Genetic Diversity and Population Structure of ...

Genetic Diversity and Population Structure of Seedling Populations of Pyrus pashia

Yu Zong, Ping Sun, Jing Liu, Xiaoyan Yue, Kunming Li & Yuanwen Teng

Plant Molecular Biology Reporter ISSN 0735-9640 Plant Mol Biol Rep DOI 10.1007/s11105-013-0680-2

1 23

Your article is protected by copyright and all rights are held exclusively by Springer Science +Business Media New York. This e-offprint is for personal use only and shall not be selfarchived in electronic repositories. If you wish to self-archive your article, please use the accepted manuscript version for posting on your own website. You may further deposit the accepted manuscript version in any repository, provided it is only made publicly available 12 months after official publication or later and provided acknowledgement is given to the original source of publication and a link is inserted to the published article on Springer's website. The link must be accompanied by the following text: "The final publication is available at link.springer.com”.

1 23

Author's personal copy Plant Mol Biol Rep DOI 10.1007/s11105-013-0680-2

ORIGINAL PAPER

Genetic Diversity and Population Structure of Seedling Populations of Pyrus pashia Yu Zong & Ping Sun & Jing Liu & Xiaoyan Yue & Kunming Li & Yuanwen Teng

# Springer Science+Business Media New York 2013

Abstract Pyrus pashia, described as an intermediate species between oriental and occidental pear groups, is one of the most important wild pears. This study used microsatellite markers at 14 loci to assay genetic diversity and differentiation within P. pashia using three complementary methods. Four hundred and seventy seedlings were obtained from 38 half-sib families from four sites in the central Yunnan Province of China. These 14 loci displayed high polymorphism, and the descriptive statistics of diversity varied significantly among seedling populations. One hundred and seventy-three different alleles were detected, with an average of 12.4 alleles per locus. The overall expected and observed heterozygosity values were 0.749 and 0.643, respectively. Allelic richness at the different sites ranged from 1.00 to 20.36, and the Shannon’s information index for each locus was from 0.35 to 2.37, with a mean value of 1.82. Genetic differentiation was detected at both family and site levels using Bayesian model and neighbor-joining clustering approaches and the results compared with that of principal coordinate analysis. Two clusters, each with a similar number of families, were detected in the data set. Analysis of molecular variation indicated that the major variation occurred within families and that the minimum partitions of genetic variation exist among families, representing 89.14 and 10.86 % of the total variety, Electronic supplementary material The online version of this article (doi:10.1007/s11105-013-0680-2) contains supplementary material, which is available to authorized users. Y. Zong : P. Sun : J. Liu : X. Yue : Y. Teng (*) Department of Horticulture, the State Agricultural Ministry Key Laboratory of Horticultural Plant Growth, Development of Quality Improvement, Zhejiang University, Hangzhou, Zhejiang Province 310058, China e-mail: [email protected] K. Li Institute of Horticulture, Yunnan Academy of Agricultural Sciences, Kunming City, Yunnan Province 650205, China

respectively. Families derived from site 2 displayed the maximum allelic richness and had a greatly mixed genetic composition. We recommend that these, especially families 9, 11, and 12, should be the focus of future preservation and usage investigations. Keywords Genetic diversity . Seedling . Population structure . Pyrus pashia

Introduction Pyrus pashia D. Don (Rosaceae) is naturally distributed throughout southwestern China (Yu et al. 1986) and the Himalayan region and has considerable morphological diversity (Krause et al. 2007). P. pashia is considered an intermediate species between oriental and occidental pear groups and may have played an important role in the evolution of the Pyrus genus (Rubtsov 1944; Challice and Westwood 1973). As P. pashia shows extensive adaptation to the environment, it is widely used as a pear rootstock in Southwest China (Yu et al. 1986). Therefore, the conservation of P. pashia is of significant importance for both commercial use and research. To conserve germplasm resources of P. pashia efficiently, it is necessary to evaluate its genetic diversity. To date, few studies of the genetic diversity and population structure of P. pashia have been reported, and this strongly hampers the conservation, management, and use of the species. Wild populations of fruit trees are often affected by humans through overexploitation and fruit choice; this results in a loss of genetic diversity and structure. For these reasons, the conservation of wild resources has received increasing attention over the last decades (Jolivet et al. 2011). Transplanting wild fruit trees into a germplasm repository, or through shoot grafting, is conducive to qualitative and quantitative trait assessment, conservation, and usage, but is

Author's personal copy Plant Mol Biol Rep

labor-intensive, costly, not always convenient (especially when collection locations are large distances from conservation sites), and also comes with an inherent risk of mortality (Richards et al. 2009). Consequently, regeneration of wild fruit trees from seed has become more crucial in comparison to other methods. The requirement to ensure a high genetic diversity, which could serve as a representation of the entire genetic diversity of the population, is a major obstacle for seed collection. The constraints of collection and management of wild germplasm have led to the development of strategies for germplasm evaluation. The assessment of genetic relationships amongst ex situ conserved populations will help breeders identify the genetic relatedness between the seedling and original population (Van Treuren et al. 2001; Potts et al. 2012). This allows the development of strategies for the management and exploitation of resources required to promote successful fruit breeding. Hence, knowledge of genetic diversity and population structure within the seedling population, after seed collection, is important for the successful use of resources (Jolivet et al. 2011). Similar studies have been carried out on other types of fruit tree. Richards et al. (2009) evaluated the genetic diversity of seedling individuals of Malus sieversii sampled from disparate locations in Kazakhstan, Uzbekistan, and Tajikistan using simple sequence repeat (SSR) markers to identify distinct genetic diversity and differentiations among populations. However, similar kinds of study on pear are limited. During the last decade, studies involving pear have concentrated on cultivar identification and genetic relationship analysis using distinct molecular markers (Teng et al. 2001, 2002; Bao et al. 2007, 2008; Yao et al. 2010). Recently, attempts have also been made to understand the genetic diversity of wild pears. Katayama et al. (2007) estimated the genetic diversity of Pyrus ussuriensis var. aromatica using SSR and chloroplast DNA markers. Wuyun et al. (2013) studied the genetic diversity of 186 accessions of wild P. ussuriensis belonging to 12 populations. Liu et al. (2012, 2013) determined the genetic diversity and population structure of Pyrus calleryana and P. pashia using cpDNA analyses. These reports focused on materials sampled from wild populations and will be of help for the in situ conservation of these species. However, when seeds are used to conserve the genetic diversity of the wild species population, estimations of genetic diversity and population structure are also required. Table 1 Location information for the four P. pashia collection sites in Yunnan Province, China

In this study, 14 pairs of SSR primers developed from apple (Malus spp.) and pear were used to detect the genetic diversity and population structure of 470 individuals from 38 half-sib families of P. pashia. The results of our study will provide essential information for the collection, conservation, and usage of P. pashia.

Materials and Methods Plant Materials We investigated four wild P. pashia populations from the central Yunnan Province of China in 2010 (Table 1 and Fig. 1). Sixty-two trees within the four populations were randomly selected to be maternal trees for seed collection. The sampling distance between adjacent trees in the same population was at least 20 m to avoid identical genotypes. The position of each tree was fixed using a portable Global Positioning System. One hundred fruits from each maternal tree were collected and brought to the lab in October. Seeds were removed from the fruits and rinsed with water. Next, the seeds were stratified in damp sand for 60 days at 4 °C and sowed in plug trays in March 2011. Seedlings were grown under greenhouse conditions. Ninety days after germination, plantlets were transplanted to the field. Seedlings from the same maternal tree were treated as a family. In September, newly expanded leaves were collected from 470 robust seedling trees and stored at −80 °C until required. DNA Extraction Genomic DNA was extracted from the leaf tissue of 470 P. pashia individuals from 38 families, following the modified CTAB protocol described by Doyle and Doyle (1987), and the DNA concentration was diluted to 10–30 ng μl−1 after assessing the quality and quantity on 1 % (w/v) agarose gels using standard DNA markers (TaKaRa, Dalian, China). PCR Amplification and SSR Analyses We used 14 previously developed polymorphic SSR primer pairs (Gianfranceschi et al. 1998; Yamamoto et al. 2002a, b; Liebhard et al. 2002; Yao et al. 2010). PCR amplifications were carried out in a total volume of 15 μl (10 ng DNA

Site

Latitude (° N)

Longitude (° E)

Elevation (m)

No. of maternal trees

Location

1 2 3 4

24.97 25.00 25.27 25.25

102.30 102.34 103.33 102.74

1,887 1,839 2,032 2,152

15 15 17 15

Anning County Anning County Malong County Kunming City

Author's personal copy Plant Mol Biol Rep Fig. 1 Geographic locations of wild populations of P. pashia. Solid circles show the location of the sampled populations; the size of solid circle correlates roughly with the number of individuals from seedling populations. Dashed circles show other wild populations surrounding the sampled populations. A, B, and C represent the three source populations of P. pashia for range expansion during interglacial periods (modified from Liu et al. 2013)

template, 0.4 μM of each primer, 200 μM dNTPs, 2 mM MgCl 2 , and 0.5 U Taq DNA polymerase; TaKaRa). Amplifications were performed in a Mastercycler Gradient PCR Thermocycler (Eppendorf, Hamburg, Germany) programmed for an initial denaturation at 94 °C for 2 min and 30 s, followed by four cycles of 94 °C for 30 s, 65 °C for 1 min, and 72 °C for 1 min (the annealing temperature was reduced by 1 °C per cycle); the preliminary cycles were followed by 30 cycles of 94 °C for 30 s, 65 °C for 1 min, and 72 °C for 1 min for primers CH02B10 and CH01B12. Amplifications using primers KA14 and KU10 were carried out under the following conditions: initial denaturation at 94 °C for 2 min, followed by 10 cycles of 94 °C for 1 min, 60 °C for 1 min, and 72 °C for 2 min (with the annealing temperature reduced by 0.5 °C per cycle); this was followed by 25 cycles of 94 °C for 1 min, 55 °C for 1 min, and 72 °C for 2 min. A final 5-min 72 °C extension was used for all of the four primers above. Reactions involving the other primers were performed as follows: 5 min at 94 °C, then 35 cycles of 40 s at 94, 55, and 72 °C, respectively, followed by a 6-min extension at 72 °C. Amplicons were pooled together with an internal size standard (GeneScan™ 500 LIZ, Applied Biosystems, USA) according to different dye color and expected fragment sizes and subsequently separated and sequenced using an ABI 3700XL Genetic Analyzer (Applied Biosystems). Genotyping data were analyzed using GeneMapper 4.0 software (Applied Biosystems).

Data Analyses Genotyping errors in the SSR profiles, including null alleles, short allele dominance, and scoring errors caused by stuttering, were checked using MICRO-CHECKER software (Van Oosterhout et al. 2004). Then, a set of measures, from different collection sites and families, were estimated using GenAlEx 6.4 (Peakall et al. 2006). The aspects measured included the number of different alleles (N a), number of effective alleles (N e), observed heterozygosity (H o), and expected heterozygosity (H e). The fixation index (F is) and genetic differences (F st, pairwise and overall) for groups (different sites, families, and clusters) were also computed using GenAlEx 6.4. A sample-adjusted metric of allelic richness (A r) was calculated using FSTAT 2.9.3 (Goudet et al. 2001) for group comparisons. Genetic differences among and within sites or families were further estimated with analyses of molecular variance (AMOVA) using ARLEQUIN 3.5 (Excoffier and Lischer 2010). The identification of the genetically homogeneous group in a data set is always challenging and is affected by several factors including the number of loci used, the magnitude and scale of gene flow, the variation at each locus, and the number of samples (Evanno et al. 2005). Therefore, we used three complementary methods to estimate the number of clusters at both non-hierarchical and hierarchical levels. Non-hierarchical genotypic clustering was performed using the genotypes obtained from all 470 individuals


independent of sites or families. The population structure was evaluated using a Bayesian approach and executed using the software STRUCTURE 2.3.3 (Pritchard et al. 2000). This revealed the genetic structure by assigning individuals or predefined groups to K clusters. In the current study, we chose the admixture model and correlated allele frequencies without using sites or families of individuals as priors as described by Richards et al. (2009). Different K values ranging from 1 to 10 were adopted to infer the number of clusters for 10 replicate runs, with 100,000 iterations burn-in period followed by 100, 000 iterations Markov chain Monte Carlo. The results were uploaded to the STRUCTURE HARVESTER web site (Earl and von Holdt 2012) to estimate the most appropriate K value. Replicate cluster analyses of the same data could result in several distinct outcomes for estimated assignment coefficients, even though the same starting condition was used. Therefore, we employed CLUMPP software (Jakobsson and Rosenberg 2007) to average the 10 independent simulations and illustrated the result graphically using DISTRUCT (Rosenberg 2004). A neighbor-joining tree of 38 families based on Nei’s genetic distance, D A (Nei et al. 1983), was constructed using POPULATION 1.2 (Langella 2000). The distance measure D A was applied as this has previously been shown to give reliable phylogenetic trees when analyzing microsatellite data (Takezaki and Nei 2008). Trees were presented and modified using DENDROSCOPE 3 (Huson and Scornavacca 2012). Principal coordinate analysis (PCA) was used to further confirm cluster analysis results; this was performed using GenAlEx 6.4. PCA is a multivariate technique that depicts relationships among studied genotypes. It allows researchers to identify and plot the major axes of variation within a low dimensional graph. The process assigned total variation to several coordinates; the first two or three coordinates will indicate most of the separation among groups.

Results

Table 2 Diversity statistics of the 14 SSR P. pashia loci Locus

Na

Ne

I

He

Ho

CH02B10a CH01B12a KA14b KU10b MES2c MES7c MES17c MES108c MES122c MES138c NB109d NH019d NH023d NH026d Mean

5.0 18.0 12.3 20.5 11.8 15.0 2.3 12.0 7.8 14.3 11.5 15.8 13.5 13.8 12.4

2.2 6.8 5.0 8.4 6.2 8.1 1.3 5.3 3.8 7.4 4.9 7.9 3.4 5.7 5.5

0.960 2.239 1.902 2.369 2.024 2.266 0.349 1.909 1.489 2.205 1.797 2.325 1.686 1.958 1.820

0.524 0.853 0.790 0.880 0.841 0.873 0.188 0.812 0.716 0.861 0.771 0.871 0.691 0.821 0.749

0.302 0.643 0.807 0.657 0.824 0.786 0.127 0.815 0.589 0.796 0.675 0.840 0.316 0.828 0.643

N a number of different alleles, N e number of effective alleles, I Shannon’s information index, H o observed heterozygosity, H e expected heterozygosity a

Gianfranceschi et al. (1998) and Liebhard et al. (2002)

b

Yamamoto et al. (2002a)

c

Yao et al. (2010)

d

Yamamoto et al. (2002b)

of locus MES17) suggested that the collection of genotypes was considerably diverse. Genetic Differentiation The optimum number of clusters (K) was estimated according to the procedure given by Evanno et al. (2005), and we identified a clear maxima for DeltaK at K = 2 (Fig. 2). Therefore, the preliminary Bayesian modeling approach result revealed that two clusters (red and green) captured the major split in 38 families. A genotype was assigned to the cluster in

Genetic Diversity The 14 SSR loci showed high polymorphism, and the number of identified alleles for each locus ranged from 2.3 (MES17) to 20.5 (KU10). One hundred and seventy-three different alleles were detected in 470 individuals, with an average of 12.4 alleles per locus. The number of effective alleles ranged from 1.3 (MES17) to 8.4 (KU10). Shannon’s information index (I) for each locus ranged from 0.349 to 2.369, with a mean of 1.820. The overall mean values of expected and observed heterozygosity were 0.749 and 0.643, respectively (Table 2). Heterozygosity deficiency was detected at loci CH02B10, CH01B12, KU10, and NH023. The high value of expected and observed heterozygosity (with the exception

Fig. 2 Modeling of cluster number for P. pashia using STRUCTURE. LnP(K) and DeltaK were calculated in accordance with the method of Evanno et al. (2005)


which it owned the highest membership coefficient (Richards et al. 2009). The bar plot revealed that the green cluster included 23 families from all of the four collection sites. Families from sites 3 and 4 were completely assigned to the green cluster. The seedling trees from sites 1 and 2 generally scattered into the red cluster; however, families 9, 11, and 12, collected from site 2, had highly mixed clusters (Fig. 3). The use of a neighbor-joining (NJ) tree enabled the 38 families to be divided into two major branches. These contained 18 and 20 members, respectively (Fig. 4). The tree showed a similar divisive pattern with that displayed by the Bayesian method (Fig. 3). The main difference seen between these two methods was the clustering of families 9, 11, and 12. The NJ tree result suggested that there was a close relationship among families 9, 10, 11, 12, and 13. However, these were sorted into distinct clusters. Families 9, 11, and 12 were grouped into the green cluster according to the modeling approach; families 10 and 13 were grouped into the red one (Fig. 3). These incongruencies further supported the notion that families 9, 10, 11, 12, and 13 had a highly mixed genetic composition. PCA was performed to further confirm the results of the Bayesian modeling and neighbor-joining tree clustering experiments. The initial two axes captured 52.55 % of all the variation. Two groups were detected; the first group was composed of 13 families from sites 1 and 2, whereas the second group included the remaining families (Fig. 5). The PCA results displayed a few differences with the results implied by STRUCTURE when using the Bayesian approach. In the PCA analysis, families 8 and 10 were grouped together with families that originated from sites 3 and 4. However, the STRUCTURE analysis suggested that they had a closer relationship with families that came from sites 1 and 2. Genetic Variation Among Sites and Clusters The two gene pools were defined by several complementary clustering methods, and moderate incongruencies between the different methods did not influence the overall clustering. The inbreeding coefficient among families ranged from −0.125 (locus KA14) to 0.438 (locus NH023), with a mean value of 0.04. Genetic differentiation among the clusters showed slightly higher F is than seen among families; it ranged from

Fig. 4 Neighbor-joining tree of the 38 families of P. pashia based on Nei’s genetic distance (Nei et al. 1983). The branches were colored according to the Bayesian assignment method. Numbers united with the branches show bootstrap values higher than 50 % based on 1,000 replications

0.008 (locus MES108) to 0.534 (locus NH023), with a mean value of 0.171 (Table 3). The mean F st among sites was 0.038. The coefficient of genetic differentiation among families ranged from 0.118 to 0.191, with a mean value of 0.154. The F st among the clusters for each locus ranged from 0.009 to 0.052, with a mean value of 0.022; this indicated a moderate difference from the average F st among the sites (Table 3). The pairwise F st values ranged from 0.013 (between sites 1 and 2) to 0.044 (between sites 1 and 3; data not shown). Site-specific diversity observations using the 14 loci revealed that allelic richness varied significantly among the four sites. A r ranged from 1.00 (locus MES17, site 3) to 20.36 (locus KU10, site 4). Site 4 showed abundant diverse genetic variation and had the largest number of private alleles, whereas site 3 had the least genetic variation, having no private alleles (Table 4). A r at the cluster level suggested that there was moderate divergence between the red and green clusters. Of the 14 loci, CH02B10, MES17, and MES122 presented severe deficiency of allelic richness across the sites. AMOVA suggested that the main variation occurred within families (Electronic Supplementary Material Table S1); this accounted for 89.14 % of all genetic variation.

Fig. 3 Genetic relationship among the 38 families of P. pashia revealed by a Bayesian modeling approach under K =2. Different families are split by vertical black lines. Sites 1–4 represent the collection sites

Author's personal copy Plant Mol Biol Rep Table 4 Allelic richness (A r) of sites and clusters in P. pashia loci Locus

Fig. 5 Principal coordinate analysis of the 38 families of P. pashia based on the genetic distance of families

Discussion In this study, 14 pairs of microsatellite primers were used to analyze the genetic diversity and population structure of seedling populations of P. pashia. Thirteen loci displayed a high degree of diversity with 5.0–20.5 alleles per locus, whereas locus MES17 displayed low polymorphism. Six EST-SSR markers with MES as prefix were developed from the Malus EST database (Yao et al. 2010); the other eight genomic SSR markers were developed from apple and pear. As the EST-derived microsatellites are within transcribed regions of DNA, they might be expected to be more conserved and less polymorphic than genomic SSR (Yao et al. 2010). Heterozygotes were significantly lower than expected for loci CH02B10, CH01B12, and NH023; this indicated the presence of null alleles or P. pashia inbreeding at the collection sites (Van Oosterhout et al. 2004). The mean gene diversity, H e, of Table 3 Genetic differentiation among and within different groups of P. pashia

F is inbreeding coefficient of families (or clusters), F it inbreeding coefficient in the total sample, F st genetic differentiation among sites, clusters (identified through Bayesian approach), and families within sites

Locus

CH01B12 CH02B10 KA14 KU10 MES2 MES7 MES17 MES108 MES122 MES138 NB109 NH019 NH023 NH026 Mean

Sites

Clusters

1

2

3

4

Red

Green

CH02B10 CH01B12 KA14 KU10 MES2 MES7 MES17 MES108 MES122 MES138 NB109 NH019 NH023 NH026

4.46 13.30 11.83 16.64 10.65 14.71 3.00 9.91 6.00 11.29 7.83 14.54 12.31 11.76

4.94 16.73 10.82 15.79 10.79 14.30 3.00 12.34 7.76 15.20 10.68 15.84 15.07 13.36

3.98 13.94 7.98 10.00 9.00 9.96 1.00 7.94 5.00 10.96 9.96 11.98 6.96 8.00

4.19 15.45 13.53 20.36 11.68 13.84 1.97 11.89 7.38 14.28 12.18 14.95 11.61 11.48

6.96 19.80 14.89 24.75 13.88 17.96 3.00 14.00 8.88 16.00 13.88 16.96 15.00 18.86

4.99 24.41 14.91 26.63 14.41 16.60 2.00 14.66 10.04 16.63 15.41 17.95 19.79 16.49

Mean Private alleles

10.59 1

11.90 1

8.33 0

11.77 2

14.63 2

15.35 3

Allelic richness is based on a minimum sample size of 48 diploids for the sites and 192 diploids for the clusters. Private alleles are the number of alleles unique to a single population

the 470 individuals was 0.749; this was similar to that (0.770) reported by Cao et al. (2012) in their analysis of 32 wild P. ussuriensis accessions and higher than that (0.653 and 0.745) estimated by Volk et al. (2006) and Liu et al. (2012), who studied 145 Pyrus communis individuals and 77 P. calleryana accessions, respectively. The high heterozygote

Among families

Among clusters

F st

F is

F it

F is

F it

Families

Sites

Cluster

0.310 0.146 −0.125 0.163 −0.118 −0.093 0.167 −0.102 −0.005 −0.024 −0.001 −0.073 0.438 −0.122 0.040

0.442 0.263 0.041 0.288 0.057 0.079 0.326 0.045 0.156 0.112 0.158 0.066 0.546 0.010 0.185

0.403 0.275 0.009 0.279 0.029 0.058 0.307 0.008 0.147 0.104 0.138 0.066 0.534 0.042 0.171

0.427 0.287 0.047 0.286 0.048 0.072 0.343 0.038 0.155 0.125 0.157 0.077 0.540 0.050 0.189

0.190 0.137 0.148 0.150 0.157 0.157 0.191 0.133 0.160 0.133 0.159 0.130 0.192 0.118 0.154

0.039 0.027 0.063 0.026 0.041 0.030 0.076 0.038 0.031 0.039 0.037 0.033 0.025 0.020 0.038

0.040 0.016 0.038 0.009 0.020 0.015 0.052 0.030 0.009 0.023 0.023 0.012 0.012 0.009 0.022


levels observed in this study may be attributed to the following: individuals were sampled from descendent populations; seeds collected from maternal trees were developed through open pollination; and more individuals were used in our studies than in previous ones. According to the study of Yamamoto et al. (2002a), alleles are frequently observed at loci KA14 and KU10 in Asian pears (Pyrus pyrifolia , P. ussuriensis , and P. calleryana ), but rarely detected in P. communis (European pear). Our results revealed that alleles were almost totally observed at the KA14 and KU10 loci in P. pashia. This result indicated that P. pashia may be genetically close to Asian pear species, especially P. calleryana (Zheng et al. 2011; Liu et al. 2012), Pyrus dimorphophylla, P. pyrifolia, and/or P.×serrulata, as indicated by Zheng et al. (2011). As the materials and regions covered in this study were limited, it is not possible to directly conclude from our data that P. pashia is a transitional species between occidental and oriental pears. Bayesian clustering analysis has been proven to be an efficient method to evaluate the genetic structure of fruit tree populations such as apple (Zhang et al. 2012), pear (Ferreira et al. 2011), grapevine (Vitis vinifera; DeAndres et al. 2012), and chestnut (Castanea sativa Mill.; Martin et al. 2012). Both STRUCTURE and PCA analyses detected genetic composition clearly. In this study, the 38 families were nearly equally divided into two clusters. Interestingly, families 9, 11, and 12 from site 2 were genetically close to families from sites 3 and 4. Of all the four locations, seedlings from site 2 displayed the highest allelic richness, and the families collected from site 2 were captured in both clusters. Previous studies based on chloroplast DNA indicated that the wild population of P. pashia at site 2 contained the largest number of haplotypes among the four locations. However, the haplotype diversity of the wild population from site 2 was lower than that of sites 1 and 3 (Liu et al. 2013). P. pashia is widely spread in central Yunnan Province, and there are no natural barriers between wild populations. The four sites where we collected the seeds were located in an area between three wild populations considered to be refugia during Quaternary glaciations (Fig. 1), and these could have served as source populations for range expansions during interglacial periods (Chiang et al. 2001; Liu et al. 2013). Therefore, range expansion played an important role in shaping the current genetic and phylogeographic structure of P. pashia (Liu et al. 2013). We may infer that the diversity of the seedling populations was formed through colonization and admixture of alleles from adjacent populations based on the results indicated by chloroplast and nuclear DNA markers. Furthermore, the analyses of chloroplast DNA can only indicate seed dispersal and provide no information about pollen gene flow, which is a main component shaping the organization of genetic diversity within and among populations. The discordance between these two markers

suggested that asymmetrical gene flow has occurred among these populations (Bai et al. 2010). However, more evidence from nuclear data is required to support this possibility, especially considering that individuals of the seedling population from site 3 were fewer than those of the other seedling populations. Alternatively, geographic relationships of the sampling sites to Kunming city should be considered. While the relative impact of human activity on the distribution and population structure of crops cannot be concluded easily from genetic data (Hunt et al. 2011), we cannot exclude that human intervention may have influenced the population structure of P. pashia. Conservation and management of the germplasm are fundamental parts of plant breeding projects (Erfani et al. 2012). Conservation priority should be given to families that display a high number of private alleles while maintaining abundent genetic variability (Martin et al. 2012). As the families derived from site 2 displayed the maximum amount of allelic richness and had a highly mixed genetic composition, we recommend that these should be paid special attention in conservation and breeding projects, especially families 9, 11, and 12. Seedling trees from families 10 and 13 were also highly heterozygous. The five families may have valuable traits suitable for rootstock breeding. In conclusion, seedling populations of P. pashia collected from the central Yunnan Province contained abundant polymorphism and a high level of genetic differentiation. Our study depicted some genetic characteristics of P. pasha seedling populations, and these will provide future strategies for the management, conservation, and use of the seedlings.

Acknowledgments This work was financed by the National Natural Science Foundation of China (no. 30871690).

References Bai WN, Liao WJ, Zhang DY (2010) Nuclear and chloroplast DNA phylogeography reveal two refuge areas with asymmetrical gene flow in a temperate walnut tree from East Asia. New Phytol 188: 892–901 Bao L, Chen KS, Zhang D, Cao YF, Yamamoto T, Teng YW (2007) Genetic diversity and similarity of pear cultivars native to East Asia revealed by SSR (simple sequence repeat) markers. Genet Resour Crop Evol 54:959–971 Bao L, Chen KS, Zhang D, Li XG, Teng YW (2008) An assessment of genetic variability and relationships within Asian pears based on AFLP (amplified fragment length polymorphism) markers. Sci Hortic 116:374–380 Cao YF, Tian LM, Gao Y, Liu FZ (2012) Genetic diversity of cultivated and wild Ussurian pear (Pyrus ussuriensis Maxim.) in China evaluated with M13-tailed SSR markers. Genet Resour Crop Evol 59:9–17

Author's personal copy Plant Mol Biol Rep Challice JS, Westwood MN (1973) Numerical taxonomic studies of the genus Pyrus using both chemical and botanical characters. Bot J Linn Soc 67:121–148 Chiang T, Chiang Y, Chen YJ, Chou C, Havanond S, Hong TN, Huang S (2001) Phylogeography of Kandelia candel in East Asiatic mangroves based on nucleotide variation of chloroplast and mitochondrial DNAs. Mol Ecol 10:2697–2710 DeAndres MT, Benito A, Perez-Rivera G, Ocete R, Lopez MA, Gaforio L, Munoz G, Cabello F, Zapater J, Arroyo-Garcia R (2012) Genetic diversity of wild grapevine populations in Spain and their genetic relationships with cultivated grapevines. Mol Ecol 21:800–816 Doyle JJ, Doyle JL (1987) A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem Bull 19:11–15 Earl DA, vonHoldt BM (2012) STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conserv Genet Resour 4:359–361 Erfani J, Ebadi A, Abdollahi H, Fatahi R (2012) Genetic diversity of some pear cultivars and genotypes using simple sequence repeat (SSR) markers. Plant Mol Biol Rep 30:1065–1072 Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol 14:2611–2620 Excoffier L, Lischer H (2010) Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Mol Ecol Resour 10:564–567 Ferreira dos Santos A, Ramos-Cabrer AM, Diaz-Hernandez MB, PereiraLorenzo S (2011) Genetic variability and diversification process in local pear cultivars from northwestern Spain using microsatellites. Tree Genet Genomes 7:1041–1056 Gianfranceschi L, Seglias N, Tarchini R, Komjanc M, Gessler C (1998) Simple sequence repeats for the genetic analysis of apple. Theor Appl Genet 96:1069–1076 Goudet J (2001) FSTAT, a program to estimate and test gene diversities and fixation indices version 2.9.3. www.unil.ch/izea/softwares/fstat.html Hunt HV, Campana MG, Lawes MC, Park YJ, Bower MA, Howe CJ, Jones MK (2011) Genetic diversity and phylogeography of broomcorn millet (Panicum miliaceum L.) across Eurasia. Mol Ecol 20:4756–4771 Huson DH, Scornavacca C (2012) Dendroscope 3: an interactive tool for rooted phylogenetic trees and networks. Syst Biol 61:1061–1067 Jakobsson M, Rosenberg NA (2007) CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics 23:1801–1806 Jolivet C, Holtken AM, Liesebach H, Steiner W, Degen B (2011) Spatial genetic structure in wild cherry (Prunus avium L.): I. Variation among natural populations of different density. Tree Genet Genomes 7:271–283 Katayama H, Adachi S, Yamamoto T, Uematsu C (2007) A wide range of genetic diversity in pear (Pyrus ussuriensis var. aromatica) genetic resources from Iwate, Japan revealed by SSR and chloroplast DNA markers. Genet Resour Crop Evol 54:1573–1585 Krause S, Hammer K, Buerkert A (2007) Morphological biodiversity and local use of the Himalayan pear (Pyrus pashia) in Central Bhutan. Genet Resour Crop Evol 54:1245–1254 Langella O (2000) POPULATIONS 1.2: population genetic software, individuals or population distance, phylogenetic trees. http:// bioinformatics.org/∼tryphon/populations/ Liebhard R, Gianfranceschi L, Koller B, Ryder CD, Tarchini R, Van de Weg E, Gessler C (2002) Development and characterisation of 140 new microsatellites in apple (Malus ×domestica Borkh.). Mol Breeding 10:217–241 Liu J, Zheng X, Potter D, Hu C, Teng Y (2012) Genetic diversity and population structure of Pyrus calleryana (Rosaceae) in Zhejiang Province, China. Biochem Syst Ecol 45:69–78

Liu J, Sun P, Zheng X, Potter D, Li K, Hu C, Teng Y (2013) Genetic structure and phylogeography of Pyrus pashia (Rosaceae) in Yunnan Province, China, revealed by chloroplast DNA analyses. Tree Genet Genomes 9:433–441 Martin MA, Mattioni C, Molina JR, Alvarez JB, Cherubini M, Herrera MA, Villani F, Martin LM (2012) Landscape genetic structure of chestnut (Castanea sativa Mill.) in Spain. Tree Genet Genomes 8:127–136 Nei M, Tajima F, Tateno Y (1983) Accuracy of estimated phylogenetic trees from molecular data. J Mol Evol 19:153–170 Peakall R, Smouse PE (2006) GENALEX 6: genetic analysis in Excel. Population genetic software for teaching and research. Mol Ecol Notes 6:288–295 Potts SM, Han Y, Khan MA, Kushad MM, Rayburn AL, Korban SS (2012) Genetic diversity and characterization of a core collection of Malus germplasm using simple sequence repeats (SSRs). Plant Mol Biol Rep 30:827–837 Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155:945–959 Richards CM, Volk GM, Reilley AA, Henk AD, Lockwood DR, Reeves PA, Forsline PL (2009) Genetic diversity and population structure in Malus sieversii, a wild progenitor species of domesticated apple. Tree Genet Genomes 5:339–347 Rosenberg NA (2004) DISTRUCT: a program for the graphical display of population structure. Mol Ecol Notes 4:137–138 Rubtsov GA (1944) Geographical distribution of the genus Pyrus and trends and factors in its evolution. Am Nat 78:358–366 Takezaki N, Nei M (2008) Empirical tests of the reliability of phylogenetic trees constructed with microsatellite DNA. Genetics 178:385–392 Teng YW, Tanabe K, Tamura F, Itai A (2001) Genetic relationships of pear cultivars in Xinjiang, China as measured by RAPD markers. J Hortic Sci Biotechnol 76:771–779 Teng YW, Tanabe K, Tamura F, Itai A (2002) Genetic relationship of Pyrus species and cultivars native to East Asia revealed by randomly amplified polymorphic DNA makers. J Am Soc Hortic Sci 127:262–270 Van Oosterhout C, Hutchinson WF, Wills D, Shipley P (2004) MICROCHECKER: software for identifying and correcting genotyping errors in microsatellite data. Mol Ecol Notes 4:535–538 Van Treuren R, Van Soest L, Van Hintum TJ (2001) Marker-assisted rationalisation of genetic resource collections: a case study in flax using AFLPs. Theor Appl Genet 103:144–152 Volk GM, Richards CM, Henk AD, Reilley AA, Bassil NV, Postman JD (2006) Diversity of wild Pyrus communis based on microsatellite analyses. J Am Soc Hortic Sci 131:408–417 Wuyun T, Ma T, Uematsu C, Katayama H (2013) A phylogenetic network of wild ussurian pears (Pyrus ussuriensis Maxim.) in China revealed by hypervariable regions of chloroplast DNA. Tree Genet Genomes 9:167–177 Yamamoto T, Kimura T, Sawamura Y, Manabe T, Kotobuki K, Hayashi T, Ban Y, Matsuta N (2002a) Simple sequence repeats for genetic analysis in pear. Euphytica 124:129–137 Yamamoto T, Kimura T, Sawamura Y, Manabe T, Kotobuki K, Hayashi T, Ban Y, Matsuta N (2002b) Genetic linkage maps constructed by using an interspecific cross between Japanese and European pears. Theor Appl Genet 106:9–18 Yao LH, Zheng XY, Cai DY, Gao YA, Wang K, Cao YF, Teng YW (2010) Exploitation of Malus EST-SSRs and the utility in evaluation of genetic diversity in Malus and Pyrus. Genet Resour Crop Evol 57: 841–851 Yu DJ, Lu LD, Gu CZ, Guan KJ, Jiang WF (1986) Flora Republicae Popularis Sinicae 38:367–370 Zhang Q, Li J, Zhao YB, Korban SS, Han YP (2012) Evaluation of genetic diversity in Chinese wild apple species along with apple cultivars using SSR markers. Plant Mol Biol Rep 30:539–546