Horses have played a pivotal role in the development of human society. As an important domestic animal, the origin and history of modern domestic horses have been studied for decades in the biology and archaeology fields. At present, it is widely believed that horses were domesticated from several wild populations 4,000 to 6,000 years ago (Vila et al., 2001; Outram et al., 2009). Subsequently, they differentiated into many types after long-term domestication (Jansen et al., 2002). In history, horses differed from other domestic animals that were kept as food sources, and attained a prominent role as animals of transport and warfare, changing societies on a continent-wide scale. Literature reports the importation of horses into China from different parts of the world throughout history (Xie, 1987). As a strategic military asset, governments paid much attention to horse breeding throughout Chinese history. Some archaeological evidence and genetic data supports China as one of the earliest areas of horse domestication (Lei et al., 2009). China has rich horse genetic resources, the number of domestic horses in China was 7.639 million in 2004 (Statistic report in China, 2004). Chinese indigenous horse breeds are classified into five types based on geographical distribution and characteristics. They are Mongolian, Kazakhstan, Hequ, Southwest and Tibetan types (Chang, 2009). Four horse types are distributed in western China and still act as an important source of animal power because regions such as Guizhou, Shaanxi and Xinjiang are mountainous and pastoral. Ningqiang horse is an ancient breed. Individuals of this breed with a height at the withers less than or equal to 106 cm are called Ningqiang pony, while the other with a height greater than 106 cm are called Ningqiang horse (Hou, 1989). With the development of mechanization, the role of horse as a provider of transport and draft power is decreasing. The amount of horses declined quickly in recent years. So it is urgent to study and protect horse genetic resources.

The mtDNA is a small extra-chromosomal genome that is maternally inherited. The equine mtDNA has an approximate length of 16,660 bp and consists of a coding region which contains 37 genes and a non-coding region called the D-loop region of approximately 1,192 bp (Wolstenholme, 1992; Boore, 1999; Bowling et al., 2000). The mtDNA sequences have been widely used to study taxonomy, phylogeny, genetic structure and biological identity (Peng et al., 2007; Ji et al., 2008; Ki et al., 2010). D-loop mtDNA is an important tool to study equine diversity, phylogenetic relationships and origins, such as the Mongolian horse (Lei et al., 2009), Cheju horse (Yang et al., 2002), Debao pony (Jiang et al., 2011), Japanese horse (Kakoi et al., 2007), Tibetan horse (Xu et al., 2007), Przewalskii’s wild horse (Ishida et al., 1995) and donkey (Lei et al., 2005). The diversity and origin of the Ningqiang pony, Guizhou horse and Kazakh horse as indicated by microsatellite markers and mtDNA D-loop have been reported (Du et al., 2009; Lei et al., 2009). However, the mtDNA D-loop sequence (247 bp) studied in the literature was short and might not fully show the diversity and origin of the three horse breeds in western China. Consequently, a 600 bp mtDNA control region of 43 animals from Ningqiang pony, Guizhou and Kazakh breeds and 81 mtDNA D-loop sequences of 6 breeds from GenBank (Table 1) were used in this present study to determine the diversity and origin of the three native horse breeds in western China.

In recent years, many researches have indicated the diversity and origin of different horse populations, and suggested that domestic horses were derived from several wild populations (Lister, 1998; Vila., 2001; Jansen, 2002). Wallner (2003) studied the unique nucleotide sequence of the Y chromosome, and found that Przewalskii haplotype was the ancestor. Lei et al. (2009) showed that the domestication of horses was not a single discrete event, multiple horse domestication events may have occurred later in other regions, and that China was one of these domestication regions.

We found 33 haplotypes in 43 horses from three western China native horse breeds based on 600 bp of mtDNA control region sequence, and all these haplotypes grouped into seven lineages. These three horse breeds shared A lineage, and also had their respective lineages. Nine haplotypes were identified in ten Ningqiang ponies, and were distributed into two lineages (A, F). There were 13 haplotypes in 14 Guizhou horses representing three lineages (A, B, D, G). In 19 Kazakh horses, 15 haplotypes were distributed into four lineages (A, C, E, F). These data showed that the three western China native horses possessed abundant mtDNA diversity and indicated multiple maternal origins. The results were in accordance with other results from Yunnan, Mongolian, Debao, Guanzhong and other horse breeds in China (Wang et al., 1994; McGahern et al., 2006; Li et al., 2009). Some Ningqiang ponies, Guizhou horses and Kazakh horses were clustered into A lineage, indicating lineage A is predominant. Ningqiang pony breed demonstrated only A and F lineages, indicating close genetic relationship within this breed, which was consistent with the results of Lei et al. (2009). However, in the study of Lei et al. (2009), the Ningqiang breed indicated A, B, D and F lineages, Kazakh breed indicated A, B, C, D, F and G lineages, and Guizhou breed indicated A to G lineages. The possible reason for this difference with the present study is that more samples were analyzed in the study of Lei et al. (2009). It is necessary to point out that the lineages in the study of Lei et al were based on 247 bp while in the present study 600 bp D-loop sequences were used in these three horse breeds. In many other mammalian species, the variation in this short segment of the mitochondrial genome is often accompanied by high levels of recurrent mutations, thus blurring the structure of the tree and rendering the distinction between some important ancient branches within the tree virtually impossible (Achilli et al., 2009). So, to improve mtDNA phylogeny, the longer D-loop sequences in mitochondrial genomes will be helpful to achieve the highest level of molecular resolution in horses (Achilli et al., 2012).

In this study, three typical horse breeds were selected to clarify the origin, evolutionary relationship and exchange background of western horses in China. Kazakh horse is an ancient breed, and distributed in the Chinese northwest region, while the Guizhou horse is a southwest horse breed. Ningqiang pony is a breed from the border land between northwest and southwest and possesses some characteristics of northern and southern horse populations. Many researchers claim that Ningqiang pony belongs to the Southwest type. The recent results of archaeological studies indicated that there was a chama gudao (Tea-Horse road in ancient times) in southern Shaanxi and that Ningqiang county was an important center of the Tea-Horse trade system. Combined with the geography, history and the data from genetics, we believe that Ningqiang ponies were affected by northern and southern horse populations, such as the Mongolian, Guizhou and Kazakh breeds. Our data demonstrated that Chinese native horse breeds had extensive genetic flow from breeds all over the world. The results were in accordance with the historical records that horses were continuously imported for military forces and trade during Chinese history (Xie, 1987).