SEARCH

SEARCH BY CITATION

Keywords:

  • ectomycorrhizal symbiosis;
  • community structure;
  • tree individual;
  • spatial autocorrelation;
  • horizontal distribution

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Ectomycorrhizal fungi (EcMF) form diverse communities and link different host plants into mycorrhizal networks, yet little is known about the magnitude of mycobiont diversity of a single tree individual. This study addresses species richness and spatial structure of EcMF in the root system of a single European aspen (Populus tremula) individual in an old-growth boreal mixed forest ecosystem in Estonia. Combining morphological and molecular identification methods for both plant and fungi, 122 species of EcMF were recovered from 103 root samples of the single tree. Richness estimators predicted the total EcMF richness to range from 182 to 207 species, reflecting the observation of 62.3% singletons and doubletons within the community. Fine-scale genetic diversity in Cenococcum geophilum indicates the presence of 23 internal transcribed spacer genotypes. EcMF community was significantly spatially autocorrelated only at the lineage level up to 3 m distance, but not at the species level. Proximity of other hosts had a significant effect on the spatial distribution of EcMF lineages. This study demonstrates that a single tree may host as many EcMF species and individuals as recovered on multiple hosts in diverse communities over larger areas.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Plants, animals and fungi establish parasitic and mutualistic interactions with microorganisms. The multitude of functions, nutritional dependence and tissue specificity may lead to high diversity of host-associated microbial communities. Individual host organisms that differ by genetic background and inhabit different environment can be considered distinct ecosystems possessing a specific assemblage of microbial taxa (Fierer et al., 2008). While pathogenic and mutualistic partners of individuals of humans and a few model organisms have been extensively addressed (e.g. Eckburg et al., 2005; Grice et al., 2008; McKenna et al., 2008), individuals of other organisms such as plants, animals and fungi have received relatively little attention. The community and diversity of microorganisms reflects the health and environmental conditions of host individuals in humans (Eckburg et al., 2005) and plant–microbial systems (Bhatia, 2008).

Mycorrhizas play an essential role in ecosystem functioning by providing mineral nutrition to autotrophic vegetation that account for the majority of terrestrial biomass. In boreal forests, ectomycorrhiza is the dominant mycorrhiza type that involves economically important woody plant families and many lineages of fungi mostly belonging to Basidomycota and Ascomycota (Brundrett, 2009; Tedersoo et al., 2010a). Many studies have assessed the community composition and diversity of ectomycorrhizal fungi (EcMF) associated with deciduous trees and particularly conifers (Horton & Bruns, 2001). Most of these studies are performed in plots or transects established in patches of vegetation usually ranging from 102 to 105 m2 and reflect the community of multiple plant individuals (Horton & Bruns, 2001). However, little is known about the whole EcMF assemblage of individual trees. While addressing other ecological hypotheses, <25 species of EcMF have been identified from single adult host trees (5 spp. in 10 samples of Vateriopsis seychellarum Heim, Tedersoo et al., 2007; 16 spp. in 22 samples of Pinus sylvestris L., Saari et al., 2005; 10–24 spp. in 10 samples of Pseudotsuga menziesii (Mirbel) Franco, Cline, 2004). Moreover, a single dominant P. menziesii individual may harbour at least eight genets of an EcMF species (Beiler et al., 2010).

Studies performed three decades ago revealed that fungi associated with trees in forest edges have a clear successional pattern that is also reflected in horizontal distribution of both fruit-bodies and mycorrhizas around the trees (Deacon et al., 1983; Last et al., 1987). In seedlings, increasing distance from adult trees is associated with reduced EcMF richness and altered community composition (Cline et al., 2005; Dickie & Reich, 2005). Pioneer EcMF colonize seedlings via spores, paving a way for species accumulation during soil development and improved carbon economy of the maturing host (Nara, 2009). Pioneer species are outcompeted, but remain in disturbed microsites (Tedersoo et al., 2008, 2009) and in the periphery of root systems (Last et al., 1987) that receive less carbon (Johnsen et al., 2005)

Populus spp. play an important role in horticulture, paper and wood industries (Stettler et al., 1996; Dickmann et al., 2001; Taylor, 2002) as well as genetics studies (Jansson & Douglas, 2007). Moreover, Populus tremula represents one of the model organisms and is subject to whole-genome sequencing and ‘symbiome’ projects (Martin et al., 2004). Populus spp. form both arbuscular mycorrhiza and ectomycorrhiza, but the latter dominates in mature trees of most species and mesic habitats (Vozzo & Hacskaylo, 1974; Lodge & Wentworth, 1990). Early experimental studies involving Populus spp. and various EcMF revealed that these trees associate with a plethora of fungi from many genera (Walker & McNabb, 1984; Godbout & Fortin, 1985). Using in situ molecular identification techniques, between 43 and 54 species of EcMF are found on root tips of Populus in planted or naturally regenerating stands (Kaldorf et al., 2004; Krpata et al., 2008).

In this study, we aimed to quantify the number of EcMF in the root system of a single P. tremula tree individual in its natural forest environment. By focusing on only one tree, the major emphasis was to study the entire EcMF community of an individual host and to compare the taxonomic richness with other relevant studies. We further hypothesized that the EcMF community has a strong spatial pattern (Lilleskov et al., 2004; Pickles et al., 2010; Tedersoo et al., 2010b) that relates to the distance from tree trunk as observed in EcMF fruiting patterns and occurrence of mycorrhizas around solitary, isolated trees (Deacon et al., 1983).

Material and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Sampling

Sampling was conducted in an old-growth mixed forest in Järvselja, southeastern Estonia. The mean annual precipitation is 620 mm and mean annual temperature is +4.5 °C. The canopy is nearly closed and consists of several ECM trees such as Picea abies (L.) Karst., Tilia cordata Mill., Betula pendula Roth and P. tremula L. Luvisol is the main soil type in the area. The average age of individual trees is 110–130 years. A 28 × 28 m plot (58°17′N, 27°19′E) was established with an approximately 100-year-old European aspen (P. tremula L.) tree located in the centre. Tilia cordata and P. abies dominated the plot, and the closest aspen was located >80 m from the sample tree. The plot size was established based on preliminary root sampling that revealed no aspen roots beyond 19.8 m from the tree trunk, although root systems of isolated aspen trees in forests may exceed 30 m (Stone & Kalisz, 1991). In September and October, 2008, 130 root samples (15 × 15 cm to 10 cm depth) were collected from the junctions of a square grid in 2-m intervals (Fig. 1). Root samples that lacked aspen roots were replaced by another sample collected 0.5 m from the initial position. This step was repeated until aspen roots were found, but up to four times and then considered to be empty. Root samples were kept at +4 °C and processed within 2 days of collection.

image

Figure 1.  Relative location and EcMF species richness on roots of a single European aspen tree (triangle). Circles along grid lines, root samples; dotted circles, samples lacking aspen roots.

Download figure to PowerPoint

Roots were detached from soil and debris by rinsing and washing in tap water. Aspen roots were preliminarily identified and separated from other EcM roots of spruce, birch and lime by their light beige colour, bitter taste, relatively smaller diameter and less branched structure. Root tips were considered to be ectomycorrhizal when they had a fungal mantle as revealed at × 10–50 magnification under a stereomicroscope. EcM root tips were sorted into morphotypes based on colour, shape, texture, occurrence of cystidia, emanating hyphae and rhizomorphs. For each morphotype, at least two healthy root tips per sample were mounted into CTAB buffer (1% cetyltrimethylammonium bromide, 100 mM Tris-HCL (pH 8.0), 1.4 M NaCl, 20 mM EDTA) for molecular analyses. In addition, the proportion of ectomycorrhizal root tips and each morphotype were estimated based on visual inspection under a stereomicroscope.

Molecular methods

A single root tip from each morphotype per sample was subjected to DNA extraction with a Qiagen DNeasy 96 Plant Kit (Qiagen, Crawley, UK) according to the manufacturer's instructions. PCR reactions were performed as described in Tedersoo et al. (2006) using a forward primer ITSOF-T (5′-CTTGGTCATTTAGAGGAAGTAA-3′) in combination with reverse primers LB-W (5′-CTTTTCATCTTTCCCTCACGG-3′), ITS4 (5′-TCCTCCGCTTATTGATATGC-3′) or ITS2 (5′-TCCTCCGCTTATTGATATGC-3′) to amplify the nuclear rRNA gene internal transcribed spacer (ITS) region. DNA samples of Cenococcum were amplified with ITSOF-T and a newly designed Cenococcum-specific primer, ITS4Cg (5′-CACATGGCAARGGCAACCG-3′) to minimize sequencing errors for the evaluation of single-nucleotide polymorphisms that represent individuals. For estimating the number of Cenococcum individuals, only high-quality sequences with no ambiguities were considered in the analysis. To improve identification, the nuclear 28S (nuLSU) rRNA gene was amplified in certain root tips using primers LR0R (5′-ACCCGCTGAACTTAAGC-3′) and TW13 (5′-GGTCCGTGTTTCAAGACG-3′). Host identification was confirmed by amplification of the plastid trnL region using primers TrnC (5′-CGAAATCGGTAGACGCTACG-3′) and TrnD (5′-GGGGATAGAGGGACTTGAAC-3′) from at least one root tip from each sample and from all singleton EcMF taxa. The presence of PCR products was checked on a 1% agarose gel by electrophoresis and visualized under UV light. PCR products were purified using Exo-Sap enzymes (Sigma, St Louis, MO). Primers ITS5 (5′-GGAAGTAAAAGTCGTAACAAGG-3′), ctb6 (5′-GCATATCAATAAGCGGAG-3′) and TrnD were used for sequencing of the fungal ITS, nuLSU and plant trnL regions, respectively. The DNA from another root tip was extracted if the PCR or sequencing failed or yielded a nonectomycorrhizal fungus. sequencher 4.9 software (GeneCodes Corp., Ann Arbor, MI) was applied to correct, trim and assemble raw sequences. Unique sequences were submitted to EMBL. blastn searches for the manually edited sequences were performed against the International Sequence Database (INSD) or the fungi-targeted database UNITE (Abarenkov et al., 2010). Sequences were grouped to species based on 97% similarity as a barcoding threshold. Species were assigned to lineages – monophyletic EcMF taxa – according to Tedersoo et al. (2010a). The relative abundance of each lineage was calculated as the number of species included divided by the number of all taxa.

Data analysis

Only samples containing EcM root tips of aspen (n=103) were used for statistical analyses. Tests for spatial autocorrelation were performed both at the level of species and lineages. Distance among samples were defined in 11 classes (1–3, 3–5, 5–7, 7–9, 9–11, 11–13, 13–15, 15–17, 17–19, 19–23, 23–35 m). Bray–Curtis and Euclidean dissimilarity indices were, respectively, applied to generate distance matrices for EcMF community and geographical coordinates. Mantel test and Moran's I-test were, respectively, performed using the ecodist and ape packages in r (R Core Development Team, 2007). A nonparametric manova was used to test the effect of distance from the aspen tree trunk, distance from other nearest host and microtopography on EcMF community composition as implemented in the Adonis routine of the vegan package in r (Oksanen et al., 2010). Because no autocorrelation was detected among samples, samples were considered to be spatially independent replicates.

Linear regression and two-way anovas followed by Tukey's post hoc comparisons for unequal sampling were applied to test the effects of distance from tree, proximity to other host trees and micro-topographical characteristics of sampling sites on percent colonization of roots (arcsine square root-transformed) and EcMF species richness.

Rarefaction curves and their 95% CI; Jacknife2, Chao2 and ICE biodiversity richness estimators were calculated in estimates 8.0 (Colwell, 2006) based on 9999 permutations. Root samples were used as sampling units. Rarefaction analyses were run twice, either shuffling individuals among samples or not to determine the effect of patchiness on accumulating species richness (Chazdon et al., 1998). Because the curves were nearly identical (i.e. very low patchiness), only curves of nonshuffled data are shown.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Diversity and community structure of EcMF

Aspen roots were recovered in 103 out of 171 samples (60.2%) based on the combined morphological and molecular identification method. Colonization of EcMF averaged 72.2±15.6% (mean±SD). The average species richness was 4.2±1.4 per sample. Within these samples, 495 of 537 (92.2%) analysed morphotypes were sequenced successfully. Based on 97% ITS sequence similarity, individual sequences were assigned to 122 EcMF species. Seventy-seven (62.3%) species occurred only once or twice. Species richness estimators suggested that the minimum true richness is 182 (Chao 2) to 207 (Jackknife 2) EcMF species on this tree (Fig. 2). However, neither the rarefaction curve nor minimum richness estimate curves approached an asymptote, suggesting underestimation of the true richness by the estimators.

image

Figure 2.  Rarefied species accumulation curve of EcMF (triangle) associated with a single aspen tree, its 95% confidence intervals (dotted lines) and the minimum species richness estimates: Chao 2 (open circle), Jacknife 2 (open square) and ICE (open triangle).

Download figure to PowerPoint

Cenococcum geophilum was the most frequent species that occurred in 59.2% of the samples. Based on minor differences within the ITS sequences, 23 distinct ITS genotypes of C. geophilum were found. The total number of ITS genotypes for this species was estimated to be between 32 (Chao 2) and 40 (Jacknife 2) (Fig. 3). Other most frequent species included /meliniomyces2 and /inocybe5 (Supporting Information, Table S1). The most species-rich EcMF lineages were the /tomentella–thelephora (30 spp.), /cortinarius (15 spp.), /inocybe (12 spp.) and /russula–lactarius (10 spp.).

image

Figure 3.  Rarefied species accumulation curve of ITS genotypes of Cenococcum geophilum (triangle) associated with a single aspen tree, its 95% confidence intervals (dotted lines) and the minimum species richness estimates: Chao 2 (open circle), Jacknife 2 (open square) and ICE (open triangle).

Download figure to PowerPoint

Spatial distribution of EcMF community

Species composition was highly variable among samples, and Mantel test revealed no significant autocorrelation among samples based on species composition. Moreover, shuffling species individuals among samples did not change the species richness estimates, suggesting poor patchiness. However, the occurrence of lineages was significantly spatially autocorrelated between 1 and 3 m distance (Mantel r=0.028; P=0.048) (Fig. 4). Only the occurrence of lineage /tomentella–thelephora was significantly autocorrelated in this distance class (Moran's I=0.145; P=0.012). Distance to aspen trunk, microtopography and proximity to the other host trees had no significant effects on species richness and ectomycorrhizal colonization (P>0.05). Proximity to the other host trees explained 2.3% of the variation in the distribution of EcMF lineages (Adonis: F1,95=2.36, P=0.038), whereas distance from the aspen trunk and microtopography had no effect. None of the factors affected species composition of the samples. No statistically significant correlation was found between the presence or absence of lineages and distance from the aspen trunk (F1,95=0.82, P=0.561).

image

Figure 4.  Mantel correlogram based on lineage data set. Closed triangles, P<0.05; open triangles, P>0.05.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

EcMF diversity and community structure

We demonstrated that EcMF community can be hyperdiverse even at the level of a host individual. The total diversity of EcMF may be still strongly underestimated, because we sampled only the top soil horizon in a single season and neglected the roots outside the plot. Both temporal (Buée et al., 2005; Courty et al., 2008) and vertical (Tedersoo et al., 2003; Lindahl et al., 2007) niche differentiation are suggested to enhance the EcMF diversity. Nevertheless, our results indicate that a comprehensive sampling in a relatively small area (784 m2) from a single tree individual can reveal a similar number of EcMF species compared with other sampling designs covering mixed forest sites (e.g. Tedersoo et al., 2006; Courty et al., 2008; Morris et al., 2009). This suggests that the spatial pattern of EcMF community can be fairly small, and the sampling design is an important factor in studying EcMF diversity. For example, Courty et al. (2008) found 75 EcMF species from 90 samples in a temperate oak forest using a different sampling scheme (Fig. S1). In particular, the current study more than doubles the number of species found in Populus stands previously (Kaldorf et al., 2004; DeBellis et al., 2006; Krpata et al., 2008) that may, however, result from the coexistence of other host species in the area, differences in age or sampling methods.

Besides diversity, the community structure of EcMF also reflects that of previous studies in the same forest site (Tedersoo et al., 2003, 2008) and boreal forests in general (Horton & Bruns, 2001). For instance, in another study conducted at the same location with study site area of 10 ha, Tedersoo et al. (2008) also reported C. geophilum and Meliniomyces bicolor as two most frequent species colonizing seedlings growing on forest floor. While the /russula–lactarius and /piloderma lineages were relatively species-rich in this study, these taxa were under-represented in previous studies on Populus EcMF (Kaldorf et al., 2004; Krpata et al., 2008). This discrepancy may be random, related to difference in the age of the study trees (Deacon et al., 1983; Visser, 1995) or the dominance of other host trees in our study site that may affect the community composition of aspen.

Moreover, single-nucleotide polymorphisms in the dominant mycobiont C. geophilum indicated that tens of individuals of EcMF species may associate with a single tree. Other studies also reported that this species can be genetically diverse at small scales (e.g. LoBuglio & Taylor, 2002; Douhan & Rizzo, 2005), although it lacks sexual reproduction. Much remains to be discovered in the role of individual EcMF species and lineages in the functioning of trees and the entire ecosystem. However, differences in vertical distribution (Taylor & Bruns, 1999; Lindahl et al., 2007), enzyme production potential (Pritsch et al., 2004) and association with particular microsites (Tedersoo et al., 2008) may be associated with specific niches and functional complementarity.

Spatial structure of EcMF

Distance from the aspen trunk and microtopography had no clear effect on richness and community composition of EcMF. Bruns (1995) suggested that distance from the trunk affects EcMF community composition due to the differential carbohydrate partitioning. This hypothesis explains the high proportion of late successional EcMF species near the tree trunk in solitary trees and forest edges (Last et al., 1987). Because most of the EcMF associated with Populus spp. seem to be host generalists (Godbout & Fortin, 1985; Molina et al., 1992; Cripps & Miller, 1993), other carbohydrate sources explain well the lack of community differentiation along with distance from the aspen trunk. Nevertheless, considering the possible insufficiency of our sampling method, replication of trees is also needed to test these hypotheses.

Proximity to other host trees is known as an important factor determining EcMF community of seedlings in sparsely wooded ecosystems (Dickie et al., 2002; Cline et al., 2005; Dickie & Reich, 2005; Teste & Simard, 2008). However, the effect of nearby host trees on EcMF community is poorly known. In our study, proximity to the other host trees affected the EcMF community composition, indicating that the nearby non-conspecific tree individuals may provide more nutrients to the shared fungi, and thus, determine the community composition. Although EcMF tend to be host promiscuous, host preference and host-biased shifts in community composition are common in boreonemoral forests (Ishida et al., 2007; Tedersoo et al., 2008).

Besides the presence of 62.3% singletons and doubletons, spatial autocorrelation in EcMF community was weak even in the shortest distance classes among root samples of the aspen tree, indicating random horizontal distribution of EcMF around the tree. Although we could not detect an effect of microtopography or distance from the tree trunk in this study, these factors and their interaction may blur the patterns and render spatial effects below the detection limit. Nevertheless, the low spatial autocorrelation at the species level suggests that the samples are spatially independent at 2-m scale (Pickles et al., 2010). This in turn indicates that EcMF genets are fairly small and little aggregated. Greater spatial autocorrelation at the EcMF lineage level (particularly within the lineage /tomentella–thelephora) compared with species level suggests that species within lineages have a clumped distribution due to soil factors – i.e. a phylogenetically determined niche (Tedersoo et al., 2003, 2009; Peay et al., 2010). These hypotheses warrant further investigation to understand the importance of evolutionary processes in shaping the structure and function of EcMF communities (Parrent et al., 2010). Alternatively, Mantle and Moran's I-tests may have more power with less zero-values resulting from lumping of species.

In conclusion, this study demonstrates that a single tree individual may host at least 200 species and tens of genets of EcMF according to species richness estimators. These estimates exceed previous findings by nearly an order of magnitude, indicating that sampling effort and species recovery are critical for comparing among studies. Our results show that rather than distance to the host tree trunk and microtopography, the proximity to other host trees have the strongest effect on EcMF community. We hypothesize that these trees modify soil nutrient resources by differential quality of litter and root exudates that consequently alter the EcMF community structure (Rumberger et al., 2004). Further studies integrating isotope probing of horizontal carbon partitioning (Johnsen et al., 2005) and molecular identification of soil hyphal communities in homogeneous habitats (Pickles et al., 2010) may improve our understanding of EcMF spatial distribution.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank Katren Mikkel and Mari-Liis Barkala for their assistance in the laboratory and field, respectively. We are grateful to anonymous reviewers for constructive comments. Financial support was provided by ESF grants 6606, 7434, 0092J and FIBIR.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Fig. S1. Comparison of species richness among this study and six previous studies in temperate forests (Izzo et al., 2005; DeBellis et al., 2006; Tedersoo et al., 2006; Courty et al., 2008; Krpata et al., 2008; Pickles et al., 2010) and different sampling design: the sample size (cm3), soil volume sampled (cm3) and area of study site (m2).

Table S1. List of EcMF species associated with a single European aspen tree.

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
FEM_1000_sm_figs1.eps4249KSupporting info item
FEM_1000_sm_tables1.doc223KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.