Development of Reproducible EST-derived SSR Markers and Assessment of Genetic Diversity in Panax ginseng Cultivars and Related Species.

Little is known about the genetics or genomics of Panax ginseng. In this study, we developed 70 expressed sequence tag-derived polymorphic simple sequence repeat markers by trials of 140 primer pairs. All of the 70 markers showed reproducible polymorphism among four Panax speciesand 19 of them were polymorphic in six P. ginseng cultivars. These markers segregated 1:2:1 manner of Mendelian inheritance in an F2 population of a cross between two P. ginseng cultivars, 'Yunpoong' and 'Chunpoong', indicating that these are reproducible and inheritable mappable markers. A phylogenetic analysis using the genotype data showed three distinctive groups: a P. ginseng-P. japonicus clade, P. notoginseng and P. quinquefolius, with similarity coefficients of 0.70. P. japonicus was intermingled with P. ginseng cultivars, indicating that both species have similar genetic backgrounds. P. ginseng cultivars were subdivided into three minor groups: an independent cultivar 'Chunpoong', a subgroup with three accessions including two cultivars, 'Gumpoong' and 'Yunpoong' and one landrace 'Hwangsook' and another subgroup with two accessions including one cultivar, 'Gopoong' and one landrace 'Jakyung'. Each primer pair produced 1 to 4 bands, indicating that the ginseng genome has a highly replicated paleopolyploid genome structure.

INTRODUCTION

Korean ginseng (Panax ginseng Meyer) is an important medicinal herb belonging to the family Araliaceae. Ginseng has been used as oriental medicine for thousands of years [1]. The major components showing pharmacological effects are the ginsenosides, which are known for their beneficial properties to the central nervous system, cardiovascular, endocrine and immune systems [2].

In ginseng research, medicinal components and their functions have been widely investigated. However, breeding, genetic and genomic studies have been rarely performed because of difficulty in maintaining plants and reproducing progenies. Approximately three to four years of growth is necessary to produce a small number of seeds, approximately 40 seeds per plant [3], thus hindering systematic management of genetic materials. Up to now, eight elite cultivars, ‘Chunpoong’, ‘Yunpoong’, ‘Gumpoong’, ‘Gopoong’, ‘Sunpoong’, ‘Sunwon’, ‘Sunone’ and ‘Chungsun’, have been bred by pure line selection and have been registered as commercial varieties since 1997 in Korea [4]. Even though the new varieties show better yields and qualities compared to that of the local landrace which is a native mixed line [3,4], they are being cultivated in less than 10% of the total ginseng cultivation fields because of the lack of well-organized seed production and supplying system. Stable seed supply systems with credible authentication method for each ginseng cultivar will promote high quality ginseng production via improvement of the ginseng breeding and seed industry.

Molecular breeding tools using DNA markers may be an alternative and indispensable way for ginseng improvement because marker assisted selection can reduce efforts and time for breeding. However, very limited numbers of DNA markers were also reported for ginseng. Random DNA markers such as random amplified polymorphic DNA [5-7] and amplified fragment length polymorphism (AFLP) [8] are used to study the diversities of local ginseng collections. However, these random primer-based markers could not be shared by common. Approximately, 60 of simple sequence repeat (SSR) markers have been produced from SSR-enriched libraries [9,10] and from bacterial artificial chromosome (BAC) end sequences [11] and are studied to determine the genetic diversity of ginseng collections. All of these SSR markers were derived from genomic sequences and these were not intensively studied between ginseng cultivars. Even though several papers described ginseng expressed sequence tags (ESTs) [12-15], there have been no reports on development of EST-derived SSR markers and their utilization in ginseng.

ESTs providing comprehensive transcript information [16] are valuable resources for development of molecular markers because they are derived from relatively conserved genic regions. EST-derived SSRs are also more advantageous than genomic SSRs because of the rich public availability of EST sequences and their high transferability to related species [17]. Thus, we are going to develop large number of EST sequence-derived SSR markers and construct a high resolution genetic map which can be utilized as a frame for genome sequencing. In this study, we tried to develop reproducible EST-derived SSR markers which can be applied for mapping and assessing a genetic similarity between registered commercial inbred varieties. And we estimated a polyploidy level of ginseng genome based on numbers of gene-based polymerase chain reaction (PCR) products.

MATERIALS AND METHODS

Plant materials and DNA extraction

Six P. ginseng accessions, four registered cultivars, ‘Chunpoong’, ‘Yunpoong’, ‘Gumpoong’ and ‘Gopoong’, bred by inbred line selection in Korea Ginseng Corporation (KGC) Natural Resources Research Institute (Daejeon, Korea), and two representative local landraces, ‘Jakyung’, mixed lines with red fruits, and ‘Hwangsook’, mixed lines with yellow fruits, and three related Panax species, P. quinquefolius originated in the USA, P. japonicus originated in Japan, and P. notoginseng originated in China, were included in the determination of genetic diversity. DNA pools derived from more than 15 individuals were used to represent each cultivar and landrace of P. ginseng. The DNA pool consisted of a mixture of the same amount of template DNA from 15 individuals of each cultivar and landrace of P. ginseng and from five individuals of P. quinquefolius. However, single individual DNA was used to represent P. japonicus and P. notoginseng because of limited materials. An F2 population that consisted of 51 individuals from a cross between ‘Yunpoong’ and ‘Chunpoong’ was used to determine the inheritability and reproducibility of the newly developed markers. All leaf samples were kindly provided from KGC Central Research Institute. Total DNA was extracted using the modified cetyltrimethylammonium bromide method [18]. DNA concentrations were measured using ND-1000 (NanoDrop Technologies Inc., Wilmington, DE, USA).

Construction of the ginseng expressed sequence tag database and repeat motif screening

A ginseng EST database was constructed by collecting sequences from public databases. After removal of poly-A tails using PanGEA [19], repeat-oriented sequences and SSR motif-containing sequences were characterized from the raw EST database using RepeatMasker ver. 3.2.6 (http://www.repeatmasker.org) which was downloaded and installed on the local computer. In the screening process, a default mode with a “-poly” option was used to select genuine SSR motifs.

Designing primers

We extracted ESTs which contained 3-6 copies of SSR motifs using Tandem Repeat Finder [20]. Primer pairs were designed from the flanking sequences of SSR motifs (Table 1) with 18 to 27 bp nucleotides using the Primer3 program (http://frodo.wi.mit.edu/primer3/). Product sizes ranged from 150 to 600 bp. Standalone BLAST executables (BLASTN 2.2.15, ftp://ncbi.nlm.nih.gov/blast/executables) were used to avoid primer design in duplicate sequences. Various sources of SSR markers reported in the previous papers were designed and tested together [9,10].

Table 1.

Simple sequence repeats found in ginseng EST sequences and summarization of SSR marker developments

Repeat unit (length)	SSR1)	Primer design2)	PCR Success (%)	Polymorphic between
				Panax ginseng cultivars (%)	Panax species (%)
Mono	855	11	4 (36.4)	0	0
Di	379	28	23 (82.1)	9 (39.1)	16 (69.6)
Tri	246	45	41 (91.1)	9 (22.0)	30 (73.2)
Tetra	87	10	8 (80.0)	0	2 (25.0)
Penta	70	12	12 (100)	1 (8.3)	11 (91.7)
Hexa	34	5	4 (80.0)	0	3 (75.0)
Degenerate	-	29	27 (93.1)	0	8 (29.6)
Total	1,671	140	119 (85.0)	19 (16.0*)	70 (58.8)

EST, expressed sequence tag; SSR, simple sequence repeat; PCR, polymerase chain reaction.

1)No. of potentially polymorphic SSRs based on the poly option of the RepeatMasker program.

2) No. of sites used for primer designing.

Polymerase chain reaction and electrophoresis

PCR amplifications were performed in a 25 μL volume with 1 U Taq DNA polymerase (Vivagen, Seongnam, Korea) according to the manufacturer’s protocol using a DNA Engine Thermal Cycler (Bio-rad, Hercules, CA, USA). Conditions for the PCR cycle were as follows: 5 min at 95℃ for denaturation, 38 cycles of 10 s at 95℃, 30 s at Tm℃, 20 s at 72℃, and 10 min at 72℃ for final extension. PCR products were separated on 2% agarose gels and on 5% denaturing polyacrylamide gels or 9% non-denaturing polyacrylamide gels.

Data analysis

Blast2GO ver. 2.4.2 was used to annotate polymorphic ESTs with default parameters [21]. Analyses of developed marker data were conducted using PowerMarker ver. 3.25 software [22]. Allele frequency data were obtained as a binary matrix and were imported into NTSYSpc 2.11X (Exeter Software; Setauket, NY, USA) for phylogenetic analysis using Dice’s coefficient [23] and the unweighted pair group method with the arithmetic mean [24]. Bootstrapping of the tree with 1,000 replications was generated in Winboot [25].

RESULTS AND DISCUSSION

Simple sequence repeat motif in ginseng expressed sequence tag sequences

A total of ca. 11 Mb, 19,578 ginseng ESTs, were collected from the public database and reconstructed in our local server (http://im-crop.snu.ac.kr). Using an EST trimming process, ca. 17.5 kb of poly-A tails were removed and 1.6% (179,540 bp) of ESTs were masked as repeat sequences. Of them, 1,584 bp showed a significant homology with 19 kinds of transposable elements, 8,619 bp had a significant homology with 32 non-coding RNA elements, and 62,348 bp derived from 1,344 regions were screened as low complexity DNA. A total of 2,158 regions spanning 108,116 bp were classified as simple sequence repeats, and 1,671 sites spanning 68,763 bp detected in 1,567 ESTs (8.0% of the total raw ESTs) were classified as potential polymorphic SSR sites by the “-poly” option (Table 1). Most of the ESTs (94.6%) contained one SSR motif, with exceptions containing two SSRs (4.6%) or three to five SSRs (0.8%).

Classification based on the nucleotide lengths of the repeat motifs showed that 51% of SSRs were composed of mono-nucleotide motifs, but mononucleotide repeat motifs showed poor success rates for PCR as well as polymorphism that may have derived from “stutter” artifacts in the PCR reaction [26]. When we ignored mononucleotide repeats, di- and tri-nucleotide repeats were the most abundant classes, including 46.4% and 30.1%, respectively. Similarly, di-nucleotide repeat motifs were also predominant in kiwifruit, spruce and coffee [27-29]. Meanwhile, tri-repeat motifs were most predominant in the EST-SSRs of grape, sugarcane, barley, wheat, maize, sorghum, rice, rye, oats, and flax [30-36].

Assessment of homogeneities in ginseng landraces and registered cultivars

During development of DNA markers polymorphic between accessions, we found that DNA markers were not homogeneous among individuals in one local landrace accession or in the registered cultivars, as shown in Fig. 1. The most popular landrace ‘Jakyung’ individuals showed up to six different genotypes among 15 individuals, even though all were derived from the same seed lot (Fig. 1A). Meanwhile, one of the elite cultivars, ‘Chunpoong’, showed a relatively uniform genotype with two off-types among 20 individuals randomly selected from the same seed lot (Fig. 1B). Screening with three markers developed in this study showed a relative range of heterogeneity of 10% to 30% in six accessions. Also, we found that previously reported polymorphic ginseng markers were not reproducible in our trial [9,10,31]. This may be due to the heterogeneity of the ginseng population because they used an individual plant DNA for representative of each accession.

Fig. 1.

Allelic variations among individuals of a landrace and a cultivar. Denaturing polyacrylamide gel electrophoresis was conducted for separation of PCR products using individual DNA and the GES0002 marker. (A) A total of 15 individuals of landrace ‘Jakyung’ were surveyed. Different genotypes were denoted as a-f and genotype d is shown as major. (B) A total of 20 individuals of cultivar ‘Chunpoong’ were surveyed. Only number 4 plant denoted by * shows heterozygous allele. L, DNA ladder; GES, ginseng expressed sequence tag-simple sequence repeat.

Approximately 56% of AFLP bands showed polymorphism among wild P. ginseng individuals in Russian Primorye area and population structure study clearly differentiated their phylogenetic relationships based on frequencies of individual alleles [37]. Similarly, more divergence detected in wild P. quinquefolius than the cultivated [38] that might be derived by possible out-crossing events even though the plants prefer self-fertilization [39]. And cultivated P. notoginseng population remained fair level of biodiversity, ranged 74% to 39% of divergence depends on location in China [40]. Our result showed abundant genetic diversity even in cultivating P. ginseng landraces (Fig. 1A) that might be derived from bulked seed harvesting from genetically unfixed lines and also from temporal out-crossing [39]. Meanwhile, eight elite cultivars were bred by pure line selection even though each showed approximately 10% of off-type allele. Therefore, we considered that utilization of a DNA pool from many individuals will be credible to identify more reproducible markers than using a single plant for representing each Korean ginseng cultivar even though the method can ignore many rare alleles. We concluded to use a DNA pool derived from 15 individual plants to represent each Korean ginseng cultivar because our purpose is development of markers which can be a general representative for each elite cultivar (Fig. 1B).

Development of simple sequence repeat markers and their transferability to related species

We designed a total of 140 primer pairs amplifying 111 SSR motifs and 29 degenerated SSR motifs (Table 1). Among them, 119 pairs produced bands of which 105 were the same as the expected sizes and 16 were larger than the expected. And 21 primers produced no clear PCR product. The PCR failure rate was 15% for the 140 trials of EST-SSR primers, which is in the 10% to 40% range reported in previous EST-SSR analyses for other species [27,30,31,35,41-44]. Seventy pairs, corresponding to 58.8% of the successful PCR primers, showed polymorphism for at least one of nine accessions including six P. ginseng accessions and three related species. Polymorphisms were mainly restricted to interspecies, and only 19 of 70 SSRs showed polymorphism between P. ginseng cultivars as well as between Panax species and were named ginseng EST-SSR (GES) (Fig. 2A, B; Tables 1 and 2). The other 51 markers were polymorphic only among Panax species. Among those, 43 were derived from intact SSR regions and were named Panax EST-SSR (Fig. 2C, D; Tables 1 and 2). Eight were derived from degenerated SSR regions and were named Panax EST (Tables 1 and 2). BLASTX analysis revealed that 51 of 70 polymorphic ESTs showed significant hits and 19 did not have any matches with the known proteins. Among the significant 51 hits, 22 showed best hits with genes in Vitis vinifera (Table 2).

Fig. 2.

Non-denaturing polyacrylamide gel electrophoresis of PCR products from nine Panax accessions using different markers, (A) GES0003, (B) GES0019, (C) PES0010 and (D) PES0034. Lanes 1-6 indicate P. ginseng accessions, Gumpoong (1), Hwangsook (2), Yunpoong (3), Chunpoong (4), Gopoong (5), Jakyung (6) and related Panax species, P. japonicus (7), P. quinquefolius (8), P. notoginseng (9). L, DNA ladder; GES, ginseng expressed sequence tag-simple sequence repeat; PES, Panax expressed sequence tag-simple sequence repeat.

Table 2.

Characteristics of 70 polymorphic expressed sequence tag-simple sequence repeat loci in Panax ginseng cultivars and related species

Marker	Primer pair (5’ → 3’)	Repeat motif	Ta (℃)	Size1)	Nr2)	Na3)	MAF	GD	PIC	Nb4)	Sequence description	Min. E-value
GES0001	GCATGGCAATTTGGAGAGAGGTACG	(GAA)n	60	196	14	7	0.3333	0.8148	0.7938	3	No hits found
	GTTCTGTGACTTGCCGGTTTGCTCC
GES0002	GTGGTGAGAAAGGGAAAGAGCAATCG	(TA)n	60	176	22	7	0.2222	0.8395	0.8194	2	No hits found
	CCCTCGATCTACAGATGATCAAATAGC
GES0003	TTTCAAATGGATCTATGAGAATAATGA	(TTC)n	54	247	20	5	0.4444	0.7160	0.6773	3	No hits found
	TGGGCACATAAAAAGACAGTG
GES0004	CTAGTCAACAACATCATCATCATCC	(TA)n	56	213	27	5	0.3333	0.7407	0.6987	4	Sugar isomerase	1.E-12
	TGCAGAATTAACTGAGACTCAAGAA
GES0005	TTCTTCCTTGCACGTTTCTACTACT	(CCA)n	56	190	14	7	0.2222	0.8395	0.8194	2	at5g14920 f2g14_40	2.E-28
	AATATAATTGCTACACTCCCCTTGG
GES0006	AGCCTAGTGTGCAGAAGTAAAGTGT	(TA)n	56	238	26	4	0.4444	0.6667	0.6072	3	Protein	2.E-15
	TGAAGTAGAACTGATCACAGAGTGC
GES0007	GGGGCTTCTCTAATTTACACCTTTA	(GA)n	55	243	16	4	0.4444	0.6914	0.6401	3	Protein	9.E-41
	AAAGATGAAAACTTGATGCTTGTTC
GES0008	TGGTCTAGAACAGAAAAGATCGAGT	(TA)n	55	187	28	5	0.4444	0.7160	0.6773	2	No hits found
	GTACTGTCTGTGGTTTGAATGATTG
GES0009	TGAACCACATGATTTACGATTAGTG	(TA)n	56	240	11	6	0.3333	0.7901	0.7615	2	Xyloglucan endotransglycosylase	9.E-13
	GACATATCTGCATGGCTTTCTTAAT
GES0010	AGGACTTCAATGCTAGAACTCAGAA	(TA)n	56	285	17	4	0.4444	0.6914	0.6401	2	Mitochondrial ribosomal protein l11	2.E-44
	CATGGGCTAAATAATAAAAGACCAA
GES0011	GTTATGACCGGTAAATTAGGTTGGT	(TA)n	56	229	11	4	0.5556	0.6173	0.5688	2	Protein	2.E-46
	CAATCCACATCAAGACCATATTACA
GES0012	TAATTATATTTGTGTTGGCAGACGA	(TC)n	54	227	15	5	0.4444	0.7160	0.6773	3	Predicted: hypothetical protein [Vitis vinifera]	4.E-13
	CTCGGCATACAACATTTAACTTACC
GES0013	ATTAGTAGAACGTACAGCCCAAACC	(CCA)n	56	230	8	5	0.4444	0.7160	0.6773	2	Carboxyphosphonoenolpyruvate mutase	5.E-58
	TATGGTAACTTTAGGCTGGTGTAGC
GES0014	GAGAAAATTGAGGAACCAAACAAG	(CTA)n	54	299	12	4	0.6667	0.5185	0.4847	4	No hits found
	GTTTTCTCCACAACTACTGGCTCT
GES0015	AAAATTCTGCTCACACTCTCTCTGT	(CTA)n	56	193	10	4	0.5556	0.6173	0.5688	2	No hits found
	CGGAGTTTTTGAAGATAAGAATCAA
GES0016	ATTTATATATCTTCACGCTGCTTCG	(TCC)n	56	230	10	4	0.4444	0.6667	0.6072	3	Beta-galactosidase a-peptide	5.E-13
	CAAAAATAAGAGATGGAGATGGAGA
GES0017	AAAATGGTTCCAAATTGTGCTTC	(TTA)n	56	239	11	5	0.3333	0.7654	0.7279	4	No hits found
	AAGGTGGAAATAAGGAGAGAAAAGA
GES0018	CTCTCTCTCTCTCTCTCTCATCTGC	(TTC)n	56	170	8	4	0.6667	0.5185	0.4847	2	Protein	1.E-46
	AAAGAAGAACCACAAACACTAAACG
GES0019	GTACTATGGATAAAGCTGGAATGGA	(TAGGG)n	56	207	6	5	0.3333	0.7654	0.7279	2	No hits found
	CGGTAAGTGACACTAAGAACAACTG
PES0001	GGAGCAGCAATAGACCAAGG	(CCCTG)n	55	356	4	2	0.7778	0.3457	0.2859	3	Predicted: hypothetical protein [V. vinifera]	1.E-04
	TTGTTTGGAAACCTGGGAAC
PES0002	TCGGAGCGAGAAGAAGAGTC	(ATG)n	56	290	9	3	0.7778	0.3704	0.3402	2	Nascent polypeptide associated complex alpha chain	1.E-68
	CGTCTTCATCATCCTGAGCA
PES0003	GGTGGAGATCACAAGGAAGG	(GAA)n	56	312	8	3	0.7778	0.3704	0.3402	2	No hits found
	TGGCAACAATCAGCATCCTA
PES0004	CGAAGGTGCACCAAAAGTCT	(CACCAT)n	56	365	5	2	0.7778	0.3457	0.2859	4	No hits found
	GGACGAAGACGTGGCTCTAC
PES0005	TGGGTTCAACTTTGGAGGAG	(CAGGT)n	56	243	11	3	0.7778	0.3704	0.3402	2	Protein	5.E-15
	CTCTTTCACCGCAACAGACA
PES0006	CAACCTTTTAATTCCTTTGCTACA	(CAT)n	54	172	10	2	0.7778	0.3457	0.2859	2	Transcription factor gt-3a	1.E-08
	CCGTCTCAATATTCACACTGATCT
PES0007	CGAGGAGTCAAAGGTGGAAG	(GAA)n	56	266	15	2	0.7778	0.3457	0.2859	2	Dehydrin	1.E-23
	CGCCTGGAAGTTTTTCTTTG
PES0008	AACGTGATGCATGTCGAGAG	(TA)n	54	176	26	3	0.7778	0.3704	0.3402	3	Catalase	1.E-24
	GCACCGAGTTTTCCCAAGTA
PES0009	GGAGGCCCGACTTACCTACT	(GGC)n	54	213	6	2	0.8889	0.1975	0.1780	1	No hits found
	CACGTTGACGTGGCTATCTG
PES0010	GTCTCGCAAAGAATGTCAGC	(CCA)n	55	189	7	2	0.7778	0.3457	0.2859	2	g1-like protein	2.E-47
	CTGCTTTTGCACCTCATAGC
PES0011	TATCCACCAAAACACTACTCATCCT	(ATG)n	54	258	7	2	0.8889	0.1975	0.1780	1	Predicted: hypothetical protein [V. vinifera]	2.E-05
	CCTCTTAGACTCGTCATTAGGTTCA
PES0012	ATTTAGCTTGGCTATATGTGAATGG	(CAG)n	54	284	6	3	0.7778	0.3704	0.3402	2	No hits found
	GGACAGAAGTGAAGCATTTCATAGT
PES0013	TCCTAAATTAGCACTAAACGCACAT	(CAG)n	54	162	8	3	0.7778	0.3704	0.3402	1	Dna binding	1.E-37
	TTGTTTACTAAATTCATGGGAGAGG
PES0014	CAACTGCAAAGTCAAAATAATACGA	(TA)n	56	180	15	3	0.7778	0.3704	0.3402	2	Myb-like transcription factor 1	5.E-21
	GTAATCTTCCAGCTATCAAAGACCA
PES0015	ACAAGAACAATTGTCAAAGGAAGTC	(TA)n	56	300	12	3	0.7778	0.3704	0.3402	1	Ribulose bisphosphate small subunit	1.E-62
	CTTTCAACACCTGAGATGAATCAGT
PES0016	GAATGATCATATATACCTCCACTGGT	(TA)n	54	252	11	3	0.7778	0.3704	0.3402	2	Dehydrin 7	2.E-50
	TAAAATAAGCATTAGCAGAGCCATC
PES0017	TCGGTAAGGATATCATCAACAAAAT	(TC)n	54	174	10	3	0.7778	0.3704	0.3402	2	Sterol carrier protein 2-like	2.E-18
	TTTTTGATAAAGACAAGGTCAAAGC
PES0018	GGTATTGCTCGTGAACTTTGTAACT	(TC)n	54	201	20	3	0.7778	0.3704	0.3402	2	Aspartyl protease	8.E-107
	CAATAGGAAGAGAAGAAAACCAACA
PES0019	GGAAACAGGGGTAGAAGAAGTGTAT	(TC)n	56	287	10	3	0.7778	0.3704	0.3402	4	Dcn1-like protein 4	2.E-70
	AGTATTTGTTGTTCCTTTCCTGGAT
PES0020	CTATACCTCAGCACCAGTTTCAACA	(CAG)n	54	298	6	3	0.7778	0.3704	0.3402	2	S-rnase-binding protein	9.E-63
	TATCTGCGAATTATTTTCCATGAAT
PES0021	GAAAACATTTGTGTTTCAGTAGGC	(CAG)n	56	155	10	3	0.7778	0.3704	0.3402	2	60s ribosomal protein	5.E-92
	AATGAGCTTCAGGTAAATATCATCG
PES0022	CCAAGCCACATAAATCTAGGAGTATC	(CAG)n	56	154	7	3	0.7778	0.3704	0.3402	2	No hits found
	AAAAACAACAAGTGCAGTTACACAA
PES0023	CACAGTGAGGAAGAAGAAGAAGAAG	(CAT)n	56	152	8	2	0.7778	0.3457	0.2859	3	60s ribosomal protein l6	1.E-54
	ACCTGGAATACTTTCCAATACCG
PES0024	TAATAATAATTTGATGCGGTTCCAT	(CAT)n	56	172	10	2	0.7778	0.3457	0.2859	2	Protein	5.E-24
	GGTGTTGTCAATTAGGAGAGAGAAA
PES0025	AAAATCAATCTCCCATAAATTTGGT	(CTG)n	56	219	8	3	0.7778	0.3704	0.3402	2	Serine threonine-protein kinase	1.E-13
	TGATGTTTTGAAACAGAATCTTCAA
PES0026	AGAATTTGAAGATGATGAAGAATCG	(GAA)n	56	214	9	3	0.7778	0.3704	0.3402	2	Protein	9.E-81
	AAAAGCTTTAGCCAAAGAAAGAGAG
PES0027	ACTTTTATCCCAAAGCATCTTTTCT	(GAA)n	55	181	7	3	0.7778	0.3704	0.3402	1	No hits found
	GTTCTACCTCTGAATTGGCACTAGA
PES0028	GATACCTCAACAAAAATCCATCAAC	(GGA)n	55	154	9	3	0.7778	0.3704	0.3402	3	Protein	3.E-70
	ATGACGTTGTGCTTTTATAGCTTCT
PES0029	TACTCCTTCCCCATTTAATTTCTTC	(TCC)n	56	183	8	3	0.7778	0.3704	0.3402	2	Ap2 erf domain-containing transcription factor	9.E-29
	GAGGGCAGTTTTGGTAGATATTTTT
PES0030	AACCTAATGTCCGGAAGGTAGTAAC	(TTA)n	56	259	9	3	0.7778	0.3704	0.3402	3	Receptor protein	2.E-42
	CAATGTATGGAGGAAGGTTTATTTG
PES0031	CGAATTCTGATTCTTGACATTTCTT	(TTC)n	55	210	7	3	0.7778	0.3704	0.3402	4	Protein aq_1857	7.E-28
	CGAAATATGGAGTAAGACGCTGTAT
PES0032	GTGAAGCTTGAACACTTAGAAGAGG	(TCCC)n	55	165	6	3	0.7778	0.3704	0.3402	3	Oligosaccharyl transferase	3.E-123
	TCATCATACTTTGCTAACACGTCAC
PES0033	AAAACGAAGAAAATGTTACAAGTGC	(TCTA)n	56	298	6	3	0.7778	0.3704	0.3402	2	No hits found
	GTATCACCACAACATCAAATTACCA
PES0034	GCTTATCTTGCCTGATAATGTCCTA	(ATCTG)n	56	259	5	3	0.7778	0.3704	0.3402	2	Protein	1.E-59
	CAACATGTAATTACGTCTCTCATGC
PES0035	CAATTTCTTCCGTATAAACCAATTT	(AATCA)n	56		5	2	0.7778	0.3457	0.2859	1	Predicted protein [Populus trichocarpa]	5.E-06
	ACTCCTACCTGCACAATTTGAATAC
PES0036	AGTGGCAACTCTAGAGAAACTATGC	(CACCC)n	55	259	5	3	0.7778	0.3704	0.3402	1	Conserved hypothetical protein [Ricinus communis]	2.E-22
	ATACGTTACATGGCAGCTGAATACT
PES0037	CAAAAGGCTTGCTCTATATTGTGAT	(CAGGC)n	56	275	4	3	0.7778	0.3704	0.3402	2	Ras-related small gtp-binding protein	1.E-93
	TCCAGTCATTAATACTTGCAACAAA
PES0038	TCTCCTTCTGCTGGTAGTAAAAATG	(CCGTA)n	56	190	4	3	0.7778	0.3704	0.3402	2	Unnamed protein product [Vitis vinifera]	5.E-38
	ATAAATTCTGCATCTACGGTGTAGC
PES0039	TGATCCAGTTCAATTTCTTGTTTTC	(GTTTG)n	56	170	4	3	0.7778	0.3704	0.3402	2	No hits found
	ATAATCTTTCAATTTCCCGTACACA
PES0040	ATTGTTGAAGAAATTGGTTGTTTGT	(TTTTC)n	56	270	6	3	0.7778	0.3704	0.3402	3	Phosphate phosphoenolpyruvate translocator precursor	1.E-36
	CTGAGTTCATTTCCTGGAACATAGT
PES0041	ACTGTTTATGACTGGCTCTACAGTG	(TTTTG)n	56	282	9	3	0.7778	0.3704	0.3402	1	Auxin-repressed protein	2.E-29
	GTACCGTCCATGACATAACATAACA
PES0042	TAAATTCACTTTGTGTGTGTGTGTG	(GGGAGA)n	56	229	5	3	0.7778	0.3704	0.3402	2	O-methyltransferase-like protein	6.E-89
	GTTCTTGATCGTGATTCTTTTCAAG
PES0043	GGGAGACTAATTTTCTTTGCTTTTC	(TTCTCC)n	56	224	3	3	0.7778	0.3704	0.3402	2	F-box family protein	2.E-29
	TTCTGATGAGTTCTGTGTAGGAGTG
PE0001	GCCCTAGCCCTAATCAATCC	-	55	314	-	2	0.7778	0.3457	0.2859	3	At4g30930-like protein	9.E-25
	GGGCCAATGACCTTATACCC
PE0002	GATCTCGAACCGACGAACTC	-	54	376	-	3	0.6667	0.4938	0.4377	1	No hits found
	AACCATACTGCCAACAATTAAGC
PE0003	GCCTTCTGAACTTCCTGGTG	-	54	154	-	2	0.7778	0.3457	0.2859	1	Eukaryotic translation initiation factor 3	3.E-11
	GTCAGGTTCTGCAGGTGGA
PE0004	GGTTCTCGGGACAATGAAAG	-	54	219	-	2	0.7778	0.3457	0.2859	2	Unnamed protein product [Vitis vinifera]	6.E-09
	ACCCCATTCCCTTCTCTCAC
PE0005	GCAACATCACCGTCAATGAG	-	56	264	-	3	0.6667	0.4938	0.4377	1	Glycine-rich rna-binding protein	2.E-37
	CACAAATTTACCAGCCACCA
PE0006	TCCTCTGCCACATTTAAGCA	-	55	309	-	2	0.7778	0.3457	0.2859	2	No hits found
	TCATGTTGCAAGAGCAAAGC
PE0007	GTGGAAGAGGCAAAACCAAG	-	56	210	-	3	0.7778	0.3704	0.3402	2	Nucleic acid binding	8.E-44
	AGCCATGCTAGGTCTGTTGG
PE0008	GGTCTTGGTCTTGGAGTTGG	-	54	221	-	2	0.6667	0.4444	0.3457	1	No hits found
	CCTCCTTGATTTCCACCTGA
Average						3.3143	0.6810	0.4568	0.4176	2.2

MAF, major allele frequency; GD, genetic diversity; PIC, polymorphic information content; GES, ginseng expressed sequence tag-simple sequence repeat; PES, Panax expressed sequence tag-simple sequence repeat; PE, Panax expressed sequence tag.

1)Expected amplicon size; 2)No. of repeats; 3)No. of alleles; 4)No. of bands around the expected size in P. ginseng.

The mean level of polymorphism was 20% at the species level that is similar to those of previous EST-SSR studies in other species [30,35,45-48]. The level of polymorphism in EST-SSR is lower than that of genomic SSR primers because the transcribed regions are more conserved than the non-coding regions [17,29,34,49-52]. Our gene-based SSR markers showed comparable levels of polymorphisms to those of the genomic SSR markers for distinguishing P. ginseng accessions. Only 22 of 189 (representing 11.6%) [10] and 11 of 94 genomic SSR primers (11.7%) [9] were polymorphic among P. ginseng accessions when the primers were designed using a microsatellite-enriched library. Meanwhile, 12 out of 31 BAC end sequence-derived genomic SSR primers (38.7%) were polymorphic among P. ginseng accessions [11,53].

Even though not all of the nucleotide repeat units were surveyed by PCR, our trials showed that big differences in the rates of successful PCR and the appearances of polymorphisms depended on the repeat unit length. SSRs with penta- and tri-nucleotide repeat motifs showed the highest degrees of PCR success and polymorphism detection between Panax species. SSRs having a di-nucleotide motif showed the highest polymorphism rates among the ginseng cultivars. Meanwhile, SSRs derived from mono-nucleotide polymers were not optimal for PCR amplification or polymorphism detection (Table 1).

Transferability among related species is considered the most important feature for EST-SSR markers that help to produce conserved orthologous markers and thus be applicable to related species which have little genomic information [17]. In this study, primers designed from P. ginseng ESTs were successfully used in the related species of P. japonicus, P. quinquefolius and P. notoginseng with 100%, 97.1%, and 75.7% transferabilities, respectively, that is similar to the previous studies shown 100% transferability between P. ginseng and P. quinquefolius [9,11].

Number of bands and estimation of the polyploidy level in Panax ginseng

Recent progress in the field of genomics has uncovered highly replicated polyploidy levels in most of the plant genome [54,55]. The P. ginseng genome is considered as tetraploid because of the chromosome number variations, 12 vs. 24 pairs [56]. However, there has been no molecular evidence to determine their ploidy level or to identify a polyploidization event in the Panax species. Polyploidy levels were previously studied in the olive complex (Olea europaea) based on the band numbers of highly polymorphic SSR markers [57]. Various allele numbers were detected in various subspecies, and maximums of four and six alleles were detected in tetra- and hexaploid subspecies, respectively, that were consistent with those of the flow cytometry analyses.

To estimate copy numbers of homologous genes and thus assume the polyploidy level, we counted the number of bands around the expected size in four relatively homogeneous inbred cultivars by assuming that different bands may have been derived from recently duplicated paralogous genes. Electrophoresis of the PCR products revealed various band patterns around the expected sizes. Out of the 119 successful SSR primer pairs, only 17 pairs yielded one specific target band, as shown in Fig. 2C, and 49, 22, and 10 pairs produced two, three, and four bands, respectively (Table 2 and Fig. 2). The other 21 pairs yielded unspecific faint bands. Overall, the data indicate that over 85% of the genes remained as duplicate genes with one to three extra paralogous gene copies, thus indicating that ginseng has a highly replicated polyploidy level which may range from a tetra- to octa-paleoploidy genome. These results are similar to or greater than the polyploidy level suggested to be a natural tetraploid [56,58].

Different band numbers were detected in two landrace accessions because of their heterogeneity among individuals, as shown in Fig. 1A. One major polymorphic band was detected in four ginseng cultivars, but two clear bands were observed in two landraces, Hwangsook (2) and Jakyung (6), presumed to be derived from different alleles in two groups of individuals in the landraces because we used a DNA pool derived from 15 individual plants (Fig. 2B). Differences in band numbers were also detected in different species, such as in lane 7 in Fig. 2B and lane 8 in Fig. 2C, that may have been derived from a difference in gene copy numbers. One clear band was amplified in most accessions, but two bands were produced in a single P. japonicus plant and a P. quinquefolius DNA pool derived from five individuals, indicating that the species are heterozygous allele or included an extra paralogous gene.

Phylogenetic analysis of ginseng cultivars and related species

Because most amplicons showed multi-band profiles, genotyping was limited only to the major bands which appeared around the expected size. Variations in amplicon sizes were manipulated as unweighted and independent characteristics. Major allele frequencies were in the range of 0.2222 to 0.8889, with an average of 0.6810. The number of alleles was in the range of two to seven, with an average of 3.3143. Gene diversity and polymorphism information content ranged from 0.1975 to 0.8395 (average, 0.4568) and 0.1780 to 0.8194 (average, 0.4176), respectively (Table 2).

A phylogenetic analysis of the nine accessions was conducted using 215 allelic data points produced from 70 markers. Three clades were separated at similarity coefficients of 0.7: P. ginseng-P. japonicus clades, P. notoginseng and P. quinquefolius (Table 3 and Fig. 3).

Table 3.

Dice’s similarity coefficient matrix for nine accessions obtained from 70 marker data

	Chunpoong1)	Yunpoong1)	Gumpoong1)	Gopoong1)	Hwangsook1)	Jakyung1)	P. japonicus	P. quinquefolius	P. notoginseng
Chunpoong*	1.0000
Yunpoong*	0.7571	1.0000
Geumpoong*	0.7571	0.8429	1.0000
Gopoong*	0.8000	0.8286	0.8000	1.0000
Hwangsookjong*	0.7857	0.8571	0.8429	0.8000	1.0000
Jakyungjong*	0.8143	0.8714	0.7857	0.8714	0.8571	1.0000
P. japonicus	0.7286	0.8143	0.8286	0.7857	0.8000	0.7857	1.0000
P. quinquefolium	0.0286	0.8143	0.0143	0.0143	0.0143	0.0143	0.0429	1.0000
P. notoginseng	0.0143	0.8143	0.0143	0.0143	0.0143	0.0143	0.0429	0.2143	1.0000

1)Panax ginseng cultivars or accessions.

Fig. 3.

Dendrogram of the nine Panax accessions, six P. ginseng accessions and three relative species. Phylogenetic tree was constructed based on the genotypes of 70 markers using unweighted pair group method with the arithmetic mean clustering analysis. Bootstrap values were calculated by 1,000 replications and only significant values were denoted on the branches.

It is notable that P. japonicus clustered among the P. ginseng accessions with high similarity coefficients which averaged 0.7905, and 94.3% of the alleles of P. japonicus were observed in six P. ginseng accessions. Similarity coefficients among P. ginseng accessions ranged from 0.7571 to 0.8714 (Table 3). P. notoginseng and P. quinquefolius were divided with P. ginseng-P. japonicus clade with similarity coefficients of 0.0209 and 0.0207, respectively. The similarity coefficient between P. notoginseng and P. quinquefolius was 0.2149 which is higher than their value with P. ginseng-P. japonicus clade. The coefficient value is not coincided with the transferability value of each marker. Even though 97.1% of markers were amplified in P. quinquefolius, only 75.7% were amplified in P. notoginseng that indicate P. ginseng-P. japonicus clade is much closer to P. quinquefolius than to P. notoginseng. The biased data might be derived from genotype scoring method because only band appearance was counted and non-amplification was treated as missing. Sequence level analyses will clearly show the phylogenetic relationships of the species such as several studies based on conserved DNA sequences such as internal transcribed spacer sequences [59-62] and chloroplast DNA [60,63,64].

Reproducibility and utility of the markers

Most of the markers reported in the ginseng genome were limited to identification of individuals instead of authentication of cultivars or accessions because of the difficulty of genetic studies and the limited utility of pure inbred lines. Therefore, no inheritance study has yet been reported in ginseng. Our purpose was to develop stable and reproducible polymorphic markers which can discriminate elite cultivars and can be used for genetic mapping. Therefore, we have selected polymorphic markers between DNA pools of 15 individual plants for representing each accession to identify major polymorphic markers by overcoming the heterogeneity. Furthermore, to determine stable and reproducible inheritance of the markers, we analyzed seven of the GES markers against 51 F2 individuals resulting from a cross between ‘Yunpoong’ and ‘Chunpoong’ which were most diverse elite ginseng cultivars (Fig. 4). All the markers segregated with a good fit to the Mendelian 1:2:1 ratio for the genotype of Yunpoong homozygote:heterozygote:Chunpoong homozygote (Table 4 and Fig. 4) indicating that these inheritable and reproducible markers can be utilized for discrimination of each cultivar and for mapping using the F2 population.

Fig. 4.

Segregation of the polymorphic marker GES0019 in a F2 population between Yunpoong and Chunpoong. polymerase chain reaction products were separated by non-denaturing polyacrylamide gel electrophoresis. Lanes and L indicate DNA ladder; 1, Yunpoong; 2, Chunpoong. F2 population includes 51 F2 individuals. Y, C, and H indicate genotype of each F2 individual which is same with Yunpoong, Chunpoong, and Yunpoong/Chunpoong heterozygote, respectively. GES, ginseng expressed sequence tag-simple sequence repeat.

Table 4.

Goodness-of-fit analysis for seven markers in a F2 population between a cross of ‘Yunpoong’ x ‘Chunpoong’

Marker	Observed value			χ²-value	p-value
	Yunpoong	Heterozygote	Chunpoong
GES0003	10	30	11	1.63	0.44
GES0010	11	28	12	0.53	0.77
GES0013	18	24	9	3.35	0.19
GES0014	14	24	13	0.22	0.90
GES0015	18	26	7	4.76	0.09
GES0018	12	23	16	1.12	0.57
GES0019	16	26	9	1.94	0.38

GES, ginseng expressed sequence tag-simple sequence repeat.

There is no report on availability of segregating population in P. ginseng because of lack of pure inbred lines and genetic study. Both parental lines, ‘Yunpoong’ and ‘Chunpoong’, showed relatively high homogeneous genotypes with less than 10% of off-type alleles. Furthermore, both cultivars show distinct agricultural characteristics such as stem numbers, root shapes, disease durability and fruit colors. ‘Yunpoong’ is known as a best cultivar for high yield of roots with vigorous growth and ‘Chunpoong’ is known as a best cultivar for red ginseng processing [4]. We have identified 14 polymorphic markers between these two cultivars and seven of them showed clear mappable genotype scores for each F2 individual. Recently, we had obtained a large scale transcriptome sequence data from both parental lines using Roche GS FLX Titanium platform and 50 Gbp of whole genome sequencing data from one of the parental line, ‘Chunpoong’, using Illumina Genome Analyzer II platform [65]. Application of fast evolving next generation sequencing technology and the utility of the mapping population may promise acceleration of high density genetic mapping and complete genome sequencing of the mysterious medicinal plant, ginseng.