Ectomycorrhizal Cortinarius species participate in enzymatic oxidation of humus in northern forest ecosystems



  • In northern forests, belowground sequestration of nitrogen (N) in complex organic pools restricts nutrient availability to plants. Oxidative extracellular enzymes produced by ectomycorrhizal fungi may aid plant N acquisition by providing access to N in macromolecular complexes. We test the hypotheses that ectomycorrhizal Cortinarius species produce Mn-dependent peroxidases, and that the activity of these enzymes declines at elevated concentrations of inorganic N.
  • In a boreal pine forest and a sub-arctic birch forest, Cortinarius DNA was assessed by 454-sequencing of ITS amplicons and related to Mn-peroxidase activity in humus samples with- and without previous N amendment. Transcription of Cortinarius Mn-peroxidase genes was investigated in field samples. Phylogenetic analyses of Cortinarius peroxidase amplicons and genome sequences were performed.
  • We found a significant co-localization of high peroxidase activity and DNA from Cortinarius species. Peroxidase activity was reduced by high ammonium concentrations. Amplification of mRNA sequences indicated transcription of Cortinarius Mn-peroxidase genes under field conditions. The Cortinarius glaucopus genome encodes 11 peroxidases – a number comparable to many white-rot wood decomposers.
  • These results support the hypothesis that some ectomycorrhizal fungi – Cortinarius species in particular – may play an important role in decomposition of complex organic matter, linked to their mobilization of organically bound N.


Northern forest ecosystems are dominated by conifers, birch and ericaceous shrubs with high concentrations of lignin, condensed tannins and other phenolic substances in their litter. At later decomposition stages in the humus layer, the relative share of nonhydrolysable organic pools increases (Berg & McClaugherty, 2014), and nitrogen (N) accumulates in organic forms, restricting its availability to the generally N-limited vegetation (Northup et al., 1995; Schmidt-Rohr et al., 2004). Thus, release of nutrients through organic matter oxidation is likely to be an essential driver of ecosystem productivity. However, the mechanisms and microbial drivers involved in the degradation of boreal forest humus layers remain poorly understood.

The only organisms known to have the capacity to completely degrade lignin to CO2 are specialized wood decomposers known as white-rot fungi (Hatakka, 1994). To break up phenolic macromolecules they use elaborate systems of extracellular oxidative enzymes, in which many different families, such as laccases, peroxidases and hydrogen peroxide-producing enzymes, interact during wood decomposition (Rayner & Boddy, 1988). The Class II family of peroxidases comprises extracellular enzymes, unique to Agaricomycetes, and encompasses lignin peroxidases, manganese peroxidases and versatile peroxidases. By reducing H2O2, these enzymes obtain particularly high redox potentials and are able to oxidize a wide range of substrates (Hatakka, 1994; Baldrian, 2008). Mn-peroxidases are the most widely spread and diverse group among the Class II peroxidases (Floudas et al., 2012). They oxidize Mn(II) to Mn(III), which is chelated with oxalic acid or other organic acids and subsequently interacts further with a wide range of phenolic and nonphenolic substrates, leading to their degradation (Martinez, 2002). Mn-peroxidases have primarily been studied in the context of wood decomposition. However, Mn-peroxidase activity may also be substantial in litter and soils (Sinsabaugh, 2010), and mRNA transcripts of Mn-peroxidase genes have been detected in hardwood forest soil (Kellner et al., 2010).

In the boreal forest, typical white-rot fungi with recognized capacity to produce Mn-peroxidases primarily colonize relatively fresh litter components on the surface, whereas basidiomycete communities in the more decomposed humus are dominated by symbiotic mycorrhizal fungi (Lindahl et al., 2007; Clemmensen et al., 2013). Recently, several cases where ectomycorrhizal fungi have lost much of their genetic capacity to degrade organic matter during evolution from saprotrophic ancestors have been highlighted (Plett & Martin, 2011; Floudas et al., 2012; Wolfe et al., 2012). Nevertheless, genes coding for Mn-peroxidases have been detected in several different ectomycorrhizal genera (Bödeker et al., 2009), and peroxidase activity has been demonstrated in pure cultures and laboratory microcosms containing certain ectomycorrhizal species (Lindahl et al., 2005). Furthermore, ectomycorrhizal fungi have been demonstrated to release 14C from labelled aromatic substances, such as lignin, lignocellulose and coniferyl alcohol (Trojanowski et al., 1984; Haselwandter et al., 1990). More recently, general positive relationships between soil peroxidase activity and ectomycorrhizal species richness (Talbot et al., 2013) and relative abundance (Phillips et al., 2014) have been demonstrated.

As root-associated symbionts, ectomycorrhizal fungi play a central role in nutrient cycling, by linking the vegetation and the soil environment. They obtain carbon (C) from their host trees and deliver N and other nutrients in return. Ectomycorrhizal fungi may aid their host trees in accessing N from a wide range of organic pools (Abuzinadah & Read, 1989; Entry et al., 1991; Lindahl et al., 2002; Read & Pérez-Moreno, 2003; Hobbie et al., 2013). In humus layers colonized by ectomycorrhizal fungi, the C : N ratio increases with humus age, presumably due to the active exchange of C for N accomplished by the ectomycorrhizal symbiosis (Lindahl et al., 2007; Wallander et al., 2010; Clemmensen et al., 2013).

Decomposition of soil organic matter often slows down in response to N deposition (Janssens et al., 2010), and several studies have also reported decreases in peroxidase activity in forest soils after N addition (Sinsabaugh, 2010), particularly in substrates at later stages of decomposition. We hypothesize that, supported by C from their hosts, certain groups of ectomycorrhizal fungi use oxidative enzymes to decompose organic matter, in order to mobilize N for themselves and their hosts. These ectomycorrhizal fungi would act as decomposers, not primarily to obtain C to support their metabolism, but to forage for organic N in the absence of more easily available, inorganic N sources (Lindahl et al., 2007; Talbot et al., 2008; Sinsabaugh, 2010; Rineau et al., 2013).

Members of the ectomycorrhizal genus Cortinarius are often a dominant component of fungal communities in boreal (Lindahl et al., 2010) and sub-arctic (Clemmensen et al., 2006) soils. Screening with degenerate primers identified several Class II peroxidase genes in different Cortinarius species (Bödeker et al., 2009). Furthermore, using 14C dating of fungal proteins, Hobbie et al. (2013) showed that Cortinarius armillatus mobilizes amino acids directly from old organic matter. Here, we seek further support for the hypothesis that Cortinarius species may be an important source of Class II peroxidase activity in forest soils, playing an active role in transformation of complex and recalcitrant organic compounds.

We employed three different approaches to test the hypothesis that Cortinarius species actively degrade complex organic matter in search for organic N. (1) At two sites – a boreal pine forest and a sub-arctic birch forest – variation in Cortinarius DNA abundance was assessed by 454-sequencing of internal transcribed spacer (ITS) amplicons (Lindahl et al., 2013) and related to Mn-peroxidase activity measured in soil extracts. The effect of local, short-term NH4+ amendments on Mn-peroxidase activity was also investigated. (2) An isolate of Cortinarius glaucopus was recently subjected to full genome sequencing (, and we screened the genome for the presence of Class II peroxidase genes. (3) Using PCR primers with a narrow target range to amplify peroxidase coding mRNA, we investigate the potential expression of Mn-peroxidase genes specifically by ectomycorrhizal Cortinarius species in undisturbed symbiosis with mature trees.

Materials and Methods

Field sites and sampling

The field experiment was conducted at two sites: a sub-arctic mountain birch (Betula pubescens spp. tortuosa Ledeb.) forest with an understorey of ericaceous dwarf shrubs (Vaccinium myrtillus L., Vaccinium vitis-ideae L., Empetrum nigrum L.) and mosses (Dicranum majus Turner, Hylocomium splendens L.), located near Abisko, northern Sweden (68°20′N, 18°49′E, altitude 425 m, mean annual temperature c. 0°C, annual precipitation 300 mm, pH = 4.3, C : N in humus = 28); and a boreal Scots pine (Pinus sylvestris L.) forest with an understorey of ericaceous dwarf shrubs (Vaccinium myrtillus L., Vaccinium vitis-ideae L., Calluna vulgaris (L.) Hull) and mosses (Pleurozium schreberi (Bridel) Mitten) located near Jädraås in mid Sweden (plot IhV, 60°49′N, 16°30′E, altitude 185 m, mean annual air temperature 3.8°C, mean annual precipitation c. 600 mm, pH = 3.75, C : N in humus = 45).

Nitrogen treated and control plots, 20 × 20 cm in size, were amended with 10 g m−2 NH4-N as (NH4)2SO4 or equimolar amounts of K2SO4, respectively. The treatments were applied as 250 ml of 57 mM solutions per plot, using a spinal needle. In order to achieve an even distribution, 10 ml of liquid was added to each of 25 points in each plot at 2–10 cm depth, (roughly corresponding to the Oe and Oa horizons). Control and N-treated plots were alternated along two parallel 112–116-m-long transects with 4 m distance between plots. Each treatment was replicated 30 times at the birch site and 29 times at the pine site, totalling 118 samples.

Two days after N amendment, soil cores (4 cm diameter) were collected from the centre of each plot. The litter layer (intact dead needles, leaves, twigs and mosses) and mineral soil were removed, and the inbetween humus (Oe and Oa) was frozen on dry ice immediately in the field, later carefully homogenized by grinding in liquid N and used for extraction of extracellular enzymes, DNA and RNA. Humus from three soil cores surrounding the central core was pooled and used for measurements of ammonium concentration in water extracts (5 g FW humus in 40 ml H2O) by flow injection analysis on a FIAstarTM5000 system (FossTecator, Höganäs, Sweden).

Analysis of Mn-peroxidase activity

Extractable Mn-peroxidase activity was measured via the oxidative coupling of DMAB (3-dimethylaminobenzoic acid) and MBTH (3-methyl-2-benzothiazolinone hydrazone hydrochloride) in the presence of Mn2+ and H2O2. In order to analyse extractable Mn-peroxidase activity, 2 g of humus was extracted with 6 ml deionized water by shaking (1 h) at 4°C and centrifuging (10 000 g for 10 min). A 50-μl aliquot of extract was added to 140 μl of reaction solution (100 mM sodium lactate and 100 mM sodium succinate, adjusted to pH 4.5 with sulphuric acid, 50 mM DMAB, 1 mM MBTH, 1 mM MnSO4-4H2O solution). At the start of the experiment 10 μl of 1 mM H2O2 was added. Background peroxidase activity was estimated by substituting MnSO4 with 2 mM Na2-EDTA 2H2O. A boiled aliquot served as a negative control. Absorption at 590 nm was measured every 10 min over a time period of 2 h at 25°C in a SynergyTM HT micro-plate reader (BIO-TEK, Winooski, VT, USA), and enzyme activities were estimated by regression of absorbance against time.

Analysis of fungal communities

Freeze-dried humus material was ground to a fine powder, and DNA was extracted from 50 mg according to Bödeker et al. (2009). The extracts were further purified using the Wizard DNA clean up kit (Promega). ITS2 amplicons for 454-sequencing were produced using the fITS9 primer in combination with the ITS4 primer provided with sample-specific extension tags according to Ihrmark et al. (2012). For every sample, three PCR reactions were pooled, in order to minimize random PCR distortion of community composition. Amplicons were purified using the Agencourt AMPure kit (Beckman Coulter Inc., Beverly, MA, USA), and concentrations were measured photometrically using a Qubit Fluorometer (Invitrogen). A total of 150 ng PCR product from each sample was pooled into a single sample, which was further purified using the GeneJET PCR-purification kit (Thermo Fisher Scientific, Waltham, MA, USA). Adaptor ligation and 454-sequencing was performed by LGC Genomics GmbH (Berlin, Germany) on a GL FLX Titanium system (Roche).

Sequences were processed using the bioinformatic pipeline SCATA ( Sequences with an average quality score below 20 or below 10 at any single position were discarded, using the high quality region (HQR) extraction option. Highly conserved 5.8S sequence was removed by cropping 80 bp from the fITS9 primer end, leaving 15 bp of the 5.8S region, 38 bp of the LSU region, and the internal ITS2 region, adding up to 200–400 bp. Sequences were compared for similarity, using BLAST as a search engine, with the minimum length of pairwise alignments set to 90% of the longest sequence. Pairwise alignments were scored using a scoring function with 1 in penalty for mismatch, 0 for gap opening and 1 for gap extension. Homopolymers were collapsed to 3 bp before clustering. Sequences were assembled into operative taxonomic units (OTUs) by single linkage clustering with a 98.5% sequence similarity required to enter an OTU. The 214 most abundant OTUs, representing between 52% and 92% of the total sequences per sample (78% on average), were identified taxonomically by comparing representative sequences to the UNITE reference database (Kõljalg et al., 2013) and NCBI's nonredundant DNA sequence database. Sequence data are stored at NCBI's Sequence Read Archive under the accession number SRP014921. Sample-specific accession numbers are included as Supporting Information Table S1.

Amplification and sequencing of mRNA from Cortinarius peroxidase genes

In order to maximise the chance of detecting RNA transcripts of Cortinarius peroxidase genes, six control samples with high Mn-peroxidase activities (four and two from the pine and birch sites, respectively) and five N-amended samples with high abundance of Cortinarius species in the DNA pool (all from the pine site) were selected for RNA extraction. From these samples, 2 g of humus was ground in liquid N and extracted with 15 ml extraction buffer (2% CTAB, 2% polyvinylphosphate, 100 mM TrisHCL-pH 8, 25 mM EDTA, 2 M NaCl, 2% β-mercaptoethanol) and 15 ml of phenol-chloroform-isoamylalcohol (25 : 24 : 1) for 30 min at 55°C. After another phenol-chloroform extraction followed by centrifugation at 10 000 g for 30 min at 4°C, RNA was precipitated from the supernatant by addition of 0.25 volumes of 10 M LiCl. The obtained RNA was further purified using the RNeasy Plant Mini-Kit (Qiagen), and residual DNA was degraded by DNaseI (Sigma-Aldrich).

Based on Mn-peroxidase sequences previously obtained from Cortinarius species (Bödeker et al., 2009), two conserved sites were chosen as targets for PCR primers. The amplicons yielded by the forward primer PC3.1 (5′-CCW TTY TTY TTY AAA CAC KC-3′) and the reverse primer PC2 (5′-AA RTC RTC IGC ICC IGC-3′) include an intron, enabling distinction of amplified cDNA from genomic DNA.

In order to amplify cDNA from mRNA transcripts of Cortinarius Mn-peroxidase genes, RT-PCR was conducted using the OneStep RT-PCR Kit (Qiagen), and PCR products were visualized on agarose gel (PCR cycling conditions after reverse transcription were the same as already described). Positive samples were cloned and re-amplified. In addition to soil RNA, DNA from selected sporocarps was amplified as reference material. Nine Cortinarius species (Cortinarius paragaudis UNITE ID: UD B000686, C. testaceofolius UDB002160, C. diasemospermus UDB002161, C. semisanguineus UDB001178, C. obtusus UD B00204, C. acutus UDB002211 and UDB002194, Carmillatus UDB002174, C. brunneus UDB01541) were selected, based on their high abundance in the samples analysed for gene transcription. Extraction of DNA from the nine Cortinarius sporocarps, PCR amplification of Mn-peroxidase genes (annealing temperature of 57°C), cloning of PCR products, and re-amplification of cloned inserts were conducted according to Bödeker et al. (2009). Sanger sequencing of all cloned amplicons (cDNA and sporocarps) was performed by Macrogene Inc. (Seoul, Korea). All sequences are archived at NCBI's GenBank under the accession numbers JX467495JX467500 for transcript sequences and JX467501JX467532 for sporocarp derived sequences.

Genome screening and phylogenetic analyses

The genomes of Cortinarius glaucopus and Hebeloma cylindrosporum ( were screened for Class II peroxidase genes using BLAST and text searches. Identified protein sequences and Class II peroxidases from PeroxiBase (Fawal et al., 2013) were aligned, using MUSCLE (Edgar, 2004) implemented in MEGA5 (Tamura et al., 2011). Phylogenetic reconstruction was based on amino acid sequences, using the maximum likelihood method under the JTT model and a gamma distribution of substitution rates with invariant sites, partial deletion (95%) of gap positions, and 1000 bootstrap replications.

Peroxidase nucleotide sequences, obtained either from sporocarp DNA or from soil mRNA, were assembled using SeqMan (DNAStar Inc., Madison, WI, USA) with a 97% sequence similarity cut-off. After manual removal of introns from sporocarp-derived gene sequences, nucleotide sequences were aligned together with already published reference sequences from Bödeker et al. (2009) using ClustalW in Megalign (DNAStar Inc.), and relatedness was investigated by neighbor-joining analysis, based on total character difference in PAUP* 4.0b10 (Sinauer Associates, Sunderland, MA, USA).

Statistical analysis

For each sample, we calculated an OTU-specific enzyme activity, by multiplying the enzyme activity of the sample by each OTU's relative abundance. Average OTU-specific enzyme activities were calculated across all nonamended samples and ranked numerically within each of the two sites, in order to identify taxa where high relative amplicon abundance was correlated with high peroxidase activity:

display math(Eqn 1)

(Ei, enzyme activity in sample i; Pi, relative amplicon abundance of the OTU in sample i (a proportion between 0 and 1); n, total number of samples.)

Correlations between Mn-peroxidase activity, Cortinarius abundance and NH4+ concentration were tested across control and amended plots for each of the two sites independently by multiple regression in STATISTICA (StatSoft Inc., Tulsa, OK, USA), with log(x + 1)-transformed Mn-peroxidase activity as the responding variable and the summed relative abundance of OTUs assigned to Cortinarius species and log-transformed NH4+ concentration as explaining variables. Due to missing data, the regressions were conducted on 55 samples from the pine site and 56 samples from the birch site. The effect of N amendment on Cortinarius abundance was tested with ANOVA for each site individually. A single birch forest sample with very high peroxidase activity was found to contain primarily DNA from the white-rot wood decomposer Resinicium bicolor, as well as other wood-inhabiting taxa, and was omitted from all analyses.

In spite of previous criticism (Amend et al., 2010), we use the relative ITS-amplicon abundance as a proxy for fungal biomass and activity. The phylogenetic composition of sequenced amplicons has been shown to reflect the template reasonably well (Ihrmark et al., 2012), and we assume that the ITS template to biomass ratio is fairly constant within a specific fungal genus. Thus, the relative abundance of Cortinarius ITS amplicons should correlate with biomass sufficiently well to be useful as a marker. Furthermore, the alternative to succumb to records of presence/absence is a less attractive option, because major data distortion would be expected, due to a low incidence of biologically nonrelevant DNA (e.g. from spores), contaminations or wrongly assigned sequences (Carlsen et al., 2012).


Samples from birch forest control plots had on average four times higher ammonium concentration than those from the pine forest (Fig. 1a), whereas the Mn-peroxidase activity was almost 20 times higher in the pine forest than in the birch forest (Fig. 1b). Two days after treatment, NH4+ concentrations were raised an order of magnitude in N-treated plots at both forest sites (Fig. 1a).

Figure 1.

(a) NH4+ concentration and (b) Mn-peroxidase activity in water extracts of humus samples from nitrogen (N)-amended (closed bars) and nonfertilized (open bars) plots in a boreal pine forest and a sub-arctic birch forest in Sweden. Error bars, ± SE. Mn-peroxidase activity is expressed as μmol reaction product of 3-dimethylaminobenzoic acid (DMAB) and 3-methyl-2benzothiazolinone hydrazine hydrochloride (MBTH) formed per hour and g OM.

Successfully sequenced amplicons representing the total fungal community were obtained from 112 out of the total 118 samples. The 454-sequencing yielded 351 000 sequences, of which 215 000 were retained after quality filtering, leading to a sequencing depth of between 187 and 11 248 sequences per sample (average 1900). The sequences were assembled into 2109 OTUs. A major fraction of the 214 identified and most abundant OTUs – on average 38% of the identified amplicons at both sites – were assigned to Leotiomycetes (Ascomycota). Basidiomycete sequences were equally or less abundant (on average 34% and 17% at the pine and birch sites, respectively), and ectomycorrhizal fungi on average accounted for 76% and 33%, respectively, of the basidiomycete amplicons. At the pine site, the ectomycorrhizal community was dominated by Cortinarius (on average 40% of ectomycorrhizal amplicons), Piloderma (35%) and Suillus (15%) species. At the birch site the ectomycorrhizal community was dominated by Pseudotomentella (39%), Cortinarius (23%), Lactarius (9%), Russula (8%), Leccinum (7%) and Cenococcum (6%) species. There was no significant effect of N fertilization on the relative abundance of Cortinarius amplicons at any of the sites (= 0.45 and 0.94 for the pine and birch sites, respectively).

The OTU-specific averages of Mn-peroxidase activities indicated that, for many of the detected Cortinarius species, high abundance of DNA co-localized with high enzyme activity. For the pine site, among the 66 OTUs with a relative abundance > 0.2% (together accounting for 69% of the amplicons), four of the five highest OTU-specific enzyme activities were assigned to Cortinarius species ranging from 2.2 to 5.1 times higher values than the site average (Table 1). However, there were other Cortinarius species – including that with the highest average relative abundance – C. caperatus – which did not show a clear correlation with Mn-peroxidase activity. For the birch site, among the 61 OTUs with a relative abundance > 0.2% (together accounting for 74% of the amplicons), three of the four highest OTU-specific enzyme activities were assigned to Cortinarius species, ranging from 1.6 to 3.5 times higher values than the site average (Table 2).

Table 1. Pine site – taxonomic affiliation, operative taxonomic unit (OTU)-specific average peroxidase activity, relative abundance, and frequency of occurrence of OTUs detected by 454-sequencing of internal transcribed spacer (ITS) amplicons from nonfertilized humus samples originating from a boreal pine forest in Sweden
OTU affiliationPhylumaMn-peroxidase activitybAbundancecOccurrenced
  1. a

    A, Ascomycota; B, Basidiomycota; M, Mucoromycotina.

  2. b

    Expressed as μmol reaction product of 3-dimethylaminobenzoic acid (DMAB) and 3-methyl-2benzothiazolinone hydrazine hydrochloride (MBTH) formed per hour per g OM (organic matter). The values represent OTU-specific averages calculated according to (Eqn 1). The average enzyme activity of all samples is marked in grey.

  3. c

    Representing the average relative amplicon abundance of OTUs in percent.

  4. d

    Representing the number of samples (out of 29) in which the OTU was recorded.

Cortinarius paragaudis B21.81.24
Cortinarius diasemospermus B12.71.03
Cortinarius biformis B10.00.58
Cortinarius semisanguineus B9.31.25
Penicillium canescens A7.63.413
Heterobasidion annosum B6.60.314
Resinicium bicolor B6.30.621
Mollisia A6.20.22
Candida A5.60.614
Mycena B5.40.612
Suillus variegatus B5.02.720
Oidiodendron A4.94.129
Piloderma olivaceum B4.50.56
Davidiella tassiana A4.50.419
Oidiodendron A4.50.313
Piloderma B4.35.528
Mortierella M4.10.523
Cortinarius vibratilis B4.00.64
Cortinarius obtusus B3.70.310
Cortinarius brunneus B3.00.810
Rhizoscyphus A3.00.721
Cortinarius caperatus B3.06.010
Rhizoscyphus ericae A2.90.511
Mytilinidion A2.90.616
Tricholoma fucatum B2.71.14
Luellia recondita B2.50.713
Sebacina B2.40.517
Cortinarius B1.60.411
Colpoma A1.50.36
Lachnellula B1.20.25
Lactarius rufus B0.00.32
Table 2. Birch site – taxonomic affiliation, operative taxonomic unit (OTU)-specific average peroxidase activity, relative abundance, and frequency of occurrence of OTUs detected by 454-sequencing of internal transcribed spacer (ITS) amplicons from nonfertilized humus samples originating from a sub-arctic birch forest in northern Sweden
OTU affiliationPhylumaMn-peroxidase activitybAbundancecOccurrenced
  1. a

    A, Ascomycota; B, Basidiomycota; M, Mucoromycotina.

  2. b

    Expressed as μmol reaction product of 3-dimethylaminobenzoic acid (DMAB) and 3-methyl-2benzothiazolinone hydrazine hydrochloride (MBTH) formed per hour per g OM (organic matter). The values represent OTU-specific averages calculated according to Eqn (Eqn 1). The average enzyme activity of all samples is marked in grey.

  3. c

    Representing the average relative amplicon abundance of OTUs in percent.

  4. d

    Representing the number of samples (out of 29) in which the OTU was recorded.

Cortinarius obtususB0.780.82
Cortinarius acutusB0.720.51
Coniozyma leucospermiA0.360.42
Cortinarius armillatusB0.350.76
Lactarius vietusB0.320.312
Luellia reconditaB0.221.325
Pseudotomentella tristisB0.200.626
Cenococcum geophilumA0.200.89
Mycena galopusB0.200.211
Mycena epipterygiaB0.130.211
Leccinum variicolorB0.130.29

Mn-peroxidase activity in the humus samples was significantly and positively correlated with the summed relative abundance of Cortinarius amplicons at both sites, and there was a negative correlation with NH4+ concentration, which was significant at the pine site but marginally insignificant at the birch site (Table 3, Fig. 2).

Table 3. Multiple regression of Mn-peroxidase activitiesa in humus samples from a pine and a birch forest with the relative amplicon abundance of Cortinarius species and NH4+ concentrationsb as explaining variables
 VariableCoefficient (± SE)P-valueR2 of model
  1. a

    Expressed as μmol reaction product of 3-dimethylaminobenzoic acid (DMAB) and 3-methyl-2benzothiazolinone hydrazine hydrochloride (MBTH) formed per hour per g OM (organic matter) and log(x + 1)-transformed.

  2. b

    Expressed as μg NH4-N g−1 OM and log-transformed.

Pine siteIntercept0.45 (0.08)< 0.00010.21
Cortinarius 0.89 (0.34)0.01
NH4+−0.13 (0.06)0.03
Birch siteIntercept0.072 (0.011)< 0.00010.28
Cortinarius 0.40 (0.10)0.0002
NH4+−0.014 (0.007)0.07
Figure 2.

Mn-peroxidase activity in relation to abundance of internal transcribed spacer (ITS) amplicons assigned to the ectomycorrhizal genus Cortinarius in (a) a boreal pine forest and (b) a sub-arctic birch forest in Sweden. Half of the plots were amended with (NH4)2SO4 solution 2 d before sampling. Mn-peroxidase activity is expressed as μmol reaction product of 3-dimethylaminobenzoic acid (DMAB) and 3-methyl-2benzothiazolinone hydrazine hydrochloride (MBTH) formed per hour and g organic matter (OM). Closed symbols and regression lines represent samples with > 20 μg extractable NH4-N g−1 OM, whereas open symbols and dashed regression lines represent samples with < 20 μg NH4-N g−1 OM. Samples where Cortinarius caperatus made up > 50% of the total Cortinarius amplicons are marked in red.

Screening of the Cortinarius glaucopus genome identified 11 Class II peroxidase genes, whereas three genes were found in the Hebeloma cylindrosporum genome. Phylogenetic analysis of protein sequences together with 174 Class II peroxidase proteins retrieved from the PeroxiBase database (Fawal et al., 2013) as well as 20 translated amplicon sequences from Bödeker et al. (2009) showed an overall correspondence with previous studies on peroxidase phylogeny (Morgenstern et al., 2008; Hofrichter et al., 2010; Floudas et al., 2012). Sequences from C. glaucopus and H. cylidrosporum grouped with the single Laccaria bicolor peroxidase in a well-supported branch. However, translated amplicon sequences from other Cortinarius spp. were not included in this branch, but instead grouped with peroxidases from Hypholoma sublateritium and Galerina marginata (Fig. 3).

Figure 3.

Phylogenetic tree of Class II peroxidase protein sequences based on maximum likelihood estimation. Sequences derive from PeroxiBase, NCBI or JGI's fungal genome sequencing program. Numbers at branch nodes represent bootstrap support (only values > 50 are given). Numbers in parentheses represent the number of sequences in condensed branches. Sequences denoted as ‘amplicons’ result from PCR amplification of DNA from sporocarps in Bödeker et al. (2009).

Positive RT-PCR amplification products of Mn-peroxidase origin were obtained from two samples: a control plot and a fertilized plot, both from the pine forest site. Sequencing of the RT-PCR products yielded six unique Mn-peroxidase encoding sequences. PCR products obtained from the reference sporocarp material yielded another 32 unique sequences from nine Cortinarius species. Three of the mRNA sequences from the humus samples grouped with sequences from Cortinarius semisanguineus (two of the mRNA sequences were identical to sporocarps sequences), two grouped with sequences from the Cortinarius acutus/obtusus species complex and one could not be assigned to any of the included species (Fig. 4).

Figure 4.

Neighbour-joining (NJ) tree based on amplicon nucleotide sequences, representing Class II peroxidase genes from Cortinarius species after the removal of introns. Sequences highlighted in grey were obtained by reverse transcription of mRNA from humus samples from a boreal pine forest in Sweden, whereas other sequences were derived from sporocarp material. Sequences with NCBI accession numbers originate from Bödeker et al. (2009). Numbers at branch nodes represent bootstrap supports (only values > 50 are given). According to 454-sequencing of internal transcribed spacer (ITS) amplicons, sample 2CJ contained C. semisanguineus (27%), C. caperatus (3%), C. obtusus (1%) and C. brunneus (0.2%). Sample 3NJ contained C. obtusus (3%).


The significant correlation between Mn-peroxidase activity and the occurrence of Cortinarius DNA in humus samples (Tables 1-3, Fig. 2) indicates that members of this genus make an important contribution to overall Mn-peroxidase activity in boreal forest soils. We were able to link mRNA transcripts of Mn-peroxidase genes, obtained from forest soil, to ectomycorrhizal species by sequence similarity to DNA obtained from sporocarps, showing that Cortinarius peroxidase genes are active under field conditions (Fig. 4). Recently it was proposed that the capacity to produce ligninolytic peroxidases has contracted in ectomycorrhizal lineages during their evolution from a saprotrophic ancestor (Floudas et al., 2012), but the 11 Class II peroxidase genes detected in the genome of C. glaucopus put this ectomycorrhizal species at the same level as most white-rot wood decomposers, challenging the generality of this hypothesis. Rather, the potential of ectomycorrhizal fungi to enzymatically attack complex soil organic matter seems to differ between different evolutionary lineages. Ectomycorrhizal fungi are, thus, likely to represent a series of different functional guilds with some species having little capacity to degrade organic matter (e.g. Laccaria bicolor; Martin et al., 2008) and others being more capable degraders, either by producing extracellular peroxidases (e.g. Cortinarius species; this study) or by other mechanisms (e.g. Paxillus involutus; Rineau et al., 2012). Our findings support the hypothesis that certain ectomycorrhizal fungi may be directly involved in soil organic matter degradation in the humus layers of northern forest ecosystems, as proposed previously on several occasions (Frank, 1894; Courty et al., 2007; Lindahl et al., 2007; Cullings et al., 2008; Talbot et al., 2008, 2013; Bödeker et al., 2009; Sinsabaugh, 2010; Rineau et al., 2012; Phillips et al., 2014), but also questioned due to a lack of empirical evidence (Baldrian, 2009).

Cortinarius species play a central role in ectomycorrhizal communities of boreal forests, in terms of high relative abundance as well as species richness, and our results suggest that some members of this genus may be key players in the mobilization of nutrients from organic matter. This is in accordance with the isotope-based evidence by Hobbie et al. (2013) that C. armillatus mobilized intact amino acids from old organic stocks. However, C. caperatus – the most abundant Cortinarius species in the ITS-DNA pool from the pine forest site – did not show a positive correlation with enzyme activity (Table 1, Fig. 2), suggesting variation within the genus. In addition, some samples had high Mn-peroxidase activity without a high proportion of Cortinarius DNA, indicating that species from other genera also contributed to the overall enzyme activity. In spite of the significant correlation between Cortinarius DNA and Mn-peroxidase activity, the explanatory power of the regression models was low (Table 3). Thus, peroxidase production in boreal forest humus is likely neither to be exclusive to the genus Cortinarius, nor a general feature of all Cortinarius species.

The phylogenetic analysis of Class II peroxidase genes from the C. glaucopus genome and amplicons derived from sporocarps of other Cortinarius species appeared in two well-supported separate clades (Fig. 3). Clearly, different gene lineages are present in different species, indicated by the several species-specific groups. Amplification with degenerate primers may easily yield false negatives (Bödeker et al., 2009), and it is possible that orthologues of the C. glaucupus genes are present in many other species but failed to amplify. However, no orthologues of the amplicon genes were detected in the C. glaucopus genome, and its 11 genes seemingly result from relatively recent duplications of a different ancestral gene than that giving rise to the amplified sporocarp and soil mRNA sequences. Clearly, a combination of parallel gene losses and paralog expansion in terminal lineages during the evolution of Class II peroxidases has resulted in complex phylogenetic patterns (Floudas et al., 2012). Originally, 11 samples were screened for Mn-peroxidase gene transcription, but only two of them yielded amplification products. This low success rate could either be caused by PCR inhibition, or gene transcription may be more variable between samples than species occurrences and enzyme activities. Another possible explanation is a too high specificity of the primers in relation to the apparent divergence of Cortinarius peroxidase genes. For example, although highly degenerated, the primers would most likely fail to amplify the genes detected in the C. glaucopus genome.

In the pine site we observed a significant decrease in Mn-peroxidase activity in response to raised NH4+ concentrations (Table 3, Fig. 1), supporting previous observations that peroxidase activity in forest soils is sensitive to increases in N availability (Sinsabaugh, 2010). By adding N at a small spatial scale and shortly before sampling, no major shifts in fungal community composition would be expected, and our results suggest that individual organisms respond to the increased N availability by rapidly decreasing their enzyme production. A similar short-term negative effect of N on peroxidase production has been observed in pure cultures of the wood decomposer Phanerochaete chrysosporium (Kirk & Farrell, 1987). Previously, repressed transcription of laccase genes has been observed in response to simulated N deposition (Edwards et al., 2011). Repression of lignin decomposition by inorganic N has also been frequently observed both in the laboratory and in the field (Fog, 1988), and generally N fertilization increases long-term C accumulation in forest soils by repressing CO2 release (Mäkipää, 1995; Janssens et al., 2010).

The evolutionary forces behind N repression of oxidative enzymes remain uncertain for saprotrophs, as the main benefit of these enzymes has been thought to be increased access to cellulose due to degradation of lignin. From a mycorrhizal perspective, however, product repression by NH4+ seems sensible, as these fungi mainly colonize well-degraded humus substrates that are depleted in cellulose (Lindahl et al., 2007) but where, potentially, large quantities of N are bound in phenolic macromolecules (Schmidt-Rohr et al., 2004). N repression of mycorrhiza-associated Mn-peroxidase activity supports the hypothesis that N mobilization is the principal driver for ectomycorrhizal fungi to attack organic matter enzymatically. Oxidative degradation by fungal peroxidases has been described as a co-metabolic process that requires input of external C from nonlignin sources (Kirk & Farrell, 1987). With direct access to an external C source in the form of host photo-assimilates, mycorrhizal fungi are well conditioned to exploit recalcitrant substrates by oxidative mechanisms (Rineau et al., 2013). Oxidation by Mn-peroxidases may expose N-containing residues for further cleavage, enabling uptake and reallocation of small organic N compounds by mycorrhizal hyphae. Such selective mining for N would leave the residual material with a higher C : N ratio, which may contribute to the progressive increase in the C : N ratio of boreal forest humus layers colonized by ectomycorrhizal fungi (Lindahl et al., 2007; Wallander et al., 2010; Clemmensen et al., 2013). Thus, although our results support the idea that some ectomycorrhizal fungi may degrade organic matter to mobilize N, this does not imply that they are saprotrophs, that is, that they obtain a significant share of their C demand from dead organic matter.

A direct role of ectomycorrhizal fungi in organic matter oxidation could explain accelerated soil C loss in response to the presence of coniferous tree roots (Dijkstra & Cheng, 2007). Elevated atmospheric CO2 concentrations generally stimulate mycorrhizal mycelial activity in the soil as photosynthesis rates increase and more C may be allocated to roots and mycorrhizal symbionts (Treseder, 2004). Congruently, increased activity of oxidative enzymes in soil was reported in response to exposure of ectomycorrhizal trees to elevated CO2 (Carney et al., 2007). Furthermore, an active mycorrhizal community with high mycelial turnover rates successively depleted humus of N and minimized long-term accumulation of humus layers in boreal forests (Clemmensen et al., 2013). These observations all fit well with an active involvement of ectomycorrhizal fungi in the decomposition of recalcitrant organic matter in forest ecosystems.


We gratefully acknowledge the Swedish University of Agricultural Sciences and the Swedish Research Council FORMAS for financial support of the project. Andy Taylor at the James Hutton Institute is gratefully acknowledged for providing sporocarp material and the sequenced C. glaucopus isolate. Annegret Kohler at INRA as well as Igor Grigoriev and Alan Kuo at JGI are acknowledged for their leading roles in the acquisition of C. glaucopus and H. cylindrosporum genome sequence data. The genome sequencing conducted by the US Department of Energy's Joint Genome Institute is supported by the Office of Science of the US Department of Energy under Contract DE-AC02-05CH11231. This work was partly supported by the French National Research Agency through the Clusters of Excellence ARBRE (ANR-11-LABX-0002-01) to F.M. We also thank Katarina Ihrmark and Wiecher Smant for practical help and advice during the laboratory work. Abisko Scientific Research Station and Anders Michelsen are gratefully acknowledged for logistic support. Anders Dahlberg, Nicholas Rosenstock and Jan Stenlid are thanked for critical reading of the manuscript.