Introduction

Invasive plants have been reported to have fewer viral and fungal pathogens, higher tolerance to new environmental stresses and more rapid growth characteristics, such as early germination and high biomass production, than noninvasive plant species (Blumenthala et al. 2009; Schlaepfer et al. 2009). Phragmites is a known invasive plant that can threaten ecosystems by quickly taking over marsh communities after introduction to a site (Chun and Choi 2009). It crowds out native plants while changing the marsh hydrology and wildlife habitat and increasing the potential for fires (Meyerson et al. 2000).

Taxonomically, the genus Phragmites belongs to the Tribe Arundineae and can be divided largely into four species: Phragmites australis, Phragmites karka, Phragmites mauritianus, and Phragmites japonicus (Clevering and Lissner 1999). Among them, Phragmites australis is the most frequently occurring worldwide. Recently, it has been reported that at least two different Phragmites species including Phragmites australis and Phragmites japonicus are present in Korea (Kim and Kim 2009). Phragmites japonicus mainly inhabits river sand while Phragmites australis can be found in wetlands. A recent study has shown genetic diversity of Korean Phragmites using RAPD markers that distinguished Phragmites australis and Phragmites japonicus (Kim and Kim 2009). Such genetic diversity of Korean Phragmites might be highly correlated with ploidy levels. However, there is no report about ploidy levels of Korean Phragmites. Interestingly, we observed that the natural habitat of Phragmites japonicus in Korea is sometimes located close to that of Phragmites australis suggesting possible hybridization between two species.

In plant molecular phylogenetic studies, approaches based on chloroplast DNA (cpDNA) sequences are widely used because of the low evolutionary rate of these sequences at the genus or species level (Soltis et al. 1993). For example, a haplotype network of Phragmites worldwide has been identified based on two noncoding regions of chloroplast DNA, revealing cryptic invasion of nonnative Phragmites into North America (Saltonstall 2002). However, it is not enough to address genetic complexity and gene flow only with cpDNA sequences, which are maternally inherited. To complement haplotype analysis, a previous study used microsatellite markers to study genetic variation of North American Phragmites (Saltonstall 2003b). Furthermore, genetic diversity of various Phragmites was also examined by means of AFLP, RAPD (Lambertini et al. 2006, 2008), and RFLP analyses (Saltonstall 2003a).

Before designing experiments, we raised several hypotheses regarding Korean Phragmites. The first is that there might be nonnative Phragmites species from other continents in Korea. If introduced species are present in Korea, we would like to determine their origins. The second is that Phragmites hybrids can be found using genetic markers. The third is that new genetic markers could be developed to facilitate the detection of hybrids between Phragmites species. To investigate those hypotheses, we collected 27 Phragmites australis and three Phragmites japonicus samples from 29 regions in Korea and determined their genetic diversity and hybrids by molecular markers including haplotypes and genotypes.

Materials and methods

Collection of Korean Phragmites

We collected 27 Phragmites australis as well as 3 Phragmites japonicus samples from 29 regions of South Korea (Fig. 1a). The samples were collected in November 2008, 2009, and 2010. A majority of the sampling regions were located along the coast and rivers.

Fig. 1
figure 1

Sampling regions and representative panicle images for Korean Phragmites. a A map of South Korea showing the sampling regions for Phragmites. The individual code indicates the region of sampling. Each code was based on a direction and latitude, i.e., seven regions in the eastern part of the country (named E1–E7), eight inland regions (from M1–M8), six southern regions (from S1 –S6), and six western regions (from W1–W6). As controls, we collected three Phragmites japonicus samples (P.ja). P.ja1 and P.ja2, indicated by a square, were collected from Kadam-ri, Hoengseong, and P.ja3, indicated by a triangle, was collected from Yeonggwang. The representative panicle images for Phragmites japonicus (b) and Phragmites australis (c) were taken from Kadam-ri, Hoeseong on November 11, 2010. As compared to Phragmites australis, Phragmites japonicus showed relatively slow senescence with green leaves and violet colored stems. In general, Phragmites japonicus grows in sandy streams and rivers and has creeping stems with silky hairs

Full size image

Genomic DNA isolation

Leaf samples were kept at −80°C. Frozen leaves were ground using a mortar and pestle in the presence of liquid nitrogen. Total genomic DNA was extracted using a DNeasy® Plant Mini Kit (QIAGEN, Hilden, Germany) according to the manufacturer’s instructions.

PCR, cloning, and sequencing for chloroplast DNAs and PhaHKT1

To get the sequences of two chloroplast intergenic regions from each Phragmites sample, we used two primer pairs, trnT (UGU) “a”-trnL (UAA) “b” (Taberlet et al. 1991) and rbcL-psaI (Saltonstall 2001), from previous studies. In addition, to determine the genotypes of Phragmites samples, the PhaHKT1 nuclear gene that encodes a high-affinity K+ transporter known to function in salt tolerance was used (Takahashi et al. 2007). Using the full-F2 and full-R4 primers, a full-length PhaHKT1 gene was amplified from the genomic DNA of each reed sample, as previously described (Takahashi et al. 2007). Each amplified PCR fragment was purified using a PCR purification kit (SolGent, Daejeon, South Korea). Purified PCR products were cloned into the pCR®2.1-TOPO® Vector, using a TOPO TA cloning® kit (Invitrogen, Karlsruhe, Germany). The plasmids were sent to the SolGent company for sequencing using M13F (−20) and M13R (−20). Genomic sequences for PhaHKT1-1 (W3), PhaHKT1-2 (M1), PhaHKT1-3 (W1), PhaHKT1-4 (S4), PhaHKT1-5 (M6), and PhaHKT1-6 (M7) were deposited in the NCBI nucleotide database with the accession numbers 1299511, 1299578, 1299584, 1299586, 1299588, and 1299590, respectively. The psaI-rbcL and trnT-trnL intergenic sequences for the Phragmites australis samples S1 and E4 and Phragmites japonicus were deposited in NCBI with the accession numbers 1299592–1299594 and 1299624–1299626, respectively.

Statistical analyses

The sequences of two chloroplast intergenic regions were downloaded from the NCBI nucleotide database (Saltonstall 2002). These data were combined with our sequence results to dissect haplotypes of Korean Phragmites. The integrated sequences were aligned by ClustalW with default parameters implemented in the MEGA4 program (Tamura et al. 2007). Then, sequences were manually edited again. The DNA alignment file was saved in a Phylip- or NEXUS-compatible format for input into the TCS program (version 1.21) that estimates gene genealogies from DNA sequences (Clement et al. 2000). Finally, a parsimony network of Korean Phragmites was created by the TCS program, applying the defined algorithms (Templeton et al. 1992).

Results

Haplotypes of Korean Phragmites based on chloroplast DNA sequences

Several studies have reported that Phragmites species are invasive. We hypothesized that nonnative Phragmites might be present in South Korea. To that end, we analyzed haplotypes of the Phragmites samples using two known chloroplast intergenic sequences (Saltonstall 2001). We also used three Phragmites japonicus samples (P.ja1, P.ja2, and P.ja3) that are morphologically different from Phragmites australis as shown in Fig. 1b. The sequence results in this study were combined with previous haplotype sequences for phylogenetic analysis (Saltonstall 2002). The neighbor-joining tree identified three groups of Korean Phragmites (Fig. 2a). Group A includes E4, three Phragmites japonicus samples, and 10 haplotypes. Group B consists solely of 15 known haplotypes. Group C contains 26 Phragmites australis samples and haplotype P. Interestingly, the P.ja1 and P.ja2 collected from the same region (Hoengseong) were closely related with E4 suggesting that E4 presumably is a natural hybrid (Fig. 2a). Previously, it was shown that two haplotypes, P and W, are found in Asia and Australia (Saltonstall 2002). Thus, most of the Korean Phragmites australis closely related with haplotype P originated from Asia.

Fig. 2a, b
figure 2

Haplotype diversity of 30 Korean Phragmites. a A phylogenetic tree was generated using neighbor-joining (NJ) method and 1,000 bootstrap replicates based on two chloroplast intergenic sequences. The tree shows haplotypes of 27 Korean Phragmites australis, 3 Phragmites japonicus, and 345 previously studied Phragmites samples worldwide. The number at each node represents the percentage bootstrap scores. Three Phragmites japonicus samples (P.ja1, P.ja2, and P.ja3) were used as out-groups. The geographic distribution of haplotypes follows a previous study: North America (haplotypes A–H, S, Z, I, and M), South America (I and Y), Europe (L–O, and T), Asia/Australia (I, J, L, M, O, P, Q, U, W, and X), and Africa (K, M, R, and V). b A parsimony network of 27 Korean Phragmites australis and 3 Phragmites japonicus samples was grouped with two haplotypes: P and W. Unlabeled nodes represent predicted samples not present in our study. Loops in the network are caused by homoplasies in the number of repeats in some indels, as shown in a previous study. The ancestral haplotypes (A, B, C, D, G, H, and Q) of the network are indicated by a square

Full size image

In order to find the origins of Korean Phragmites, we generated a parsimony network using the TCS program (Fig. 2b). The parsimony network identified that the three samples Busan (S1), Icheon (M2), and Ansan (W2) as well as haplotype P are the origins of Phragmites australis in Korea (Fig. 2b). Most of the Korean Phragmites australis seems to have diverged directly from three origins but M4, M5, M6, E5, and E7 samples have an additional branch point (Fig. 2b). Again, E4 and three Phragmites japonicus samples were highly connected with haplotypes W and X in a parsimony network. Based on parsimony network, P.ja1, P.ja2, and P.ja3 might be diverged from haplotype X whereas E4 originated from haplotype W.

Genotypes of Korean Phragmites based on the PhaHKT1 gene sequences

A previous study reported that the sequence of the PhaHKT1 gene of Phragmites australis can be used to distinguish between Phragmites with salt sensitivity growing in riverside areas and Phragmites with salt tolerance growing in saline regions (Takahashi et al. 2007). This gene might be used to explain the ecological differences among Phragmites. To detect hybrids, we performed PCR with primers specific for chloroplast and PhaHKT1 genes, respectively. We could get PCR products from all Phragmites tested, including three Phragmites japonicus samples, using the chloroplast gene primers, demonstrating the high quality of the genomic DNAs (Fig. 3a). Using primers specific for the PhaHKT1 gene, we obtained PCR products from all of the Phragmites samples except M3 and the three Phragmites japonicus samples, although we tried PCR reactions several times, changing annealing temperatures (Fig. 3a). This result confirmed that that the primer pairs of PhaHKT1 gene are very specific to Phragmites australis. Therefore, we assumed that the M3 might be the natural hybrid between Phragmites australis and Phragmites japonicus. Sequencing results revealed that several nucleotide polymorphisms are present among Phragmites samples (Fig. 3b). For instance, 16 Phragmites australis samples have a 78-nucleotide deletion (polymorphism 1). In addition, we detected two 3-nucleotide deletions (polymorphism 2 and 3) and one 8-nucleotide deletion (polymorphism 4) (Fig. 3b). According to sequence analysis of the PhaHKT1 gene, we could detect six different genotypes (noted as PhaHKT1-1 to PhaHKT1-6) in the 26 Korean Phragmites samples (Fig. 3c).

Fig. 3a–c
figure 3

A phylogenetic tree of 26 Korean Phragmites samples based on the PhaHKT1 nuclear gene sequences. a PCR results using trnT/trnL, rbcL/psaI, and PhaHKT1 primer pairs. Lanes 1 and 2 P.ja1 and P.ja2 from Phragmites japonicus, lanes 3 and 4 S1 and E4 from Phragmites australis. Using PhaHKT1 primer pairs, we could amplify the PhaHKT1 gene only from Phragmites australis and not from Phragmites japonicus, implying specificity of the PhaHKT1 primer pairs for Phragmites australis. b Alignment of PhaHKT1 sequences showing four polymorphisms (P1–P4) within 26 Korean Phragmites australis indicated by boxes. c A phylogenetic tree was created by neighbor-joining method and 1,000 bootstrap replicates using the MEGA4 program. The number at each node represents the percentage bootstrap scores. Colored and white boxes indicate the presence and absence of polymorphic DNA regions, respectively. P1 Polymorphism 1, P2 polymorphism 2, P3 polymorphism 3, P4 polymorphism 4

Full size image

Discussion

Haplotypes of Korean Phragmites

In this study, haplotype analysis provided several interesting results. The phylogenetic tree showed that Korean Phragmites can be largely divided into two groups: Phragmites australis and Phragmites japonicus (Fig. 2a). Group A contains three Korean Phragmites japonicus samples along with haplotypes W and X, indicating a possible strong genetic relationship of both haplotypes with Phragmites japonicus (Fig. 2a). In contrast, 26 Korean Phragmites australis samples and haplotype P were grouped together by phylogenetic tree and parsimony network analyses suggesting that the potential origin of Korean Phragmites australis might be haplotype P (Fig. 2a, b). Furthermore, the bootstrap values in the phylogenetic tree were relatively low implying considerable haplotype diversity of Korean Phragmites australis (Fig. 2a). All Korean Phragmites samples in this study seem to have originated from Asia and Australia based on their phylogenetic relationship with haplotypes P, W, and X. These results suggested that there are no introduced Phragmites species from other continents in South Korea. Taken together, we propose that two different Korean Phragmites species with different origins have been separately adapted in South Korea.

Parsimony network revealed that the Korean Phragmites australis has at least three origins including Busan (S1), Icheon (M2), and Ansan (W2) (Figs. 1a, 2b). These data suggest that Phragmites australis in Korea might have been introduced by land or by the prevailing westerlies from China. For instance, Busan (S1) is historically the major port city of Korea where foreign plants can be easily introduced by ships. Also, the Han river flows through Icheon (M2) to the Western Sea implying the Korean Phragmies could be distributed via the Han river. It is possible that extending the sampling size to include Phragmites in North Korea might show a different origin; however, our data in the current study suggest that the origins include both coastal and inland areas.

Hybrid speciation of Korean Phragmites

Previous studies have suggested that the genus Phragmites can be divided into four or five species (Clevering and Lissner 1999). Among them, Phragmites japonicus has a similar morphology to Phragmites australis, and it is distributed mostly in East Asia. The morphological differences between Phragmites australis and Phragmites japonicus have been clearly shown in a previous study (Ishii and Kadono 2001). Although many research groups have tried to find hybrids in their Phragmites collections, they were not successful (Clayton 1967; Gordon-Gray and Ward 1971). Integrative studies using morphology, ploidy level, and flowering period indicated that Phragmites species could not hybridize (Ishii and Kadono 2001). However, in a recent study, researchers showed that such a hybrid of native and introduced samples of Phragmites australis was possible by manually cross-pollinating introduced and native Phragmites australis to create offspring and confirming the genotypes of the offspring with microsatellite markers (Meyerson et al. 2009).

Although the collections in this study included hybrids between Phragmites species, it was very difficult to distinguish them by morphological traits. However, we could demonstrate that two Phragmites australis samples were hybrids as revealed by their haplotypes and genotypes. For example, sample E4, which showed the typical morphology of Phragmites australis, was grouped with haplotype W, which is closely related to Phragmites japonicus (Fig. 2a, b). The genotype of the M3 sample was revealed as Phragmites japonicus based on PCR results with specific PhaHKT1 primers (Fig. 3a). The three different Phragmites japonicus samples and repeated experiments proved that the PCR results were very reliable. As a result, we suppose that the phenotype of Phragmites australis might be dominant in the hybrids. Furthermore, the present study is the first report that identified natural hybrids between different Phragmites species in the wild. Moreover, we propose that such hybridization plays an important role for genetic variation of Korean Phragmites enabling them to be more adaptable in given environmental conditions during plant evolution.

Application of genetic markers to dissect speciation

To study the genetic diversity of Phragmites species, many kinds of genetic markers have been used. The most popular genetic markers are the chloroplast intergenic regions used to define haplotypes; dominant markers, including RAPD and AFLP; and co-dominant markers, such as RFLP and microsatellites. In many cases, researchers use one type of genetic marker to dissect genetic diversity. However, such an approach hampers the search for hybrids. An integrated analysis using several genetic markers can facilitate the discovery of hybrids. For instance, in a previous study, researchers attempted to find Phragmites hybrids worldwide based on haplotypes and microsatellites (Saltonstall 2003b). However, that study could not identify any hybrids. Similarly, we also used haplotypes for identifying a hybrid. The main reason for our success in the present study is that we used Phragmites japonicus as a control in which morphological traits are well characterized. In addition, for the first time, we used a nuclear gene to elucidate hybrids in Korean Phragmites. Indeed, genetic markers using nuclear genes very often have shown high variances caused by high gene copy numbers. But if the primers are designed specifically for a certain species, the gene can be an effective molecular marker, such as PhaHKT1, which distinguishes Phragmites australis and Phragmites japonicus.

In summary, this study provides a wide range of information about ecological, molecular, and genetic characteristics of Phragmites species in Korea. We also suggest that the various morphological traits of Phragmites samples in Korea are caused by hybrid speciation. Furthermore, our results highlight the importance of newly developed genetic markers that can reveal hybrids.