ORIGINAL ARTICLES
Estimation of Phylogeny of Nineteen Sedoideae Species Cultivated in Korea Inferred from Chloroplast DNA Analysis
2018 Volume 87 Issue 1 Pages 132-139
Details
2018 Volume 87 Issue 1 Pages 132-139
The genetic diversity and relationships among plants belonging to the subfamily Sedoideae (Crassulaceae), some of which are indigenous to Korea or introduced from other countries, were determined using chloroplast (cp) nucleotide sequence analysis. To analyze genetic diversity and variation among 19 plants including species belonging to Sedum, Hylotelephium, and Phedimus, the tRNA-Leucine gene (trnL [UAA]) and adjoining spacer in chloroplast DNA (cpDNA) were sequenced and compared across species. Species were divided into two main groups based on the cpDNA sequence comparison. The generated phylogeny indicated that many native Sedum species had diverged from S. album. Members of the Phedimus and Hylotelephium species, and several Sedum species analyzed here, clustered distinctively in different groups. Using cpDNA sequence analysis, we successfully discriminated Sedoideae plants cultivated in Korea from each other, even at the intraspecific level, and the results were reflective of the morphological and biogeographical characteristics. These findings could be useful for classifying samples for proper naming, choosing breeding materials for new cultivars, or identifying species for conservation of horticultural crop resources.
The Crassulaceae family includes 35 genera in six subfamilies that are divided into two lineages, a Crassula lineage and a Sedum lineage. The latter lineage is composed of three subfamilies (Echeverioideae, Sedoideae, and Sempervivoideae) (Berger et al., 1930). Despite the general consensus that a substantial part of it is highly artificial, Berger’s classification for Crassulaceae has been widely used because of its comprehensiveness and great practical value (Van Ham and ’T Hart, 1998). Sedum, the largest genus of Crassulaceae, is a polyphyletic genus that encompasses much of the morphological diversity present in the family. The Sedum genus is, however, still poorly characterized and requires further investigation (Mort et al., 2001). Plants of the genus Sedum occur across the Northern Hemisphere, including in the Mediterranean countries, Mexico, and Far Eastern countries like China, Japan, and Korea (Stephenson, 1994). Sedum is known to contain about 470 species (Mayuzumi and Ohba, 2004), 22 of which are indigenous to Korea (http://www.nature.go.kr). The taxonomy of the subfamilies in Crassulaceae is based on the number and arrangement of floral parts, the degree of sympetaly, and the phyllotaxis, with these factors featuring in each of the larger genera (Van Ham and ’T Hart, 1998). Plants in the genus Sedum have succulent leaves, and the flowers usually have five petals, and only occasionally four or six. These features, along with their chromosome counts, define the genus Sedum (Ohba, 1977). While the basic genome chromosome number in Sedum species is eight (x = 8), chromosome doubling has occurred repeatedly, and some Sedum species contain x = 16, 32, and 64. Polyploidy appears to be common throughout the Sedum genus, and some members also show euploidy with chromosome counts of x = 7 or 31 (Stephenson, 1994). Especially in Korea, Sedum taxonomy is mainly based on morphological characteristics, and there is currently no accurate method of classifying Sedum species by genetic variation and divergence from the closest relative. Hylotelephium and Phedimus species have been previously classified as Sedum species (e.g., Sedum telephium) (Ohba, 1977). However, there is still confusion in the common or public naming system in Korea. Sedum species, such as S. sarmentosum, which is known as ‘Dolnamul’ in Korean, have been used traditionally as food. Recently, Sedum species were recognized as useful plant resources for landscape architecture or materials for ground cover in Korea. Despite these uses, a classification system for Sedum and related species based on genetic identification has not been properly established in Korea. To avoid naming confusion in the commercial market and improper identification, we aimed to identify indigenous Sedum and related species and those previously imported and now cultivated in Korea using molecular genetics data from chloroplast DNA (cpDNA) sequences to facilitate the conservation of horticultural resources and breeding materials.
Since DNA sequence data has led to reconstructions of classifications that often conflict with traditional taxonomy, cpDNA sequences have proven to be a valuable tool for phylogeny construction in plants because of their small size, high copy number, and maternally inherited nature (Chase and Palmer, 1989; Dong et al., 2013; Wu et al., 2010). The simultaneous alignment of many nucleotide sequences is an essential tool in molecular biology to detect homology among species, and the use of cpDNA in multiple sequence alignment to examine relationships and evolutionary changes is advantageous. To resolve phylogenetic relationships at several taxonomic levels, the maturase K (matK) gene of cpDNA was used (Soltis et al., 1996). Recently, the phylogeny of the Dendrobium species in Thailand was reconstructed using data from a partial matK gene and rDNA internal transcribed spacer (ITS) sequences (Srikulnath et al., 2015). Using sequences in a noncoding region in cpDNA, Crassulaceae species have been analyzed by several research groups for their molecular evolution or discrimination within the genus (Van Ham and ’T Hart, 1998; Van Ham et al., 1994). In particular, Mayuzumi and Ohba (2004) evaluated the phylogenetic positions of Eastern Asian Sedoideae plants using trnL-F (spacers between tRNA-Leucine and Phenylalanine) data. Recently, trnL-F regions were applied to estimate the relationships and track the origins among 31 hostas (Lee and Maki, 2015). Random amplified polymorphic DNA (RAPD) has also been used previously to set conservation priorities and design management strategies for taxa in Sedum species (Olfelt et al., 2001). Korean Sedum plants, in particular, are thought to have diverse phenotypes at both the species and interspecies levels (Kwon and Jeong, 1999). When investigating the relationships between 31 ecotypes of S. sarmentosum distributed in Korea, researchers found that the results from RAPD markers significantly correlated with morphological characteristics (Kim et al., 2008). Specifically, our study included 19 newly generated cpDNA sequences of 11 Sedum, 3 Hylotelephium, and 5 Phedimus species, and used molecular markers to examine the relationships between them. Sedoideae species used in this study either grow naturally or have been recently introduced to Korea and are cultivated. To discriminate native species from imported species and to develop hybrid cultivars showing useful new traits in breeding programs, the genetic characteristics and relationships need to be assessed. Therefore, we aimed to estimate the genetic relationships between Sedoideae plants from various origins by examining cpDNA sequences. The information obtained in this study could be useful for discriminating the native plants from each other for identification during resource conservation programs. The evaluation of molecular cpDNA markers would allow the development of proper discrimination methods among Sedoideae plants cultivated in Korea and their application for various practical uses, including in breeding programs.
Nineteen sample plants were obtained from the National Institute of Horticultural and Herbal Science located in Gyeonggi province, Korea, in May, 2014. Samples comprised eleven Sedum species, five Phedimus species, and three Hylotelephium species (Table 1). The closely related genera Aeonium and Greenovia were used as outgroups for the analysis. The species used and their overall characteristics are shown in Table 1. As the imported Sedoideae plants have been cultivated in Korea over several years, they were used in addition to the indigenous plants. Genomic DNA was extracted using an DNeasy Plant Mini Kit (Qiagen, USA). For each sample, 100 mg FW of plant leaves was used. The DNA quantity and quality were measured by absorbance at 260 nm with an ND 2000 spectrophotometer (Nanodrop Technologies, USA), and isolated DNA was checked by visualization on a 1% agarose gel with ethidium bromide (EtBr).
Morphological, phenological, and biogeographical characteristics of Sedoideae species used in this study.
To compare cpDNA sequence variation, regions consisting of a tRNA-UAA gene (trnL) and trnL-trnF intergenic spacer (trnL-F) located in the large single-copy region of cpDNA were amplified. Samples were analyzed using primers encompassing trnL and flanking two intergenic spacers, trnL-F and trnT-L, as previously described by Taberlet et al. (1991). PCR amplification reactions contained 10–20 ng of genomic DNA, 20 pM of each primer (Taberlet et al., 1991), 10 mM Tris-HCl at pH 8.3, 50 mM KCl, 1.5 mM MgCl2, 0.25 mM each of dATP, dGTP, dCTP, and dTTP, one unit of Taq DNA polymerase (Takara, Japan), and sterilized water in a total volume of 20 μL. PCR amplifications were performed at 94°C for 3 min, followed by 35 cycles at 94°C for 1 min, 50°C for 1 min, and 72°C for 2 min, and finally incubated at 72°C for 5 min. PCR fragments were electrophoresed on a 1% agarose gel in 0.5×TAE buffer at 135V for 18 min and then stained with EtBr. Amplification products (60 ng) were ligated into pGEM T-easy vector (Promega, USA), and the ligation mixture was transformed into Escherichia coli, strain DH5α. The recombinant plasmids were extracted by alkaline lysis and clones sequenced (Macrogen Inc., Korea) using T7/SP6 universal primers. A BLASTN search program in the National Center for Biotechnology Information (NCBI) database confirmed that these sequences were cpDNA.
Multiple alignment and generation of phylogeny based on cpDNA sequencesNucleotide sequences of the cpDNA region in each sample were used to establish a distance matrix for analysis. Sequences of two outgroup species, A. viscatum and G. aizoon, were added to create a multiple alignment matrix; the sequences of A. viscatum and G. aizoon to be compared were retrieved from the NCBI database (Table 2). The cpDNA sequences were aligned using the CLUSTAL W multiple sequence alignment algorithm in MEGA6 with the following parameters: gap-opening penalty, 15; gap extension penalty, 6.66; transition weight, 0.5; delay divergence cut-off value, 30% (Tamura et al., 2013). A pairwise distance matrix was generated using the Kimura 2-parameter model (Kimura, 1980) and a phylogenetic tree was constructed using the Maximum-Likelihood (ML) method. Each bootstrap test was performed with 1000 replicates.
Nucleotide length analyzed from the region of the trnL(UAA) gene and trnL-F spacer located in the cpDNA of Sedoideae plants used in this study.
Newly determined sequences of plant samples used in this study were approximately 1.4 kb long, but only the region of one spacer and trnL (UAA) gene of approximately 800 bp was investigated and compared between plants (Table 2). Sequences obtained from the cpDNA regions in 19 plants were analyzed and assigned the accession numbers shown in Table 2. Variations in the size of the sequenced region were observed with nucleotide site variation. The sequence of this region has been frequently used to investigate interspecific relationships in many species belonging to Crassulaceae and other plants because of the high and low intraspecific variations that occur at intervals in the noncoding regions (Taberlet et al., 2007). The length of trnL ranged from 568 bp in S. sexangulare to 617 bp in S. album (Table 2). The length of the trnL-F ranged from 246 bp in three Phedimus species to 286 bp in H. spectabile (Table 2). Most of the variations in length were caused by nucleotide insertion events, including some serial insertions. Serial insertions of nucleotides in the trnL sequenced region were found in S. album, where 19 and 29 nucleotides were inserted, respectively, in the middle of the gene (data not shown). An insertion of 23 nucleotides was observed in the trnL-F region of H. spectabile. Interestingly, serial deletions were also observed. Sixteen nucleotides in the trnL-F were missing at the same position in P. kamtschaticus, P. middendorffianus, P. takesimense, and in two species of P. spurius. Yang et al. (1999) theorized that this phenomenon was caused by rapid evolution leading to the removal of non-functional sequences. In this study, spacer trnL-F was relatively uniform in size, unlike the trnL gene (Table 2). Multiple-alignment generated a corresponding matrix according to the algorithmic calculations used. Distance matrix analysis from cpDNA sequences were performed using the estimated sequence divergence values by the K80 substitution model in the MEGA6 program (Table 3). According to the distance matrix, genetic distance was observed from 0.002 between P. kamtschaticus and P. takesimense to 0.108 between S. oryzifolium and H. spectabile. The ML tree-building method was used to construct a phylogenetic tree. The clustering of species in the phylogenetic tree showed two main clades (Fig. 1). Clade I included three Hylotelephium species, five Phedimus species, S. rupestre, and outgroup species, while clade II included only nine Sedum species. In the phylogenetic tree, the compared Sedum species diverged from S. album, except for S. rupestre (Fig. 1). Interestingly, S. rupestre was clustered differently with other Sedum species, but it was clustered in the main group comprising Hylotelephium and Phedimus species (Fig. 1). S. rupestre is also known as ‘Petrosedum rupestre’ or ‘S. reflexum’ and this may explain its separation from other Sedum species. Five genera, Hylotelephium, Petrosedum, Phedimus, Rhodiola, and Sinocrassula, are thought to be separated from the genus Sedum (Horvath, 2014). In the phylogenetic tree, two different cultivars of P. spurius clustered together, as did S. sarmentosum and S. lineare. In Korea, these latter two species have been called ‘Dollamul’, meaning stone plant. The phylogenetic tree showed that three Phedimus species, P. kamtschaticus, P. middendorffianus, and P. takesimense, were grouped together and these species are classified as the ‘Kirincho’ group in the common classification method used in Korea. The ‘Kirincho’ group comprises nine species, including P. aizoon and its two varieties and P. kamtschaticus, P. takesimense, P. latiovalifolium, P. middendorffianus, P. zokuriense, and P. selskianum according to the classification of Korean natural resources database (http://www.nature.go.kr). However, in some studies and in commercial markets, these species are still typically named and recorded as Sedum, not Phedimus, despite the fact that these are traditionally thought to be different from other types of Sedum plants. Hylotelephium species are easily distinguished from other Sedum species by their taller upright and leafy stems (Horvath, 2014). In addition, the three species of Hylotelephium used here are indigenous to Korea, and these three Hylotelephium species were grouped together in Group I (Fig. 1). In the main group II, outgroup species were isolated from other Sedoideae species and positioned at a basal level.
Pairwise distances of Sedum and related Sedoideae species estimated by the Kimura 2-parameter model.
Phylogenetic tree generated using the Maximum-Likelihood method based on sequences of the trnL(UAA) gene and trnL-F intergenic spacer of the analyzed Sedoideae species including two outgroups. Numbers at nodes indicate corresponding bootstrap values. Leaf morphology and distributional features are represented by symbols as characterized above.
The habitats of Sedum species are widely distributed throughout the Northern hemisphere, with species traditionally classified by morphological characteristics. However, some ambiguity still exists in the conventional classification and naming of Sedum species, especially in Korea, where both indigenous and species introduced from other countries exist. Therefore, standardized methods for Sedum classification are required. In some cases, discrepancies between RAPD analysis and morphological based classification have been observed, especially in polyploid plants, because of the presence of multiple alleles and environmental factors. Furthermore, cytological criteria for Sedoideae systematic treatment are not available due to the high degree of diversity (Mayuzumi and Ohba, 2004). A sequence-based method was therefore applied to generate the Sedum phylogeny in this study.
According to the distance matrix, S. oryzifolium and H. spectabile showed the highest distance values (Table 3). This indicates that there may be considerable genetic distance between these species as indicated by the differences in leaf morphology, type of phyllotaxis, and phenology (Table 1). Phedimus spurius and its cultivar ‘Tricolor’, which showed relatively low distance values, are very similar in their morphology except in their leaf colors. In contrast to the variegated leaf color of P. spurius ‘Tricolor’, P. spurius leaves are uniformly green. Based on the matrix, a phylogeny was generated, showing two major clades. Distinctive topological features of the tree were a basal branching of S. album and monophyletic grouping of other Sedum species with the exception of S. rupestre (Fig. 1). Using cpDNA genome sequence analysis of Sedum, Van Ham and ’T Hart (1998) analyzed four species, three of which were also used in the present study (S. album, S. acre, and S. sarmentosum). Interestingly, among the four species in that previous study, S. acre and S. sarmentosum were the least related to each other, consistent with our generated phylogeny. This may be due to the different biogeographic distribution, considering that S. sarmentosum is indigenous to Asian countries, including Korea, while S. acre originated from Europe (Table 1). Sedum album was shown to be distant from the other Sedum species in the phylogeny according to cpDNA (Van Ham and ’T Hart, 1998), which also consistent with our findings (Fig. 1). Additionally, Phedimus species were always grouped with Hylotelephium species in this study, with reliable bootstrapping values (50% majority rule), consistent with the cpDNA genome analysis by Van Ham and ’T Hart (1998). Recently, cpDNA variations have been used to trace the origin of Hosta cultivars by Lee and Maki (2015) who previously analyzed two noncoding regions of cpDNA in wild populations of Hosta to identify specific regional features (Lee and Maki, 2013). Based on this analysis, they reported that the genetic diversity of wild species was lost, and hypothesized the occurrence of spontaneous mutations. The RAPD-based grouping of 31 S. sarmentosum ecotypes distributed in Korea was strongly related to morphological characteristics (Kim et al., 2008). This pattern has been found in many previous studies of Sedum phylogeny conducted using RAPD analysis (Kwon and Jeong, 1999), matK sequence data (Mort et al., 2001), and chloroplast data (Van Ham and ’T Hart, 1998). In the matK sequence analysis by Mort et al. (2001), the phylogeny revealed the polyphylic nature of Sedum species, with S. sexangulare, S. oryzifolium, and S. sarmentosum classified into the Acre clade in the Sedoideae subfamily, while S. rupestre was included in the Sempervivum clade in the Sedoideae subfamily. The polyphyletic position of S. rupestre, as observed in this study, may be explained as follows—in morphological, phenological, and biogeographical contexts, S. rupestre was distinguished with other Sedum species classified in the Acre clade. Based on cytogenetic features, species in the Sempervivum clade had a basic chromosome number of 28 (x = 28). However, the Sedum species examined in this study are classified in the Acre clade, and have various chromosome numbers, although the most common number is x = 10 (Mort et al., 2001; Uhl and Moran, 1973). In Korea, P. takesimense, confined to the island ‘Uleung-do’ in Korea, has been reported to have a similar phenotype to P. aizoon, and has similar branching patterns to P. kamtschaticus (Kwon and Jeong, 1999). However, P. kamtschaticus is decumbent while P. takesimense is procumbent. Our study also showed similar grouping features to this earlier report (Fig. 1). The phylogeny also indicated that groupings could reflect morphological characteristics, especially in leaf cross-section types. Sedum species with a flat leaf type can be grouped in a minor group of the main group II (Fig. 1). Sedum species distributed in the European region can also be differentiated from Asian and American originated species (Fig. 1). Overall, our study showed that Sedum species could be characterized by geographic origin or native habitat (Table 1; Fig. 1). These results, based on the cpDNA analysis (Fig. 1), suggest that phenological characteristics affected the genetic relationships between these species.
Classifications for pragmatic reasons have sometimes confused the true phylogeny or origin of plants, providing incorrect information for crossing, breeding, and other uses. In this study, data from cpDNA analysis supported the phylogenetic linkages among 19 species in the Sedoideae subfamily. In conclusion, the phylogeny generated in this study using species cultivated in Korea indicated that cultivated plants such as Phedimus or Hylotelephium classified in the Telephium clade could be distinctively differentiated from most Sedum species in the Acre clade, and classification using cpDNA analysis can be explained based on the morphological, phenological, and biogeographical features. This study found that cpDNA sequences obtained from nineteen Sedoideae species provided phylogenetic information that addressed the polyphylous (Sedum) and monophylous (Hylotelephium and Phedimus) nature of, and genetic variability among, the samples analyzed. However, further sampling with exact sampling locality information and further investigations combining nuclear information such as rDNA sequences or chromosome analysis are needed to better understand taxonomy and evolutionary divergence. We hope that the results and techniques revealed in this study will be useful in providing a better means of efficiently and accurately clarifying the taxonomic relationships among various Sedum species and related species that are native to Korea or introduced from abroad. Moreover, the DNA sequences analyzed here could be used as molecular markers for identification in breeding programs for new hybrid cultivars and for the conservation of horticultural crop resources.
We appreciate the support of Dr. Seung Won Han (NIHHS).
