepherd Dog (BShD) is a traditional mountain livestock guard dog breed, usually named Karakachan Shepherd. Other names are Ovcharsko Kuche and Thracian Mollos. It is possible that the Balkan shepherd dogs had a direct predecessor, known as the now extinct Molossus dog (Kitchell Jr & Kenneth, 2014). The Molossus originated on the Balkans as described by many old authors as Schöps (1937) who wrote about it in the German cynological magazine ''Zeitschrift für Hundeforschung''. The nearest members of this breed are the Tornjak (Bosnia and Herzegovina and Croatia), the Sharplaninac (FR Macedonia), and the Akbash Dog (Turkey).
A newly created dog breed based on the Karakachan Dog (BShD) is the Bulgarian Shepherd Dog (BOK). Other breeds, such as the Central Asian Shepherd Dog (Alabai), the Caucasian Shepherd dog, the St. Bernhardshund, the Newfoundland, etc., have also taken part in the development of the breed. Bulgarian Shepherd Dog was created mainly as a very large companion dog, suitable for urban environment and the exhibition ring. There is no evidence that it was used as a traditional shepherd dog.
The aim of this study is to reveal mtDNA variation in three Bulgarian native dog breeds with respect to their phylogenetic origin. The obtained genetic profile and distribution of the main subclades among these breeds are compared with other related geographic populations.

Population structure and sample collection
The experiments had been carried out under permissions and the guidelines of the Bulgarian Academy of Sciences and the Bulgarian Ministry of Environment and Waters (no. 627/30.03.2015). Hair from 130 animals belonging to four different breeds or populations of dogs were collected from breeding kennels, and from distinct geographic locations in their historical breed regions (Table 1, Fig. 1). The hair samples from the new breed of the Bulgarian Shepherd Dog (BOK, n = 14) were taken from three kennels under the   Animals were selected based on morphological standards and on information about their pedigree in order to exclude related animals. All dog lineages originate from different regions of the country, which is confirmed by kennels and dog associations.

DNA extraction, PCR amplification and sequencing
Total DNA was isolated from hair follicles by using a GeneMATRIX Tissue and Bacterial DNA purification Kit (Cat. No. E3551-01; EURx Ltd., Gdansk, Poland) according to the manufacturer's instructions. Briefly, hair was cut into pieces (up to 100 roots). After that, the hair was digested in a lysis buffer (a component of the DNA purification kit), 20 µL of 1M DTT and proteinase K, and incubated overnight at 56 • C. The extracted DNA was resuspended in 50 µL of an elution buffer. The DNA concentration was determined spectrophotometrically, and the quality of the DNA samples was examined on 1% agarose gel electrophoresis stained with Greensafe premium (Cat. No. MB13201; Nzytech, Lisbon, Portugal). The purified DNA was stored at −20 • C until PCR assay.
The successfully amplified products were purified by a PCR purification kit (Gene Matrix, PCR clean-up kit, EURx, Poland) and sequenced in both directions by a PlateSeq kit (Eurofins Genomics, Ebersberg, Germany).

Statistical and phylogenetic analysis
All 130 sequences were manually edited and aligned by MEGA software version 7.0 (Kumar, Stecher & Tamura, 2016) using the dog mtDNA sequence NC_002008 as a reference (Kim et al., 1998). The obtained sequences (about 730 bp in length from covered tRNA-Pro genes and the beginning of the D-loop region, HV1) included in this study were deposited in the National Center for Biotechnology Information (NCBI) GenBank database under accession numbers (NCBI: MG920357-MG920486). Sequences were analyzed by polymorphic SNPs, and haplogroups were determined according to Duleba et al. (2015) and Song et al. (2016) as well as MitoToolPy program (Peng et al., 2015), (http://www.mitotool.org/mp.html) with reference sequence EU789787 (Pang et al., 2009). The statistical quantities for the DNA sequences, including number of haplotypes and haplotype diversity, nucleotide diversity and Fu and Li's D and F test were performed by using DnaSP 5.10.1 (Librado & Rozas, 2009). Phylogenetic relationships of mtDNA haplotypes were explored by a All positions containing gaps and missing data were eliminated. Principal component analyses (PCA) were performed using Excel software implemented by XLSTAT, as described elsewhere (Achilli et al., 2007). The PCA were carried out: one-by considering only our sample.

Genetic diversity and differentiation of Bulgarian dogs
The primers: HVI-F15453/HVI-R16107 amplified a sequence comprising tRNA-Pro genes and the beginning of the D-loop region (HVR1) at the 5 end of the control region (about 630 bp). All sequences covered from 15,361 to 16,092 bp according to Ref. sequence accession number NC_002008 (Kim et al., 1998). We also observed in all dog sequences an insert at the position 15,514 bp according to Ref. sequence accession number EU789787 (Pang et al., 2009). This insert covered 23 bp similar to Ref. sequence accession number NC_002008. Thirty-eight different haplotypes were obtained from all 130 individuals (accession number MG920357-MG920486) ( Table 1). We identified 19 unique haplotypes in all 130 sequences (Table 1, Supplemental Information 1). The largest number of them, seven, was found in the Bulgarian Hound Dog: Barak, while the least-three unique haplotypes-were found in the Bulgarian Scenthound Dog: Gonche (Table 1, Supplemental Information 1).
The coefficient of diversity within all samples was calculated to be 0.014 ± 0.0022 (Kumar, Stecher & Tamura, 2016). The mean group distance varied from 0.019 in the Bulgarian Shepherd Dog: BOK to 0.009 in Bulgarian Shepherd Dog: Karakachan. The highest value of mean distance between populations was calculated as 0.017 between the Bulgarian Shepherd Dog: BOK, the Bulgarian Hound Dog: Barak, and the Bulgarian Scenthound Dog: Gonche. The lowest value 0.013 of the mean distance was observed between the Karakachan and the Gonche.
The Bulgarian Shepherd Dog (BOK) showed the highest value of haplotype diversity (0.934 ± 0.0037), followed by the Bulgarian Hound Dog Barak (0.916 ± 0.0006) ( Table 2). Among the remaining dog breeds, none showed haplotype diversities higher than 0.90. Nucleotide diversity per site (π n ) was on average high, ranging from 0.019 (Bulgarian Shepherd Dog: BOK) to 0.0087 (Bulgarian Shepherd Dog: Karakachan). According to Vila, Maldonado & Wayne (1999), this statistic is not highly sensitive to sample size. The mean number of pairwise differences (π ) between haplotypes within populations varied from 13.83 (Bulgarian Shepherd Dog: BOK) to 5.45 (Bulgarian Scenthound Dog: Gonche). The highest number of haplotypes (Hn), 15, was observed in the Bulgarian Hound Dog: Barak, and the lowest, 10, in the Bulgarian Shepherd Dog: Karakachan and the Bulgarian Scenthound Dog: Gonche.
The Fu and Li's FL-D and FL-F tests showed the lowest values in the BOK dog, which was not surprising, because other dog breeds were included in the creation of that breed ( Table 2).
We defined 38 haplotypes in all dog breeds, 19 of which were unique ( Table 2). All unique haplotypes were represented with only one sample, while haplotype H36 and H37 were represented with four samples belonging to subhaplogroup D2b (Supplemental Information 1).
In order to graphically display (and summarize) the mitochondrial relationships among the analyzed breeds, we performed a principal component analysis (PCA).
The PCA plot clearly separated the two investigated dog breeds: hunting and shepherd (Supplemental Information 1). The hunting dogs: Gonce and Barak are grouped closely, while the BOK and Karakachan dogs are rather distant. The PCA also supported the hypothesis of the independant origin of the investigated hunting and shepherd dog breeds.

Phylogenetic analysis
We defined each (sub)clade by a specific mutation motif encompassing control region (HVR1). We observed four main clades-A, B, C and D. Within clade A, we identified eight subclades; within clades B and C, we identified a lower diversity compared to clade A, i.e., clade B (B1a1 and B1a4) and clade C (C1b3 and C2). (Fig. 2, Supplemental Information 1).
There were significant differences in the distribution in haplogroups between hunting and shepherd dog breeds (Figs. 2 and 3 and Supplemental Information 1). Clade C was observed in the Bulgarian hunting dog breeds (Bulgarian Scenthound Dog: Gonche and Barak), while subhaplogroup D2b was observed only in shepherd dogs (Bulgarian Shepherd Dog: BOK and Bulgarian Shepherd Dog: Karakachan). Additionally, it was also found that some of the subhaplogroups showed different presence in different dog breeds (Fig. 2).

Balkan dog breeds-origin and history
Up to date information on the genetic diversity of dogs in South-Eastern Europe is missing. It is not known how much the Balkan dog population is influenced by crossbreeding along migration routes between the Middle East (Anatolia), West Asia and Europe. On the Balkans, native breeds are typical hunting (Bulgarian Barak Hound, Bulgarian Scenthound Dog (Gonche), Hungarian hound (Transylvanian Scenthound), Serbian Hound and Greek Harehound) and shepherd dogs (Bulgarian Shepherd Dog, Greek Sheepdog (Greece), Tornjak (Bosnia and Herzegovina and Croatia), Sharplaninac (FR Macedonia), Akbash Dog (Turkey). All these have a common predecessor from historic or prehistoric time. These breeds diverged are not morphologically distinctive but they are classified separately because of historical reasons related to changing state borders (in the last 200 years). There is no historically described pedigree for these breeds, but it is believed that their predecessor dates back to the times of the Ottoman, the Roman and the Bulgarian Empires, and even to the Hellenic (Greek) and Thracian kingdoms. It should be noted that hunting dogs have been painted during all historic periods as aristocratic dogs, visually similar to scenthound breeds.  et al., 2015;Pires et al., 2006;Pang et al., 2009;Adeola et al., 2017). We identified 38 haplotypes in all dog breeds. Among them, 19 haplotypes (85%) are identical to the ones available on GeneBank worldwide (Supplemental Information 1) (Cairns et al., 2017;Boyko, Boyko & Boyko, 2009;Duleba et al., 2015;Strakova et al., 2016;Imes et al., 2012;Bekaert et al., 2012;Pires et al., 2006;Verscheure, Backeljau & Desmyter, 2014). Specific for all Bulgarian dog populations is the higher frequency of clade C (about 20%) compared to European dog population (about 10%) (Duleba et al., 2015;Ardalan et al., 2011). Clade C, together with clade D, was predominant in ancient European dogs about 7,000 years ago (63 and 20%, respectively), while most modern European dogs have sequences with haplogroups A and B (64 and 22%, respectively) (Frantz et al., 2016). As a comparison, during the same historical period in South-Western Asia (Anatolia), the frequencies of clades C and D2 were about 7% and 2%, respectively (Ardalan et al., 2011). Otherwise, the subclade D2b from the Bulgarian sets is represented with new specific haplotypes H36 and H37, which differ from other available D2 sequences by parsimony informative substitution T/C at position 15,955 bp, according to ref. sequence accession number NC_002008. Moreover, from all five defined D2 sequences, D2b haplotypes were found in 80% of this set (Supplemental Information 1), which suggests that these haplotypes have been restricted to the Balkans and were isolated a long time ago (probably several millennia ago).

Phylogenetic analysis of the Bulgarian dog population and relationships with surrounding dog populations
The results, especially the frequency of clade C, may be interpreted as a proof of the conservation of ancient European mitotypes in the studied native Bulgarian dogs. a typical mtDNA profile different from that of the Bulgarian hunting dogs. Subhaplotype D2b is typical for these breeds-about 10% for the Karakachan Dog and about 7% for BOK (Supplemental Information 1). Subclade D2 has also been observed in mountain and shepherd dogs like those in Spain as well as the Estrela Mountain Dog and the Alentejo Shepherd Dog (Portugal), and the Turkish Shepherd Dog: Kangal (Turkey) (Pires et al., 2006;Savolainen et al., 2002;Adeola et al., 2017). All these breeds belong to the European and South-West Asian Molossus group.

CONCLUSION
In conclusion, the mtDNA profile of native Bulgarian dogs in the underinvestigated so far South-Eastern Europe has shown typical European haplogroup dissemination. The Bulgarian native dog population is characterized by the highest frequency of clade A (55%), followed by clade B and C (about 18%) and the specific South-European clade D2 (about 10%). Hunting dogs (Gonche and Barak) have a specific mtDNA profile characterized by the presence of clade C and the absence of the Mediterranean subclade D2, as compared to the mountain Shepherd Dog (Karakachan). These data may explain the different phylogenetic origin of two geographically separated European dog populations: Central Europe (hunting dog breeds) and Mediterranean region (mountain shepherd dog).
• Georgi Radoslavov analyzed the data, contributed reagents/materials/analysis tools, prepared figures and/or tables, authored or reviewed drafts of the paper, approved the final draft.
• Peter Hristov conceived and designed the experiments, analyzed the data, prepared figures and/or tables, authored or reviewed drafts of the paper, approved the final draft.

Field Study Permissions
The following information was supplied relating to field study approvals (i.e., approving body and any reference numbers): The experiments were carried out under permissions and the guidelines of the Bulgarian Academy of Sciences and the Bulgarian Ministry of Environment and Waters (no. 627/30.03.2015).