Introduction

Root fungal endophytes are ubiquitous, mainly ascomycetous, intraradical associates of healthy plant roots (Rodriguez et al. 2009), their functional significance for the host plant varying from mutualistic to antagonistic (Sieber and Grünig 2006; Purahong and Hyde 2010). A subgroup of root endophytes called dark septate endophytes (DSE; Jumpponen and Trappe 1998; Addy et al. 2005), which have been under active research during last decades, are likely involved in host nutrient uptake (Newsham 2011).

Various factors are known to attribute to endophytic fungal community richness (Saikkonen et al. 2010; Botella and Diez 2011; Ghimire et al. 2011). The fungal communities of roots are affected by, for example, host plant phenology (Waid 1957), season (Wilberforce et al. 2003; Perez-Naranjo 2009), and root carbohydrate status (Hadacek and Kraus 2002). In addition, several soil factors, such as pH, soil structure, nutrient availability, and water regime, play important roles (Wilberforce et al. 2003; Summerbell 2005; Sieber and Grünig 2006; Perez-Naranjo 2009). Biotic interactions of the endophytic fungi with the host plant and within the microbial and plant communities have been shown to play a role (Ahlich and Sieber 1996; Carlsen 2002; Narisawa et al. 2002; Sieber and Grünig 2006). Although the prevalence of root fungal endophytes is well documented and several papers exist on the diversity of root fungal endophytes in specific host plants (e.g., Summerbell 2005; Sánchez Márquez et al. 2007, 2010; Khidir et al. 2010; Sakayaroj et al. 2010; Tejesvi et al. 2010), only limited information can be found on designed comparison of endophyte communities of adjacent host plants (but see Carlsen 2002; Grünig et al. 2004; Tejesvi et al. 2010).

The boreal forests form a characteristic vegetation zone in the northern hemisphere. The understory in these forests is dominated by ericaceous dwarf shrubs with only few herbaceous plant species being present (Nilsson and Wardle 2005). Of the herbaceous plants, the perennial grass Deschampsia flexuosa is omnipresent and the pseudoannual forb Trientalis europaea is typical, although more patchily distributed (A.L. Ruotsalainen, personal observation). These plant species often grow intermixed and their roots are inhabited by root endophytic fungi, as proven, e.g., by Zijlstra et al. (2005) and Ruotsalainen et al. (2004, 2007). Despite the ubiquity of these plant species, the fungal species inhabiting their roots are surprisingly poorly known. Two reports on root endophytic fungal species of D. flexuosa exist (Tejesvi et al. 2010; Zijlstra et al. 2005), whereas, to our knowledge, the root-inhabiting fungal species of T. europaea have not been studied to date. D. flexuosa and T. europaea are phylogenetically distant and represent different life cycle (e.g., Piqueras and Klimes 1998) and life form (Raunkiær 1934) strategies. As both the host identity and the site-specific factors affect the root endophytic community, it is warranted to ask how similar or dissimilar the root endophytic fungal communities these neighboring, but otherwise quite distinct, plants harbor.

We compared the root endophytic fungal communities of D. flexuosa and T. europaea growing in a close proximity at two growth sites to gain an insight into the interdependencies of root endophyte community structures of herbaceous plants in the understory of boreal mixed forest. More specifically, we asked: (1) Do these adjacent plants have similar root fungal communities? (2) Do the communities differ between two growth sites? (3) Do the fungal isolate numbers parallel with the root fungal colonization percentages of septate hyphal types? The study was carried out by analysis of pure culture isolates and root fungal colonization percentages. The fungal isolates were identified based on internal transcribed spacer (ITS) sequencing and phylogenetic analysis. A culture-based method was chosen for this study because it allows the comparison of fungal abundance, although we are aware of the disadvantage that it is selective towards specific fungal species (Hyde and Soytong 2008; Tejesvi et al. 2010).

Materials and methods

Root sampling and isolation of the fungi

The study plants, T. europaea L. (Primulaceae) and D. flexuosa L. (Poaceae), were collected as whole, 24 specimens per plant species, with a borer of 15 cm in diameter in late autumn 2008 from two sites in Oulu, Northern Finland (65°01′ N; 25°30′ E). The plant specimens at both sites were collected from an area of 1–2 acres and they consisted of a mixed sample of different heights and flowering stages of the focal plants. Both sampling sites were Pinus sylvestris L.- and Betula pubescens Ehrh.-dominated midboreal mixed forests experiencing slight cultural impact from neighboring properties and situated at a distance of 10 km from each other. These particular sites (hereafter named as site 1 and site 2) were selected to represent two vegetationally similar, but physically separate locations. The ground layer of the sites was dominated by D. flexuosa, Vaccinium myrtillus L. (Ericaceae), and Vaccinium vitis-idaea L. (Ericaceae). Soil pH of site 1 was 3.9 ± 0.1 and of site 2 was 4.4 ± 0.2, conductivity was 69 ± 21 and 27 ± 14, respectively, and total nitrogen percentage was 1.50 ± 0.40 and 0.15 ± 0.13 (average ± SD), respectively (n = 5).

The plants were brought to the laboratory and roots were washed clean of debris by subsequent washes in tap water. Roots were carefully cleaned under a dissection microscope, surface-sterilized by 3.5 % NaOCl for 15 min, and rinsed three times with sterile water. Care was taken to include only roots of good quality with no visible signs of senescence. A root sample of 10 cm in total length per study plant was cut aseptically in a laminar flow hood into 1-cm pieces, which were placed on an agar plate (9 cm in diameter). Modified 1/2 MMN medium with malt extract was used (Marx 1969) as it has been found suitable in an earlier study (Tejesvi et al. 2010). The plates were incubated at room temperature in the dark and checked frequently for fungal growth. When a newly emerged colony was observed, it was immediately transferred to a fresh plate and maintained as a pure culture in similar conditions.

Root fungal colonization assessment

The root fungal colonization of dark septate and hyaline septate type hyphae (hereafter referred to as DSE and HSE, respectively; Barrow 2003) and dark septate sclerotia (DSE sclerotia) were studied. These colonization types can be formed by several ascomycete species and can, therefore, be assumed to associate with outgrowing root endophytic community. For the study of the intraradical colonization of roots by these fungi, and especially colorless hyphae of HSE, we used the staining method described by Phillips and Hayman (1970). Although this method is commonly used for the study of arbuscular mycorrhizal (AM) fungi, we are not reporting AM colonization because AM fungi are not culturable on artificial media (Smith and Read 2008) and, therefore, are not detected by the methodology used in this study. The staining was carried out as follows: The roots were rinsed with tap water, cleared overnight in KOH (1 % for D. flexuosa and 10 % for T. europaea) at room temperature, and rinsed with tap water. The bleaching was only performed for T. europaea in alkaline H2O2 for 15 min, after which the roots were again rinsed with tap water. Then, the root samples were acidified in 1 % HCl (1 h for D. flexuosa and 2.5 h for T. europaea) and stained in 0.01 % trypan blue in lactoglycerol for 1 h at 80°C. To obtain successful staining of fungal structures, roots of T. europaea required longer exposure times (Ruotsalainen et al. 2004). The colonization percentages of roots were determined under a compound microscope by using the method described by McGonigle et al. (1990) where 50 intersects per one root system were studied at the magnification of × 150–600.

Isolation of DNA from pure cultures

All pure culture isolates obtained were first classified by macromorphological characteristics (growth, color, and surface structure of the colony) into 16 isolate types. ITS sequencing and phylogenetic analysis were performed for further identification instead of micromorphological investigation. Microscopic examination of the cultures was abandoned because it would have required specialist knowledge, prolonged incubation (up to several years in cold; Ruotsalainen, personal observation) of cultures, and/or that typically root endophytic fungi do not produce identifiable conidial structures in culture (Jumpponen and Trappe 1998; Addy et al. 2005). DNA of sufficient quality was obtained for 12 isolate types, with 2 duplicates yielding in total 14 DNA samples. From five isolate types representing single isolates, DNA of sufficient quality was not obtained.

DNA was extracted from 0.5 to 1.0 g fresh mycelia according to the method described by Pirttilä et al. (2001). The target rDNA region including ITS1 and ITS2 regions and 5.8 S gene was amplified using primers ITS1 (TCCGTAGGTGAACCTGCGG) and ITS4 (TCCTCCGCTTATTGATATGC) (White et al. 1990). Amplifications were performed in a total reaction volume of 25 μl containing 2 mM of each dNTP, 2.5 mM MgCl2, 5 pM of each primer, 1 unit of Taq DNA polymerase (Dynazyme, Finnzymes, Espoo, Finland), and 100 ng of template DNA. PCR amplifications were performed in a thermal cycler (PTC 200, MJ Research, Watertown, MA, USA) with an initial denaturing step of 94°C for 3 min, followed by 35 amplification cycles of 94°C for 60 s, 50°C for 60 s, and 72°C for 120 s and a final extension step of 72°C for 10 min. The amplification products were separated by electrophoresis at 100 V for 1 h in 1× TAE buffer on 1.4 % (w/v) agarose gel, which was stained with ethidium bromide (0.5 μg/ml) and visualized under 300 nm UV light and photographed. A 100-bp size marker (MBI Fermentas, Vilnius, Lithuania) was used as reference.

Sequencing of fungal ITS region

Amplification products obtained from PCR reactions with unlabeled ITS primers (ITS1 and ITS4) were used for sequencing. Sequencing reactions were performed with Big Dye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) according to the manufacturer’s instructions. Extension products were then purified using the ethanol/EDTA precipitation protocol and analyzed on ABI 3100 Avant Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) as recommended by the manufacturer. DNA sequences obtained for each isolate from each forward (ITS1) and reverse (ITS4) primer were inspected individually for quality. Both strands of the DNA were then assembled to produce a consensus sequence for each isolate using the Sequencher 4.7 software (Gene Codes Corporation, Ann Arbor, MI, USA). The sequences were submitted to the National Center for Biotechnology and Information and accession numbers were obtained.

Molecular phylogenetic analysis

All sequences were compared with ITS sequences available in the GenBank by BLASTn search. The closest matches in the GenBank were included in a Clustal alignment and aligned using ClustalX with default settings (Thompson et al. 1997). A phylogenetic analysis was performed by the maximum parsimony method using Molecular Evolutionary Genetics Analysis (MEGA) (Tamura et al. 2007). Confidence in specific clades from the resulting topology was tested by bootstrap analysis with 1,000 replicates. Tree scores including consistency index (CI), retention index (RI), and rescaled consistency index (RCI) were also calculated for all trees. Branches corresponding to partitions reproduced in <50 % bootstrap replicates were collapsed. All positions containing gaps and missing data were eliminated from the dataset (Complete-Deletion Option).

Statistical analysis of the number of isolates and number of isolate types

The effect of host species (D. flexuosa and T. europaea) and growth site (two sites) on the total number of root fungal isolates per plant and on the number of isolate types were analyzed by using generalized linear models with Poisson error (McCullagh and Nelder 1989). The effect of host species and site on DSE hyphal colonization was studied by two-way analysis of variance (ANOVA). On HSE hyphal and DSE sclerotial colonization data, a generalized linear model with negative binomial error was applied because these were zero-inflated (Crawley 2007). The conditions were controlled for all analyses by using residual plots (Crawley 2007). Correlation between DSE hyphae and isolate number was checked with Pearson correlation coefficient separately for both plant species (for HSE and DSE sclerotia, the Pearson correlation was not applied because the conditions for these analyses could not be met due to zero inflation). Shannon–Wiener diversity index was calculated based on the number of fungal isolates for both host plants in a site-specific manner. The statistical analyses were carried out using R statistical program version 2.10.1 (Ihaka and Gentleman 1996; R Development Core Team 2009).

Results

Colonization frequency of root fungal endophytes

One hundred and eleven pure culture isolates were obtained altogether; 84 from D. flexuosa and 27 from T. europaea. The total number of isolates per study plant was higher for D. flexuosa, but there was also a significant interaction between the host species and the study site. The number of isolates in D. flexuosa was lower at site 2, whereas in T. europaea, it was higher at site 2 (Figs. 1 and 2; Table 1). The colonization percentages of root endophytic fungi in the study plants, determined by the number of DSE and HSE hyphae, were higher in D. flexuosa (Fig. 3). In contrast, DSE sclerotia were more common in T. europaea, but there was also an interaction with the study sites (Fig. 4). There was no correlation between DSE hyphal colonization and the total isolate number in either of the study plants (D. flexuosa r = −0.19, N = 22, p = 0.374 and T. europaea r = 0.07, N = 22, p = 0.730).

Fig. 1
figure 1

Number of root fungal isolates (±SE) in D. flexuosa and T. europaea roots at two studied sites

Full size image
Fig. 2
figure 2

Number of fungal isolate types (±SE) in D. flexuosa and T. europaea

Full size image
Table 1 Impact of host plant species (D. flexuosa and T. europaea) and study site on total number of fungal isolates and the number of different isolate types and colonization parameters
Full size table
Fig. 3
figure 3

The colonization percentages of DSE and HSE in roots of D. flexuosa and T. europaea

Full size image
Fig. 4
figure 4

The colonization percentage of DSE sclerotia in roots of D. flexuosa and T. europaea at two studied sites

Full size image

DNA-based identification and phylogenetic analysis

Out of the 14 sequences which were BLAST-searched (Table 2), six sequences could be identified to the species level, namely, two isolates of Phialocephala fortinii (DF7 and TE14), Meliniomyces variabilis (TE2), Phialocephala sphaeroides (TE5), Umbelopsis isabellina (DF10), Mycena sanguinolenta (DF11), Mycena galopus (DF4), and Umbelopsis sp. (DF3). Six isolates (DF1, DF6, DF8, TE9, DF13, and DF15) were classified at the family level, all belonging to ascomycetes.

Table 2 Culturable endophytic fungal ITS sequences obtained from D. flexuosa (DF) and T. europaea (TE) closely matching to the GenBank sequences
Full size table

The phylogenetic analysis was done using MEGA4. Some of the GenBank sequences that were closely related to the endophytic fungi of the present study were included in the phylogenetic analysis. In the phylogenetic tree, altogether 39 strains, which included 14 root endophytic fungi of T. europaea and D. flexuosa and 25 sequences from GenBank, were analyzed. Base frequencies across taxa (mean frequencies: A = 0.242; T = 0.286; C = 0.233; G = 0.239), as determined by MEGA4, were homogeneous. There were a total of 317 positions in the final dataset, out of which 164 were parsimony informative. The most parsimony rooted tree of the 14 most parsimonious unrooted trees (CI = 0.732; RI = 0.904; RCI = 0.662) is shown in Fig. 5. Similar results were obtained using neighbor-joining analyses (data not shown).

Fig. 5
figure 5

Phylogenetic analysis of ITS and 5.8 S rDNA sequences of root endophytic fungi of D. flexuosa and T. europaea. The tree was derived from 14 endophytic fungal sequences and 25 sequences were retrieved from GenBank. Branch lengths are scaled in terms of expected numbers of nucleotide substitutions per site, number at branches are bootstrap values (1,000 replicates, values below 50 % are not shown). The sequences from GenBank are given with accession numbers followed by names

Full size image

Most of the endophytic fungi isolated in this study were distributed across the phylogram into five clades. Acephala applanata was used as an outgroup. The first cluster consisted of 13 fungi distributed into different subgroups. DF11 and DF4 were grouped together with M. galopus, M. sanguinolenta, Mycena haematopus, Mycena purpureofusca, and Mycena rubromarginata, whereas DF10 and DF3 grouped with U. isabellina and another Umbelopsis species. DF15 did not cluster with any genera or species included in the study. TE9, DF1, and DF8 that were only classified at the family level as ascomycetes, grouped with other two ascomycetes species in the second clade. M. variabilis (TE2) grouped with other M. variabilis strains (EF093178 and EF093172) in the third clade with a high confidence. DF13 and DF6 that were classified as ascomycetes species grouped with other ascomycetes species strains in clade 4, whereas P. sphaeroides (TE5) grouped together with another P. sphaeroides strain from GenBank. P. fortinii (DF7 and TE14) clustered together with two Phialocephala subalpina (EF093161), one P. europaea, and six P. fortinii strains from GenBank in the fifth clade of the phylogram.

Diversity of isolate types in host plants

The number of isolate types per study plant was higher in D. flexuosa than in T. europaea (Figs. 1 and 6; Table 1). D. flexuosa roots also had higher total species richness than T. europaea roots, as they harbored 16 different isolate types in total, whereas only 4 types were found in T. europaea (Fig. 6). According to the Shannon–Wiener index (H), both D. flexuosa and T. europaea had higher fungal diversity at site 2 (H = 2.50 and 1.21, respectively) compared to site 1 (H = 2.21 and 0.67, respectively). In addition, the sequencing revealed inconsistencies resulting from visual typification. One isolate type of D. flexuosa consisted of two separate fungal species, identified as M. galopus and M. sanguinolenta (Fig. 6). Similarly, another isolate type of D. flexuosa consisted of two separate fungal taxa based on sequencing, identified as ascomycete species 5 and 6 (Fig. 6).

Fig. 6
figure 6

Number of host plant specimens which yielded fungal outgrowth at studied sites 1 and 2. Isolate names are based on the closest match of ITS in GenBank, isolates U1–U5 not identified due to poor DNA quality. M. galopus and M. sanguinolenta were visually classified into one isolate type but could be separated by sequencing. Ascomycete species 5 and 6 were visually classified as one isolate type, but sequencing indicated heterogeneity within this type, which cannot be separated in the figure

Full size image

Several isolate types were specific for D. flexuosa (Fig. 6). The only isolate type found frequently from T. europaea was the DSE P. fortinii, which was present in seven specimens (in 50 % of T. europaea samples; Fig. 6). In contrast, P. fortinii was found in 21 % of D. flexuosa samples (Fig. 6). In addition to P. fortinii, one ascomycete species (isolate number 4) was found from both host plant species at both sites (Fig. 6). P. sphaeroides, which belongs to the DSE fungi, was specific for site 2, being relatively abundant there in both study plants (Fig. 6).

Discussion

Markedly higher numbers of root fungal endophytic isolates were obtained from D. flexuosa than from T. europaea. This finding was supported by the higher colonization percentage of hyphae of septate endophyte types in the roots of D. flexuosa. In earlier studies, the DSE-type colonization percentage of T. europaea has been more or less higher in field conditions, although variations were found between sites (Ruotsalainen et al. 2004). DSE hyphal colonization percentage of D. flexuosa has earlier been reported, on average, to be 20–50 % of the root length (Zijlstra et al. 2005; Ruotsalainen et al. 2007), which corresponds well with the present results. These results suggest that D. flexuosa has a higher number of root fungal endophytes than T. europaea, estimated both by root colonization percentage and fungal outgrowth.

The number of different fungal isolate types was higher in D. flexuosa than in T. europaea, which is in accordance with the results on total fungal isolate numbers and root fungal colonization percentage. Similarly, the diversity, estimated as the Shannon–Wiener index, was also higher in D. flexuosa. These results altogether indicate that the root fungal community of D. flexuosa is more diverse compared to neighboring T. europaea. Based on the sequencing results, our isolate morphotypes may include multiple, unrelated species and, therefore, the absolute numbers of fungal taxa given in this study should be treated as indicative. The root endophyte communities of D. flexuosa have earlier been studied by Vrålstad et al. (2002) and by Zijlstra et al. (2005). Although those studies reported several helotialean fungi in the roots of D. flexuosa, the same fungal taxa were not found in our study, except for P. fortinii, a fungus belonging to the DSE complex (Jumpponen and Trappe 1998). This can be partly due to markedly increased accuracy of fungal endophyte identification by sequences available in the GenBank today or because of differences in methodology, as culturing methods are always selective (Hyde and Soytong 2008). On the other hand, P. fortinii dominates the root endophyte community of D. flexuosa when analyzed by a culture-based method (Tejesvi et al. 2010).

Besides differences in root fungal abundance between the plant species studied, the number of fungal isolates and the percentage of dark septate fungal sclerotia in roots were affected by the sampling site. These differences were most distinctive for T. europaea. Based on general vegetation analysis, the sampling sites were relatively identical (A.L. Ruotsalainen, personal observation), but the soil parameters indicate subtle differences. Generally, substrate pH and nutrient availability strongly affect the performance of fungi (Deacon 2006). High nutrient availability has been reported to decrease the diversity of root endophytic community (Wilberforce et al. 2003). According to Postma et al. (2007), DSE colonization increased in relation to AM colonization along a decreasing pH gradient. The DSE P. fortinii is also known to prefer acidic soils (Ahlich et al. 1998). Therefore, such soil factors may contribute to differences observed in the fungal communities at each site, which also were more profound in T. europaea than in D. flexuosa. However, as the number of sampling sites was only two in our study, this conclusion must be interpreted cautiously. Therefore, there is preliminary evidence that the growth site has a modifying impact on intraradical fungal communities and more distinct differences can be found between adjacent plant species within these sites.

To our knowledge, there are no earlier studies on root endophyte communities of T. europaea, although root fungal colonization rates have been reported earlier (Ruotsalainen et al. 2002). No fungal isolates specific only for T. europaea were identified in this study, when compared to D. flexuosa. The fungal isolates of T. europaea were identified as M. variabilis, P. fortinii, P. sphaeroides, and ascomycete species 4 and they were found in both study plant species, and P. fortinii and ascomycete species 4 were also present at both study sites. P. fortinii is an ubiquitous DSE fungus found in plants of wide taxonomic range (Jumpponen and Trappe 1998, Addy et al. 2005; Tejesvi et al. 2010).

A few strains were isolated from D. flexuosa in this study for the first time. M. variabilis has earlier been reported to colonize several plant hosts (Ohtaka and Narisawa 2008), especially ericaceous plants (Hambleton and Sigler 2005), but to our knowledge, this is the first report from a grass. Similarly, P. sphaeroides, which is considered to belong to the DSE group, has been described quite recently from trees and forbs from Sphagnum-dominated, acidic wetland sites (Wilson et al. 2004). To our knowledge, our finding is the first from a grass. The presence of fungal isolates identified as M. galopus and M. sanguinolenta in D. flexuosa roots is somewhat surprising, as the members of this basidiomycete genus are generally considered saprophytic (Emmett et al. 2008). However, Sánchez Márquez et al. (2006) reported a Mycena sp. as an endophyte in the grass Dactylis glomerata, and Tejesvi et al. (2010) also found a Mycena species (Mycena galericulata) as the root endophyte in D. flexuosa. Some Mycena species can form an association with nonphotosynthetic orchids, being parasitized in this case (Ogura-Tsujita et al. 2009). These observations support our finding of Mycena species being a root endophyte and may indicate a new, yet undiscovered, association between this species and D. flexuosa. Two strains were identified as Umbelopsis (Mortierella) isabellina, which also is a soil saprophyte belonging to Mucorales (Meyer and Gams 2003), but earlier, Summerbell (2005) has reported U. isabellina as a root endophyte Picea mariana (Pinaceae). It is also worth mentioning that the interpretation of our results at a detailed taxon level should be done with slight caution, as the names of sequences in GenBank may be erroneous (Ko Ko et al. 2011).

To conclude, our data show that D. flexuosa and T. europaea living in close proximity within a relatively small area harbor root fungal endophytic communities which differ in diversity and richness and are, to some extent, affected by the growth site. Grasses and forbs have major differences in their life cycles and survival strategies (Raunkiær 1934). For example, their adaptive responses to herbivory and to the successional changes in the environment have marked differences (Grime 1979; Tilman 1982). The variability in growth responses may be related to species-specific differences in the plant metabolism and phenology, which are, in turn, known to affect the endophytic community (Hadacek and Kraus 2002; Wilberforce et al. 2003). For example, grasses depend more on sugar accumulation, whereas forbs depend on both sugar and K + for osmotic adjustments, which may also influence the presence of various fungal endophytes inside the plant (Barnes et al. 2007). Furthermore, T. europaea has a pseudoannual life cycle (Taylor et al. 2002), which means that the fungal inoculation of new emerging roots must be received via surrounding soil and roots.