Blood from 30 CS individuals were sampled from two local institutes (Gangwon Provincial Livestock Research Center and Chungbuk Veterinary Service Center). Blood was collected from HU (n = 30) and JB (n = 30) from Chungbuk Veterinary Service Center and Jeju Special Self-Governing Provincial Livestock Institute, respectively. Blood samples of these three breeds were randomly collected, while avoiding parent-offspring or sib pairs where possible according to pedigree information of each institute. The HW (n = 30) was additionally sampled for blood from 11 farms (1 to 7) in Yeongju City.

Genomic DNA from blood samples was extracted using the DNeasy Blood Kit (Qiagen, Germany) following the manufacturer’s procedure. Genomic DNA of two exotic breeds, Holstein (n = 30) and Charolais (n = 26) was obtained from the Animal Genetic Resources Station, National Institute of Animal Science, Rural Development Administration.

Thirty microsatellite markers (BM1818, BM1824, BM2113, CSRM60, CSSM66, ETH3, ETH10, ETH152, ETH185, ETH225, HAUT24, HAUT27, HEL1, HEL5, HEL9, HEL13, ILSTS005, ILSTS006, INRA005, INRA023, INRA032, INRA035, INRA037, INRA063, MM12, SPS115, TGLA53, TGLA122, TGLA126, and TGLA227) were analyzed to estimate various parameters of genetic diversity. Microsatellites were amplified in multiplexes (3 to 4 co-amplified loci), with the exception of ETH185. Polymerase chain reaction amplification was performed in a 15 μL reaction mixture, which contained 1.5 mM MgCl₂, 1 U of AmpliTaq Gold (Applied Biosystems, Foster, CA, USA), 3–10 pmol of each forward (labeled with a fluorescent-colored dye) and reverse primer and approximately 10 ng genomic DNA as a template.

The PCR amplification comprised an initial denaturation at 95°C for 15 min, followed by 35 cycles of denaturation at 95°C for 30 s, annealing at 55°C to 63°C (respective optimal annealing temperature) for 90 s, extension of starters at 72°C for 90 s and a final extension of starters at 72°C for 40 min, using the GeneAmp PCR System 9700 (Applied Biosystems). Electrophoresis was carried out using an ABI 3130xl Genetic Analyzer (Applied Biosystems). Allele sizes of each microsatellite were determined using GeneMapper ver. 4.0 (Applied Biosystems). The formulated allele data was used for statistical analyses.

Cervus ver. 3.0.3 (Kalinowski et al., 2007) was used to estimate allele frequencies, total number of alleles (TNA), mean observed (H_Obs) and expected (H_Exp) heterozygosities ad mean polymorphism information content (PIC) per locus and breed. Allelic richness (A_R) for each breed was calculated to correct distortion by sample size difference using FSTAT ver. 2.9.3.2. (Goudet, 2002). The D_A genetic distance (Nei et al., 1983) was calculated with MSA (Dieringer and Schlötterer, 2003). The three dimensions of a multivariate factorial correspondence analysis (FCA) were computed using GENETIX ver. 4.05 (Belkhir et al., 2004) from allele doses for each individual. The FCA was carried out from all animals and for the 30 loci. Genetic structure and the degree of admixture of KNC breeds were investigated using the Bayesian clustering procedure of STRUCTURE ver. 2.3 (Pritchard et al., 2000). Fifty independent runs were performed for each K between 2 and 10, with a burn-in period of 50,000 iterations followed by 100,000 iterations of the Markov chain Monte Carlo algorithm. To identify the most probable groups (K) that best fit the data, we used the STRUCTURE HARVESTER (Earl and von Holdt, 2012), which implements the Evanno method (Evanno et al., 2005). The program CLUMPP ver. 1.1 (Jakobsson and Rosenberg, 2007) was used to align the 50 repetitions of each K. The CLUMPP out files were visualized using DISTRUCT ver. 1.1 (Rosenberg, 2004).

Allele ranges, number of alleles, heterozygosity and PIC per locus are summarized in Table 1. A total of 276 alleles were detected at 30 microsatellite loci across four KNC breeds. The TNA per locus ranged from 4 (ILSTS005) to 17 (TGLA122), with a mean of 9.20±0.58 alleles. The mean of H_Exp across loci was 0.733±0.018, with estimates per locus ranging from 0.473 (ILSTS005) to 0.893 (TGLA53). For H_Obs, the mean for all loci was 0.667±0.028, and the range was between 0.174 (INRA035) and 0.855 (CSRM60).

The 30 microsatellite markers used were recommended from by International Society for Animal Genetics/FAO (Hoffmann et al., 2004). Most of the loci were highly informative (PIC>0.5), with the exception of ILSTS005 (0.375) and HEL13 (0.413). Similarly, ILSTS005 and HEL13 have been reported to be the relatively low informative markers (Padilla et al., 2009). According to Botstein et al. (1980), PIC of >0.5 indicates a highly informative locus for chromosomal mapping and genetic diversity. Therefore, most microsatellite marker sets are highly informative and useful for evaluation of genetic diversity and population structure in KNC breeds.

The various indices of genetic diversity across the four KNC breeds are shown in Table 2. The mean number of alleles and A_R ranged from 4.73±0.32 (HU) to 6.97±0.41 (HW) and from 4.39±0.28 (HU) to 6.42±0.37 (HW), respectively. The mean value of H_Exp was highest in HW (0.713) and lowest in JB (0.604). The mean value of H_Obs ranged from 0.613 (HU) to 0.683 (CS). Generally, the levels of genetic diversity of HW and CS breeds were higher than those of HU and JB. The difference of the levels of genetic diversity among Korean cattle breeds could be explained by population sizes and distribution region. Populations of HW and CS are relatively large (currently 3,000,000 animals) and medium-sized (currently 1,700 animals), respectively. These two breeds are widely distributed throughout Korea. On the other hand, HU and JB are numerically small populations, with 300 to 400 animals each. Moreover, JB are restricted in their distribution to Jeju Island. HU and JB are considered endangered breeds (IUCN, 2000). The H_Exp and H_Obs values observed in KNC breeds are reportedly similar or higher than those of British and Spanish cattle breeds (H_Exp = 0.56 to 0.68 and H_Obs = 0.59 to 0.67) (Wiener et al., 2004; Martin-Burriel et al., 2007; Padilla et al., 2009). The mean values of H_Obs in HW and CS breeds were lower than those of H_Exp. Generally, the mating of the four KNC breeds has been non-randomly performed by using limited bulls. In addition, the number of bulls used is fewer in HW and CS than in HU and JB based on population size.

To verify the genetic relationship among the four Korean native and two exotic cattle breeds, we calculated the D_A distance and constructed a phylogenetic tree. Among the Korean native breeds, HW and CS were closest (D_A = 0.129), with the largest difference observed for JJ and HU (D_A = 0.265) (Table 3). The phylogenetic tree was constructed from a neighbor joining (NJ) clustering based on the D_A distance matrix among breeds (Figure 2). The genetic relationship between HW and CS was relatively close among the Korean breeds, whereas HU and JJ were distinctly separated. The NJ tree indicated a clear separation of two exotic breeds (HT, CR) from the four KNC breeds. Yoon et al. (2005) emphasized that genetic similarity was high among Hanwoo, Heugu, two Japanese breeds, and a Chinese breed (Yanbian). Additionally, the bootstrap values among these breeds ranged from 40 to 50. Thus, the low values in this study were caused by high genetic similarity among Korean native breeds.

The FCA analysis revealed the very clear separation between the HT, CR, HU, JJ, and other two Korean breeds (HW and CS) (Figure 3). About 75% of the variance was accounted for by the first to three dimensions of the FCA Axis 1 (33.33% of total variance explained), which separated HT and CR from the Korean native breeds. Axis 2 (24.37%) further separated HU, JJ and the HW and CS Korean breeds. Axis 3 (17.20%) distinctly separated HT and CR. These results indicated that individuals from HW and CS were mixed, suggesting a closer relationship between them. D_A distance, phylogenetic tree and FCA results provide genetic evidence for the differentiation of the four Korean cattle breeds.

STRUCTURE software was used to determine the unbiased structure without prior knowledge regarding the number of breeds (Figure 4). At K = 2, two main groups that accurately corresponded to Korean native and exotic cattle breeds were formed. As K increased, the contributions of the assumed populations resulted in the progressively complete separation of the 6 breeds. The largest delta K (ΔK) value was calculated as previously described (Evanno et al., 2005). The optimum ΔK value (ΔK = 92.94, data not shown) was found at K = 5. These results failed to differentiate between the HW and CS breeds. However, each breed grouped in its own cluster with an estimated membership >0.893 at K = 6 (ΔK = 52.54) (Figure 4, Table 4). According to Leroy et al. (2009), the highest ΔK values can potentially reveal the optimal K, but some weakly defined substructures can be found when only a small number or breeds are analyzed. In addition, Kim et al. (2013) suggested that CS and HW have a genetic difference based on sequence variation and phylogenetic analysis of mtDNA cyt b gene. Based on these reports, we considered that HW and CS could have distinct genetic characteristics.

In this study, the level of genetic diversity among KNC breeds differed according to their population sizes (endangered, vulnerable or not at risk). However, these were higher than those of several native cattle breeds of other countries. In analyses of genetic relationship and clustering, all KNC breeds were genetically differentiated from the two exotic breeds. In addition, the scientific evidence supports the genetic differentiation among four KNC breeds. The results suggest that each KNC breed had distinct breed-specific genetic characteristics. The results of this study may be useful as scientific evidence to design plans for future conservation, improvement and breed management of each KNC breed.