Portal de Periódicos da CAPES

Stable isotope probing implicates a species of C ortinarius in carbon transfer through ectomycorrhizal fungal mycelial networks in A rctic tundra

2015; Wiley; Volume: 210; Issue: 2 Linguagem: Inglês

10.1111/nph.13797

1469-8137

Julie R. Deslippe, Martin Hartmann, Susan J. Grayston, Suzanne W. Simard, William W. Mohn,

Lichen and fungal ecology

Mycorrhizal networks (MNs) occur when the mycelium of one or more mycorrhizal fungus colonizes two or more nearby plants (Molina et al., 1992). MNs can serve as pathways for the transfer of carbon (C), nitrogen (N), phosphorus (P), water, defense signals and allelochemicals among plants (see Simard et al., 2012, for a review) but, to the best of our knowledge, the fungi involved in these resource transfers among green plants in the field have never been directly identified. This information is important because it would reveal the unique roles of mycorrhizal fungal species, shedding light on some of the functions of these complex belowground networks, and potentially providing additional ecological context for the increasingly large body of molecular community data that is accumulating for mycorrhizal fungi (Blaalid et al., 2014; Horn et al., 2014; Morgado et al., 2015). MNs mediate plant–plant interactions with potential implications for plant diversity at local and regional scales (Perry et al., 1989; McGuire, 2007; Deslippe & Simard, 2011) and are likely to have foundational roles in the structure and regeneration of terrestrial ecosystems (Simard, 2009; Simard et al., 2012). Thus, an improved understanding of MNs may lead to more appropriate and effective land conservation and ecological restoration practices. In recent decades, regional warming associated with anthropogenic climate change has led to increased plant biomass across the Arctic tundra biome (Jia et al., 2003; Chapin et al., 2005; Macias-Fauria et al., 2012). Differences in the relative productivity of plant species have led to altered plant community compositions, with ectomycorrhizal (EM) shrub species increasingly dominating in many regions (Myers-Smith et al., 2011; Bonfils et al., 2012). In the moist-acidic tundra of Arctic Alaska, the EM shrub Betula nana has increased most strongly (Sturm et al., 2001) and this effect is further enhanced through experimental warming (Chapin et al., 1995; Sistla et al., 2013). When warmed, the EM fungal (EMF) community associating with B. nana shifts from being dominated by members of the Russulaceae to one dominated by Cortinarius spp. (Deslippe et al., 2011). Cortinarius spp. are C-demanding, rhizomorph-forming basiodimycetes that grow extensive mycelia in soil (Agerer, 2001, 2006). Some Cortinarius species produce highly efficient oxidative enzymes (Bodeker et al., 2009, 2014), which they utilize to mobilize growth-limiting N for their host from complex soil organic matter (Lilleskov et al., 2002; Hobbie & Agerer, 2010; Hobbie et al., 2013). Recent work has highlighted the importance of N mobilization by Cortinarius spp. in maintaining rates of C-cycling in other high-latitude ecosystems (Clemmensen et al., 2013, 2015; Lindahl & Tunlid, 2015) and similar processes could be involved where shrubs are spreading in Arctic tundra as climate warms. Indeed, the transition from ericaceous tundra heath to EM shrub tundra is accompanied by significantly higher rates of fungal hyphal growth and C turnover and lower soil organic C stocks in Swedish Lapland (Parker et al., 2015). Studies of C-transfer through EMF mycelial networks have often utilized stable or radiocarbon isotopes and employed 'pulse-chase' methodology to trace photosynthetic C from a labeled 'donor' plant to nearby 'receiver' plants. These studies have revealed that the magnitude of C transfer through EMF mycelial networks is highly variable in nature, ranging from < 1% to 10% of the donor plant's net photosynthesis (Simard et al., 1997a; Teste et al., 2010). This variability reflects plant physiological factors that affect the magnitude of source–sink gradients for C among networked plants, and fungal factors, such as the extent of fungal colonization of roots, the composition of EMF community and the continuity of the hyphal pathway (see Simard et al., 2012, for a review). Indeed, small but statistically significant C-transfer to receiver plants through soils and a discontinuous hyphal pathway have been observed (Philip et al., 2010; Deslippe & Simard, 2011), suggesting a possible additional role for bacteria, saprotrophic fungi, or other members of the rhizosphere community in C-transfer among plants. Bacterial-mediated C transfer among plants could occur, for example, if a donor plant's rhizodeposit C was acquired by a rhizosphere bacterium which subsequently became an endophyte of the receiver plant (Rosenblueth & Martinez-Romero, 2006). Previously, through 13CO2 pulse-chase labeling of B. nana plants in Arctic tundra we showed significant transfer of C through EMF mycelial networks to aboveground tissues and rhizomes of B. nana receivers in conspecific pairs only. We also found low but nonzero C transfer among B. nana pairs through soil pathways. Here we report the use of stable isotope probing (SIP) of phospholipid fatty acids (PLFAs) paired with a DNA-SIP-pyrosequencing approach of root samples collected in that study to test the hypotheses that: (1) EMF were the primary conduits for C among B. nana individuals, and (2) Cortinarius were more important conduits for C than other members of the EMF community associating with B. nana. A significant and unique role for one or more species of Cortinarius in C-transfer among B. nana individuals would constitute evidence that EMF species in mixed communities perform particular functions for their host, one step towards elucidating the functions of these diverse relationships. This study took place in Low Arctic tussock tundra at Toolik Lake, Alaska, USA (68°38ʹN, 149°34ʹW). The plant community at this site is co-dominated by the EM dwarf shrub B. nana L. and the nonmycorrhizal sedge Eriophorum vaginatum, with an understory mainly of ericaceous plant species and mosses (Supporting Information Table S1). B. nana 'donor' plants were sealed in gas-tight chambers and pulsed with 3.2 mmol 13CO2 (Fig. 1a). B. nana donors were removed from labeling chambers when CO2 concentrations inside the chamber fell below ambient concentrations (c. 4 h; Fig. 1b). After a 7-d chase period, stems, leaves and rhizomes of the B. nana donor and all potential 'receiver' plant species present in the 55 cm-diameter study plots were harvested and 13C-tissue content analyzed (Fig. 1c). We found statistically significant 13C-transfer only among pairs of B. nana plants (Deslippe & Simard, 2011). We sampled the roots of each donor B. nana plant as well as one independent receiver B. nana plant per plot. Independent receiver B. nana plants were not connected to the donor plant via belowground plant tissues (e.g. rhizomes or root grafts). The independence of the receiver plant was determined at the time of destructive sampling of the plot. Where more than one B. nana receiver occurred, the largest plant was selected for root sampling. Likewise, we sampled the roots of an unlabeled B. nana plant growing a minimum distance of 2 m upwind from each 13C-CO2 labeling plot (12C control plant; Fig. 1c). Fine roots occurred predominantly in the organic soil horizon. We collected B. nana roots by tracing the main stem to belowground rhizomes and these to clusters of root tips, which could then be sampled in an intact state. Considerable effort was made to sample the entire fine root system of each plant, but it is possible that some fine root clusters were missed. Each sample consisted of a minimum of three terminal root clusters and adhering organic soil particles, these were 5–10 cm3 in volume, and all were visibly EM. Samples were placed immediately on ice and frozen at −80°C within 2 h. They remained frozen during transport to the laboratory. For PLFA-SIP, we selected root samples from pairs of B. nana (one donor, one receiver) in six experimental plots from the Deslippe & Simard (2011) study, selecting the plots where the greatest total 13C-enrichment of the receiver plant tissues (sum of leaf, stem and rhizome 13C-contents) had been observed. We also sampled the roots of 12C control plants adjacent to each plot (Fig. 1c). The purpose of the 12C-control plant root sample was to provide estimates of the natural abundance of 13C in microbial PLFAs. Total lipids were extracted from c. 1 g freeze-dried rhizosphere soil, fractionated, and the phospholipid fraction was trans methylated as described by Bengtson et al. (2009). Fatty acid methyl esters (FAMEs) were analyzed by capillary GC-combustion-isotope ratio mass spectrometry (GC-C-IRMS) at the Stable Isotope Research Unit in the Department of Crop and Soil Science, Oregon State University, Corvallis, OR, USA. FAMEs were identified as described by Williams et al. (2006) and Butler et al. (2003) and quantified using 19:0 methyl ester as the internal standard. We identified 21 FAMEs containing 14–20 C atoms. The δ13C values of individual PLFAs were determined according to Williams et al. (2006), where the atomic 13C excess of PLFAs in samples was calculated relative to the mean δ13C value for the corresponding PLFA for the root samples of the six unlabeled control B. nana plants. Values were expressed as ng PLFA 13C incorporation g−1 dry weight soil. We used the 13C enrichment of 18:2ω6,9 as an indication of fungal 13C incorporation in each sample. As a measure of bacterial 13C incorporation, we used the sum of the enrichment of the following PLFAs: i15:0, a15:0, 15:0, i16:0, 10Me16:0, i17:0, a17:0, cy17:0, 17:0, 18:1n7, 10Me18:0 and cy19:0 (Frostegard & Baath, 1996). We assessed correlations among the 13C-enrichments of fungal and bacterial PLFAs and dry weight of plant tissue 13C content using Pearson's product-moment correlation in SPSS v.22 (IBM SPSS Statistics, Armonk, NY, USA). For each PLFA, the proportion of 13C-enrichment per sample was then calculated. Means are reported ± 1 standard error (SE). For DNA-SIP, we selected root samples from the receiver B. nana in three experimental plots together with root samples from each adjacent unlabeled 12C-control plant (Fig. 1c). DNA was extracted twice from 0.5 g of fresh-frozen root sample using a FastDNA Spin Kit for Soil (MP Biomedicals, Solon, OH, USA). The replicate DNA extractions were then pooled and quantified using a NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). DNA-SIP followed the 'fractionation' method without EtBr described by Neufeld et al. (2007). Gradients were formed using a Beckman Coulter Ultracentrifuge fitted with a Vti 65.2 rotor (Beckman Coulter Canada LP, Mississauga, ON, Canada). Following Gallagher et al. (2005), we included 10 μg '13C-carrier DNA' within each ultracentrifuge tube. 13C-DNA was generated by extracting DNA from Escherichia coli grown in Minimal Media liquid culture, with 13C-glucose as a sole C source. Gradients were displaced and separated into 12 fractions with water colored with toludine blue using a Razel Clinical Syringe Pump (Razel Scientific Instruments Inc., Georgia, VT, USA). The total DNA content of each fraction was quantified by agarose gel electrophoresis through comparison to a known quantity of 1 kb Trackit™ quantification ladder (Life Technologies, Waltham, MN, USA) using AlphaView imaging software (Protein Simple, Toronto, ON, Canada). 13C-enriched 'heavy' DNA was contained in fractions 7 and 8, while fraction 10 contained 'light' 12C-DNA. We pooled fractions 7 and 8 from each receiver or control plant for a single 'heavy' 13C DNA fraction per plant. Eukaryotic ribosomal internal transcribed spacers (ITS2) were amplified from the heavy and light fractions using the primers ITS3/ITS4 (White et al., 1990) and sequenced using 454-pyrosequencing GS-FLX Titanium technology (Roche 454 Life Sciences, Branford, CT, USA) at the Genome Quebec Innovation Centre, Montreal, Canada. We pyrosequenced heavy and light fractions of the receiver plants from the three experimental plots individually, but pooled the heavy fractions (7 and 8) of all three unlabeled control plants for a total of seven pyrosequencing samples (i.e. 'Biosamples'). The fungal community of the biosample derived from the pooled heavy fractions of the unlabeled control plant roots was compared to that derived from the heavy fractions of receiver plants from experimental plots. This served to confirm that the fungi identified as being enriched in the heavy fractions of receivers from experimental plots did not also occur in significantly higher abundances in samples with only natural abundance of 13C. This situation could arise, for example, for taxa with higher than average G + C contents as the lower buoyant densities of high G + C content sequences could cause them to migrate to the heavier fractions. Details of the PCR and sequencing methodologies are described by Hartmann et al. (2012). Our DNA-SIP pyrosequencing approach yielded 82 599 raw sequence reads, which were submitted to the European Nucleotide Sequence Read Archive under the study accession no. PRJEB8276. Procedures for pyrotag processing, operational taxonomic unit (OTU) delimitation, and assigning taxonomic affiliations to OTUs are described by Hartmann et al. (2014). Briefly, curated sequences were clustered into OTUs at an identity threshold of 97% using the unsupervised Bayesian clustering algorithm CROP (Hao et al., 2011). OTUs were queried against the UNITE database (Abarenkov et al., 2010) and assigned taxonomy using the naive Bayesian classifier (Wang et al., 2007) and a minimum bootstrap support of 60% in Mothur (Schloss et al., 2009). These taxonomic assignments were used for all statistical analyses, although the taxonomic resolution achieved through analysis of the ITS2 amplicon differed among OTUs. We then used manual GenBank-BLASTn searches of sequences to gain additional insight to the taxonomic affiliations of OTUs that differed in abundance among the heavy and light fractions. Additionally, we aligned all unique Cortinarius-affiliated sequences with closely related INSD sequences that were derived from vouchered Cortinarius specimens and created a phylogenetic tree using Geneious v.8.0.5 (Kearse et al., 2012) (Fig. S1). We used indicator species analysis to determine the fidelity of OTUs to the heavy and light fractions. Indicator values were calculated with the method of Dufrene & Legendre (1997) using Monte Carlo tests of significance with 5000 permutations, as implemented in PC-ORD v.6.19 (McCune & Mefford, 1999). In addition, we used indicator species analysis to examine the fidelity of the sum of all Russulaceae-affiliated sequences to heavy or light fractions. Similarly, we examined the fidelity of the sum of all Cortinarius-affiliated sequences that did not belong to the dominant Cortinarius OTU, to heavy and light fractions. We focused on members of the Russulaceae and on Cortinarius spp. because they are common members of the EMF community associating with B. nana at this site (Deslippe et al., 2011) and because the identification of several Russulaceae-affiliated, and Cortinarius sp. OTUs could have reduced our ability to detect significant 13C-enrichment of these groups overall. All means are reported ± 1 SE. PLFA-SIP analysis of the root samples of six B. nana pairs indicated greatest enrichment of the saturated fatty acid 16:0, which is common to both prokaryotes and eukaryotes (Fig. 2), and thus reflects 13C-incorporation by plant roots as well as by soil organisms. The fungal biomarker 18:2ω6,9 showed the next greatest enrichment, with the roots of receiver plants containing on average, one-third of the absolute 13C-enrichment (0.016 ± 1.2 × 10−3 ng 13C-PLFA g−1 dry weight (DW) soil) as 18:2ω6,9 in the roots of donor plants (0.058 ± 4.2 × 10−3 ng 13C-PLFA g−1 DW soil). For all samples, 18:2ω6,9 showed much greater 13C-enrichment than did the sum of 13C-enrichment of all bacterial PLFAs. Mean proportional 13C-enrichment of the fungal biomarker was nearly seven-times that of the bacterial biomarkers, and these means did not differ significantly between root samples of donor and receiver plants (Fig. 2). These results suggest a prominent role for fungi relative to bacteria in 13C-transfer among pairs of B. nana in the field, providing strong support for our first hypothesis. Plant tissue 13C enrichment of donors and receivers was significantly positively correlated to the 13C enrichment of 18:2ω6,9 (Kendall's tau = 0.455, P = 0.04), but not significantly correlated to the sum of 13C-enrichment of all bacterial PLFAs, or to the sum of total PLFA 13C-enrichment. These findings suggest that fungal, rather than bacterial 13C enrichment is more closely linked to plant 13C tissue content. In the case of donor plants, this is consistent with the observation that plant C is acquired first by mycorrhizal fungi before being dissipated through the saprotrophic fungal and bacterial community (Treonis et al., 2004; Drigo et al., 2010; Churchland et al., 2012; Kaiser et al., 2015). In the case of receiver plants, this observation provides support for a direct hyphal link as a pathway for C-transfer among B. nana pairs. The 13C enrichment of 18:2ω6,9 was also significantly positively correlated to both total and bacterial PLFA 13C-enrichment (Kendall's tau = 0.818, P < 0.001; Kendall's tau = 0.545, P = 0.01, respectively), indicating that fungi dominated acquisition of 13C from plants and suggesting that they may have mediated subsequent acquisition of 13C by the bacterial community, a finding that agrees with other recent work (Kaiser et al., 2015), and provides additional support for our first hypothesis. Our DNA-SIP pyrosequencing approach yielded 31 996 fungal sequences that clustered into 640 OTUs at a 97% sequence identity threshold (Table S2). A curated file in FASTA format of all fungal ITS sequences generated in this study is appended to this manuscript (Methods S1). Overall, the rhizosphere fungal community was moderately even (Pielou's J ′ = 0.56), with the most abundant OTU comprising only 15% (4734) of all fungal sequences, while the seven next most abundant OTUs comprised between 5% and 9% each (Table S2). The proportional abundances of most OTUs were similar in the heavy and light fractions of samples. The most abundant OTU, an unclassified member of the Ascomycota with 100% identity to the uncultured fungus from boreal Alaska (INSD accession no. JN889866), was a prominent example of this (Fig. 3; Table S3). However, the second most abundant OTU17 (11.2%, 3017) was on average two orders of magnitude more abundant in the heavy than light fractions (0.12 ± 0.033 vs 0.0012 ± 0.00057; Fig. 3; Table S3), while it was absent from the heavy fractions derived from the unlabeled 12C-control plant rhizospheres (Fig. S2). OTU17 was detected as the sole significant indicator of the heavy fraction at α = 0.1 (P = 0.093; Fig. 3; Table S4). This finding provides support for our second hypothesis that Cortinarius sp. was an important conduit for C among B. nana plants relative to other members of the root-associated fungal community. The 3017 sequences that clustered into OTU17 represented 45 different ITS2 sequences, which had 98–100% sequence identity to ITS2 amplicons derived from vouchered specimens of Cortinarius collinitus, C. favrei, C. fennoscandicus, C. muscigenus, C. septentrionalis and C. stillatitius, with C. collinitus returned as the closest match for 41 of 45 sequences. Phylogenetic analysis indicated that ITS2 is conserved among these morphological species (Fig. S1) and we are therefore unable to provide more taxonomic detail. However, C. fennoscandicus and C. septentrionalis are considered to be exclusively associated with Betula (Bendiksen et al., 1993). We observed numerous Cortinarius spp. fruiting at the study site. One of these, which had a very slimy cap, a character consistent with Cortinarius section Colliniti (Bendiksen et al., 1993), was much more common than the others (Fig. S3). This Cortinarius sp. occurred in very high abundance from early August until snowfall each year. Here we show that one or more species of Cortinarius was very likely the pathway for C-transfer among B. nana plants in Arctic tundra. While C-transfer through EM networks to receiver plant tissues has been well documented for nearly two decades (Simard et al., 1997a,b; Wu et al., 2001; Bingham & Simard, 2011; Deslippe & Simard, 2011), its ecological relevance has been the subject of much debate (Fitter et al., 1999; Simard et al., 2002; van der Heijden & Horton, 2009). The crux of this debate has hinged on the counterintuitive notion that a heterotrophic soil fungus would relinquish growth-limiting C to its photoautotrophic host plant (Simard et al., 2012). While this situation is the norm for fungi and plants in mycoheterotrophic relationships, green plants are typically strong carbohydrate sources relative to their symbiotic fungi and the movement of C against this concentration gradient is expected to be minute (Simard et al., 2012). By contrast, EMF are often N sources relative to their host plants, and significant C-transfer among green plants through MNs would be expected if organic N was a dominant form in which C moves. We propose that C-transfer among B. nana through EM networks may reflect the movement of amino acid N from Cortinarius to B. nana. In the moist acidic tussock tundra near Toolik Lake Alaska soil N in inorganic and free amino acid forms is typically < 4 μg g−1 DW soil over the growing season (Weintraub & Schimel, 2005a) and surface soil layers contain > 72% organic matter (Marion et al., 1997). In these soils, seasonal peaks in proteolysis coincide with periods of maximal root growth, when soil NH4 and free amino acid concentrations are at their lowest values (Weintraub & Schimel, 2005b). EMF oxidative exoenzymes (Bodeker et al., 2009, 2014) are likely to contribute significantly to these cycles, securing a high proportion of the 61–86% of foliar N that EMF supply to plants (Hobbie & Hobbie, 2006). Once taken up by plant and EMF transporters, amino acids undergo rapid metabolic conversions involving a wide range of compounds within roots (Nasholm et al., 2009), and presumably hyphae. These metabolic conversions may include deamination via the GS/GOGAT (Glutamine synthetase/Glutamate-2-oxoglutarate aminotransferase) cycle, yielding NH4. In plant roots, most NH4 is incorporated to new amino acids before transport occurs (Nasholm et al., 2009, and references cited therein). If similar conversions occur within the EMF mycelium, it seems plausible that an EMF with proteolytic activity could generate relatively high mycelial concentrations of newly formed amino acids, that is, amino acids containing soil-derived N and recent plant photosynthate C. These new amino acids could then move down concentration gradients to be acquired by plants connected in the MN. While studies to elucidate the forms of N transferred from EMF to their host plants are few, glutamine, which contains five C atoms for every two N atoms, was found to be the primary molecule through which N is transferred from members of the Russulaceae to Fagus sylvatica (Martin et al., 1986). Moreover, amino acid transporters are abundant in plant genomes (Rentsch et al., 2007) and strongly upregulated during symbioses with EMF (Martin et al., 2010). Thus, while the high 13C enrichment of Cortinarius spp. DNA we observed in the roots of receiver B. nana plants constitutes evidence that these plants are linked to donors through Cortinarius MNs, the transfer of C through these MNs into plant tissues (Deslippe & Simard, 2011) likely reflects mycelial N uptake, conversions and the transfer of newly formed glutamate, to B. nana receivers. This situation contrasts with that of arbuscular mycorrhizal fungi (AMF), which take up inorganic N and incorporate it to arginine, but transfer it to the plant without C (Govindarajulu et al., 2005). If amino acids are the dominant form of N transferred from EMF to their host plants, then differences in N-transfer at the symbiotic interface among EMF and AMF may actually underpin the controversy around C-transfer through MNs, as studies that have found large and significant C-transfer through MNs to plant tissues have been exclusively in EM systems (Simard et al., 2002, 2012). Previously, we found that on average 4.1% of a donor plant's net photosynthetic C was transferred to receiver B. nana tissues through MNs (Deslippe & Simard, 2011). Significant C-transfer occurred only within pairs of B. nana and not between B. nana and other plant species, including the other EM plants (Salix pulchra, S. reticulata). However, we observed large variability in C-transfer through the MN pathway relative to other belowground pathways (rhizomes, root grafts and the soil solution). Given that Cortinarius sp. was exceptional among root-associated fungi in being highly enriched in 13C, and that the proportional abundance of Cortinarius-affiliated EMF on B. nana can vary from 4% to 58% of the population in response to warming treatment at this site (Deslippe et al., 2011), it seems reasonable that differences in the relative abundance of Cortinarius sp. on B. nana roots may account for at least some proportion of the high variability in C-transfer that we observed. Indeed the number of hyphal links among EM plants is significantly positively correlated to P-transfer through MNs (Li et al., 2004). To further assess support for this notion we used indicator species analysis to independently examine the 13C-enrichment of all Russulaceae-affiliated sequences and of all Cortinarius-affiliated sequences that did not cluster into OTU17. However, even when the proportional abundances of these OTUs were summed they did not display higher fidelity for the heavy fractions. This indicates no significant 13C-enrichment among members of the Russulaceae or among members of other species of Cortinarius (Table S5). Thus, in addition to the more common and abundant EMF on B. nana receivers that were not enriched in 13C in this study (e.g. Rhizoscyphus ericae) it appears that members of the Russulaceae and other Cortinarius species were not active conduits of C-transfer among B. nana. The unique enrichment of OTU17, which is likely a single Cortinarius sp. supports the notion that MNs dominated by rhizomorph-forming fungal species are better facilitators of resource transfer than those dominated by shorter-distance exploration types such as those we identified in the Russulaceae. Here we show that a Cortinarius sp. was the likely pathway for C-transfer among B. nana plants in Arctic tundra. Our data suggest that a Cortinarius sp. performs a particular function for its host that other EMF do not do to any significant extent. While the mechanism for C-transfer through MNs has yet to be elucidated, one possible explanation is that C-transfer among B. nana individuals through MNs reflects the movement of amino acid N from Cortinarius to B. nana. This view is consistent with the evidence that some Cortinarius species (Bodeker et al., 2009, 2014) are important in mobilizing growth-limiting N for their host from complex soil organic matter (Lilleskov et al., 2002; Hobbie & Agerer, 2010; Hobbie et al., 2013), which in turn is important in maintaining relatively high rates of C-cycling (Clemmensen et al., 2013, 2015; Lindahl & Tunlid, 2015). Thus C-transfer among B. nana individuals through mycelial networks of Cortinarius sp. may add to the evidence linking certain EMF to the maintenance of C-turnover in high latitude ecosystems (Clemmensen et al., 2015). Given the profound response of Arctic EM shrubs (Callaghan et al., 2011; Myers-Smith et al., 2011; Elmendorf et al., 2012) and EMF communities (Deslippe et al., 2011, 2012; Morgado et al., 2015) to climate warming, it is possible that these unique activities of Cortinarius sp. are influencing ecosystem-scale processes. The authors are grateful to the Editor and to five anonymous reviewers for constructive feedback on earlier versions of this manuscript. The authors thank Emma Harrower of the University of Tennessee for helpful advice regarding phylogenetic treatment of Cortinarius ITS2 sequences. The authors are grateful to R. Yarwood and D. Myrold of the Stable Isotope Research Unit, Oregon State University, for GC-C-IRMS analysis of microbial PLFAs. J.R.D. is grateful to J. Neufeld at the University of Waterloo for training in DNA-SIP methodologies. The authors thank G. Shaver, M. S. Bret-Harte, and J. E. Hobbie and the Arctic LTER for project support. This work was made possible by a National Science and Research Council of Canada International Polar Year grant (NSERC-IPY 350392-07) to S.W.S, W.W.M. and S.J.G., and a University of British Columbia (UBC) Post doctoral fellowship to J.R.D. J.R.D., S.J.G., S.W.S. and W.W.M. designed the research. J.R.D. conducted the fieldwork and laboratory analyses. J.R.D. and M.H. analyzed the data, J.R.D. wrote the manuscript with input from M.H., S.J.G., S.W.S. and W.W.M. Please note: Wiley Blackwell are not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office. Fig. S1 Phylogenetic tree containing OTU17 and other Cortinarius-affiliated ITS2 sequences. Fig. S2 Proportional abundances of the fungal ITS-OTUs in 12C-control plant root samples. Fig. S3 Cortinarius sp. sporocarp fruiting at the study site at the time of sampling. Table S1 List of all vascular plant species in study plots Table S4 Indicator species analysis for fungal OTUs Table S3 Proportional abundances and taxonomic affiliations of fungal ITS-OTUs derived from samples Table S5 Proportional abundances and taxonomic affiliations of Russulaceae and non-OTU17-Cortinarius-affiliated fungal ITS-OTUs Methods S1 A curated file of the 31 996 fungal ITS sequences generated in this study in FASTA format. Please note: The publisher 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.