Enzymatic activities and stable isotope patterns of ectomycorrhizal fungi in relation to phylogeny and exploration types in an afrotropical rain forest


Author for correspondence:
Leho Tedersoo
Tel: +372 56654986
Email: leho.tedersoo@ut.ee


  • Ectomycorrhizal (ECM) fungi obtain both mineral and simple organic nutrients from soil and transport these to plant roots. Natural abundance of stable isotopes (15N and 13C) in fruit bodies and potential enzymatic activities of ECM root tips provide insights into mineral nutrition of these mutualistic partners.
  • By combining rDNA sequence analysis with enzymatic and stable isotope assays of root tips, we hypothesized that phylogenetic affinities of ECM fungi are more important than ECM exploration type, soil horizon and host plant in explaining the differences in mineral nutrition of trees in an African lowland rainforest.
  • Ectomycorrhizal fungal species belonging to extraradical mycelium-rich morphotypes generally displayed the strongest potential activities of degradation enzymes, except for laccase. The signature of 15N was determined by the ECM fungal lineage, but not by the exploration type.
  • Potential enzymatic activities of root tips were unrelated to 15N signature of ECM root tip. The lack of correlation suggests that these methods address different aspects in plant nutrient uptake. Stable isotope analysis of root tips could provide an additional indirect assessment of fungal and plant nutrition that enables enhancement of taxonomic coverage and control for soil depth and internal nitrogen cycling in fungal tissues.


Mineral nutrition of most terrestrial plants relies on mycorrhizal fungi that take up both mineral and simple organic compounds from soil solution. While the majority of plant taxa form arbuscular mycorrhizal (AM) symbiosis with members of the phylum Glomeromycota, the ectomycorrhizal (ECM) symbiosis involving various lineages of mostly Basidiomycota and Ascomycota dominates in economically important trees such as Pinaceae, Fagaceae, Dipterocarpacae and Myrtaceae (Brundrett, 2009; Tedersoo et al., 2010a). These ECM fungal lineages have probably evolved their mycorrhizal habit independently and may play a differential role in nutrition and protection of their plant hosts (Lindahl & Taylor, 2004; Bödeker et al., 2009; Tedersoo et al., 2010a). Fungal species belonging to different lineages display substantial differences in ECM morphology – particularly the presence and abundance of extraradical mycelium including rhizomorphs – that are responsible for exploration and transportation of nutrients (Agerer, 2001, 2006). These so-called ECM exploration types are believed to differ in their efficiency of carbon (C) storage, enzymatic activities, nutrient uptake and translocation (Courty et al., 2010; Hobbie & Agerer, 2010; Lilleskov et al., 2011; Pritsch & Garbaye, 2011).

Potential activities of extracellular enzymes involved in the degradation and nutrient release from soil organic matter have been used to address functional diversity among ECM fungi in situ (Courty et al., 2010; Pritsch & Garbaye, 2011). Activity measurements rely on the enzymatic cleavage of specific substrates by functional enzymes present and probably bound at the surface of ECM root tips (Pritsch et al., 2004, 2011). Although ECM root tips support diverse populations of bacteria and microfungi (Frey-Klett et al., 2007; Tedersoo et al., 2009), consistent differences among ECM fungal species and negligible activities in nonmycorrhizal (NM) roots suggest that most of the potential enzymatic activities in root tips are exerted by ECM fungi rather than by other microbes and the host plant itself (Courty et al., 2010, 2011). There are substantial differences in potential enzymatic activities among ECM fungal species, seasons, soil horizons and their interactions (Courty et al., 2005, 2006, 2010; Buée et al., 2007; Jones et al., 2010), supporting the view that ECM fungi display at least some degree of functional complementarity in their benefits to host plants (van der Heijden & Kuyper, 2003; Nara, 2006; Rineau & Courty, 2011).

The relative abundance of stable isotopes in food webs is based on the discrimination against heavier isotopes in several biochemical processes and therefore reflects patterns of water availability, photosynthesis activity and nutrient cycling (Dawson et al., 2002). In ECM symbiosis, nutrient transport from fungi results in depletion of 15N in plants, while ECM fungi become more enriched in the process (Högberg et al., 1999a). Fruit bodies of ECM fungi are therefore typically more enriched in 15N compared with soil and humus saprotrophs (Gebauer & Dietrich, 1993; Taylor et al., 1997). In addition, ECM fungi access recent photosynthesis products rather than C from soil and debris (Högberg et al., 2001; Hobbie & Horton, 2007; Baldrian, 2009), which render them more depleted in 13C compared with saprotrophic fungi. These differences in the natural abundance of stable isotopes enable ECM symbionts to be distinguished from saprotrophs (Gebauer & Taylor, 1999; Hobbie et al., 1999; Kohzu et al., 1999; Mayor et al., 2009) – a phenomenon called the ‘ECM–sap divide’ (cf. Henn & Chapela, 2001). However, stable isotope patterns vary widely among ECM fungal species and differ both spatially and temporally, depending on nutrient availability (Taylor et al., 1997, 2003). While the local 13C abundance in fungal fruit bodies is mostly attributable to host species (Högberg et al., 1999b), 15N abundance depends on exploration depth by the extraradical mycelium (Agerer et al., 2012), N source (Högberg et al., 1999a) and the ratio of chitin to amino acids in fungal tissues (Taylor et al., 1997). Furthermore, fruit bodies of species with greater amounts of extraradical mycelium, including rhizomorphs, are relatively more enriched in 15N, which has been at least partly ascribed to greater internal nitrogen (N) cycling and accumulation of 15N-depleted chitin in the extraradical mycelium (Handley et al., 1996; Zeller et al., 2007; Hobbie & Agerer, 2010). Natural abundance of 15N in fruit bodies and soil mycelial biomass is positively correlated with the ability of the fungal taxa to take up proteins in pure culture experiments (Lilleskov et al., 2002, 2011).

Stable isotope natural abundance has been rarely addressed in root tips because of difficulties in identifying mycobionts and obtaining sufficient material for analysis of root tips; however, these constraints have been overcome along with technological advances in the last decades. Early studies demonstrated that the natural abundance of 15N in tree roots is intermediate between that of foliage and soil (Gebauer & Dietrich, 1993). Owing to a substantial fungal proportion, ECM root tips are more enriched in 15N than NM root tips (Högberg et al., 1996; Zeller et al., 2007). Information about the effect of fungal species and exploration types on stable isotope natural abundance in ECM root tips is still lacking, although 15N natural abundance in fine roots is considered a good indicator of N availability across fertilization gradients and soil horizons (Högberg et al., 1996), and the identity of mycobionts explains a large proportion of root tip respiration and fine root traits in temperate forests (Ostonen et al., 2009; Trocha et al., 2010).

Because of great differences in seasonality, temperature and humidity, soil processes in tropical ecosystems differ substantially from those in temperate and boreal forests (Six et al., 2002; Zhang et al., 2008; Hyodo et al., 2010). While taxonomic information about tropical ECM fungal communities is rapidly accumulating (Tedersoo et al., 2007, 2010b,c; Diedhiou et al., 2010; Peay et al., 2010; Smith et al., 2011), the functional aspects of tropical fungi remain poorly understood. Preliminary research indicates that certain species of tropical ECM fungi are able to access organic N sources that render them functionally similar to fungi in boreal and temperate ecosystems (Brearley et al., 2005). Here we focused on potential enzymatic activities and stable isotope patterns of ECM fungi in an African tropical lowland rainforest. We postulated three main alternative hypotheses: potential enzymatic activities of root tips are stronger in ECM than in NM root tips; potential enzymatic activities and 15N natural abundance in ECM root tips are mostly related to the phylogenetic origin of fungi (lineage); and potential enzymatic activities and 15N natural abundance are correlated and reflect similar aspects of ECM tree nutrition.

Materials and Methods


Field work was carried out in the primary lowland rainforest of the Crystal Mountains at the edge of the Wolue-Ntem Plateau close to the Tchimbele hydropower dam in the Mbé National Park, in northwestern Gabon (0°37′N; 10°24′E). Topography of the dissected plateau is steep, with elevations ranging between 300 and 843 m. The rainforest represents a former Pleistocene forest refugium and it belongs to the Atlantic Equatorial Coastal Forests ecoregion (Burgess et al., 2004). The vegetation is described in detail by Sunderland et al. (2004; cf. plots 1 and 3). The ECM hosts include Bikinia le-testui (Pellegr.) Wieringa, Gilbertiodendron ogoouense (Pellegr.) J. Leonard, Anthonotha spp., Aphanocalyx spp., Tetraberlinia bifoliolata (Harms) Hauman, Tetraberlinia sp. (all Caesalpiniaoideae), Uapaca guineensis Mull.Arg., Uapaca heudelotii Baill. (Phyllanthaceae) and Marquesia excelsa R.E.Fr. (Dipterocarpaceae). The dominant AM trees include Dichostemma glaucescens (Euphorbiaceae), Santiria trimera (Burseraceae) and Dialium sp. (Caesalpinioideae). The climate is moderately seasonal with a mean annual temperature of 23.6°C and average rainfall of c. 3000 mm that peaks in June–July (Sunderland et al., 2004). The soils are formed on highly weathered basement rock formations dating back to the Archean (2500 million yr) and include a thin (1–3 cm) organic layer composed of partly decomposed leaves. The soil profile below the organic layer consists of yellow clay and granite and is not differentiated to at least 1 m depth. Sampling of 86 soil cores (15 × 15 cm to 5 cm depth) from various hosts was performed in a 20 ha plot in May 2009 as described in Tedersoo et al. (2011), where host specificity among ECM fungi was addressed. Root samples were transported to the Tchimbele village and processed within 36 h of collection. ECM roots were separated from nonECM roots and cleaned from soil particles in water. ECM roots were further sorted by host taxa based on the colour and branching pattern of root systems. Two to four uncut root fragments (10–20 cm in length) of each potential host species were subjected to morphotyping of ECM root tips based on their colour, mantle texture and structure of hyphae, cystidia and rhizomorphs. All morphotypes were further assigned to ECM exploration types (i.e. contact, short distance, medium-distance fringe and long-distance types; cf. Agerer, 2001). The ‘mat’ type distinguished here lacked rhizomorphs, but possessed very dense extraradical mycelium that completely embedded ECM root tips and nearby soil particles, corresponding to a variant of short-distance exploration type. Clusters of ECM and NM root tips and single ECM tips from each abundant (> 7) morphotype per root fragment and soil core were separated from the remaining soil particles and extraradical mycelium, including rhizomorphs, using tweezers. NM root tips were not assessed for colonization of AM fungi. Root tips were frozen in 1.5 ml Eppendorf tubes for transportation and further enzymatic and isotope analyses. In ECM root tips, proportions of fungal mantle, Hartig net (i.e. cortical cells) and central root core were estimated based on cross-sections of ECM root tips assuming that roots are of cylindrical shape and that different root tip tissues have similar densities. One to five root tips from each morphotype were kept in 0.2 ml Eppendorf tubes containing 200 μl 1% CTAB DNA extraction buffer (1% cetyltrimethylammonium bromide, 100 mM Tris–HCl (pH 8.0), 1.4 M NaCl, and 20 mM EDTA) for molecular identification. Single root tips were subjected to DNA extraction with a DNeasy 96 Plant kit (Qiagen). PCR and sequencing of both fungal and plant symbionts are described in Tedersoo et al. (2011). Species were separated based on 97% internal transcribed spacer (ITS) sequence similarity threshold. Sequences of ECM root tips were released in International Sequence Databases (INSD) under accessions FR731846FR731947.

To study the ecosystem-level nutrient cycling, we collected fruit bodies of fungi with ECM, saprotrophic or unknown (potentially saprotrophic or parasitic) lifestyle. We also collected foliage of eight ECM trees (= 1–11) and three AM tree species (Podococcus barteri G.Mann & H.Wendl., Santiria trimera (Oliver) Aubrev, Polyalthia suaveolens Engl. & Diels; = 4) as well as samples of organic (0–2 cm depth) and mineral (2–5 cm) soil horizons (= 32). We focused on understorey foliage (0.2–1.0 m above ground) for all trees to account for age differences (2–300 yr) and avoid within-canopy vertical differences in δ13C patterns (Cerling et al., 2004). Fruit bodies of fungi were dried using an air dryer at 35°C to preserve their DNA. Leaves and soil were dried on silica gel immediately after collection. All collections are vouchered in the herbarium of Tartu University (TU). Fruit bodies of fungi were further subjected to DNA extraction and sequencing following the same procedures as for root tips, and are available from INSD (accessions JQ657762JQ657797).

Soil samples were subjected to analyses of mineral nutrients. Phosphorus was extracted using ammonium lactate and measured by flow injection analysis (Ruzicka & Hansen, 1981). Available potassium was determined from the same solution by the flame photometric method (AOAC956.01). Exchangeable calcium and magnesium contents in the soil were measured in ammonium acetate extract (pH = 7.0). The organic matter content was determined based on the loss of gases after ignition for 2 h at 360°C. Soil pH was measured in 1 M potassium chloride solution.

Enzymatic and isotope assays

Ectomycorrhizal and NM root tips were subjected to measurement of potential activities of the following enzymes: leucine aminopeptidase, β-xylosidase, β-glucuronidase, cellobiohydrolase, N-acetylglucosaminidase, β-glucosidase, acid phosphatase and laccase. Briefly, activities of leucine aminopeptidase and N-acetylglucosaminidase indicate degradation of the N-containing compounds proteins and chitin, respectively. Acid phosphatase separates orthophosphate residues from phosphomonoesters. Cellobiohydrolase, β-xylosidase, β-glucuronidase and β-glucosidase hydrolyse aliphatic bonds, whereas laccase is a phenoloxidase involved in depolymerization and ring cleavage of aromatic compounds. Particularly leucine aminopeptidase and laccase have alternative functions in signal transduction and anabolic pathways in many organisms (Baldrian, 2006; Matsui et al., 2007).

Seven replicate root tips (cut to 2 mm in length) of each ECM morphotype or NM roots per sample were subjected to sequential assays of enzymatic activities in 96-well plates (AcroPrep™ 96-filter plate with 30–40 μm mesh size; Pall Life Sciences, Crailshelm, Germany) as outlined in Pritsch et al. (2011). Briefly, each filter plate in which assays were performed comprised 84 wells of root tips, six replicates (wells) of negative control and six replicates for calibration. Both root tips and controls were incubated in 150 μl incubation buffer (except 120 μl in laccase assay) on a microplate shaker at room temperature. At the end of each enzymatic assay, incubation solutions were transferred to measurement plates using a vacuum manifold. For fluorescence measurements, black microplates (Nunc, Langenselbold, Germany) were used, while a clear 96-well microplate (Nunc) was used for the colorimetric determination of laccase activity. In between enzyme assays, root tips where rinsed with 150 μl of rinsing buffer (Pritsch et al., 2011) by use of a vacuum manifold. Fluorescence was measured at 364 nm excitation and 450 nm emission in the Infinite M200 (Tecan Group Ltd, Männedorf, Switzerland) microplate reader. The assay for laccase was measured spectrophotometrically at 420 nm. After eight enzyme tests, root tips were transferred into a clear 96-well microplate with 50 μl water in each well. Root tips were scanned and the projected surface area of each root tip was evaluated using software WinRhizo (Regent Instruments, Inc., Quebec, Canada). Enzyme activities were calculated from fluorimeter and photometer readings as described by Pritsch et al. (2011) and are expressed as mol mm−2 min−1 of released substrate.

To measure the natural abundance of stable isotopes 13C and 15N, at least 0.1 g of NM and ECM root tips, central parts of stipes of fruit bodies (or nonhymenium tissue of other fruit body types), leaves and soil were ball-milled, and subsequently analysed for total C, 13C, total N and 15N content using an elemental analyzer (Eurovector, Milan, Italy) coupled with an isotope ratio mass spectrometer (MAT 253; Thermo Electron, Bremen, Germany). Stable isotope results were expressed as δ13C and δ15N following international standards (Werner & Brand, 2001). Isotope measurements were performed in two to six replicates (except there was no replication in 60.0% of root tip samples owing to paucity of material) and average values were used throughout the study.

Statistical analyses

Using the nlme package of R (R Core Development Team, 2010), a series of generalized least-squares (GLS) models were constructed to address the relative importance of ECM status (df = 1), ECM lineage (df = 7), ECM exploration type (df = 5), host plant (df = 6) and soil horizon (df = 1) on stable isotope content and potential enzymatic activities of ECM root tips. The effect of ECM colonization on δ13C and δ15N and potential enzymatic activities of root tips was studied by including mycorrhizal status and host plant species as fixed factors. In ECM root tips, the effects of ECM lineage, exploration type, host species and soil horizon on stable isotope composition and potential enzymatic activities were subsequently addressed. In GLS models involving stable isotope ratios as dependent variables, the values of all eight potential enzymatic activities, proportion of fungal mantle and inner root core and N concentration were used as covariates to test for significant relationships between stable isotope and enzymatic patterns and to partial out the effects of potential confounding variables. Differences in stable isotope composition between ECM and AM tree species, ECM and NM root tips and trophic groups of fruit bodies were tested with one-way ANOVAs. Variables were log-transformed before analyses when necessary. Tukey tests were further applied to detect significant differences among multi-level fixed factors such as exploration type, ECM lineage and host plant species. The significance level α = 0.05 was used throughout the study. The best models were chosen based on the smallest corrected Akaike (AICc) values.

We further assessed whether the potential enzymatic activities depend on genetic distance within ECM lineages. To accomplish this, we built genetic distance matrices using Jukes–Cantor distance for all ECM fungi and for the /russula-lactarius and /tomentella-thelephora lineages (the most common and species-rich groups) separately. A distance matrix involving the potential activities of eight enzymes was generated based on the Euclidean dissimilarity index. Partial Mantel tests were performed to test the relative effects of genetic distance and stable isotope ratios on all enzymatic activities taken together, using exploration types and soil horizon as additional explanatory (confounding) factors. Mantel correlograms were adopted to address distance decay from conspecific isolates (< 3% distance) to relatively closely related species (3–8%; 8–13%) and two groups of more distantly related species (13–20%; > 20%).


Root tip morphology and identification

Altogether, 287 root tips were subjected to molecular identification that revealed 101 species of ECM fungi (including 95 basidiomycete and six ascomycete species) based on 97% ITS similarity threshold. The same morphotypes sampled from different root fragments in the same sample belonged to one species in > 90% of cases. Among species, assignment to ECM exploration types was generally concordant in different samples, except in a few cases where senescent root tips were devoid of rhizomorphs. Placement of these root tips was therefore adjusted before statistical analyses. All root fragments in samples were successfully assigned to plant species by combining morphological and molecular features.

The largest number of species was assigned to contact and short-distance exploration types. All ECM exploration types included species from more than one ECM fungal lineage and all major lineages comprised several exploration types (Supporting Information, Table S1). The /russula-lactarius and /tomentella-thelephora were the most species-rich ECM lineages. The /russula-lactarius lineage was also particularly rich in exploration types. In addition to the dominant smooth or cystidiate contact type, several Russula spp. belonged to the medium-distance fringe exploration type. The long-distance ECM exploration type comprised mostly species of the /boletus lineage.

Enzymatic patterns of ECM root tips

In total, 1176 root tips were subjected to potential enzymatic activity assays of eight degradation enzymes. There were significant differences between ECM and NM root tips in six potential enzymatic activities. The average activities of acid phosphatase (t1,163 = 4.56; < 0.001), N-acetylglucosaminidase (t1,163 = 4.06; < 0.001), β-glucosidase (t1,163 = 4.77; < 0.001), cellobiohydrolase (t1,163 = 4.61; < 0.001), laccase (t1,163 = − 3.26; = 0.001) and β-xylosidase (t1,163 = 2.73; = 0.007) were, respectively, 2.64, 2.37, 2.94, 4.12, 7.99 and 1.74 times higher in ECM compared to NM root tips. The activities of leucine aminopeptidase and β-glucuronidase were highly variable among ECM root tips and did not differ significantly between ECM and NM root tips.

According to the best GLS model, ECM exploration type rather than host species or ECM lineage affected potential enzymatic activities of ECM root tips. Activities of six out of eight tested enzymes differed among ECM exploration types. The activities of β-glucosidase (among all types: t4,115 = 2.70; < 0.008), cellobiohydrolase (t4,119 = 3.92; = 0.002), acid phosphatase (t4,108 = 1.98; = 0.049), β-glucuronidase (t4,121 = 2.84; = 0.005) and leucine aminopeptidase (t4,121 = 4.04; < 0.001) were generally higher in the long-distance exploration type than in the contact, short distance and mat types (Fig. 1a–e). By contrast, the activity of laccase was significantly higher in the contact exploration type than in all other exploration types (t4,121 = 2.95; = 0.004; Fig. 1f).

Figure 1.

The effect of ectomycorrhiza exploration type on potential enzymatic activities of mycorrhizal root tips (mean ± 95% CI): (a) β-glucosidase; (b) cellobiohydrolase; (c) acid phosphatase; (d) β-glucuronidase; (e) leucine aminopeptidase; (f) laccase. Significantly different activities as revealed by Tukey test are indicated with different small letters. Abbreviations for exploration types (n, number of replicates): LD, long-distance (= 6); MDF, medium-distance fringe (= 12); mat (= 4); SD, short-distance (= 75); C, contact (= 28).

Besides exploration types, potential activities of three enzymes – acid phosphatase, leucine aminopeptidase and cellobiohydrolase – differed significantly among phylogenetic lineages of ECM fungi. Acid phosphatase activity was significantly higher in the /tomentella-thelephora lineage and Atheliales compared to the /russula-lactarius and /coltricia lineages (t7,108 = 2.47; = 0.015; Fig. 2a). Leucine aminopeptidase activity was significantly higher in the Atheliales compared to all other lineages except /boletus (t7,114 = 8.79; < 0.001; Fig. 2b). The /boletus and /tomentella-thelephora lineages had significantly higher cellobiohydrolase activities than the /clavulina, /coltricia and /marcelleina-peziza gerardii lineages (t7,114 = 3.09; = 0.003; Fig. 2c). ITS-based genetic distance across all ECM fungi and the /tomentella-thelephora lineage had no effect on the potential enzymatic activities taken together. By contrast, there was a strong effect of genetic distance on both the enzymatic activities (rMantel = 0.197; = 0.002) and distribution of exploration types (rMantel = 0.184; = 0.008) in the /russula-lactarius lineage, but exploration type had no significant direct effect on enzymes in this group. The effect of genetic distance was significantly phylogenetically autocorrelated in the first three genetic distance classes of the /russula-lactarius lineage (Fig. 3).

Figure 2.

The effects of ectomycorrhizal fungal lineage (a–c) and host plant species (d) on potential enzymatic activities of mycorrhizal root tips (mean ± 95% CI): (a) cellobiohydrolase; (b) acid phosphatase; (c) leucine aminopeptidase; (d) acid phosphatase. All significantly different activities as revealed by Tukey test are indicated with different small letters. Abbreviations for lineages (n, number of replicates): Ath, undefined lineage of Atheliales (= 4); Bol, /boletus (= 14); Cla, /clavulina (= 6); Col, /coltricia (= 5); Mar, /marcelleina-peziza gerardii (= 3); Rus, /russula-lactarius (= 67); Seb, /sebacina (= 5); Tom, /tomentella-thelephora (= 21). Abbreviations for host species (n, number of replicates): ApSp, Aphanocalyx sp. (= 9); BiLe, Bikinia le-testui (= 6); GiOg, Gilbertiodendron ogoouense (= 6); TeBi, Tetraberlinia bifoliolata (= 27); MaEx, Marquesia excelsa (= 12); UaGu, Uapaca guineensis (= 44); UaHe, Uapaca heudelotii (= 21).

Figure 3.

Distance decay in potential enzymatic activities along increasing phylogenetic distance in the /russula-lactarius lineage of ectomycorrhizal fungi. Circles and error bars represent median values and their 95% CIs. Within phylogenetic distance classes (< 3%; 3–8%; 8–13%; 13–20%; 20–30%), enzymatic profiles are more similar than expected, when CIs lie above the zero-value.

The activities of acid phosphatase (t6,108 = 3.66; < 0.001; Fig. 2d) and β-glucosidase (t6,115 = 3.15; = 0.002) differed significantly among host tree species, but Tukey test conservatively rejected significance in pairwise differences for β-glucosidase. The activities of both enzymes were somewhat higher in Aphanocalyx sp. Soil horizon had no significant effect on the potential activities of any studied enzyme.

Stable isotopes

The fungal mantle and inner root core (excluding cortical cells with Hartig net) contributed 32.7 ± 9.5% (mean ± SD) and 43.1 ± 8.4% to the volume of ECM root tips, respectively, but these proportions were not related to plant or fungal taxa or exploration types. Based on the N and C concentrations in NM and ECM root tips and the formula for volume of a cylinder, we calculated that 58.1 ± 10.8% N and 42.7 ± 9.6% C are distributed in fungal tissues assuming that all root tips have similar tissue density; NM roots and the core of ECM roots have similar N concentrations; and fungi contribute 50% to the mass of Hartig net.

The understorey foliage of both ECM and AM trees was strongly depleted in 13C, but only slightly depleted in 15N compared with soil, ECM root tips and fungal fruit bodies (Fig. 4). The foliage of ECM plant species was significantly more enriched in 15N than that of AM plants (t1,9 = 6.61; < 0.001), but there were no significant differences in δ13C values between the foliage of ECM and AM trees. In paired samples of foliar and soil isotopes, ECM plants were significantly more enriched in 15N than were AM plants with respect to both organic (F1,21 = 15.7; < 0.001) and mineral soil horizons (F1,21 = 21.4; < 0.001).

Figure 4.

Stable isotope signatures (13C, 15N) of plant foliage, soil and fungi. Error bars and values in parentheses indicate 95% standard errors and the number of replicate samples, respectively. Diamonds indicate fungi with unknown trophic strategies. AM, arbuscular mycorrhiza; ECM, ectomycorrhiza.

Ectomycorrhizal colonization of root tips had a significant effect on their stable isotope patterns. NM root tips of ECM trees were, on average, 4.7‰ enriched in 13C compared with the foliage of ECM trees, but had similar concentrations of 15N. Conversely, ECM root tips were significantly more enriched in both 13C (5.3‰) and 15N (2.0‰) compared with ECM foliage. Thus, ECM root tips had, on average, 2.5‰ higher δ15N values than NM root tips. Soil and ECM root tips had similar stable C and N isotope signatures. The top mineral soil was slightly enriched in both 13C and 15N compared with organic soil. On average, ECM fruit bodies showed 2.8‰ higher δ13C values and 1.2‰ higher δ15N values than ECM root tips. Fruit bodies of nonECM fungi had highly variable δ13C and δ15N values. By excluding Hygrocybe spp. and four other fungal species as outliers, the putatively saprotrophic fungi had only 1.4‰ higher δ13C values than the ECM fungi (F1,28 = 7.07; = 0.013), but there was no difference in δ15N values (0.9‰, F1,28 = 0.504; = 0.484). Members of the genus Hygrocybe and two specimens of the orders Agaricales and Gomphales were highly enriched in 15N compared with both ECM and saprotrophic fungi and other ecosystem components. However, their δ13C values were similar to or slightly below those of respective roots and organic soil (Fig. 4). Comparison of ITS-large subunit rRNA gene (LSU) sequences of these fungal groups with INSD did not reveal their potential lifestyle, because they shared < 85% ITS sequence similarity with available sequences (not shown).

Generalized least-squares models revealed that both the δ13C and δ15N values of root tips were significantly dependent on their ECM colonization (δ13C: t1,126 = 2.29, = 0.024; δ15N: t1,126 = 6.15, < 0.001) and soil horizon (δ13C: t1,126 = −2.29, = 0.026; δ15N: t1,126 = −2.70, = 0.008). Host tree species had an additional, highly significant effect on δ13C values of all root tips (t5,126 = 3.51; P < 0.001) and ECM root tips (t5,92 = 3.19; = 0.002). In particular, root tips of U. heudelotii had significantly lower δ13C values than those of U. guineensis and B. le-testui. By contrast, ECM lineage had a significant effect on δ15N of ECM root tips (t5,92 = 3.46; = 0.001). The lineages of /russula-lactarius, /tomentella-thelephora and /sebacina were significantly more enriched compared to /boletus and /coltricia according to a Tukey test (Fig. 5). The variation of stable isotope ratio was generally high among ECM fungal lineages. The proportion of fungal mantle and inner root core and root tip N concentration had no effect on stable isotope patterns. Potential enzymatic activities of ECM root tips were not significantly related to the patterns of stable isotopes either in separate or in combined analyses.

Figure 5.

Stable isotope signatures of fruit bodies of saprotrophic fungi (circles), fruit bodies of ectomycorrhizal (ECM) fungi (squares) and ECM root tips (triangles). (a) Isotope ratio of each sample; (b) isotope ratio of the most frequent ECM fungal lineages. Error bars denote 95% CI. Different colours indicate different lineages: yellow, undefined lineage within Atheliales (= 4 for ECM; = 0 for fruit bodies); orange, /clavulina (= 1;3); pink, /inocybe (= 0;2); red, /russula-lactarius (= 43;11); green, /marcelleina-peziza gerardii (= 2;0); blue, /boletus (= 13;7); cyan, /hysterangium (= 0;1); purple, /elaphomyces (= 1;0); dark brown, /coltricia (= 5;0); grey, /sebacina (= 5;0); black, /tomentella-thelephora (= 18;0); white, /amanita (= 1;0).


Enzymatic patterns

Ectomycorrhizal colonization of root tips was associated with greater potential activities of six out of eight tested extracellular enzymes involved in degradation and release of nutrients from organic materials. This is in accordance with results from a pot experiment, with ECM roots of Populus spp. possessing one to two orders of magnitude greater potential enzymatic activities than NM roots (Courty et al., 2011). The activities of leucine aminopeptidase and β-glucuronidase did not differ between NM and ECM root tips, indicating that most ECM fungi do not produce these two enzymes in significant amounts in this rainforest ecosystem.

Most potential enzymatic activities of ECM root tips were dependent on exploration type, particularly the presence or absence of rhizomorphs, rather than on fungal lineage or host species. This suggests either convergent evolution in fungi with similar ectomycorrhiza morphology or possible leaking of cytoplasm and enzymes in mycelium-rich types in which more mycelium was removed during sample preparation. The latter possibility is unlikely, because the assayed enzymes are extracellular (leucine aminopeptidase and laccase may have additional intracellular functions). Although extraradical mycelium may have an important role in secreting these enzymes (Finlay, 2008), the potential enzymatic activities were relatively higher in the ECM root tip surface of long- and medium-distance fringe exploration types than in types producing no rhizomorphs (i.e. contact and short-distance types). By contrast, the potential activity of laccase was relatively higher in the contact exploration type, suggesting that the mantle surface may have an important physiological role in species with no or very sparse extraradical mycelium. Laccase activity has been suggested to play a role in nutrient release directly from rotting leaves or dead wood for fungi with contact exploration types, such as Lactarius quietus and Lactarius subdulcis (Courty et al., 2007; Rineau & Garbaye, 2009). Tropical ECM fungi may thus have a similar function in the rapidly degrading leaf litter and organic horizon (Brearley et al., 2003; Mayor & Henkel, 2006). It is possible that ECM fungi in topsoil and litter also access nutrients from saprotrophic mycelium (Lindahl et al., 2002) or target simple organic compounds degraded by saprotrophic fungi (Tedersoo et al., 2003) rather than investing in the costly enzymatic complex by themselves.

Genetic similarity among ECM fungi had contrasting effects on enzymatic patterns, depending on both the phylogenetic scale and the enzyme. On a higher taxonomic scale, phylogenetic origin of ECM fungi (i.e. lineage) played an important role in determining the potential activities of acid phosphatase, leucine aminopeptidase and cellobiohydrolase. Leucine aminopeptidase was produced by members of only a few fungal lineages. The activities of both leucine aminopeptidase and acid phosphatase were particularly high in an undefined lineage within Atheliales that has so far been found exclusively in tropical forests (Peay et al., 2010; Tedersoo et al., 2010b; Smith et al., 2011; Phosri et al., 2012) and lacks representative fruit body sequences in INSD. The activity of cellobiohydrolase was relatively higher in the /boletus and /tomentella-thelephora lineages than in other groups. In comparison, Buée et al. (2007) found high activity of cellobiohydrolase in only two out of six species of Tomentella, but not in Xerocomus sp. ( /boletus lineage) in upper mineral soil of a temperate deciduous forest. This indicates that congeneric species may differentially adapt their enzymatic capacities in different ecosystems (Buée et al., 2007; Rineau & Courty, 2011). On a lower taxonomic scale, enzymatic patterns were more similar in conspecific isolates and closely related species within the /russula-lactarius lineage, but not within the /tomentella-thelephora lineage or across all fungi. This was not, however, related to the evolutionary conservatism in ECM morphology in the /russula-lactarius lineage (Eberhardt, 2002; Agerer, 2006; our analysis comprising local fungi).

The significant effect of plant species on enzymatic activities of ECM root tips was related to the relatively higher activity in Aphanocalyx sp. and lower activity in M. excelsa in our study. Because this trend was observed in six enzymes (significant in two cases), this phenomenon could be interpreted as differential root vitality or C investment among tree species (Mosca et al., 2007). However, more data are clearly required to address the relative importance of genetic and environmental components of plant influence on enzymatic activities in ECM fungi (Courty et al., 2011).

Stable isotopes

Vegetation in the rainforest of Gabon had higher δ15N, but similar δ13C values compared with other African rainforest and woodland communities (Högberg, 1990; Högberg & Alexander, 1995; Cerling et al., 2004). The relatively low δ13C values of foliage could be explained by high stomatal conductance because of nonlimiting water supply (Bowling et al., 2008; Diefendorf et al., 2010). In addition, understorey leaves may take up 13C-depleted CO2 from soil respiration (Gebauer & Dietrich, 1993; Cerling et al., 2004). The relatively high 15N enrichment in foliage is characteristic of the generally open N cycling in tropical ecosystems (Pardo & Nadelhoffer, 2010) and reflects the high δ15N values (6.5‰ on average) of the shallow organic horizon that may result from the paucity of the highly 15N-enriched protein complex in these clayey soils (Hobbie & Ouimette, 2009).

In boreal and arctic ecosystems, AM and NM plants are generally more enriched in 15N than ECM and are ericoid mycorrhizal plants (Michelsen et al., 1996; Bowling et al., 2008), whereas ECM and AM trees of tropical ecosystems usually have similar stable isotope patterns (Högberg & Alexander, 1995; Schmidt & Stewart, 1997; Cerling et al., 2004). In this study, the leaves of ECM trees from three families were, on average, 2.6‰ more enriched in 15N than AM trees. Our results are in accordance with previous observations in miombo woodlands of East Tanzania, where ECM trees were enriched by 1.5–2.7‰ in 15N compared with AM trees (Högberg, 1990). Both study sites are characterized by predominately clay-rich soils, whereas sandy soils dominate in African sites where foliage of ECM and AM trees have similar δ15N values (Högberg & Alexander, 1995; Cerling et al., 2004). Clay and silt particles accumulate more organic matter and reduce the turnover rate of soil organic C and N compared with sandy soils (Six et al., 2002), suggesting that ECM and AM plants may have differential access to these clay-bound, relatively 15N-rich sources of organic matter.

Except for the high δ15N values of saprotrophic fungi (8.1‰ on average), the stable isotope patterns of fruit bodies of ECM fungi and saprobes are generally consistent with previous findings from other rainforest ecosystems in Southeast Asia and South America (Kohzu et al., 1999; Mayor et al., 2009). The high δ15N values of saprotrophic fungi probably reflect their relatively 15N-enriched substrate – soil and leaves (Gebauer & Taylor, 1999; Trudell et al., 2004). The relatively low difference in δ15N in fruit bodies of ECM fungi compared with saprotrophic fungi (0.9‰ on average) and foliage of ECM host trees (3.7‰ on average) suggest that a relatively low proportion of fungal N is transported to the host plants. In addition, ECM and saprotrophic fungi may target the same highly 15N-enriched organic N sources that remain inaccessible for AM fungi and AM plants as suggested by the higher δ15N values in the foliage of ECM plants compared with AM plants in clayey tropical soils. Furthermore, soil horizon had a negligible effect on both the patterns of stable isotopes and enzymatic activities and taxonomic distribution of ECM fungi in the study site (Tedersoo et al., 2011). Taken together, our results are consistent with the hypothesis that N cycling in the poorly stratified, phosphorus-limited tropical soils may differ substantially from N cycling in the generally N-limited boreal soils with slow decomposition processes and well-differentiated soil profile (Six et al., 2002; Lindahl et al., 2007; Hobbie & Ouimette, 2009; Pardo & Nadelhoffer, 2010). Because of multiple interacting determinants of the δ15N signature as well as differences in ecosystem types and N availability (Pardo & Nadelhoffer, 2010), interpretation of the observed ecosystem-level stable isotope patterns requires direct analyses of N fluxes and therefore remains open to debate.

Many fungi with fruit bodies arising from soil, such as Hygrocybe and unidentified members of Agaricales and Gomphales, had very high δ15N values and low δ13C values that strongly deviate from known saprotrophic and mycorrhizal fungi. In the genus Hygrocybe, these discrepant stable isotope patterns are consistent with previous findings from grasslands and swamps and may indicate as yet undescribed biotrophic interactions or mechanisms of isotope fractionation (Griffith et al., 2002; Seitzman et al., 2011).

Ectomycorrhizal status, host and fungal identity had differential effects on δ13C and δ15N signatures of root tips. Ectomycorrhizal and NM root tips of ECM trees had a similar δ13C pattern, but ECM root tips had, on average, 2.6‰ higher δ15N values than NM root tips, which is consistent with findings from temperate and boreal forests (Högberg et al., 1996; Zeller et al., 2007). We estimated that the fungal mantle and inner root core (excluding Hartig net region) contribute 33 and 43%, respectively, to the volume of ECM root tips, but these proportions were not related to plant or fungal species or exploration type. Furthermore, the fungal to plant ratio and N concentration of root tips had no effect on stable isotope patterns, suggesting that the observed differences in δ15N patterns among ECM fungal lineages are not confounded by these variables.

Internal N cycling in fungal mycelium and N transport to host plants are among the key factors affecting δ15N patterns in tissues of ECM fungi (Handley et al., 1996; Högberg et al., 1996, 1999a). Fungal tissues become gradually more enriched in 15N from ECM root tips to soil extraradical mycelium, fruit body stipes and further on to gills by up to 12‰ (Zeller et al., 2007), which is attributable to the high protein to chitin ratio in reproductive organs (Taylor et al., 1997) and the relative biomass of extraradical mycelium (Hobbie & Agerer, 2010). Because of these constraining processes and variables such as soil exploration depth (Agerer et al., 2012) that influence the 15N natural abundance in fruit bodies, stable isotope analysis of root tips could provide a useful additional proxy to address nutritional aspects of ECM fungi and plants by allowing one to control, at least partly, for the effects of soil horizon and internal N cycling. Most importantly, sampling of root tips provides access to all ECM fungal taxa, including species lacking sexual reproduction or producing resupinate fruit bodies that often dominate in various ecosystems in terms of both biomass and species richness (Horton & Bruns, 2001; Tedersoo & Nara, 2010). Access to different sources of N and differential efficiency in N transfer to host plants, but also slight differences in the plant to fungal ratio in the ECM tissues, potentially contribute to the δ15N patterns of ECM root tips.

The lack of a significant relationship between stable isotopes and potential enzymatic activities indicates that natural abundance of stable isotopes does not reflect the potential activities of these particular enzymes individually or taken as a complex. Assays of potential enzymatic activities assume that enzyme abundance rather than substrate availability determines its function, but in natural conditions both properties may be important. Furthermore, in predominately phosphorus-limited tropical forests, ECM fungi may be targeting organic phosphorus or other microelements instead of N (Pardo & Nadelhoffer, 2010). Many other enzymes are involved in degradation of organic compounds and release of nutrients, which might better explain the δ15N signature than the eight enzymes studied by us. The relative contribution of N source partitioning and degradation abilities of organic compounds to 15N natural abundance still remains poorly understood in both the natural and experimental conditions (Lilleskov et al., 2002; Brearley et al., 2005).


Ectomycorrhizal fungi are mostly responsible for producing degradation and nutrient-releasing enzymes in root tips. Fungal species that possess abundant mycelium in soil are relatively more active in producing extracellular enzymes, except for laccase, at the surface of their mycorrhiza. The full enzymatic potential of these fungi can only be addressed in studies on extraradical mycelium that provide relatively larger surface area for enzymes to act on potential substrates and for nutrient uptake (Finlay, 2008). Unfortunately, such analyses cannot be performed in natural conditions because hyphae of multiple species intermingle and compete for nutrients. Our study demonstrates that the fungal partner may affect both potential enzymatic activities and δ15N profiles of ECM root tips at various phylogenetic scales, although there is no apparent correlation between the two indirect measures of nutrient acquisition. δ15N signature of root tips could provide an additional proxy for nutrition, because root tip-based assays enable the improvement of taxonomic coverage and control for the constraining exploration depth and relative mycelial biomass.


We thank A. Sadam, M. Öpik, M. Moora, M. Zobel, and H. Tamm for assistance in the field and laboratory. We are indebted to three anonymous referees and the editor, M-A. Selosse, for constructive comments on earlier versions of the manuscript. This study received support from the Estonian Science Foundation (grants no 7374, 8235, 9286) and FIBIR.