INTRODUCTION
Approximately 700 species of Asplenium L. (Aspleniaceae) are widely distributed throughout the temperate and tropical regions of all the continents except Antarctica (Schneider et al., 2004; Chang et al., 2013). The genus Asplenium known in Korea contains 15 to 20 taxa (Park, 1975; Lee, 1980; Lee, 2006; Kim et al., 2007; Lee and Lee, 2015, 2018).
Asplenium anogrammoides Christ was originally described from Jejudo Island, South Korea by Christ in 1908 based on collection made on September 25, 1906 by Faurie. In 1952, it was treated as a variety, A. sarelii var. anogrammoides (Christ) Tagawa (Tagawa, 1952). Recent investigations revealed that the species are distributed in China, India, Japan, Korea, and Vietnam (Lin and Viane, 2013; Ebihara, 2016).
Asplenium anogrammoides has been hypothesized as a hybrid origin between A. sarelii Hook. (2n = 2x = 72) and A. tenuicaule Hayata (2n = 2x = 72 based on its chromosome number counted from Japanese populations) (allotetraploidy) (Kurita, 1960; Wang, et al., 2003; Lin and Viane, 2013). A phylogenetic study using nuclear DNA also supports A. sarelii as the maternal progenitor and A. tenuicaule as the paternal progenitor (Liang et al., 2021). The morphological similarity between the tetraploid A. anogrammoides and diploid A. sarelii has led to great confusion, and identification based on morphology alone is very difficult when A. sarelii grows with A. pekinense Hance (the autotetraploid arisen from A. sarelii) and A. anogrammoides (Lin and Viane, 2013).
In Korea, A. anogrammoides, Ba-yui-zom-go-sa-ri in Korean, has been collected in Jejudo Island (Park, 1949, 1961), and A. sarelii with much confusion and misapplication, Dol-dam-go-sa-ri in Korean, has been recognized (Park, 1975; Lee, 1980; Lee, 2006; Kim et al., 2007; Korean National Arboretum, 2008; Lee and Lee, 2015, 2018; Hong et al., 2024). However, there is some confusion between A. anogrammoides and A. sarelii. Recently, A. sarelii individuals found in Korea without containing the exact collection data have been treated as A. anogrammoides, which is tetraploidy (Lin and Viane, 2013). In Korea, morphologically variable individuals of A. anogrammoides have been found. In addition, A. sareii has recently been recognized as an endemic plant in China, while A. anogrammmoides is known to be distributed in Japan and Korea (Lin and Viane, 2013). The natural hybrid polyploid taxa (autopolyploidy and/or allopolyploidy) of Asplenium often occur in sympatric regions with putative parental species. Among the members of the genus Asplenium, autopolyploidy through chromosome doubling from the ancestral species and allopolyploidy after chromosome doubling from the interspecific hybrid in Asplenium are common and have been investigated on the possibility of hybrid speciation in some taxa (Rumsey et al., 2004; Ekrt and Štech, 2008; Chang et al., 2013).
Interspecific gene flow may proceed only to the formation of F1 due to factors such as hybrid sterility. If hybrids reproduce, they may backcross with parents at the site of hybridization and create a hybrid swarm. Long-term hybrid swarms can be maintained in different ways (Anderson, 1949; Arnold, 1997). Spontaneous hybridization between the parental taxa might occur at a sufficiently high frequency to counterbalance selection against the hybrids and hybrid derivatives (Ellstrand et al., 2007; Lee et al., 2019). Alternatively, hybridization might be very rare, but the individuals of hybrid ancestry might be maintained largely by selective advantages (Ellstrand et al., 2007; Lee et al., 2012). Research has shown the reticulate evolution of the A. monanthes L. complex (Dyer et al., 2012), the A. normale D. Don complex (Chang et al., 2013, 2018), and the A. paleaceum R. Br. complex (Ohlsen et al., 2015), and the A. pekinense Hance complex and the A. varians Wall. ex Hook. & Grev. complex (Lin and Viane, 2013; Liang et al., 2021). Asplenium pekinense is an autotetraploid of A. sarelii derived from chromosome doubling in diploid A. sarelii (Mitui, 1965; Wang, 2003). Hybridization events make taxon boundaries obscure and make originally distinguished taxa a “species complex”, rendering the taxonomic treatment very difficult (Wang et al., 2003).
DNA content is often used as a proxy for ploidy levels in comparative studies of chromosome number variation (e.g., Ceccarelli et al., 1995; Suda and Trávníček, 2006; Suda et al., 2006). Based on the assumption that DNA content varies with chromosome numbers when chromosome number increases due to whole- or partial-genome duplications (polyploidy or aneuploidy), this coupling of genome size with chromosome number should help to explain the ploidy levels of taxa and potential parental taxa of hybrids (Chung et al., 2012; Lee et al., 2019). Flow cytometry estimates reliable, comparative DNA content among samples using fresh leaves (Doležel and Bartoš, 2005).
DNA sequence data have been utilized to clarify parental taxa of hybrid origin groups. Chloroplast DNA has been instrumental in revealing interspecific hybridization and in documenting the hybrid origin regions of several vascular plant species (Rieseberg, 1995; Arnold, 1997; Cronn et al., 2003; Lee et al., 2012; Lee et al., 2019). To infer maternal genetic histories, rbcL and rps4 gene and rps4-trnS intergenic spacer (IGS) regions have provided critical information in many fern groups (Murakami et al., 1999; Skog et al., 2004; Li et al., 2009; Wei et al., 2013).
We aimed to evaluate the taxonomic delimitations and phylogenetic relationships of A. anogrammoides and its related taxa using chloroplast DNA sequences. Based on variable chromosome numbers in this complex, we estimated DNA content using flow cytometry to infer ploidy levels and morphological features. Finally, we discuss the role of reticulated hybridization processes in the A. anogrammoides complex.
MATERIALS AND METHODS
The sources of plant materials for morphology and DNA analysis, including GenBank accession numbers, are listed in Appendix 1. All voucher specimens were deposited at Ewha Womans University Herbarium (EWH). Seven morphological characters should be listed or described. It was analyzed seven morphological characters observed and measured from mature individuals of A. anogrammoides and relative taxa.
Materials
Leaves used for DNA analyses were collected from the natural populations encompassing all morphological and habitat diversity. A total of 19 accessions (rbcL region) and 19 accessions (rps4 gene and rps4-trnS IGS region) of A. anogrammoides complex and the relative taxa, A. pekinense and A. tenuicaule, were sampled newly from Korea (Fig. 1, Appendix 1). From the GenBank, DNA sequences of four species of Asplenium already uploaded were used: A. anogrammoides (from Korea MK774642/MK774643; from Japan AB014693/AB574873/AB853878/AB853879), A. sarelii (from China MK826776/MK826864/MK826553; MK827381/MK827079/MK827300), A. pekinense (from Korea MK774623/MK774624/MK774640/MK774641; from Japan AY545479/AB574866; from Singapore GU929864; from China MK827356), A. tenuicaule (from Japan AB014694/AB574878, AB853885; from Australia KP835444), and A. varians (from Germany, AY300147/Japan AB853875; from China MK827141/MK826553/MK82864/MK826776; from England AY549802). The additional hybrid accessions, A. anogrammoides × A. varians (AB853883/AB853881) and A. tenuicaule × A. varians (AB853886), were compared with our data. Two species, A. incisum Thub. and A. trichomanes L., were used as outgroups, we obtained each three accessions of rbcL and rps4 gene and rps4-trnS IGS regions from the GenBank (Schneider et al., 2004; Schneider et al., 2005; Ebihara et al., 2010, 2014; Xu et al., 2019).
Cytological data (flow cytometry and chromosome counting)
Young, fresh leaves were stored in plastic bags at 4°C before their DNA ploidy levels (Suda et al., 2006; Chung et al., 2012) were determined. The diploid sample of A. tenuicaule, verified by chromosome counting, was used as an internal standard. About 100 mg of fresh young leaf tissue was chopped using a fresh razor blade in a Petri dish containing 500 μL ice-cold nucleic extraction buffer (solution B kit, Partec, Münster, Germany). The suspension was filtered through CellTris (Partec). After the incubation period (for 10 min at room temperature), the staining solution containing 2 mL of DAPI (solution B kit, Partec) was added. After staining, data analysis was recorded using the cytometer CyFlow Ploidy Analyser (Partec GmbH).
Ploidy levels were estimated as the mean of individual counts for A. anogrammoides, and related taxa comparing with means values of standard (A. tenuicaule, 2x) (Lee et al., 2019). The coefficients of variation for each analyzed sample were calculated. Moreover, we selected less than three of coefficient of variation (CV), and less than 0.2 of relative standard error. Measurements were performed at least two times per an individual on different days, and in each run 5,000 nuclei were recorded within both the standard and sample peaks. Mean and CV values of each peak were calculated using WinMDI 2.8 (Purdue University Cytometry Laboratories, West Lafayette, IN, USA).
For chromosome counting, we collected individuals from natural populations. After fixation, pinnae with young sporangia were preserved 70% ethanol, then macerated 1N HCl for 10 min before 1% acetic-orcein staining. Stained young sporangia were squared and observed at 1,000× magnification (Nikon Eclipse 50i, Nikon, Tokyo, Japan). Drawing and photographs were made from meiotic cells.
Molecular data (DNA extraction, sequencing, and analyses)
Total genomic DNA was extracted using the DNeasy Plant Mini Kit (Qiagen Inc., Valencia, CA, USA) following the manufacturer’s instructions. Polymerase chain reaction (PCR) amplification, purifications, and DNA sequencing strategies followed (White et al., 1990). The amplification reaction mixture was prepared using TaKaRa Ex Taq DNA polymerase (Takara Bio, Kusatsu, Japan). For the chloroplast DNA (cpDNA) phylogeny, a total of 31 accessions, including the individuals of A. anogrammoides, A. pekinense, A. tenuicaule, and A. varians for two cpDNA regions, rbcL and rps4-trnS, were analyzed.
Primers used for amplification and sequencing were trnS (TACCGAGGGTTCGAATC) (Souza-Chies et al., 1997; Smith and Cranfill, 2002) the rps4F1 (RPS4F1 GCCGCTAG ACAATTAGTCAATC) (Hennequin et al., 2003) for part of the rps4 and trnS genes and rps4-trnS IGS (Schneider et al., 2004), rbcL-TKT-F1 and rbcL-TKT-R3N-2 for the rbcL gene (Ebihara et al., 2005). Amplification was conducted using a PTC-100 thermal cycler (MJ Research, Reno, NV, USA) with the following temperature profile for all regions: a 32 to 37 cycle reaction with denaturalization at 94°C for 1 min, annealing at 54°C for 1 min and extension at 72°C for 2 to 3 min. In addition, an initial denaturalization at 94°C for 2.5 min and a final extension at 72°C for 10 min were performed. PCR products were purified with AccuPrep PCR Purification Kit (Bioneer Inc., Daejeon, Korea). Automated sequencing analysis was performed using a sequencer (Base station, MJ Research). Chloroplast DNA regions were each aligned with gap adjustments, using the Clustal X program, followed by manual adjustment (Thompson et al., 1997). Each cpDNA region was analyzed using maximum parsimony (MP) by PAUP* (Swofford, 2002). Bootstrap values were calculated from 1,000 replicate analyses using TBR branching swapping and simple stepwise addition of taxa (Felsenstein, 1985). PAUP* evaluated congruence of the cpDNA sequence data using the incongruence length difference test (Farris et al., 1994, 1995).
RESULTS AND DISCUSSION
Morphological analysis
Asplenium anogrammoides complex, A. pekinense, and A. tenuicaule are shown in Fig. 2. Typical tetraploid A. anogrammoides (allotetraploid of A. sarelii and A. tenuicaule) firstly reported from Jejudo Island, South Korea (Tagawa, 1952), has the distinct characters of lanceolate scales with long-tailed apexes in stipe bases, thinly to firmly herbaceous lamina, and shortly toothed margins of segments. Major characteristics of A. anogrammoides and relative taxa are summarized in Table 1 (Iwatsuki, 1995; Wang et al., 2003; Lin and Viane, 2013; Lee et al., 2019). It is hard to distinguish A. anogrammoides from A. sarelii because they have very similar characters such as lamina texture, pinna, and pinnule shapes, entire indusium margins and shapes, scales shapes, and habitat preference (Table 1). These results have supported the morphological similarity between two species, which has led to great confusion, not only between them but also with true A. pekinense originated from A. sarelii by Lin and Viane (2013).
In Korea, A. anogrammoides complex rarely occurs from Seoul to Jejudo Island. The most similar taxon with the A. anogrammoides complex, A. pekinense, has different characters such as firmly herbaceous, needle shaped long toothed pinnae or segment margins, lacerate indusium margins. Asplenium tenuicaule has triangular to narrowly triangular scales in stile bases and serrate to cunate pinnae or segment margins (Fig. 2, Table 1).
Morphological similarity complicates as the species delimitation of population individuals of A. anogrammoides complex in South Korea. Hybrids are known to have morphological intermediate characters between their parents as well as combining characters from each of the parents, and sometimes the hybrids display characters not found in either parent (Grant, 1981; Rieseberg et al., 1993; Lee et al., 2019). Our morphological results make it very difficult to conserve the hybridization from putative parental taxa. This result is similar with A. sarelii complex and A. normale complex (Wang et al., 2003; Chang et al., 2013). Moreover, as Lin and Viane (2013) pointed out, the morphological similarity between the tetraploid A. anogrammoides and diploid A. sarelii are highly confounding, and identification based on morphology alone is very difficult when A. sarelii grows with A. pekinense and A. anogrammoides.
Ploidy level analysis
Flow cytometry analysis detected diploid, tetraploid, and hexaploid individuals from Asplenium plants (Fig. 3, Table 2). Ploidy levels were estimated as the mean of individual counts for each taxon.
One sample of A. tenuicaule from Jeongseon, Gangwon-do province was determined to be a diploid (Fig. 3, Table 2) by flow-cytometry data. This species has been reported as A. varians (Korean name ‘ae-gi-kko-li-go-sa-ri’) in Korea (Park, 1975; Lee, 1980; Korean National Arboretum, 2008). Recently it has been proposed as A. tenuicaule based on a diploid ploidy level and spore shapes (Lee, 2006; Lin and Viane, 2013; Lee and Lee, 2015, 2018). Therefore, it supports that the Korean name ‘ae-gi-kko-li-go-sa-ri should be treated as A. tenuicaule.
Two samples in two localities, Andeok-myeon, Jejudo Island and Nakanyupsung, Jeonnam province, among 13 samples of A. anogrammoides were determined as tetraploids (Fig. 3, Table 2). We observed about n = 72 chromosome numbers of A. anogrammoides from one population of Bongwha, Kyungbuk province (Fig. 4B). This result supported that tetraploids of A. anogrammoides have been known to occur in Korea although it is very rarely shown (Lin and Viane, 2013). However, it may be tetraploids of A. × kidoi. The remaining ten samples from Seoul, Jeonbuk province, Jeonnam province, and Jejudo Island were estimated as hexaploids. These results confirm that most populations known as ‘Dol-dam-go-sa-ri’ in Korea are hexaploids. Moreover, we observed n = 108 chromosome number from one population of Jejudo Island (Fig. 4A). This hexaploid occurs mostly throughout the Korean peninsula, and most populations known as A. anogrammoides in Korea are hypothesized as one of the hybrids based on morphological characters as middle colored leaf blade, some needle-shaped segment margin. The flow cytometry analyses and chromosome observations clarified the ploidy levels of the species as tetraploidy or hexaploidy.
These results of comparative genome size (ploidy levels) by flow cytometry analysis could conserve the hybridization as that tetraploidy or hexaploidy confirms their polyploid nature.
Molecular data analysis
The regions combined of chloroplast genes, rbcL and rps4 genes and rps4-trnS IGS for 35 accessions examined ranged from 1,888 bp in A. incisum to 1,854 bp in A. trichomanes. Of these chloroplast genes (combined 1,900 bp), the aligned length of rbcL region is 1,205 bp, and the aligned length of rps4 gene and rps4-trnS IGS region is 703 bp. Of the 1,205 aligned positions from rbcL region, 1,121 sites (95.5%) were identical, 30 sites (2.49%) were parsimony uninformative, and 54 sites (4.48%) were phylogenetically informative. MP analysis of the entire rbcL sequences found two equally most parsimonious trees with a tree length (TL) of 91, a consistency index (CI) of 0.9451 (0.9180 excluding uninformative characters), and a retention index (RI) of 0.9769. One of 91 equally most parsimony trees is shown in Fig. 5 and identical to the ones treating gaps with missing data.
In the rbcL region, it includes 22 accessions of widely distributed populations from Ansan, Seoul to Jejudo Island in Korea, and six accessions from Japan gotten from the GenBank. Among them, there are hybrid accession, including two accessions of A. anogrammoides × A. varians (AB853883) and A. pekinense (AB574866) from Japan and one accession of A. pekinense × A. anogrammoides (A. × kidoi) based on morphological characters from Korea (Fig. 5).
In the 703 bp aligned rps4 gene and rps4-trnS IGS region, 519 sites (73.8%) were identical, 91 sites (12.9%) were parsimony uninformative, and 93 sites (13.2%) provided phylogenetic information. MP analysis of the entire rps4 gene and rps4-trnS IGS sequence found 225 equally most parsimonious trees 1 TL, with high CI (0.9244 excluding uninformative characters) and RI (0.9503). Fig. 6 presents one of parsimony trees.
The maximum sequence divergence values of cpDNA among A. anogrammoides clade (28 accessions); and A. sarelii (3 accessions), A. pekinense (3 accessions), A. varians (2 accessions), and A. tenuicaule (2 accessions) regions were 1.653, 0.426, 2.306, and 2.131, respectively. In addition, important informative nucleotide sites in cpDNA of A. anogrammoides complex, A. sarelii, A. pekinense, A. varians, and A. tenuicaule were 9, 3, 1, 8, and 8, respectively. Between A. anogrammoides clade and A. pekinense shared five specific character states at position 713 (T), 1149 (A), 1195 (C), 1210 (T), 1312 (A). Therefore, our molecular data support that A. anogrammoides clade has very closer taxonomic relationships with A. pekinense to the other species, and also, A. anogrammoides clade are distinctly different from A. pekinense, A. varians, and A. tenuicaule clade by 100% bootstrap value (Fig. 6; Appendix 2).
Asplenium sarelii had been used confusingly with A. anogrammoides in Korea, formed distinctly different grouping from A. anogrammoides clade. From MP trees, A. anogrammoides clade was strongly supported by 100% bootstrap value. Moreover, this species possesses seven specific character states and one delete character (Figs. 5, 6, Appendix 2). None of samples collected from Korea belongs to the clade of A. sarelii. This result supports that A. sarelii is not distributed in Korea (Lian and Viane, 2013).
Twenty-two accessions of A. anogrammoides clade showed very small sequence variations among the populations in the rbcL region and rps4 gene and rps4-trnS IGS region (Fig. 6, Appendix 2). Moreover, this clade is nested with strong support values, a 99% bootstrap value in the rbcL analysis and a 94% bootstrap value in the cpDNA contained data of rps4 gene and rps4-trnS IGS region. Especially, in MP trees, A. anogrammoides clade is strongly nested with Korean A. pekinense by a 100% bootstrap value.
Two accessions of A. tenuicaule from Korea (individuals 109, 124) nested as a same clade in the rbcL region with four accessions gotten from Japan and Singapore (Fig. 5, Appendix 2). It supports that Korean name ‘ae-gi-kko-li-go sa ri’ should be treated as A. tenuicaule in Korea (Lee, 2006; Lian and Viane, 2013; Lee and Lee, 2015; Lee and Lee, 2018).
In the cpDNA (rbcL and rps4 genes and rps4–trnS), A. anogrammoides accessions formed a monophyletic group although there were a few polyphyletic phenomena within populations. Genome size (flow cytometry) and chromosomes number data supported that A. anogrammoides complex from Korea are mostly hexaploids, rarely tetraploids, and no diploid plants found yet. Based on these results, we proposed that A. anogrammoides complex in Korea was mostly fertile hexaploids arisen by chromosome doubling of A. × wudangshanense, A. × huawuense, or A. × mitsutae, as sp1 (6x), sp2 (6x), and sp3 (6x) (Fig. 7, Table 3). We hypothesized that there are A. x kidoi (4x), sp1 (6x) or sp2 (6x), in order from A. x wudangshanense (3x) from between A. sarelii and A. pekinense, from A. × huawuense (3x) formed between A. sarelii and A. anogrammoides, and from A. × mitsutae (3x) formed between A. tenuicaule and A. anogrammoides. In addition, they share the morphological similar characters and grow together in most same region with its parent species, A. sarelii, A. tenuicaule, A. anogrammoides, and A. pekinense. A. × kidoi (4x), is known to be distributed in China and Japan, formed by hybridization between A. anogrammoides and A. pekinense (Lin and Viane 2013; Ebihara, 2017), and is also distributed in Korea, which we demonstrated in this study. The cpDNA, genome size, and chromosome data of most A. anogrammoides complexes (4x, 6x) distributed in Korea have different data from A. anagrommaoides (4x) in Korea, counted only from the Japanese population.
To clarify taxonomic statuses of A. anogrammoides complex [A. × huawuensis (3x), A. × kidoi (4x), sp1 (6x), sp2 (6x) or sp3 (6x)] with A. sarelii, further investigations with broader samplings nationwide and well-interpreting molecular markers should be conducted.