Population genetic study of the raccoon dog ( Nyctereutes procyonoides ) in South Korea using newly developed 12 microsatellite markers

The raccoon dog (Nyctereutes procyonoides) is distributed from southeastern Siberia to northern Vietnam, including Korea and Japan, as well as Europe. In Korea, most of its predators and competitors are extinct, which has resulted in rapid growth of the raccoon dog population. This population increase has raised concerns about its role in the ecosystem and the zoonotic transfer of various contagious diseases, and thus an effective method of raccoon dog population control in Korea is required. To investigate the genetic diversity and structure of raccoon dog populations, 12 polymorphic microsatellite loci were identified and characterized. These novel microsatellite markers were employed to obtain basic population genetic parameters for 104 N. procyonoides specimens from five locations in South Korea. The mean allele number of 12 loci across samples was 8.7, and the number of alleles per locus ranged 2–13. Mean expected and observed heterozygosities were 0.723 and 0.619, respectively. Genetic differentiation, estimated by pairwise FST, was significant for all population pairs excepting Seoul/Gyeonggi and Gangwon pair, with a moderate level of genetic differentiation for all the population pairs (mean FST = 0.054), but little differentiation between Seoul/Gyeonggi and Gangwon (FST = 0.024). Bayesian-based clustering analysis predicted that Korean raccoon dog population is composed of four distinct genetic subpopulations. These genetic information and structure of raccoon dog will be very useful to prevent spreading infectious diseases.

N. p. koreensis is an endemic subspecies and is abundant throughout South Korea.In Korea, reduction of its predators and competitors as well as high adaptability to diverse environments, have led to rapid raccoon dog population growth (Hyun et al., 2005).Consequently, the raccoon dog has become a top predator in Korea despite its modest size.The role of predator in the ecosystem is very crucial, moderating prey densities and contributing to biodiversity (Estes et al., 2001;Ripple and Beschta, 2004;Roemer et al., 2009).In addition, this population increase has raised public health concerns due to the potential for zoonotic transfer of various contagious diseases (Botvinkin et al., 1981;Cherkasskiy, 1988;Hyun et al., 2005;Oh et al., 2012).In several cases, raccoon dogs have been the core species spreading rabies and canine distemper between domestic dog and wild Canidae species in Korea (Hyun et al., 2005;Han et al., 2010;Yang et al., 2011;Oh et al., 2012).After a long break, a rabies outbreak has been ongoing since 1993 in the provinces of Seoul/Gyeonggi and Gangwon in northern Korea (Hyun et al., 2005;Yang et al., 2011;Oh et al., 2012).To prevent spreading rabies to the other parts of Korea and also con-trol other contagious diseases, it is important to understand the population structure of the raccoon dog in Korea.Studies of genetic structure provide useful information on dispersal patterns or range expansion of the population (Gillespie, 1975).However, there are currently no data on the structure or demographical trends of N. p. koreensis (Kahula and Saeki, 2004).To investigate genetic diversity and population structure, we analyzed raccoon dog populations in South Korea using polymorphic genetic markers.
Microsatellite markers have been widely used to study individual identification, phylogeographic relationships, and genetic variation, as well as population structuring and genetic differentiation of various wildlife species (Beaumont and Bruford, 1999;Polziehn et al., 2000;Hu et al., 2007;Lee et al., 2011;Park et al., 2011).Although plentiful ecological studies and some phylogenetic research with mitochondrial DNA on raccoon dog have been conducted, there are no previous data on the population structure using microsatellite markers (Kahula and Saeki, 2004).Moreover, canine-derived microsatellite markers applied in raccoon dog (N.p. procyonides) (Rogalska-Niznik et al., 2003;Szczerbal et al., 2003) were not profitable to apply to our raccoon dog samples as some markers failed for amplification or the other amplified products were monomorphic.In this study, we report the isolation and characterization of 12 novel polymorphic loci from N. p. koreensis, and their use in determining the genetic structure of five raccoon dog populations in South Korea.

Microsatellite marker development
Twelve raccoon dog microsatellite loci were isolated and characterized using microsatellite biotin-enrichment methods (Sarno et al., 2000;Kim and Sappington, 2004).Genomic DNA was extracted using a Qiagen DNeasy Blood & Tissue Kit (QIAGEN, Valencia, CA, USA).Extracted genomic DNA was digested with NdeII and fragments in the size range of 250-700 bp were isolated by electrophoresis on a 1.2% TBE agarose gel.Products were purified using a DNA Clean & Concentrator-5 kit (Zymo Research, Irvine, CA, USA).
Aliquots (1 μL) of biotinylated capture probe (5′ biotin (CA) 15 , 5′ biotin (CT) 15 , and 5′ biotin (AGC) 7 ) were annealed to 10 μL volumes of linker-ligated DNA in 89 μL of 5 × SSC.Reactions were then heated at 95°C for 10 min, cooled on ice for 30 sec, then incubated for 5 min at room temperature.One hundred microliters of washed magnetic beads (1 mg/mL) were added to the DNA and the reactions were incubated for 15 min at room temperature to attract biotinylated probes.Residual unattached fragments were removed by washing three times with 200 μL 2 × SSC at room temperature, and three times with 200 μL of 1 × SSC at optimized temperature for 3 min (65°C for (CA) 15 , 61°C for (CT) 15 , and 67°C for (AGC) 7 ).The DNA was eluted from the beads into 50 μL of water by incubation for 5 min at 95°C.The DNA containing repeat sequences was amplified with the EP-3 single primer.The PCR products were ligated to pGEM-T Vector (50 ng/μL) and transformed into Escherichia coli JM109 competent cells (Promega, USA).Positive clones were identified by screening with M13 forward and reverse primers using the method described in Schuelke (2000).The presumptive microsatellite marker amplification was conducted in 10 μL reaction volumes containing 2.5 mM MgCl 2 , 0.2 mM dNTPs, 1 μM each primer (M13 forward, M13 reverse, each internal repeat primer; (CA) 12 , (CT) 12 , (AGC) 6 ), 0.25 U i-Star Taq DNA polymerase (iNtRON Inc), and 10-50 ng of template.The PCR was performed under the following conditions: initial denaturation for 5 min at 96°C followed by 35 cycles of 96°C for 1 min, 50°C for 1 min, and 72°C for 1 min; a final extension at 72°C for 1 min was performed.Subsequent agarose gel electrophoresis enabled selection of repeated sequence inserts, seen as a smeared band pattern, which were sequenced in both directions using M13 primers.Primers for microsatellite amplification were designed using Primer 3 (Rozen and Skaletsky, 2000).

DNA samples and genotyping
We analyzed tissue samples from 104 raccoon dogs from five provinces of South Korea (Seoul/Gyeonggi, n = 22; Gangwon, n = 22; Chungcheong, n = 16; Gyeongsang, n = 22; Jeolla, n= 22, Fig. 1), collected by the Conservation Genome Resource Bank for Korean Wildlife (CGRB) associated with rescue centers throughout the South Korea.Samples of each region were collected from scattered points in a region.All genomic DNA was extracted using a DNeasy Tissue Kit (Qiagen).DNA was amplified with specifically designed genotyping primers (Table 1) using a touchdown profile for PCR amplification: initial denaturation for 3 min at 94°C , followed by 20 cycles of 94°C for 1 min, annealing temperature of 60-50°C decreased by 0.5°C per cycle, 72°C for 1 min, followed by 20 cycles of 94°C for 1 min, 50°C for 1 min, 72°C for 1 min, and a final extension at 72°C for 10 min.The PCR reaction mixture contained 1.5 mM MgCl 2 , 200 μM each dNTP, 0.5 U i-Star Taq DNA polymerase (iNtRON Inc), 0.3 μM each of the fluorescently labeled forward primer and unlabeled reverse primer, and 10-50 ng template DNA.Alleles were genotyped using an ABI Prism 3730 XL DNA Analyzer (Applied Biosystems, Foster City, CA, USA) using a GENESCAN-500 (Rox) size standard, and analyzed with GeneMarker version 1.9 software (SoftGenetics, State College, PA, USA).

Data analysis
Tests for deviation from Hardy-Weinberg equilibrium (HWE) and linkage equilibrium between loci were conducted with GENEPOP version 4.0 (Rousset, 2008).A Bonferroni correction was applied for significance levels for multiple tests (Rice, 1989).The number of alleles (A) and the probability (P) that the difference between expected (H E ) and observed (H O ) heterozygosities was due to chance were obtained using CERVUS version 3.0 (Kalinowski et al., 2007).Allelic diversity was calculated using GenAlEx v6.1 (Peakall and Smouse, 2006) to characterize the polymorphism of each marker.The plausible occurrence of null alleles was tested using Brookfield method in MICROCHECKER (Oosterhout et al., 2004), and Arlequin 3.11 (Excoffier et al., 2005) was employed to test selective neutrality of each locus using the Ewens-Watterson-Slatkin exact test of allele frequency distribution (Slatkin, 1996).
To analyze population structure, FSTAT 2.9.3.2 (Goudet, 1995) was used to calculate pairwise F ST (Weir and Cockerham, 1984).We then investigated genetic structure using the Bayesian clustering function of STRUCTURE 2.2 (Pritchard et al., 2000).Assuming that the data were represented by K separate clusters, the log posterior probability of the data for a given K, ln Pr(X/K), was generated for each of the 20 STRUCTURE runs at K values of 1-8 for the raccoon dog sample locations.The initial burn-in period was 10,000 followed by 10,000 replications after burn-in (Evanno et al., 2005).Admixture and allele frequency correlated models were chosen for the analysis and ΔK, index which is based on the rate of change in the ln likelihood of the data between successive K (1-8), was also calculated to estimate K (Evanno et al., 2005).The membership coefficients were estimated from 100 replicate runs at K = 4 with permutation analysis using CLUMPP version 1.1.2(Jakobsson and Rosenberg, 2007) and the output of genetic clustering was visualized using software DISTRUCT version 1.1 (Rosenberg, 2004).

RESULTS
We screened a total of 384 colonies that were presumed to contain repeat units, and of these, 301 were sequenced using the M13 forward primer.Eighty-two unique sequences (24 of CA repeat, 39 of CT repeat, and 19 of AGC repeat) with more than eight repeats for each sequence unit were selected and used to design primers for PCR amplification.Using these primers, 13 of CA repeat, 14 of CT repeat, and 15 of AGC repeat targeted primers resulted in positive PCR amplification from four raccoon dog DNA samples.Forward primers from each primer set were labeled with two different fluorescent dyes, 6-FAM and 6-HEX, for multiplex PCR applications.Finally, 12 polymorphic and consistently discernible microsatellite markers for Korean raccoon dog populations were obtained.
Characterization of each marker for the 22 Seoul/ Gyeonggi raccoon dog samples is presented in Table 1.No significant linkage disequilibrium was found among any of the loci pairs (not shown) and all 12 loci did not show deviation from HWE (Adjusted α = 0.004, Table 2).Although MICROCHECKER showed that Nyct 2 revealed evidence of the presence of null alleles, the Brookfield frequency indicates the heterozygote deficiency is very low (0.108), so we conclude Nyct 2 has adequate resolution to serve as a marker as well (Dakin and Avise, 2004).We applied these 12 developed markers to other four populations of South Korea and analyzed genetic structure of them as well.A total of 104 alleles were observed in 104 raccoon dog samples from five provinces for the 12 polymorphic microsatellite markers, with 77 alleles from Seoul/Gyeonggi samples, 64 alleles observed in samples from Gangwon, 73 alleles from Chungcheong samples, 65 alleles from Gyeongsang samples and 66 alleles from Jeolla samples.Descriptive statistics of raccoon dog samples from the five regions are shown in Table 2.All 12 microsatellite markers were polymorphic at all locations, with the number of alleles per locus ranging from 2-11 (mean = 8.7).The mean value of expected and observed heterozygosities were 0.723 (0.333-0.864) and 0.619 (0.227-0.733), respectively.Some loci were not in HWE after Bonferroni correction (Nyct 6 in Gangwon, Nyct 3 and 5 in Chungcheong, Nyct  2); 0 < low < 0.4, 0.4 ≤ moderate < 0.8, 0.8 ≤ high < 0.9.b Presence of null allele discovered based on Brookfield frequency.c 'Yes' denotes the null hypothesis of selective neutrality against the presence of selection was not rejected for that locus at P = 0.05, and 'No' denotes the null allele hypothesis of neutrality was rejected for that locus.Population genetic study of Korean raccoon dog 10 in Gyeongsang), however, we confirmed that exclusion of theses markers did not affect the result of STURCTURE analysis.According to the STRUCTURE analysis, the ΔK value was the highest when K was set at 4, implying that raccoon dog species in the five areas in Korea consist of four distinct populations: 1. Seoul/Gyeonggi and Gangwon, 2. Chungcheong, 3. Gyeongsang and 4. Jeolla (Fig. 2).Pie chart showing proportion of the STRUCTURE clusters (Fig. 1) and F ST also showed that the raccoon dogs collected from the five locations represent four groups.Furthermore, Nei's genetic distance between locations suggested significant genetic differences among populations (Table 3).According to the analysis, only Seoul/Gyeonggi and Gangwon populations showed a relatively close relationship.The highest level of genetic differentiation was revealed in the pairs involving Chungcheong and this population was significantly differentiated from all the other populations with similar level of F ST value (0.034-0.054).There was higher differentiation between 1. Seoul/Gyeonggi and Gyeongsang, 2. Seoul/Gyeonggi and Jeolla, and 3. Gyeongsang and Jeolla.

DISCUSSION
Twelve novel microsatellite markers were developed in this study to investigate the genetic structure and diversity of Korean raccoon dogs from five sample locations.A F ST , Nei's genetic distance, and structure analysis implied that Seoul/Gyeonggi (north-western) and Gangwon (north-eastern) populations share genetic characters enough to suggest migration between them, resulting in similar genetic profiles (Fig. 1).However, both populations showed significant differentiation from Chungcheong (mid), Jeolla (south-western) and Gyeongsang (mid and south-estern) populations.Moreover, Seoul/ Gyeonggi showed higher differentiation from southern populations, Gyeongsang and Jeolla.Geographical distance is thought to be the main reason for this differentiation.Chungcheong population is located in the center of Korea and moderately differentiated from all the other populations, and its central location could lead to higher genetic diversity than in the other locations.
Although, Jeolla and Gyeongsang populations are located next to each other, they showed little evidence of gene exchange with high F ST and distance values (Table 3).The Sobaek Mountains can explain this big difference between two populations.The Sobaek Mountains split off from the Taebaek Mountains along the boundary between Gangwon and Gyeongsang, trending southwest across the center of the Korean Peninsula to Mt. Jiri, the highest peak (over 2,000 m) in the range (Fig. 1).They mark the geographical border between the Jeolla and Gyeongsang regions (Kim et al., 2009).Meanwhile, the Sobaek Mountains separate Gyeongsang from the Chungcheong and Gangwon areas as well.However, the lower altitude of the mountain range (under 1,500 m) and several man-made roads for trading (Park et al., 2003) might be the main causes preventing more pronounced differentiation between Gyeonsang and Chungcheong/ Gangwon.This theory is in accord with Fig. 2. Genetic composition of first four Gyeongsang individuals (Pop 4) from northern part of Gyeongsang is more similar to  Chungcheong/Gangwon.We found five individuals, two from Gangwon and three from Jeolla, respectively, whose origins were questionable (Fig. 2).Over 50% of genetic composition of two Gangwon and one Jeolla individuals was of Gyeongsang origin.According to the collection record, the sampling location of the Jeolla individual is close to the border of Gyeongsang province.Therefore, this individual might have migrated from Gyeongsang area.However, in the case of the two Gangwon individuals, unless they were transported by humans, it is hard to explain what affected their genetic profiles so greatly, and more intensive study will be necessary.The other two Jeolla individuals also showed over 50% of northern part (Seoul/ Gyeonggi and Gangwon) origin.They are sure from northern part of Jeolla though, specific local information is lacking.Then, we can assume they might come from Seoul/Gyeonggi or Gangwon province by transportation.Or southern part of Chungcheong near the Jeolla may be the origin of them.Since, Chungcheong has more admixed genetic composition as we mentioned earlier in Fig. 1.
Because our data indicate Korean raccoon dogs represent four distinct groups, we should take this into consideration when formulating plans to manage this species.Preventing the spread of rabies is a case in point.Up to now, the outbreak of rabies has been limited to Seoul/ Gyeonggi and Gangwon, and our data confirm that raccoon dogs from these areas essentially represent a single population.Moreover, we could suspect that shared genetic structure facilitated transfer of rabies within this population of raccoon dogs.
Therefore, we suggest that Seoul/Gyeonggi and Gangwon should be treated as a single management unit.Likewise, the other three populations should be treated as three different management units.Finally, to protect the other three populations from the spread of rabies for the sake of public health, genetic information and structure of the raccoon dog populations should be considered.
The microsatellite markers developed in this study are firstly described specifically for N. procyonoides koreensis.Even though a null allele was presented in Nyct 2, we used it for the analysis because a very low value of Brookfield frequency of heterozygote deficiency (< 0.2, Table 1) causes an unnecessary exclusion for analysis (Dakin and Avise, 2004).Finally, all 12 microsatellite markers can be successfully applied to population genetics studies.We confirmed their utility for application in other raccoon dog subspecies (data not shown).To develop conservation management strategies and to understand the evolutionary history of N. procyonoides, studies of phylogenetic relationships, genetic diversity and population genetics using individuals from more geographical locations are required.The markers developed here should facilitate future studies of population genetics of raccoon dog species in Korea, as well as populations in other areas.The genetic information gathered in these studies will contribute to establishment of management strategies for Korean raccoon dog species.

Fig. 1 .
Fig. 1.Map of study area and sampling information.A. Map of East Asia, study area is circled; B. map of South Korea, each colored area indicates each population region and each field circle shows each sampling locality.Individuals without specific local information were described with empty circle and pie charts to show proportion of inferred clusters (summary of the STRUCTURE clusters) within population.Green triangles: The Taebaek Mountains; Green diamonds: The Sobaek Mountains.

Table 1 .
Characteristics of the 12 microsatellite loci developed and optimized for Nyctereutes procyonoides (Representative Seoul/ Gyeonggi population, n = 22).Locus name, primer sequences, repeat motif, size range of PCR products, GenBank accession numbers, polymorphism, information of null alleles and neutrality are reported a Based on the expected heterozygosity (H E, shown in Table

Table 2 .
Descriptive statistics for 12 microsatellite loci from samples of Korean raccoon dogs across five locations: Seoul/Gyeonggi, Gangwon, Chungcheong, Gyeongsang and Jeolla provinces

Table 3 .
Pairwise F ST (below diagonal) and Nei's D A genetic distance (above diagonal) between Korean raccoon dogs sampled from five locations *Significant after pairwise F ST test, with pairwise comparisons with p < 0.005.Population genetic study of Korean raccoon dog