A Molecular Phylogenetic Study of the Genus Phedimus for Tracing the Origin of “Tottori Fujita” Cultivars

It is very important to confirm and understand the genetic background of cultivated plants used in multiple applications. The genetic background is the history of crossing between maternal and paternal plants to generate a cultivated plant. If the plant in question was generated from a simple origin and not complicated crossing, we can easily confirm the history using a phylogenetic tree based on molecular data. This study was conducted to trace the origin of “Tottori Fujita 1gou” and “Tottori Fujita 2gou”, which are registered as cultivars originating from Phedimus kamtschaticus. To investigate the phylogenetic position of these cultivars, the backbone tree of the genus Phedimus needed to be further constructed because it retains inarticulate phylogenetic relationships among the wild species. We performed molecular phylogenetic analysis for P. kamtschaticus, Phedimus takesimensis, Phedimus aizoon, and Phedimus middendorffianus, which are assumed as the species of origin for “Tottori Fujita 1gou” and “Tottori Fujita 2gou”. The molecular phylogenetic tree based on the internal transcribed spacer (ITS) and psbA-trnH sequences showed the monophyly of the genus Phedimus, with P. takesimensis forming a single clade. However, P. kamtschaticus and P. aizoon were scattered in the tree. It was verified that “Tottori Fujita 1gou” and “Tottori Fujita 2gou” were embedded in a clade with P. takesimensis and not P. kamtschaticus. Therefore, origination from P. takesimensis was strongly supported. Based on these results, molecular phylogenetic analysis is suggested as a powerful tool for clearly tracing the origin of cultivated plants.


Introduction
Crassulaceae belong to the order Saxifragales of the core eudicots. It is also called the "stonecrop family" because it mostly comprises perennial herbaceous plants and many members have fleshy leaves. In this family, the Angiosperm Phylogeny Group (APG) IV system [1] currently contains three subfamilies, namely, Crassuloideae, Kalanchoideae, and Sempervivoideae, and it includes around 29-34 genera and 1400 species. Crassulaceae is distributed worldwide and has high species diversity in South Africa, Mexico, and mountainous areas in Asia. The family lives mostly in dry locations and is often found in saline areas. Crassulaceae is an important plant group that has been developed as horticultural cultivars for various morphological characteristics and vigorous growth in dry environments.
Regarding Crassulaceae classification, Berger [2] recognized 35 genera and 15,000 species in six subfamilies, but the largest genus, Sedum L., contains over 500 species with unclear phylogenetic relationships. To understand the taxonomic limitations of Sedum, numerous approaches have been Plants 2020, 9,254 2 of 15 undertaken by many researchers. Among them, Rafinesque [3] described an independent genus Phedimus Raf. from the genus Sedum, including the two species Phedimus uniflorus Raf. and Phedimus stellatus (L.) Raf., which have five-part calyx, unequal sepals longer than the petals, five equal petals, 10 stamens, and five ovaries. The author 't Hart [4] examined new combinations in Phedimus using phylogenetic analysis. They divided Crassulaceae into two subfamilies, namely, Crassuloideae and Sedoideae, and separated the subfamily Sedoideae into tribes Kalanchoeae and Sedeae. The Sedeae were again divided into subtribes Telephiinae and Sedinae, with Telephiinae including the genera Rhodiola L., Hylotelephium H. Ohba, and Orostachys Fisch. as well as Meterostachys Nakai and Phedimus, which were newly combined. Later, Ohba [5] divided Phedimus into subgenera Phedimus and Aizoon and established the current concept of the genus Phedimus, including the variants of P. aizoon and Phedimus hsinganicus (Y.C. Chu ex S.H. Fu and Y.H. Huang) H. Ohba, K.T. Fu, and B.M. Barthol. Since then, numerous molecular phylogenetic studies on Sedinae have been continuously conducted, and Phedimus has been separated from existing Sedum species. However, researchers have often continued to use a broad concept of intermixed "Sedum" and the newly defined "Phedimus" until now, even though the monophyly of Phedimus has been proven by molecular phylogenetic analysis [6][7][8]. Therefore, we have treated the genus Phedimus in accordance with the views of 't Hart [4] and Ohba [5] here.
For a long time, Phedimus has been primarily used to develop horticultural cultivars. The increasing of various developed and cultivated plants has led to the issue of their origination due to ambiguous information about their genetic origins. For example, P. takesimensis (Nakai) 't Hart is an endemic species to South Korea and was first introduced as "Sedum takesimense Nakai" at the UK Horticultural Society magazine "Sedum Society" in the early 1990s [9,10]. P. takesimensis was called the "evergreen Sedum" in this magazine because, unlike other plants in the Aizoon group that were widespread in Europe at the time (such as Sedum aizoon L., Sedum kamtschaticum Fisch., Sedum middendorffianum Maxim., Sedum kurilense Vorosch., and S. kamtschaticum subsp. ellacombianum (Praeger) R.T. Clausen), it retains green leaves even in the winter season. Because of these characteristics, P. takesimensis has been recognized as an important material for developing a horticultural plant based on its potential value, and it has entered the European market for greening and gardening. It has also been used as road landscaping plants since it was introduced to Japan in 2011 [11] as well as sold in Japanese markets in the form of seeds or young potted plants. In the case of the newly registered cultivars of "Tottori Fujita 1gou" and "Tottori Fujita 2gou", the applicant reported P. kamtschaticus (Fisch.) 't Hart as the plant of origin to the National Korea Forest Seed and Variety Center in South Korea. However, this was controversial as the cultivated plants were morphologically closer to P. takesimensis than P. kamtschaticus ( Figure 1). In the present study, we traced and clarified the definite origin of these cultivars using a molecular phylogenetic approach to overcome the limitations of discrimination via morphological characteristic comparisons. Since the controversy over the origin of these cultivars arose, no attempts have yet been made to resolve the issue using molecular phylogenetic analysis. Therefore, we suggest that it is possible to specifically trace the cultivar origin using molecular phylogenetic analysis in the case study of Phedimus.

Sequence Variations within Phedimus
The nuclear ribosomal internal transcribed spacer (ITS) region (including ITS1, ITS2, 5.8S, and the partial 28S ribosomal gene) and the psbA-trnH spacer region of chloroplast DNA sequences of 85 individuals of the genus Phedimus were determined in this study.
Within the genus, the ITS region comprised 606-607 base pairs (bp), and the total aligned length was 608 bp. There were 471 constant sites (76.84% of all sites) and 115 parsimonious informative sites. In the aligned ITS region sequence, P. kamtschaticus and P. aizoon (L.) 't Hart showed similar nucleotide polymorphisms, although there was sequence variation within the same species. In P. takesimensis, excluding sample J6, the sequence was 607 bp in length, and "Tottori Fujita 1gou" and "Tottori Fujita 2gou" were also 607 bp. There were no species-specific variations in P. takesimensis; however, when comparing P. takesimensis sequences, they were divided into three types: α type (according to nucleotide substitution at 549 and 564 bp), β type (549 bp only), and γ type (564 bp only). "Tottori Fujita 1gou" and "Tottori Fujita 2gou" were both included in these three types ( Figure 2). P. middendorffianus (Maxim) 't Hart, which showed a similar nucleotide sequence to P. takesimensis, had two species-specific nucleotide substitutions at 342 and 578 bp, unlike other taxa.

Origin and Phylogenetic Relationship of Cultivated Phedimus
Maximum likelihood (ML) trees of the ITS and psbA-trnH regions showed the monophyly of the genus Phedimus with 100% bootstrap support and 100% Shimodaira-Hasegawa (SH)-like approximate likelihood ratio test (SH-aLRT) support (Figures 3-5).
The tree length of the ITS region was 0.3632, and the sum of the internal branch lengths was 0.2592, accounting for 71.36% of the tree length. This was divided into two clades ( Figure 3). P. aizoon and P. kamtschaticus formed Clade I with 75% bootstrap value and 90% SH-aLRT support. Clade II, comprising P. takesimensis and P. middendorffianus, was supported with 69% bootstrap value and 89% SH-aLRT support. P. middendorffianus formed a subclade within Clade II with 95% bootstrap value and 91.8% SH-aLRT support. P. takesimensis formed three subclades, which matched the three types of P. takesimensis sequence variation (α, β, and γ) of the ITS region. "Tottori Fujita 1gou", "Tottori Fujita 2gou", and Phedimus sp. were included in all three subclades of P. takesimensis. Regarding K31, K41, J6, J8, and J9, each showed a position that differed from their identification. "Tottori Fujita 2gou" were also 607 bp. There were no species-specific variations in P. takesimensis; however, when comparing P. takesimensis sequences, they were divided into three types: α type (according to nucleotide substitution at 549 and 564 bp), β type (549 bp only), and γ type (564 bp only). "Tottori Fujita 1gou" and "Tottori Fujita 2gou" were both included in these three types ( Figure  2). P. middendorffianus (Maxim) 't Hart, which showed a similar nucleotide sequence to P. takesimensis, had two species-specific nucleotide substitutions at 342 and 578bp, unlike other taxa.
In the psbA-trnH region, the length variation was 375-400 bp, and a small inversion was identified ( Figure S1). There were 383 constant sites (90.12% of all sites) and 29 parsimonious informative sites in an aligned 413bp matrix. The shortest-length sequence of 375 bp had a 7 bp ML analysis for psbA-trnH was performed in two cases: corrected and uncorrected small inversion. The ML tree length was 0.1818, and the sum of the internal branch lengths was 0.0991, which was 71.36% of the tree length. The tree Baysian Information Criterion (BIC) score was 2128.3260 ( Figure 4). The highly supported long branch of the tree (100% bootstrap value and 99.7% SH-aLRT support; taxa within the box in Figure 4) collapsed into the tree after correction of the inversed sequence ( Figure 5). The ML tree length with corrected inversion was 0.1619, and the sum of the internal branch lengths was 0.0805, which was 49.70% of the tree length. The BIC score was 1943.4605. In the tree, only P. takesimensis composed a unique clade, unlike other species, with 80% bootstrap value and 86.6% SH-aLRT support. On the other hand, P. kamtschaticus, P. aizoon, and P. middendorffianus were scattered in the tree. The genus Phedimus was separated from Sedum, the largest genus in Crassulaceae, and recombined by 't Hart [4] and Ohba [5]. However, there were uncertain phylogenetic relationships among wild species, which caused confusion in applying species names and identifying taxa. Thus, we expected that the difficulty of tracing the origin of the registered cultivars would be resolved based on the analysis of the molecular phylogenetic relationships of Phedimus.

Discussion
Molecular phylogenetic study showed that the genus Phedimus separated from Sedum was monophyletic. Although there were several exceptions, closer relationships were found between P. takesimensis and P. middendorffianus as well as P. kamtschaticus and P. aizoon in the genus. Interestingly, three haplotypes were found in P. takesimensis, and all cultivars were embedded in those types.

The Origin of Cultivar "Tottori Fujita"
The controversy surrounding the origin of the two Phedimus cultivars "Tottori Fujita 1gou" and "Tottori Fujita 2gou" arose due to the lack of an established correct phylogenetic relationship of Phedimus. Although the reliance of cultivar origin on external morphological characteristics tends to depend on the arguments of the applicant, there is a problem with the authenticity of the claims due to the absence of definite classification characteristics, such as morphological variations within the species of the genus Phedimus.
The genus Phedimus was separated from Sedum, the largest genus in Crassulaceae, and recombined by 't Hart [4] and Ohba [5]. However, there were uncertain phylogenetic relationships among wild species, which caused confusion in applying species names and identifying taxa. Thus, we expected that the difficulty of tracing the origin of the registered cultivars would be resolved based on the analysis of the molecular phylogenetic relationships of Phedimus.
The results showed that the cultivars "Tottori Fujita 1gou" and "Tottori Fujita 2gou", which were registered as new cultivars at the National Korea Forest Seed and Variety Center, formed a clade with P. takesimensis. The results also strongly indicated that the parental plant samples (marked as Phedimus sp. in data) submitted by the applicant as the species of origin P. kamtschaticus when registering these two new cultivars at the National Korea Forest Seed and Variety Center were also embedded in the P. takesimensis clade rather than P. kamtschaticus. Therefore, contrary to the information provided by the applicant, it was demonstrated that the cultivars "Tottori Fujita 1gou" and "Tottori Fujita 2gou" originated from P. takesimensis.

Application of Molecular Phylogenetic Analysis to Evaluate Misidentified Plants
We could also reidentify several Phedimus samples that were used in the present study and suspected of misidentification. For example, in the cases of J6 and J7, they were germinated and grown from purchased P. takesimensis seeds, which are currently sold at seed stores in Nagano Prefecture, Japan. However, unlike other P. takesimensis samples, which have a 1bp gap at 515 bp in the ITS region, J6 and J7 have an additional 1bp gap at 398 bp and a 1bp nucleotide substitution at 113 bp. This shows the same pattern as P. kamtschaticus and P. aizoon; however, not all sequence variances are perfectly consistent with these two species. Furthermore, samples K41, J8, and J9, which were identified as P. kamtschaticus, were exactly the same as the J6 and J7 sequences, and they formed a single clade in all trees. Therefore, it is reasonable to recognize them as P. kamtschaticus or P. aizoon rather than P. takesimensis. Accordingly, Phedimus seeds released on the market should be handled carefully due to the probability of misidentification to avoid confusion.

Hybridization Leading to Discordance on the Phylogenetic Tree
Hybridization is an important mechanism that obtains new genotypes through a combination of different genomes. And then, the plant species is generally diversified into other species that are evolutionarily competitive [12]. Some plants are also generated by selective breeding that suits useful values, such as building a new lineage that maximizes the specific character manifestation [13]. In natural conditions, when an interspecific hybridization has been occurred, it may be used to recognize a new species and give a caution to carefully define the boundary of species [14]. Therefore, researchers have investigated hybridization in wild species and cultivars by morphological, physiological, and genetic approaches [15][16][17].
Hybrids originating from wild species of the genus Phedimus have been reported in Korean taxa [18,19]. In addition, various hybrid plants have been reported in the subspecies P. aizoon, and it has caused a lot of confusion in the recognition of Japanese species [20]. This is particularly the case for P. kamtschaticus and P. aizoon, which are generally identified by their leaf type. The boundary of species is still unclear because there is no apparent morphological difference between them, and interspecific hybrids have been reported.
The present study also showed that these two species were not distinguishable from each other in the phylogenetic tree. Clade I containing P. kamtschaticus and P. aizoon and the clade of P. middendorffinus were clearly recognizable in the ITS tree ( Figure 3). However, all three species were scattered in the psbA-trnH tree ( Figure 5). It is assumed that hybridization has occurred among species in Phdimus except for P. takesimnesis.
The K52 and K54 samples (P. middendorffianus) formed a subclade with another P. middendorffianus in the ITS tree. However, these samples were separated from them in the psbA-trnH tree and formed one subclade with several P. kamtschaticus (Figures 3 and 5). This finding made us speculate that the K52 and K54 samples had different maternal origin than the rest of the P. middendorffianus samples.
In the case of sample K31 (P. kamtschaticus), a hybrid origin was also suspected because it showed a similar pattern to P. takesimensis in the ITS region but a similar tendency to P. kamtschaticus in the psbA-trnH region.
Meanwhile, the J10 sample (P. aizoon) was included in Clade I of the ITS tree, but in the case of the psbA-trnH tree, the J10 sample was separate from the clade that included all the Phedimus samples (Figures 3 and 5). Besides, a unique sequence variation (A) was found at 337 bp position of the psbA-trnH sequence of J10. This suggests the possibility of de novo mutation in the J10 sample.

Molecular Marker Development for Further Study
In addition, we found the species-specific variations for P. middendorffianus at 342 bp in the ITS region ( Figure 2) and at 275 bp in the psbA-trnH region for P. takesimensis ( Figure S1). It is anticipated that these polymorphic sequence sites may be used to develop molecular markers to identify the two species among the different Phedimus species.
In the horticultural and forestry industries, all countries encourage and promote development and sale based on cooperation with the International Union for the Protection of New Varieties of Plants (UPOV). However, the importance of providing accurate information for cultivated plants has often been ignored, including the history of their generation and genetic background to evaluate the characteristics of plants in the market. In fact, in most cases, a plant can be accepted as a new cultivar if it maintains just one unique characteristic compared to the plant of origin. To establish a system to protect plant cultivars, the genetic background must be provided along with the morphological features in the future. Based on the results of this study, it is suggested that the technique of molecular phylogenetic analysis be used to trace the definite origin of cultivars and to define the phylogenetic boundary between species. Molecular phylogenetic analysis can be used as a powerful tool to produce a genetic backbone tree for tracking the origin of cultivars.

Plant Materials and Sampling
In total, 85 individuals (69 from Korea and 16 from Japan) were used in this study, including the "Tottori Fujita 1gou" and "Tottori Fujita 2gou" cultivars, which were provided from the National Korea Forest Seed and Variety Center, and four species of Phedimus, which occur in wild habitats in Korea and Japan (Figure 1). In addition, the sequence data of the related taxa, i.e., Pseudosedum and Rhodiola [21][22][23], were downloaded from the National Center for Biotechnology Information (NCBI) database (Table 1).

DNA Extraction, PCR Amplification, and Sequencing
Total genomic DNA was extracted from silica-gel-dried leaf tissues using a modified CTAB method [24] or the DNeasy Plant Mini Kit (QIAGEN, Germany) following the manufacturer's instructions. The quantity and quality of the extracted DNA were analyzed using 1% agarose gel electrophoresis with 1X Tri-acetate-EDTA (TAE) buffer and a spectrophotometer (Thermo Scientific™ NanoDrop 2000, Thermo Fisher Scientific, Waltham, MA, USA). We amplified the ITS region (rDNA) and the psbA-trnH region (cpDNA). To amplify the target regions, we used primer pairs of ITS1 and ITS4 [25] and of psbA and trnH [26] ( Table 2). The PCR reaction for the ITS region was initialized at 95 • C for 1 min, followed by 35 cycles of denaturation at 95 • C for 30 s, annealing at 51 • C for 30 s, and extension at 72 • C for 1 min; the final extension step was performed at 72 • C for 5 min. For the psbA-trnH region, PCR was initialized at 95 • C for 1 min, followed by 35 cycles of denaturation at 95 • C for 30 s, annealing at 55 • C for 30 s, and extension at 72 • C for 1 min; the final extension step was performed at 72 • C for 5 min. The PCR products were run on 1% agarose gels in 1X TAE buffer and purified using the Expin PCR SV Mini Kit (GeneAll, Seoul, South Korea). Sequencing was carried out by the 3730XL Automated DNA Sequencing System (Applied Biosystems, Foster City, CA, USA).

Molecular Phylogenetic Analyses of ITS and Plastid DNA Regions
The nuclear ITS region sequence and plastid noncoding regions, psbA-trnH, were used to compose the phylogeny. Geneious 7.1.9 [27] was used to assemble the DNA sequences resulting from the PCR products. Sequences were initially aligned with the MAFFT v7.017 alignment program [28] in Geneious 7.1.9 using the default parameter values. Then, all sequences were manually checked, and amendments were directly made.
Phylogenetic analysis was performed using maximum likelihood methods with W-IQ-TREE 1.6.11 [29], and gaps were treated as missing data. The best evolutionary substitution model identified by the Bayesian information criterion [30] was selected at TIM3e + G4 for the ITS region, F81 + F + I for the psbA-trnH region after correcting for the small inversion, and F81 + F + G4 for the psbA-trnH region with uncorrected inversion using W-IQ-TREE 1.6.11. The supporting value for each clade was estimated from 1000 bootstrap replicates [31], and we performed 1000 replications for the SH-aLRT as the branch test [32].