Papaveroideae (Papaveraceae) consists of 23 genera and approximately 240 species, characterized by having milky or colored latex sap, sepals fully enclosing the bud, actinomorphic corollas without spurs, numerous stamens, and a multicarpellate, usually unilocular gynoecium (Kadereit, 1993). Members of Papaveroideae are usually herbs and are distributed mainly in the Northern Hemisphere, extending into Central and South America (Kadereit, 1993). The subfamily was traditionally recognized as a separate family, Papaveraceae (Cronquist, 1988; Takhtajan, 1997; Kim, 2007), but Angiosperm Phylogeny Groups II, III, and IV (The Angiosperm Phylogeny Group, 2003, 2009, 2016) adopted a broad circumscription of Papaveraceae to include Fumariaceae and Pteridophyllaceae (Thorne, 1992; Mabberley, 2008). The more inclusive circumscription of Papaveraceae is supported by recent phylogenetic analyses using molecular and morphological data (Hoot et al., 1997, 2015; Wang et al., 2009), and it has been used in a broad range of recent literature (Ohwi, 1965; Zhang et al, 2008; Simpson, 2010; Hannan and Clark, 2011).
The diversity of Papaveroideae in Korea is relatively low with eight species in the four genera of Papaver L, Chelidonium L., Hylomecon Maxim., and Coreanomecon Nakai (Kim, 2007). The genus Papaver consists of approximately 80 species distributed in Central and southwestern Asia, with minor representatives in East Asia, Central and Southern Europe, and North Africa (Kadereit, 1993). Papaver somniferum L. is the best known species given its use in the production of opium, the dried latex from the fruits. Opium is the source of many important pharmaceutical drugs, including morphine, heroin, thebaine, codeine, and papaverine. Four species, including cultivated P. somniferum, are distributed in Korea. Native species of Papaver in Korea, P. coreanum Nakai and P. amurense (N. Busch) N. Busch ex Tolm., are distributed in the northern part of Korea, and their phylogenetic relationships have remained unclear. Owing to the importance in forensic botany to identify P. somniferum, the utility of DNA regions, in this case the nuclear Internal Transcribed Spacers (ITS) regions, chloroplast trnL-trnF regions and ISSR markers, were evaluated (Lee et al., 2010). These studies, however, did not include any native species of the Korean Papaver.
Chelidonium is a monotypic genus distributed in Europe and East Asia, and it is commonly distributed in Korea. Hylomecon is also a small genus with two species in East Asia, and H. vernalis Maxim. is commonly found in moist places in forests in Korea. Coreanomecon is a monotypic and endemic genus in Korea distributed mainly in the southern regions (Son et al., 2012). Coreanomecon is morphologically similar to Hylomecon by producing red latex, and it is easily distinguished from Chelidonium, which produces yellow latex. The taxonomic status of Coreanomecon has been controversial. Some authors merged Coreanomecon into Chelidonium (Ohwi, 1953), and others merged it into Hylomecon (Park, 1974; Kadereit, 1993). Palynological studies (Lee and Kim, 1984) and a cladistic analysis of the morphological characters (Kim et al. 1999) have suggested that Coreanomecon is distinct from Chelidonium and Hylomecon. The phylogenetic relationship of Coreanomecon according to molecular data have remained unclear.
DNA barcoding is a molecular technique which is used to identify a specimen with DNA sequences of short regions (Hebert et al., 2003; Kress et al., 2005; Xiang et al., 2011). The rbcL and matK regions in the chloroplast (cpDNA) genome are widely utilized in angiosperms (Hebert et al., 2004; CBOL Plant Working Group, 2009). The utility of other regions of cpDNA, such as trnH-psbA, atpF-atpH, and psbK-psbI, have been shown to be useful for increasing the resolution power (Lahaye et al., 2008; Kim et al., 2014). The ITS regions in nuclear ribosomal DNA have also been used to supplement cpDNA barcodes (Tripathi et al., 2013). DNA barcode regions have been primarily used for identifying unknown specimens at the species level. This type of molecular identification of a species is particular useful when the materials to be identified are fragmentary without having sufficient structures, such as flowers and fruits. Papaveroideae provides an excellent case, as it includes P. somniferum, often cultivated in Korea. A regional DNA barcode database for all known species is thus useful for not only applied fields, such as forensic botany, trade control at customs offices, conservation biology, and for the utilization of plant resources in bio-industries, but also for discovering cryptic species that are difficult to differentiate with morphological features and for reconstructing the phylogenetic relationships of the group of interest.
The objectives of the present study are (1) to infer the phylogenetic relationship of Coreanomecon, (2) to examine the implications of the molecular phylogeny on the evolution of the morphological characters of Coreanomecon, and (3) to assess the level of differentiation of the species and evaluate the resolution power of DNA barcodes to identify a species of Papaveroideae in Korea, with an emphasis on species of Papaver.
Materials and Methods
Taxon sampling
Analyses of Papaveroideae were conducted in three tiers. Analysis 1 includes Korean taxa only to evaluate the level of molecular differentiation among the Korean taxa. All but one species of Papaveroideae distributed in Korea were included in this study, with a total therefore of seven species in four genera (Table 1). Papaver amurense, distributed in North Korea and in northeast China, including Inner Mongolia, was not included because samples were not available. Two to six accessions per species were included to represent their distributional range, except for another North Korean species, P. coreanum Nakai, in which one sample from Baekdu Mt. was included. All materials for molecular studies were obtained from herbarium specimens deposited at the National Institute of Biological Resources (KB).
Analyses 2 and 3 expanded taxon sampling to infer the phylogenetic relationships of Coreanomecon and P. coreanum accurately. The chloroplast data from groups closely related to Coreanomecon, in this case Bocconia L., Dicranostigma Hook. f. & Thomson, Glaucium Mill., Eomecon Hance, Macleaya R. Br., Sanguinaria L., and Stylophorum Nutt., were obtained from GenBank and added to our data set for Analysis 2. Analysis 3 included the nuclear ITS sequences of the tribe Chelidonieae, including Chelidonium and Hylomecon from foreign populations and those of the Papaver clade representing major subclades within the genus (Carolan et al., 2006). Analysis 3 was conducted to infer the placement of Coreanomecon and P. coreanum, as this has not been accomplished previously.
Gene sampling
Total DNA was extracted from dried leaves of herbarium specimens using a DNeasy Plant Mini Kit (QIAGEN, Hilden, Germany). Three DNA barcode regions, specifically chloroplast rbcL and matK regions and nuclear ribosomal ITS regions, were amplified via polymerase chain reactions (PCR). Primers for the cpDNA barcodes were published in previous studies (Soltis et al., 1992; Sang et al., 1997; Cuénoud et al., 2002; Kress and Erickson, 2007) and are summarized in Kim et al. (2014). The ITS regions were amplified using the primers its6 and its9 (Youm et al., 2016). For each PCR reaction, 1 μL of total DNA was included in a 20 μL reaction mixture with Solg EF-Taq DNA polymerase (Solgent, Daejeon, Korea). Amplification of the target regions was conducted with a Veriti thermal cycler (Applied Biosystems, Foster City, CA, USA) under the following conditions: initial denaturation at 95°C for 2 min, 35 cycles each at 95°C for 30 s, at 55°C for 30 s, and at 72°C for 1 min, followed by a final extension at 72°C for 5 min. PCR products were examined on a 1% agarose gel in 1X TBE buffer, purified, and were sent to Solgent (Daejeon, Korea) or Macrogen (Seoul, Korea) for sequencing, where the sequencing reaction was prepared using the same primers used in PCR with BigDye Terminator v3.1 Cycle Sequencing Kits (Applied Biosystems). Sequences were determined using a 3730xl DNA analyzer (Applied Biosystems). Sequences were edited in Sequencher version 5.0 (Gene Codes Corporation, Ann Arbor, MI, USA), aligned manually.
Phylogenetic analysis
The chloroplast DNA barcode data were concatenated and analyzed simultaneously. The ITS data were analyzed separately and in combination with the cpDNA data. Dicentra and Corydalis were used as outgroups. Phylogenetic analyses were conducted using the maximum parsimony (MP) method in PAUP* (Swofford, 2002). All characters were treated as unordered and were weighted equally in the MP analyses. Gaps resulting from multiple alignments of indels were treated as missing data. Heuristic searches were used with 100 replicates of random sequence additions with tree bisection-reconnection branch swapping, with all of the best trees saved at each step (MulTrees). Bootstrap analyses (Felsenstein, 1985) of 1,000 pseudoreplicates were conducted with heuristic searches, a simple sequence addition in PAUP* to evaluate the support for each clade.
Bayesian phylogenetic analyses (BI) were performed with the program MrBayes version 3.2.5 (Huelsenbeck and Ronquist, 2001). A Markov chain Monte Carlo (MCMC) algorithm was employed for 2,000,000 generations, sampling trees every 100 generations, with four independent chains running simultaneously. Priors were set based on the Akaike information criterion (AIC) using modeltest (Posada and Crandall 1998). The first 4,000 trees (400,000 generations) were discarded as “burn-in,” and the remaining trees for which the log-likelihood values had reached a plateau were imported in PAUP* to calculate the posterior probability of each clade.
Hypothesis testing
A specific phylogenetic hypothesis was tested using the Shimodaira-Hasegawa (1999; SH) test, as implemented in PAUP*. In this analysis, MP trees in extended ITS data in Analysis 2 were generated while constraining the topology using the heuristic search method described previously. Three constraints were tested: (1) forcing Coreanomecon and Hylomecon monophyletic, (2) Coreanomecon and Chelidonium monophyletic, and whether (3) Coreanomecon is a sister to the clade of Chelidonium and Stylophorum. The topology within the Papaver/Meconopsis and Eomecon/Glaucium clades was collapsed in the constraint, leaving no effect on the reconstruction of the MP tree in the given data set. The resultant MP trees with the constrained relationships were evaluated with the original MP trees without the constraint. The objective of the constraint analysis was to test whether each data set does not reject the hypothesis that Coreanomecon is congeneric to Hylomecon or Chelidonium. For the SH test, 10,000 bootstrap replicates were re-sampled using the re-estimated log likelihood (RELL) method.
Assessment of utility of DNA barcodes
The resolution ability of the species for each DNA barcode and a combinatory DNA barcode were examined based on the percentage of monophyletic species given the MP and BI trees (Kim et al., 2014; Youm et al., 2016). When only one accession in a species was included, it was treated as monophyletic if the branch leading to the accession was greater than zero.
Results
The final alignment of the cpDNA data (rbcL and matK) includes 1,371 sites, among which 146 were parsimoniously informative (10.6%) (Table 1). When outgroups were excluded, 113 sites were informative (8.2%). The ITS data consist of 659 sites, of which 221 were parsimoniously informative (33.5%). For the ingroups, only 191 were informative (29.0%). The alignment of the combined data of the Korean taxa included 24 sequences with 2,030 sites, among which 367 were parsimoniously informative. In the extended data of cpDNA and ITS, which had more taxa of Papaveroideae for Analyses 2 and 3, the alignments include 1,358 and 753 sites, respectively, of which 167 (12.3%) and 359 (47.7%) were parsimoniously informative (Table 2).
A phylogenetic analysis of the Korean cpDNA data resolved well-supported clades of Coreanomecon, Chelidonium, Hylomecon, and Papaver (Fig. 1A). Coreanomecon was supported as a sister to Hylomecon. The ITS data also showed that each genus was distinct and well supported (Fig. 1B). However, Coreanomecon is sister to Chelidonium in the nuclear data (Fig. 1B). A maximum parsimony analysis of combined cpDNA and ITS data produced an unresolved relationship regarding Coreanomecon (Fig. 1C). One of the three MP trees placed Coreanomecon as a sister to Hylomecon and Chelidonium and the remaining two trees generate the sister relationship between Coreanomecon and Chelidonium (Fig. 1D). A Bayesian analysis of the combined data supported the contention that Coreanomecon is sister to Hylomecon (Fig. 1D). Within the Papaver clade, P. coreanum is sister to the remaining species, and each species was clearly supported as monophyletic (Fig. 1).
The extended cpDNA data, in which more closely related taxa of Coreanomecon and Papaver were added to gain a better understanding of the phylogenetic relationship of the endemic genus, are consistent with the cpDNA data with Korean taxa only (Fig. 2). Stylophorum, not included in Analysis I (Fig. 1), was sister to Chelidonium, and Coreanomecon was placed to a sister of Hylomecon with strong support. The Glaucium/Dicranostigma clade was sister to the clade of Chelidonium/Coreanomecon (Fig. 2). Eomecon and Sanguinaria formed a strongly supported clade. The phylogenetic relationships among Macleaya, the Eomecon and Sanguinaria clade, and the Chelidonium/Coreanomecon clade were unresolved. Monophyly of these genera consisting of the tribe Chelidonieae was weakly supported (Bootstrap support of 64% and Bayesian post-probability of 0.6) (Fig. 2). Papaver was strongly supported as sister to the tribe Chelidonieae.
Phylogenetic analysis of the extended ITS data showed that Coreanomecon is sister to the clade of Chelidonium and Stylophorum with weak support (Fig. 3). Stylophorum was not supported as monophyletic, as S. diphyllum Nutt. is more closely related to Chelidonium than it is to S. lasiocarpus (Oliv.) Fedde. Hylomecon is sister to the clade of Chelidonium, Stylophorum, and Coreanomecon. Other genera in the tribe Chelidonieae, in this case Glaucium, Eomecon, and Dicranostigma, formed a clade with moderate support. Papaver and Meconopsis together formed a strongly supported monophyletic group (Fig. 3).
Three phylogenetic hypotheses (Fig. 4) were tested using the Shimodaira-Hasegawa test. Enforcing a sister relationship of Coreanomecon and Hylomecon (Fig. 4A) in the ITS data, both for Korean taxa only and extended samples, required five additional steps (Table 3). The Shimodaira-Hasegawa test did not reject the alternative topology. Enforcement of the sister relationship of Coreanomecon and Chelidonium (Fig. 4B), as suggested in the ITS data with the Korean taxa only (Fig. 1B), required three additional steps in the cpDNA data with Korean taxa only, and the alternative topology was compatible with the cpDNA data in the S-H test. However, eight additional steps were required to generate the sister relationship of Coreanomecon and Chelidonium in both cpDNA and ITS data with extended taxon samples, and this relationship was significantly rejected in the S-H test. Placement of Coreanomecon as a sister to the clade of Chelidonium and Stylophorum (Fig. 4C) was not available in the data sets with Korean taxa only because Stylophorum is absent in the data. This topology required only three additional steps in the cpDNA data, and the S-H test indicated that it was congruent with the cpDNA data.
The chloroplast rbcL and matK data and the ITS data clearly resolved all of the species of Papaveroideae in Korea (Fig. 1). MP and Bayesian analyses of separate and combined data suggested that all of the species in Korea, including the four species of Papaver, are distinguishable according to their DNA barcodes. The resolution power of the DNA barcodes for the Korean Papaveroideae was 100% (Table 2). The minimum value of the interspecific distances among the species in the cpDNA barcode regions were up to 34 times higher than the maximum value of the intraspecific distance in the three species (Table 2). Similarly, the minimum value of the interspecific distances among the species in the ITS barcode were nine times higher than maximum intraspecific value.
Discussion
The chloroplast rbcL and matK nucleotide sequences and nuclear ITS sequences in a phylogenetic analysis resulted in well-resolved and well-supported phylogenetic trees. Members of Chelidonieae, including the Korean endemic genus Coreanomecon, are strongly supported as monophyletic (Figs. 1–3). In the phylogenetic trees with the Korean taxa only (Fig. 1), all of the species were resolved as distinct lineages. This suggests that the two cpDNA barcode regions and ITS regions are useful to identify incomplete materials that do not bear adequate characters, such as flowers and fruits.
Our phylogenetic analyses of nuclear and chloroplast data show that the position of Coreanomecon is ambiguous. Coreanomecon was consistently placed as a sister to Hylomecon in the cpDNA data (Figs. 1, 2). The nuclear data indicated that it is sister to the Stylophorum and Chelidonium clade (Fig. 3). The constraint analysis results suggest that the two different positions of Coreanomecon (Fig. 4) are not strongly in conflict with each other. The sister relationship between Hylomecon and Coreanomecon represented in the cpDNA data was not rejected in the ITS data by the S-H test (Table 3). Likewise, placement of Coreanomecon as a sister to the clade of Stylophorum and Chelidonium as represented in the extended ITS data was not rejected in the cpDNA data by the S-H test (Table 3).
It may be that phylogenetic signals from the cpDNA and ITS data are not enough to resolve these relationships. In the extended cpDNA data, only three additional steps are needed to place Coreanomecon as a sister to the Stylophorum and Chelidonium clade (Table 3). In the extended ITS data, the bootstrap support value for the placement of Coreanomecon as a sister to the Stylophorum and Chelidonium clade is less than 50%, and the posterior probability is very low (Fig. 3). There are five additional steps are required to enforce Coreanomecon as a sister to the Hylomecon (Table 3). Concatenation of data does not help to resolve this issue. Our analysis of combined cpDNA and ITS data generated equivocal topologies (Fig. 1C and D).
Alternatively, the weakly conflicting results from the phylogenetic analysis between the cpDNA and the ITS data may have been associated with hybridization. In this scenario, Coreanomecon may have derived from hybridization between Hylomecon and an ancestor of the Stylophorum and Chelidonium clade. Plastid DNA is shown to be inherited maternally in Chelidonium, Macleaya, and Papaver in Papaveroideae (Zhang et al., 2003). If we assume the maternal inheritance of cpDNA in Hylomecon and Coreanomecon, Hylomecon would have been an ovulate parent for Coreanomecon, as the cpDNA trees placed Coreanomecon as a sister to Hylomecon (Fig. 2). The paternal parent would have been a stem lineage of the Stylophorum and Chelidonium clade, as the ITS trees suggest that Coreanomecon is a sister to the clade (Fig. 3). The branch length leading to Coreanomecon is as long as that to Hylomecon (Fig. 1A) or to Chelidonium (Fig. 1B), suggesting that the hybridization event may be old.
The morphology of Coreanomecon is unique in the tribe Chelidonieae. Coreanomecon has 12-pericolpate pollen, whereas Hylomecon, Chelidonium, Stylophorum lasiocarpum have the 3-colpate type (Lee and Kim, 1984; Kim et al., 1999). Stylophorum diphyllum possesses 12-colpate, rugate, or formaninous pollen (Lee and Kim, 1984). Coreanomecon is the only taxon that has scapose inflorescence in the tribe Chelidonieae. Although a mosaic pattern having both of parental characters or an intermediate form is commonly found in hybrids at early generations, the exhibition of novel characters is common in hybrid taxa at high proportions in angiosperms (Rieseberg and Ellstrand, 1993; Rieseberg et al., 2000). More data are needed for further testing of the origin of Coreanomecon.
The S-H test indicates that a hypothesis of a sister relationship between Coreanomecon and Chelidonium significantly conflicts with both the cpDNA and the ITS data when more taxa are included. The S-H test and phylogenetic analyses of DNA sequence data indicate that inclusion of Coreanomecon in Chelidonium, as was done by Ohwi (1965), is not supported (Table 3). The two genera differ in terms of several morphological characteristics. The leaves of Coreanomecon form a rosette, while those of Chelidonium are attached onto a branched stem. The inflorescence of Coreanomecon is borne from the base of the shoot system and does not have leaves, forming a scape, whereas the inflorescence of Chelidonium is an umbel borne at the terminal of the branch or at the axil of the leaves. Coreanomecon produces orange to red sap and Chelidonium has yellow sap.
Coreanomecon is morphologically similar to Hylomecon; both show similar leaf shapes and sap colors, which lead many authors (Lee, 1973; Park, 1974; Kadereit, 1993) to recognize Coreanomecon as a part of Hylomecon. However, there are significant morphological differences in terms of the inflorescence and pollen type between the taxa, as described above. Our molecular data show that Coreanomecon form a separate clade. Therefore, the morphological and molecular data indicate that Coreanomecon should be recognized as a genus endemic to Korea (Nakai, 1935; Lee, 1980; Kim, 2007; Chung et al., 2017).
The phylogenetic analysis of the extended ITS data show that P. coreanum is sister to P. anomalum Fedde, belonging to the clade of sect. Meconella (Fig. 3). Members of the sect. Meconella are widely distributed at high latitudes in eastern Asia, Greenland, in northern Canada, and in mountainous regions of Europe and the western North America (Rändel, 1974; Kadereit, 1988). The sect. Meconella is characterized by filiform filaments, pale anthers, and valvately dehiscent capsules (Carolan, 2006). Papaver coreanum match the sectional diagnostic characters well.
The cpDNA and ITS markers provided a high degree of species resolution and resolved all of the species of Papaveroideae in Korea, separately or in combination (Table 2). Adding more taxa (Figs. 2, 3) led to identical results. Considering the economic and forensic importance of Papaveroideae, especially P. somniferum, the results of our study are consistent with Lee et al. (2010), showing that DNA barcode regions are highly useful in Papaveroideae.
PDF Links
PubReader
ePub Link
Full text via DOI
Download Citation 
Print 
