Genetic diversity of the Hungarian Gidran horse in two mitochondrial DNA markers

The Gidran is a native Hungarian horse breed that has approached extinction several times. Phylogenetic analysis of two mitochondrial markers (D-loop and cytochrome-b) was performed to determine the genetic characterization of the Gidran for the first time as well as to detect errors in the management of the Gidran stud book. Sequencing of 686 bp of CYTB and 202 bp of the D-loop in 260 mares revealed 24 and 32 haplotypes, respectively, among 31 mare families. BLAST analysis revealed six novel CYTB and four D-loop haplotypes that have not been previously reported. The Gidran mares showed high haplotype (CYTB: 0.8735 ± 0.011; D-loop: 0.9136 ± 0.008) and moderate nucleotide (CYTB: 0.00472 ± 0.00017; D-loop: 0.02091 ± 0.00068) diversity. Of the 31 Gidran mare families, only 15 CYTB (48.4%) and 17 D-loop (54.8%) distinct haplotypes were formed using the two markers separately. Merged markers created 24 (77.4%) mare families, which were in agreement with the mare families in the stud book. Our key finding was that the Gidran breed still possesses high genetic diversity despite its history. The obtained haplotypes are mostly consistent with known mare families, particularly when the two mtDNA markers were merged. Our results could facilitate conservation efforts for preserving the genetic diversity of the Gidran.


INTRODUCTION
Conservation of domestic animal breeds has played an ever increasing part in biodiversity preservation. Hungary realized relatively early the unique value of maintaining the genetic diversity of endangered species (Bodó 1985). The Gidran is one of the smallest horse populations among Hungarian horses (Pataki 1996). It was crossed with and influenced many other breeds in Eastern Europe and is closely associated with the Kisbér Half Breed, which is another traditional Hungarian breed. Chestnut is the usual colour of the Gidran, but other colours common to the Arab horse occur in this breed. In addition to having cultural and genetic value, the Gidran is well known for its athletic speed, agility, endurance, well-balanced temperament, and robust build (Glyn 1971). Due to its unique characteristics, the Gidran is widely used in many equestrian sports, such as horse jumping and carriage driving, and achieves outstanding results in international competitions. Its origin goes back to 1816, but the Gidran was only registered as a separate breed in 1885 by the Austrian Ministry of Defence (Jónás et al. 2006). During its uncertain history, two major bottleneck effects drove the breed to the edge of extinction. Fortunately, the reconstruction and regeneration of the Gidran began because of a few dedicated breeders in the 1990's. Retrenching the number of mare families was a key moment in this preservation, which took into account that each mare family has more than twenty generations of breeding history (Jónás et al. 2006). As a consequence, the final number of mare families was determined, and the latest version of the official Gidran stud book was published in 2005(Mihok 2005. The date of the establishment of each mare families and the name of the founder mares are summarized in the Online Supporting Information (Tables S5.). Although, the ongoing restoration of the Gidran is a notable example of gene pool protection (Mihok & Bodo 2003), the status of this horse is still endangered. According to the Domestic Animal Diversity Information System database of the Food and Agriculture Organization (FAO DAD-IS), the Hungarian Gidran population is composed of 298 mares and 21 stallions, but smaller populations also exist in Romania and Bulgaria (DAD-IS, 2014). Therefore, maintaining Gidran biodiversity is an important challenge not only for gene preservation but also from the point of view of the World Heritage (Mihok & Bodo 2003). Within the framework of breed conservation, genetic characterization acts an important aspect of maintaining breed integrity and managing genetic resources (Glowatzki- Mullis et al. 2006). Mitochondrial DNA (mtDNA) analysis has been used in phylogenetic and domestication studies and displays a high level of genetic variation among maternal lineages in horses (Achilli et al. 2012;Cieslak et al. 2010;Jansen et al. 2002 allowing an evaluation of maternal line assignment accuracy (Wan et al. 2004). Several investigations have shown that using two or more mtDNA markers might be more robust and powerful for genetic diversity analysis (Pedrosa et al. 2005). Therefore, analysis of sequence variations of mtDNA such as CYTB or D-loop region is an outstanding approach for the mapping of the Gidran's maternal lineage.
The main goal of this study is to examine the genetic diversity and relations among the Gidran maternal lines. We present the first phylogenetic characterization of the Gidran for the identification of rare or distinct mtDNA haplotypes using 686 bp and 202 bp sequences of the mitochondrial CYTB and D-loop, respectively. The second aim of the recent study was to recognize the overlapping haplotypes or errors in the management of the stud book to gain a better understanding of the genetic variability among the Gidran mare families. Our results could complement the recent conservation strategies to maintain the genetic diversity of this traditional horse breed.

PCR amplification and sequencing
Based on the reference Equus caballus mtDNA sequences (GenBank accession nr.: primers were also used as sequencing primers. The obtained DNA sequences were compared with the reference sequences from GenBank using Clustal X (Thompson et al. 1997).

Data analyses
Phylogenetic analysis was conducted for a total of 250 (CYTB), 246 (D-loop) individuals.
Combined CYTB and D-loop analysis was limited to those 242 horses where the PCR amplifications were successful for both markers. BioEdit v7.2.5 (Hall 2004) sequence alignment editor software was used to proof and correct individual electropherograms of the sequences. All sequence alignments were performed using a general reference sequence (GenBank accession nr.: X79547) and a latterly used reference sequence (GenBank accession nr.: JN398377) is also used in DomeTree, wich is toolkit for mtDNA analyses in domesticated animals (Peng et al. 2015).
Complementary sequences were assembled and truncated to a length of 686 bp (CYTB) and 202 bp (D-loop) to allow for maximum sample size. A BLAST search in the NCBI database was used to determine any previously unreported haplotypes. Standard diversity measures, such as polymorphic sites (Ps), haplotype (h) and nucleotide diversity (p), were calculated by DNASP 5.0 software (Rozas et al. 2003). A pairwise distance matrix between the mtDNA haplotypes was independently calculated for the CYTB and D-loop by the nucleotide p-distance (Nei & Kumar 2000). Maximum likelihood (ML) phylogeny was constructed using the Hasegawa-Kishino-Yano (HKY) plus gamma (CYTB) and Tamura 3parameter (T92) plus gamma model (D-loop) by the inbuilt model generator in MEGA5 (Tamura et al. 2011). An Equus asinus sequence (GenBank accession no.: NC001788) was used as an outgroup. Bootstrap analyses (1000 replications) were used to assess the confidence of each node.
According to the polymorphic sites, both haplotypes were assigned to the DomeTree (Peng et al. 2015) and D-loop haplotypes were also clustered into the haplogroups had been defined by Jansen (2012). A phylogenetic network based on merged CYTB and D-loop regions was constructed by use of a median-joining algorithm (Bandelt et al. 1999) as implemented in the Network 4.1 program.

Based on the sequence comparisons of the mitochondrial CYTB and D-loop markers, the
Gidran horses showed high genetic variability. Twenty-three polymorphic sites were detected in the CYTB sequences, corresponding to two indels (e.g., insertion and deletion) and 21 single nucleotide polymorphisms (SNPs) with two transversions, and representing 3.35% of the analysed DNA sequence. Within the D-loop region, 26 polymorphic sites were found (one indel and 25 SNPs with a transversion) representing 12.9% of the analysed DNA sequence (Table 1.).
Both mtDNA regions were A/T rich with the following nucleotide frequencies: T: 27.7%, C: 31.5%, A: 27.3% and G: 13.5% in CYTB and T: 30.8%, C: 24.7%, A: 33.3% and G: 11.2% in Dloop. The A and T content was richer (55% and 64.1%) in both the CYTB and D-loop regions, respectively. These data were in accordance with the order of nucleotide composition in the vertebrate mitochondrial genome. The calculated haplotype diversity (h) of the CYTB and D-loop markers was 0.8735 ± 0.011 and 0.9136 ± 0.008, whereas the nucleotide diversity was 0.00472 ± 0.00017 and 0.02091 ± 0.00068, respectively. The paired genetic distances between the haplotypes were 0.001-0.013 (CYTB) and 0.005-0.063 (D-loop). All phylogenetic analyses were performed for both separate and combined mtDNA markers; a summary of the calculated genetic diversity parameters and the haplotypes of the Gidran mares are presented in Table 1.  (Table S1 & Table S2). Ht1CYTB (n=54), Ht2CYTB (n=49) and Ht6CYTB (n=44) (Table S1 & Table S2). The combined CYTB and D-loop haplotypes could be clustered into ten ( haplogroup E was not present in the Gidran native horses considered (Table S3 & Table S4). Our additional key objectives were to identify errors in the Gidran stud book and to test The Gidran is a rare and endangered native Hungarian horse breed; therefore, the development of an effective conservation strategy is extremely urgent. A determination of phylogenetic relationships and the verification of stud book accuracy could be the first steps in the maintenance of the genetic pool (Bodó et al. 2005). To date, contrary to the situation with the Hucul, which is another, but well-studied native Hungarian horse breed (Czerneková et al. 2013;Georgescu et al. 2011;Kusza et al. 2013), no data are available on the genetic structure and diversity of the Gidran. In this regard, through DNA sequence comparisons of CYTB and the D-  (Bowling et al. 2000) and higher than in Kiso (7 haplotypes/ 136 horses) (Takasu et al. 2014). The calculated D-loop haplotype and nucleotide diversities are inconsistent with earlier horse mtDNA studies. The calculated nucleotide diversity was 0.02091 ± 0.00068, which is quite similar to the Iranian horse population (0.02 ± 0.000) reported by (Moridi et al. 2013). This relatively high number indicates, that the Gidran is genetically more diverse than, for example, the Kerry Bog (0.0155 ± 0.0040) and Sulphur Mustang breeds (0.001 ± 0.002), but not more diverse than the Marwari (0.03973 ± 0.01262) or Sorraia breeds (0.104 ± 0.012) (Devi & Ghosh 2013;Luís et al. 2006;Prystupa et al. 2012). Data from the CYTB sequences also confirmed the abundant genetic diversity of the Gidran. In the case of CYTB, the nucleotide diversity was lower in comparison to the D-loop, but it is similar to that observed in Chinese domestic horses, where the nucleotide diversity was between 0.00488 and 0.00186 while haplotype diversity was between 0.706 and 0.975 (Yue et al., 2011). Quin et These data suggest that although the Gidran is one of the smallest Hungarian horse population, the genetic diversity of the maternal lineage is preserved (Takasu et al. 2014).
BLAST showed that almost all the haplotypes found in the Gidran samples are identical to other domestic horse haplotypes in GenBank except for six CYTB and five D-loop haplotypes that have not been described for any other horse breed yet. Because uncommon haplotypes have an increased risk of extinction (Lopes et al. 2005), these distinct haplotypes support the importance of maintaining rare individuals and also emphasize the genetic diversity of the Gidran. Pedigree analysis also plays a key role in breeding programmes, which aim to maintain genetic diversity of endangered populations (Bokor et al. 2013). Our further aim was to recognize overlapping haplotypes among mare families and thereby detect incidental errors in the Gidran stud book. Using the mtDNA markers alone was not sufficiently effective because neither CYTB nor D-loop sequences made the alignment of the haplotypes to each of the mare families reasonably acceptable. To improve the efficiency of our data, starting from the study of Pedrosa et al. (2005), a phylogenetic analysis with the combined CYTB and D-loop markers was also performed ( Figure 3). Although, the results showed several inconsistencies in the distribution of the 49 common CYTB and D-loop haplotypes within the 31 mare families, we found this approach more effective to screen for registry errors rather than using only one mtDNA marker. Altogether, pedigree records were problematic in seven mares (2.89%) registered in the Gidran stud book. This number is small compared with Lipizzan and Polish horses where the discrepancies were 11% between pedigree data and mtDNA haplotypes in studies on both breeds (Głażewska et al. 2007;Kavar et al. 2002). The data collected in the present study indicates the management of the Gidran stud book was appropriate over the years. The seven problematic mares belonged to a different cluster (independent of the DNA markers and their combination) than suggested according to the Gidran stud book. Several reasons could explain the described inconsistencies. However, errors may have been made in the management of the stud book because the approximately 200 years of existence of the maternal line is a very short time for the formation of a distinct haplotype of each mare family (Devi & Ghosh 2013). Four horses did not possess their own cluster. Among them, two animals from the borodi 7 mare family formed a common cluster with borodi 1. Furthermore, two individuals of the mezőhegyesi 5 mare family shared a haplotype with animals of mezőhegyesi 11. These discrepancies are in concordance with the mare family's history (Mihok 2005). Individuals of borodi 1 and 7 shared the same haplotype. Possibly, mare families sharing the same haplotypes belong to the same mare family, which is, as a consequence of the incomplete pedigree data, now split into different mare families (Kavar et al. 2002). In contrast, based on the analysis of the two mitochondrial markers separately and combined, individuals of the mezőhegyesi 4 mare family formed two distinct haplotypes. This observation was confirmed by the available historical data, which suggests that the mezőhegyesi 4 mare family diverged over the years (Mihok 2005).
New haplotypes are emphasized by gray colour.   (*Unpublished data; **Haplotypes were clustered into haplogroups) Table S5. Summary of the 32 mare families according to the Gidran studbook. Figure S1. Classification tree of combined CYTB and D-loop haplotypes of Gidran mares according to the DomeTree. The polymorphic sites considered relative to the JN398377 reference sequence, which is identical with Haplotype 1 (H1).