The fleshy shrimp, Fenneropenaeus chinensis , is the family of Penaeidae and one of the most economically important marine culture species in Korea. However, its genetic characteristics have never been studied. In this study, a total of 240 wild F. chinensis individuals were collected from four locations as follows: Narodo (NRD, n = 60), Beopseongpo (BSP, n = 60), Chaesukpo (CSP, n = 60), and Cheonsuman (CSM, n = 60). Genetic variability and the relationships among four wild F. chinensis populations were analyzed using 13 newly developed microsatellite loci. Relatively high levels of genetic variability (mean allelic richness = 16.87; mean heterozygosity = 0.845) were found among localities. Among the 52 population loci, 13 showed significant deviation from the Hardy–Weinberg equilibrium. Neighbor-joining, principal coordinate, and molecular variance analyses revealed the presence of three subpopulations (NRD, CSM, BSP and CSP), which was consistent with clustering based on genetic distance. The mean observed heterozygosity values of the NRD, CSM, BSP, and CSP populations were 0.724, 0.821, 0.814, and 0.785 over all loci, respectively. These genetic variability and differentiation results of the four wild populations can be applied for future genetic improvement using selective breeding and to design suitable management guidelines for Korean F. chinensis culture. The fleshy shrimp, Fenneropenaeus chinensis , is an economically important species in the family Penaeidae. F. chinensis is distributed from the west coast of the Korean Peninsula to the east coast of northern China [3] and is characterized by long-distance migrations of up to 2,000 km between spawning and feeding sites [18] . The F. chinensis South Korean aquaculture industry started in 1960, and production increased very quickly to reach 1,533 metric tons in 1997. However, a white spot syndrome virus (WSSV) disease outbreak decreased production dramatically to about 998 metric tons in 1998 [13] . Disease outbreaks cause mass mortality among cultured F. chinensis worldwide, particularly in Asian countries. To solve this problem, F. chinensis was replaced with Litopenaeus vannamei as the major cultured species in South Korea in 2003 [11] , but damage to shrimp farming associated with WSSV has been increasing in recent years. The West Sea Mariculture Research Center (Taean, Korea) conducts a breeding program to produce shrimp strains that are more resistant to WSSV. If a large number of cultured shrimp are released from aquaculture facilities, they could alter the genetic composition of wild populations by either displacing them or interbreeding. This could reduce the population’s ability to adapt to new environments. Accordingly, basic knowledge about the geographic distribution, genetic diversity, and population differences of F. chinensis is important. It is now widely recognized that this information can be obtained through recently developed molecular genetics techniques. However, few reports have been published about F. chinensis population genetics in South Korea. Because microsatellite markers and simple sequence repeats have high levels of polymorphism, co-dominant inheritance, genome-wide distributions, and high reproducibility, they have been applied widely in population genetics, genetic linkage map, genetic diversity, and phylogenetic studies [10 , 16] . However, only a few microsatellite markers can be used for each shrimp species because of poor conservation of the microsatellite flanking sequences among different crustacean species, particularly in F. chinensis [15 , 27] . Several microsatellite markers have been isolated from F. chinensis [5 , 9 , 24] . Unfortunately, despite the commercial importance of this shrimp species in Korea, studies describing its genetic background are scarce. In the present study, we assessed genetic diversity within and among wild Korean F. chinensis populations and examined the genetic structure among these populations using new microsatellite DNA markers. A total of 240 wild F. chinensis individuals were collected from four locations as follows: Narodo (NRD, n=60), Beopseongpo (BSP, n=60), Chaesukpo (CSP, n=60), and Cheonsuman (CSM, n=60) ( Fig. 1 and Table 1 ). All muscle tissue samples were stored in 100% ethanol prior to DNA extraction. The tissue was homogenized in lysis buffer (MFX-2000; Toyobo, Osaka, Japan) containing 20 mg/ml proteinase K. Total DNA was isolated using the MagExtractor MFX-6100 automated DNA extraction system (Toyobo Co., Tokyo, Japan). The extracted genomic DNA was quantified using a Nanodrop ND-1000 spectrophotometer (Thermo Fisher Scientific, Barrington, IL, USA) and stored at −20℃ until use. Thirteen microsatellite loci that were amplified in F. chinensis were used in this study. Detailed information about the primers is presented in Table 2 . Each PCR contained three primer sets that were labeled at the 5‘ end of the forward primer with 6-FAM, HEX, or NED dyes (Applied Biosystems, Foster City, CA, USA). PCR amplification was carried out in a 10-μL reaction mixture containing 0.25 U Ex-Taq DNA polymerase (TaKaRa Biomedical Inc., Shiga, Japan), 1× PCR buffer, 0.2 mM dNTP mix, 10 pM of each primer, and 100 ng template DNA, using the PTC 200 DNA Engine (MJ Research, Waltham, MA, USA). The PCR conditions were: 5 min at 95℃, followed by 35 cycles of 30 sec at 95℃, 45 sec at 58℃, and 45 sec at 72℃, with a final extension of 10 min at 72°C. Microsatellite polymorphisms were screened using an ABI PRISM 3130 XL automated DNA sequencer (Applied Biosystems), and alleles were designated according to PCR product size relative to molecular size markers (GENESCAN 400 HD [ROX]; Applied Biosystems). The number of samples ( n ), expected heterozygosity ( He ), and observed heterozygosity ( Ho ) were calculated using the Arlequin software package (ver. 3.0; [8] ). Tests for allelic richness ( A_R ), number of alleles per locus ( Na ), and deviations from Hardy–Weinberg equilibrium (HWE) were estimated using GENEPOP ver. 4.0 (http://kimura.univmontp2. fr/~rousset/Genepop.htm), and the adjusted P-values for both analyses were obtained using a sequential Bonferroni test for multiple comparisons. MICRO-CHECKER 2.2.3 was used to test for the presence of null alleles. F_ST values (1,000 permutations; [25] ) were calculated using Arlequin 3.0. The population structure patterns were further investigated using a model-based Bayesian clustering procedure in STRUCTURE (ver. 2.3; [21] ), which assigns individuals to K populations based on their multilocus genotype. STRUCTURE was run for K = 2 using a burn-in length of 50,000 and a run of 50,000 steps. The overall inbreeding coefficient (F _IS ; [25] ) was also estimated using GENEPOP ver. 4.0. Analysis of molecular variance (AMOVA; [7] ) was used to test for population structure with Arlequin 3.0. A principal coordinate analysis (PCoA) was performed using GENALEX (ver. 6.0; [19] ). A neighbor-joining (NJ) tree was constructed based on the chord distances (D _CE ; [2] ) to evaluate the genetic population relationships using the POPULATION program (ver. 1.2.30). Bootstrap values were calculated using 1,000 replicates. Microsatellite markers have been used in various shrimp species, and the pedigree of mixed populations can be determined using a few microsatellite markers [14] . However, little information is available on the use of microsatellite markers in F. chinensis [27] for understanding population structure or genetic diversity [15] . In this study, F. chinensis , a commercially and recreationally valuable species, were collected from the west coast of the Korean Peninsula (four separate collections in 2006) to illuminate its population structure and genetic diversity. The genetic variability indices for the four populations are summarized in Table 3 . Allelic richness per locus ( A_R ) ranged from 3 to 31 across all populations. Mean A_R values were in the following order: NRD (16.2), CSM (17.5), BSP (16.8), and CSP (16.9). The total number of alleles per locus ( N_a ) ranged from 3 to 31. KFcg959 had the highest A_R and N_a values in the NRD population. Mean N_a ranged from 16.2 to 17.5. The mean expected and observed heterozygosities ranged from 0.835 to 0.861 and from 0.724 to 0.821, respectively. Mean heterozygosity was highest in CSM ( H_e = 0.861, H_o = 0.821), followed by BSP ( H_e = 0.843, H_o = 0.814), CSP ( H_e = 0.839, H_o = 0.785), and NRD ( H_e = 0.835, H_o = 0.724). The mean He value was higher than the mean H_o value in all populations. Among the 52 population-locus cases (4 populations × 13 loci), 13 cases showed significant deviations ( p <0.01). All populations departed from HWE at the KFcg614 locus. The NRD population departed from HWE at 6 of the 13 microsatellite loci, and the CSM, BSP, and CSP populations departed from HWE at three, two, and two microsatellite loci, respectively. The F _IS value estimated for all populations at the KFcg614 locus was significantly different from zero ( p <0.05); all populations were in an excess heterozygosity condition for the KFc673 and KFc959 loci in NRD, the KFc42, KFc438, KFc657, and KFc959 loci in CSM; the KFc37, KFc42, KFc438, KFc568, and KFc658 loci in BSP; and the KFc37 and KFc658 loci in CSP. Our results reveal that the newly developed microsatellite markers were a powerful approach to monitoring genetic diversity between the four geographically different wild Korean F. chinensis populations investigated. Heterozygosity is an important measure of population diversity at the genetic level. The mean observed heterozygosity values of the NRD, CSM, BSP, and CSP populations were 0.724, 0.821, 0.814, and 0.785 over all loci, respectively. These values were lower than the mean expected heterozygosity for the four populations. The overall heterozygosity in this study was high and differed from the reported mean heterozygosity, which is known to be low (7.3%), within crustacean populations as a whole [12] . In the present study, 18 of the 52 population-locus cases deviated significantly from HWE after applying the Bonferroni correction. However, the KFc37, KFc41, and KFc42 loci in the NRD; the KFc614 locus in the CSM; and the KFc614 locus in the CSP population also deviated from HWE without excess heterozygosity. These deviations may have been caused by selection, population mixing, nonrandom mating, presence of null alleles, or the limited sample size used in our analysis [20 , 22] . The presence of null alleles, for example, is a classical source affecting the accuracy of microsatellite loci during parentage assignment [1] and null allele frequencies > 5% can compromise pedigree estimates [17] . F_ST is the proportion of total genetic diversity that separates groups, and values range from 0 to 1. If there is no population substructure (i.e., no stable groups), F_ST will approach 0. An F_ST range of 0.00-0.05 indicates little genetic differentiation [26] . Significant pairwise F_ST values ( p <0.05) were observed between the NRD and CSM, NRD and BSP, NRD and CSP, CSM and BSP, and CSM and CSP populations ( Table 4 ). In our study, pairwise F_ST tests detected low levels of genetic differentiation among the populations, particularly between the CSP and BSP populations, indicating that a geographical barrier was not effectively maintaining genetic integrity among the populations in the four areas. Genetic distances (D _CE ) were also calculated for all possible population pairs. The D _CE measure ranged from 0.0380 to 0.3582. The smallest estimate for D _CE was between BSP and CSP (0.0380), whereas the highest estimate was between NRD and CSP (0.3582); the genetic distances between NRD and CSM, NRD and BSP, CSM and BSP, and CSM and CSP were 0.1864, 0.3030, 0.0991, and 0.1216, respectively ( Table 4 ). The result of the genetic distances among the F. chinensis populations was further confirmed by the findings of PCoA ( Fig. 2 ). The NJ tree constructed based on D _CE , indicated that the four populations were allocated into three major groups ( Fig. 3 ); that is, one group included the CSM population, one group included the BSP and CSP populations, and one group included the NRD population. The PCoA produced a result similar to that of the cluster analysis ( Fig. 3 ) and showed clear separation of the four populations into three clusters. The results of hierarchical AMOVA tests to form putative groups estimated from each NJ tree topology provided additional evidence to support previous findings in F. chinensis populations. The NJ tree topology included a group for the NRD population and another group for the CSM, BSP, and CSP populations. In this case, the fixation index among groups ( F_CT ) was not significant ( F_CT =0.0147, P=0.2494), whereas it was significant among populations within groups ( F_SC =0.0099, P=0.0000). The NJ tree topology defined a group that included the NRD, BSP, and CSP populations and another group that included the CSM population. In this case, the fixation index among groups ( F_CT ) was not significant ( F_CT =0.0010, P=0.4943), whereas it was significant among populations within groups (F _SC =0.0167, P=0.0000). These results suggest that the populations remained structured within at least one group in each test case, as the F_SC values estimated in both test cases were significant. Therefore, we separated them into three groups. The first group included the NRD population, the second included the CSM population, and the third included the BSP and CSP populations. These results show that the fixation index among groups ( F_CT =0.0220, P=0.1643), and among populations within groups (F _SC =−0.0012, P=0.8603), were not significant ( Table 5 ). The NJ tree topology and AMOVA results suggest that the four populations could be assigned to a group that included the NRD population, a group that included the CSM population, and a group that included the BSP and CSP populations ( Table 5 ). An evaluation of the evolutionary relationships among the four wild populations showed that BSP and CSP had the highest degree of genetic identity, followed by CSM and BSP, CSM and CSP, NRD and CSM, and NRD and BSP, whereas NRD and CSP were most distantly related. A previous study demonstrated the most likely explanation for this result is that wild populations of Chinese F. chinensis , which are extensively distributed in the Yellow and Bohai Seas, are two independent populations [15] . These two populations share the same wintering ground that lies in the mid-depth waters of the Yellow Sea [4] . An independent population of F. chinensis has been found off the west and south coasts of the Korean Peninsula [18] ; thus, further attempts are necessary to evaluate the genetic population relationships in this species. The goal of this study was to examine the importance of conservation and further genetic improvements in wild Korean F. chinensis . Information on genetic variation and differentiation in these four wild populations can be applied for future genetic improvement thorough selective breeding and to design suitable management guidelines for Korean F. chinensis .