Introduction

The habitats of plants range widely even within a single species, and plants use various mechanisms to colonize and survive in a specific environment (Daubenmire 1974; Larcher 2003). Since mycorrhizal fungi enable plants to access organic and inorganic sources of nutrition that are difficult for plants to gain by themselves (Smith and Read 1997; Aerts 2002), mycorrhizal associations are expected to play a crucial role in plant colonization. Although it seems certain that the mycorrhizal association is one of the key mechanisms for plants to colonize a new environment, our knowledge about the role of mycorrhizal association in plant colonization is still limited.

Epipactis helleborine (L.) Crantz, one of the most widespread orchid species in the world, has radiated extensively throughout Eurasia and North Africa (Delforge 1995) and has widely invaded North America (Luer 1975). This species occurs in a broad range of habitat types, such as dense forest floors, urban areas, open grasslands with scattered trees and calcareous soils from temperate to boreal zones (Salmia 1986; Buttler 1991; Delforge 1995; Hollingsworth and Dickson 1997). Although this species has been studied using various morphological, genetic and botanical approaches (Scacchi et al. 1987; Light and MacConaill 1991; Tyteca and Dufrene 1994; Squirrell et al. 2001; Ehlers et al. 2002; Brzosko et al. 2004), the mechanism ensuring its habitat diversity is poorly understood. All orchids, including E. helleborine, are complete myco-heterotrophs that depend upon mycorrhizal fungi for their carbon supply during germination and early developmental stages because their small, dust-sized seeds contain minimal nutritional reserves (Arditti et al. 1990; van der Kinderen 1995; Arditti and Ghani 2000). After development of green leaves, terrestrial orchids become autotrophs that fix carbon by photosynthesis. However, the nutritional mode of E. helleborine is partial myco-heterotrophy; the plants depend upon both photosynthesis and mycorrhizal fungi for their carbon supply (Bidartondo et al. 2004). Such dependence on mycorrhizal fungi suggests that the mycorrhizal partner may be one of the key factors to determine whether E. helleborine can thrive in a specific environment.

Although E. helleborine typically appears in mountainous areas, it occasionally occurs in coastal dunes (Weijer 1952; Pedersen and Ehlers 2000). Japanese E. helleborine also appears commonly in cool mountainous areas, while the coastal populations occur in Japanese black pine (Pinus thunbergii Parl.) forest (Makino 1918; Maekawa 1971; Satomi 1982) that is widespread naturally and is also planted artificially along coastal sand dune (Satake 1989; Vidakovic 1991). In warm temperate areas of Japan, E. helleborine often appears in alpine or subalpine regions and has never been found in low-altitude forests except coastal pine. In this study, we focused on the coastal habitat of Japanese E. helleborine and investigated the correlation between its habitat and mycorrhizal fungi. We surveyed mycorrhizal fungi from geographically different coastal populations and compared these fungi with those from other habitats. Coastal E. helleborine is occasionally separated at the variety level as E. helleborine var. sayekiana and other populations inland are recognized as E. helleborine var. papillosa because of the habitat differences (Hashimoto 1987). Thus, we also compared plant chloroplast and nuclear DNA sequences between coastal and inland populations to examine whether the plant genotype is correlated with habitat. We recognize these two varieties as a complex in this study because these taxa cannot be distinguished clearly by morphological characters.

Materials and methods

Sampling

Roots of Epipactis helleborine were obtained from a total of twelve plants from seven seaside populations (5–400 km apart) and a total of 30 plants from 12 inland populations (10–900 km apart) between 2005 and 2007 (Table 1). Leaves or flowers were also collected for plant DNA analysis. Collected roots were sectioned with a razorblade after thorough washing in water, and fungal colonization was confirmed with a compound microscope. Mycorrhizal tissues were surface-sterilized with 0.25% NaClO for 2 min and kept at −80°C or dried in silica gel. Voucher specimens of E. helleborine were deposited with the Herbarium, National Museum of Nature and Science, Tsukuba, Japan (TNS765889, 765892, 765893, 765917–765919, 771101–771115, 8500019, 8500034, 8500035).

Table 1 Samples of Japanese Epipactis helleborine used in this study
Full size table

Molecular identification of mycorrhizal fungi

DNA was extracted from root tissues using a DNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA). Fungal internal transcribed spacer (ITS) and large subunit (LSU) rDNA sequences were amplified using primer combinations ITS1F/ITS4 (Gardes and Bruns 1993; White et al. 1990) and LR0R/LR5 (Moncalvo et al. 2000), respectively. PCR amplification and sequencing were carried out as described by Ogura-Tsujita and Yukawa (2008). PCR products that were difficult to sequence directly were cloned using the pGEM-T Vector System II (Promega, Madison, WI, USA). GenBank (http://www.ncbi.nlm.nih.gov) accession numbers of the sequences of mycorrhizal fungi determined in this study are shown in Fig. 1 and Supplementary Fig. 1. Sequences were analyzed using a BLAST search (Altschul et al. 1997) against the NCBI sequence database (National Center for Biotechnology Information, GenBank) to find the closest sequence matches in the GenBank database.

Fig. 1
figure 1

Phylogenetic placement of mycorrhizal fungi of Epipactis helleborine (black squares) based on large subunit rDNA sequences of Pezizales available in GenBank. Sequences of ectomycorrhizal fungi (ECM) from GenBank are shown as open circles and all GenBank accessions are shown to the right. Phylogenetic analysis was conducted using neighbor-joining with 1,000 bootstrap replicates (values greater than 70% are shown)

Full size image

Plant DNA analysis

DNA was extracted from fresh leaves or flowers using a DNeasy Plant Mini Kit. The trnL intron, trnK intron, matK and rpL16 introns of chloroplast DNA and the ITS region of nuclear DNA were respectively amplified as described by Hidayat et al. (2005) using the following primer combinations: primers c/f (Taberlet et al. 1991), trnK-3914F/OMAT1R (Johnson and Soltis 1995; Hidayat et al. 2005), OMAT1F/trnK-2R (Johnson and Soltis 1995; Yukawa et al. 1999), F71/R1661 (Jordan et al. 1996) and 17SE /26SE (Sun et al. 1994). The purification and sequencing of the PCR products followed the method described by Ogura-Tsujita and Yukawa (2008). Our preliminary comparison of the sequences between individuals from coastal (C-1) and inland (I-8) populations found that the individuals from C-1 had three substitution sites and one deletion site in the trnL intron, one substitution site in matK and no polymorphisms in the rpL16 intron or ITS sequences. Hence, the trnL intron sequences were used for further analysis. The GenBank accession numbers of the plant sequences reported in this paper are AB428732–AB428736.

Phylogenetic analysis

The LSU sequences of each ITS type (Table 2) and the ITS sequences of Wilcoxina types 1 and 2 were used for analysis. Sequences were aligned using Clustal X software (Thompson et al. 1997), followed by manual adjustment. Phylogenetic analyses were conducted with PAUP* version 4.0b6 software (Swofford 2001). Distance trees were obtained using the neighbor-joining (NJ) method (Saitou and Nei 1987) with a Kimura two-parameter correction (Kimura 1980). For assessing the relative robustness of branches, the bootstrap method (Felsenstein 1985) was used with 1,000 replicates. LSU sequences were initially analyzed with the data matrix employed in Hansen and Pfister (2006), which clarified the phylogenetic relationship of the Pezizales, and these sequences belonged to the clade of Pyronemataceae, Tuberaceae, Discinaceae, and Helvellaceae. The sequences of related species found from Tedersoo et al. (2006) and Perry et al. (2007) were added to these results for the analysis. The LSU sequences of Peziza vesiculosa and Idophanus carneus were selected as the outgroup taxa.

Table 2 Patterns of fungal ITS sequence types from mycorrhizal roots and plant haplotypes in Epipactis helleborine
Full size table

To determine the detailed phylogenetic placement of Wilcoxina types from the mycorrhizal roots of the coastal populations, ITS sequences of Wilcoxina types 1 and 2 were used. These ITS sequences were initially combined with LSU sequences and analyzed with the sequences (encompassing 18S, ITS1, and 28S) employed in Egger (1996) that clarified the phylogenetic relationship of Wilcoxina species. These results showed that Wilcoxina types from E. helleborine and Wilcoxina mikolae became a sister group to Wilcoxina rehmii. From these results, ITS sequences of Wilcoxina types (encompassing the ITS1, 5.8S, and ITS2 regions) were analyzed with the closest matching sequences in the GenBank database. The sequences of W. rehmii were used as the outgroup taxon because no sequences were found for an appropriate outgroup to the genus Wilcoxina.

Results

A total of 103 orchid roots were analyzed and fungal ITS sequences were successfully obtained from 97 root tips. The predominant fungal sequences were classified into ten ITS types, i.e., Wilcoxina (2), Tuber (3), Hydnotrya (2), Helvella (1), Genea (1) and Exophiala (1), by BLAST analysis (Table 2). All ITS types were the dominant fungi in the roots of E. helleborine, except for Exophiala, which co-occurred with other ITS types. Mycorrhizal fungi of the coastal populations were restricted to Wilcoxina and other fungal types were found from the inland individuals. Besides these ITS types, Leptodontidium, Ceratobasidium, Cenoccocum, Russula, Nectria, and Trichoderma were also found. Root samples in which fungal sequences were not detected produced only plant DNA sequences.

Since all ITS types except Exophiala belonged to the Pezizales, the LSU sequences from each ITS type were analyzed with representative sequences belonging to the Pezizales from GenBank to confirm their phylogenetic placement (Fig. 1). All LSU sequences are nested within the Pyronemataceae, Tuberaceae, Helvellaceae, and Discinaceae and are closely related to the sequences of pezizalean ectomycorrhizal fungi in GenBank. Wilcoxina 1 and 2 form a monophyletic group with W. mikolae and W. rehmii with 100% bootstrap support (BS). Additionally, the ITS sequence of Wilcoxina 2 is 99% identical to that of an ectomycorrhizal fungus of P. thunbergii (AB250949). Tuber 1 and 2 are nested among Tuber species (95% BS) and Tuber 3 clusters with Tuber rufum var. rufum (96% BS). The ITS region of Tuber 1 showed near identity (99%) to the ectomycorrhizal fungus of Tilia cordata (AJ534706). Hydnotrya 1 forms a distinct lineage with Hydnotrya cubispora (100% BS), sharing 99% LSU sequence identity. Hydnotrya 2 and two fungal sequences from ectomycorrhizae in GenBank form a sister group to Hydnotrya 1. The ITS sequences of Hydnotrya 1 and 2 are, respectively, 84 and 89% identical to an ectomycorrhizal fungus of Fagus sylvatica (AJ969616). The ITS sequence of Hydnotrya 2 is 93% identical with that of an ectomycorrhizal fungus of Abies homolepis (AB218071). The sequence of the Helvella type is nested within the Helvella lineage (97% BS) and its ITS sequence is 85% identical with that of a Helvella elastica ascocarp (AF335455). The Genea type is nested within lineages of Genea and Humaria species and the ITS sequence of the Genea type closely matched that of a pezizalean ectomycorrhizal fungus (AJ968670; 83% identical) by BLAST analysis.

The ITS sequences of Wilcoxina 1 and 2 were analyzed with closely related sequences from GenBank to resolve deep-level phylogenetic relationships (Supplementary Fig. 1). Wilcoxina 2 forms a monophyletic group with an ectomycorrhizal fungus of P. thunbergii (AB250949; 100% BS) and is closely related to the ectomycorrhizal fungus of Picea abies (DQ069002). Wilcoxina 1 is highly supported as a sister taxon to the clade, which includes Wilcoxina 2 (93% BS). Wilcoxina 1 and 2 form a sister group with W. mikolae (100% BS) from ectomycorrhizal tissues of Pinus and Picea species.

The plant haplotypes were classified into five distinct types using plastid trnL intron sequences; one or two haplotypes were observed in each population (Table 2). Coastal populations consisted of haplotypes I and II while inland populations contained all haplotypes detected. Coastal populations were indistinguishable from inland populations based on plant haplotypes.

Discussion

The results of the present study show that E. helleborine is mainly associated with several taxa of the pezizalean ectomycorrhizal fungi. Phylogenetic analysis of fungal LSU sequences from E. helleborine revealed that all dominant fungal types were included within the Pezizales, especially the Pyronemataceae, Tuberaceae, Helvellaceae, and Discinaceae (Fig. 1). Additionally, the LSU sequences from E. helleborine were closely related to those of pezizalean ectomycorrhizal fungi reported by Tedersoo et al. (2006). Within fungal ITS types, Wilcoxina 1 and 2 form a strongly supported monophyletic group with other Wilcoxina species, which are known as ectomycorrhizal fungi of Pinaceae (Yang and Korf 1985; Piché et al. 1986; Scales and Peterson 1991a,b; Baar et al. 1999; Bidartondo et al. 2001; Kernaghan et al. 2003; Fujimura et al. 2005; Menkis et al. 2005; Trocha et al. 2006). Tuber species, which cluster with Tuber 1, 2 and 3, form ectomycorrhizae with a variety of trees, such as Quercus, Corylus, Tilia, Salix, and Picea (Singer and Harris 1987; Sisti et al. 1998; Selosse et al. 2004; Trocha et al. 2006; Smith et al. 2007; Morris et al. 2008; Hrynkiewicz et al. 2008). Each LSU sequence of Hydnotrya, Helvella, and Genea from E. helleborine roots clusters with those of ectomycorrhizal fungi from the GenBank database. The ITS region also supports the evidence that these lineages are ectomycorrhizal fungi since their closest sequence matches in the GenBank database always include ITS sequences from ectomycorrhizal roots. Exophiala was detected at three sites but was not a dominant fungus within any individual or site (Table 2). The ITS sequences of Exophiala closely matched those of an endophytic fungus (EF495231, 99% identity) and a fungal isolate from orchid roots (AY833042, 99% identity), suggesting that Exophiala, and also other minor fungi from E. helleborine, were probably present as minor endophytes, spores, or rhizosphere hyphae. Failure to obtain fungal PCR products from a small number of roots may be due to poor colonization or primer mismatch.

Coastal individuals were specifically associated with Wilcoxina. A total of 18 plants from six coastal sites were analyzed and Wilcoxina 1 or 2 were identified from all individuals (Table 2). Phylogenetic analysis of Wilcoxina ITS sequences showed that the fungal sequences from E. helleborine are highly similar to those from ectomycorrhizal fungi of Pinus and Picea species (Supplementary Fig. 1). In particular, the ITS sequence of Wilcoxina 2 is 99% identical to that of a mycorrhizal fungus from P. thunbergii. These results suggest that Wilcoxina 1 and 2 are primarily distributed along the Japanese coast by forming an ectomycorrhizal association with P. thunbergii, and E. helleborine grows only through association with these fungi.

In our data, the coastal populations do not differ genetically from the inland populations at the trnL intron sequence level. The distribution of the trnL intron haplotypes showed that coastal individuals consisted of two haplotypes but these haplotypes were also found in inland populations (Table 2). Plant haplotypes I and II from coastal populations were exclusively associated with Wilcoxina, while these haplotypes from inland populations were associated only with Tuber, Hydnotrya, or Helvella.

Our results demonstrate that mature Japanese E. helleborine have a strong preference for several taxa of pezizalean ectomycorrhizal fungi. Similarly, it was shown that Tuber and Wilcoxina are the dominant fungi in roots of European and American E. helleborine (Bidartondo et al. 2004). These results suggest that mycorrhizal association with pezizalean ectomycorrhizal fungi is common among E. helleborine worldwide. Secondly, we found that E. helleborine growing on coastal dunes was exclusively associated with the Wilcoxina species group, suggesting that E. helleborine successfully colonizes Japanese coastal dunes by associating with Wilcoxina. This hypothesis can be tested via in situ germination experiments. In our data, plant trnL intron haplotypes did not discriminate between coastal and inland individuals. All of our results indicate that the plant characteristics concerned with the mycorrhizal association determine the habitat pattern of E. helleborine. The results of Bidartondo et al. (2004) also suggest that E. helleborine associates with the ectomycorrhizal fungi of forest trees and successfully thrives in the low-irradiance understory of forests with nutritional supplementation from its mycorrhizal partners. Mycorrhizal association with pezizalean ectomycorrhizal fungi appears to be the mechanism behind how the globally widespread orchid, E. helleborine, colonizes its widely ranging habitat.