The ectomycorrhizal fungus Paxillus involutus converts organic matter in plant litter using a trimmed brown-rot mechanism involving Fenton chemistry

Soils in boreal forests contain large stocks of carbon. Plants are the main source of this carbon through tissue residues and root exudates. A major part of the exudates are allocated to symbiotic ectomycorrhizal fungi. In return, the plant receives nutrients, in particular nitrogen from the mycorrhizal fungi. To capture the nitrogen, the fungi must at least partly disrupt the recalcitrant organic matter–protein complexes within which the nitrogen is embedded. This disruption process is poorly characterized. We used spectroscopic analyses and transcriptome profiling to examine the mechanism by which the ectomycorrhizal fungus Paxillus involutus degrades organic matter when acquiring nitrogen from plant litter. The fungus partially degraded polysaccharides and modified the structure of polyphenols. The observed chemical changes were consistent with a hydroxyl radical attack, involving Fenton chemistry similar to that of brown-rot fungi. The set of enzymes expressed by Pa. involutus during the degradation of the organic matter was similar to the set of enzymes involved in the oxidative degradation of wood by brown-rot fungi. However, Pa. involutus lacked transcripts encoding extracellular enzymes needed for metabolizing the released carbon. The saprotrophic activity has been reduced to a radical-based biodegradation system that can efficiently disrupt the organic matter–protein complexes and thereby mobilize the entrapped nutrients. We suggest that the released carbon then becomes available for further degradation and assimilation by commensal microbes, and that these activities have been lost in ectomycorrhizal fungi as an adaptation to symbiotic growth on host photosynthate. The interdependence of ectomycorrhizal symbionts and saprophytic microbes would provide a key link in the turnover of nutrients and carbon in forest ecosystems.

. Changes in C, N, and nutrients during the conversion of soil organic matter by Pa.
involutus. The fungus was grown for 7 days using three different organic matter extracts as substrates; forest litter extracted with hot water (Forest Hot), and a maize compost extracted with cold (Maize Cold) or hot (Maize Hot) water, respectively. To avoid carbon limitation, glucose was added to the extract. Reference is the composition in the extracts before being added to the fungus, Inoculated is the concentration of the elements after 7 days of incubation.  Reference This paper (Ferrari and Mingazzini, 1995) This paper

This paper
This paper (Ferrari and Mingazzini, 1995) (Ferrari and Mingazzini, 1995) This paper This paper (Ferrari and Mingazzini, 1995) Table S3. Putative identification of the pyrolysis compounds of the forest hot (FH) extract that were most affected by the inoculation of Pa. involutus. Results are presented as the ratio between average relative peak size in the 3 incubated samples (size of the peak divided by the sum of all peak sizes in the sample) and the relative peak size in the reference sample ("Fold" column; NF=Newly formed). The dominant masses of each pyrolysate are also reported.

Supporting Information
Organic matter degradation by ectomycorrhizal fungi 13 a Prediction of secretion signals using the SignalP "*" or the Secretome "**" algorithm. "(?)" indicates that the sequences are likely to have an incomplete or comprised N-terminus; thus it is not possible to identify any secretion signal in the N-terminus.

Appendix S1
Supplementary experimental procedures Preparation of organic matter extracts. The cold water fraction for the Maize compost (MC) was extracted by shaking 120 g of material in 600 ml of MQ water for 24 h. The hot water extracts for the maize compost and forest soil (MH and FH, respectively) were obtained by boiling 120 g of material in 600 ml of MQ water for 1 h. The MC, MH and FH extracts were: 1) filtered through a nylon mesh; 2) sequentially vacuum-filtered with GF-D, GF-A and GF-F filters (Millipore, USA), respectively; 3) centrifuged 1h at 17 000 rpm at 4 o C; 4) filtered through 0.45 μm using Millex-HV non-sterile Filter units (Millipore, USA); 5) vacuum-filtered through 0.2 μm using Stericup system (Millipore, USA) to remove particles and to sterilize the extracts.
To decrease the concentration of low molecular weight compounds, the extracts were finally ultra-filtered with an Amicon Stirred Ultrafiltration Cell using regenerated cellulose membranes of a cut-off value of 10 kDa and 1 kDa, respectively (Millipore, USA). The volume of the extract was adjusted to 400 ml. The same preparations of the FH, MH and MC extracts were used in all experiments. Analysis of the chemical composition of the MH extract (total N content and synchronous fluorescence spectra) before and after being added to the glass bead plates did not reveal any detectable adsorption of macromolecules onto the glass beads.
Chemical analysis. Total organic carbon concentrations were measured using a TOC Analyzer (Shimadzu), and total nitrogen (TN) using the same apparatus with a TNM-1 detector.
Concentrations of P, K, Ca, Mg and Fe were measured using ICP-AES (Perkin-Elmer Optima 3000 DV). Glucose concentration was measured using the glucose (GO) assay kit provided by Sigma-Aldrich (Germany).
Samples for Fourier transform infrared (FTIR) spectroscopy were prepared by drying (vacuum over night at 4 o C) 5 ml of the organic matter extracts. One volume of powder was then ground and homogenized with three volumes of KBr in a mortar. The FTIR spectrum of the sample-KBr mixture was recorded using a Bruker IFS66 v/s spectrometer under mild vacuum (ca. 4 mbar)

Supporting Information
Organic matter degradation by ectomycorrhizal fungi 15 and at 25.0 ± 0.2 ºC. Data were collected in diffuse reflectance mode using a praying mantis diffuse reflectance attachment (Harrick Sci.). Each spectrum was the result of 1000 consecutive scans at a resolution of 4 cm -1 . Pure and ground KBr was used as background.
Synchronous fluorescence spectra were obtained using a Perkin-Elmer LS50B fluorescence spectrophotometer. Samples (750 μl) were kept at room temperature (20 o C) and processed at a 10 nm bandwidth and 25 nm offset (Δλ=25 nm) between excitation and emission. Samples were contained in a 1ml quartz cuvette with a 1 cm path length. The sample spectra were recorded at different dilutions in MQ water (1, 1/5, 1/10, 1/20, 1/40, 1/100) because fluorescence peaks of SF-detectable compounds have different sensitivities to dilution (Tuan, 1982). A fifty times dilution was chosen as it gave the best signal/noise ratio and as its optical density was low enough to interfere with fluorescence spectra (Senesi et al., 1991). Assignments of these regions were based on previously published data (Artz et al., 2008;Stevenson and Goh, 1971;Senesi et al., 2003).
Pyrolysis gas chromatography-mass spectrometry (Py-GC/MS) was conducted to analyze the detailed changes in chemical composition occurring during the degradation of the FH extract. A volume of 5 ml of FH reference sample and 3 ml FH inoculated samples were vacuum dried overnight at 4 o C. Samples were then analyzed by Py-GC/MS (Perkin Elmer TurboMass/ Autosystem XL with Frontier Lab double Shot pyrolyser). Each (GC/MS) peak was attributed to a pyrolysis product (molecule, i.e vanillic acid), and, when possible, the broad molecular origin (lignin, phenolic compounds, aromatic compounds, polyaromatics, N-containing compounds and polysaccharides) of each product was inferred according to literature (Table S2). Then, for each sample, we calculated the sum (S) of the peak areas (A) of the pyrolysis products of the same broad molecular origin (for example: S Lignin =A Lignin compound 1 +A Lignin compound 2 +…+A Lignin compound n ), and normalized it by the total peak areas in the sample N Lignin =S Lignin / (S phenoliccompounds +S Aromatics +S Polyaromatics +S N-containingcompounds +S Polysaccharides ). To compare reference (n=1) and inoculated (n=5) samples, we calculated the ratio (R) of this normalized value in the incubated samples compared with the reference one (R Lignin =N Lignin Incubated / N Lignin Reference ).
However, this ratio compares only relative amounts of compounds. To evaluate enrichment or

Supporting Information
Organic matter degradation by ectomycorrhizal fungi 16 loss in a class of compounds, we compared this ratio to the loss in total organic carbon measured in the samples ("No change level").
Iron-reducing compounds. The capacity of Pa. involutus to produce iron-reducing compounds when grown on organic matter extracts was examined using a ferrozine assay ( RNA isolation, cDNA synthesis, and Transposon Assisted Signal Trapping (TAST) cDNA library construction were essentially carried out as described previously (Grell et al., 2011), except that two parallel ligations were set up with different vector/cDNA ratios: one tenth of the cDNA preparation was cloned into 15 and 45 ng, respectively, of the vector pMHas7i (kindly provided by K. Schnorr, Novozymes) using one unit T4 DNA ligase from the SMART cDNA Library Construction Kit (Clontech). The cDNA library was transposon-tagged, transformed into E. coli cells, and signal-trapped clones were selected (Becker et al., 2004;Grell et al., 2011).

Supporting Information
Organic matter degradation by ectomycorrhizal fungi 19 The sequences were assembled into 348 contigs. The contig sequences were annotated using the Blast2GO tool (Götz et al., 2008), and information in the PFAM database (Finn et al., 2006) and the CAZy database (Cantarel et al., 2009). For a predicted glycoside hydrolase (GH) 16 family protein, the subfamily was predicted using the structured GH16 glycoside hydrolase database (Strohmeier et al., 2004). Signal peptide-triggered secretion and non-classical secretions signals were identified using the SignalP 3.0 and SecretomeP algorithms (Emanuelsson et al., 2007).
TAST-sequences with missing or compromised N-terminus were filled-in using sequence information from the corresponding isotig in the Pa. involutus EST database. Matches between TAST-sequences and isotigs were performed using the Blastn tool (Altschul et al., 1990). Assembly of EST sequences was performed in three steps. First, ESTs generated in earlier experiments using the Sanger method (Johansson et al., 2004;Morel et al., 2005) were assembled. Second, the EST sequences from the two 454 libraries were assembled. Third, the two sets of consensus sequences were assembled. The assembly result was a set of contig sequences representing various splice variants (isotigs in Roche vocabulary).
In step 1, the 21,163 collected Sanger sequences had an average length of 713 nucleotides; 15 of the reads contained rRNA sequences and were removed. The remaining sequences were assembled by using the Mira assembler, version 3.0.3 (Chevreux et al., 2004) with default parameters. The result was an assembly of 2,298 contigs (isotigs) representing 13,015 sequences.

Supporting Information
Organic matter degradation by ectomycorrhizal fungi 20 The remaining sequences were trashed mainly due to low quality scores. In step 2, an 454 assembly made with the gsAssembler software (Newbler) (Margulies et al., 2005) was kindly provided by the Department of Energy, Joint Genome Institute. The assembly contained 13,641 isotigs with sequence lengths >50 nucleotides. These isotigs had a mean length of 1,736 nucleotides and represented 2,114,871 reads.
In step 3, consensus sequences from the Sanger and 454 assemblies were assembled by using the CAP3 assembler (Huang and Madan, 1999). This resulted in an assembly of 13,654 isotigs. All read sequences were then mapped against the consensus sequences of the isotigs with the aid of the ssaha2 software (Ning et al., 2001) using default settings. Probably due to a higher stringency in the ssaha2 algorithm as compared to the assembly programs, there were 781 isotigs that no reads mapped to. These isotigs represented 98,281 reads. After removal of these isotigs and reads, the final numbers in the assembly was 12,873 isotigs representing 2,029,605 reads.
The average isotig length was 1,563 nucleotides. The vast majority of reads derive from the 454 sequencing (>99.9%). Based on the gsAssembler output, the 12,873 isotigs mapped to 8,620 genes (isogroups in Roche vocabulary). EST sequences are available at GenBank SRA046093.
Bioinformatic analyses of EST sequences. An EST database was constructed based on the assembly data that may be accessed from http://mbio-serv2.mbioekol.lu.se/Paxillus/Hybrid/. The database allows access to UniProt annotations, BLAST output, PFAM domains (Finn et al., 2006) and Gene Ontology (GO) (Ashburner et al., 2000). Similarity searches against UniProt (Apweiler et al., 2004) database were performed for each isotig consensus sequence. With an evalue threshold of 1e-10, 10,178 of the isotigs showed similarity with sequences from UniProt.
In the database, isotig annotations are inferred from the best Blastx hit.
To get isotigs encoding for proteins potentially involved in the degradation of lignocellulose, we compiled information on such proteins being identified from in the genome, transcriptome and secretome analysis of the wood decaying brown-rot fungus Po. placenta and the white-rot fungus Ph. chrysosporium (Martinez et al., 2004;Martinez et al., 2009;Vanden Wymelenberg et al., 2010;Vanden Wymelenberg et al., 2006;Kersten and Cullen, 2007) and other species

Supporting Information
Organic matter degradation by ectomycorrhizal fungi 21 (Vaillancourt et al., 2006;Gilbert, 2010;Hofrichter et al., 2010).The analysis generated a list of 123 "target proteins" involved in polysaccharide or lignin degradation, iron metabolism, oxalic acid metabolism and H 2 O 2 production. Identification of putative homologs of these sequences among the Pa. involutus isotigs was done in two steps. During step 1, the names of the 123 target proteins were used as queries to search for related sequences among the annotated Pa. involutus isotigs. The search generated 212 putative homologs. During step 2, a more extensive search for homologs was done by retrieving a broad range of query sequences related to those of the 123 "target proteins". First, 1,981 sequences having the names of the 123 "target proteins" in their annotations were obtained from the UniProt database of fungal proteins. Second,3,198 sequences displaying a high similarity to those of the 123 target proteins were retrieved from the UniProt database of fungal proteins by blastp searches. The two sets of sequences were compiled to a list of 5,179 protein sequences. Using this list as a query, a tblastn search (E-value threshold=1e -10 ) against Pa. involutus EST database generated 556 putative homologs among the isotigs. The putative Pa. involutus homologs identified during the 2 steps were manually annotated. Carbohydrate-active enzymes were annotated based on PFAM searches and similarity to such enzymes in the UniProt and GenBank databases, and they were classified using the CAZyme nomenclature (Cantarel et al., 2009). The resulting list contained 266 Pa. involutus isotigs encoding for proteins potentially involved in the degradation of lignocellulose. Prediction of the presence secretory signal peptides and non-classical secretion signals were done using the SignalP 3.0 and SecretomeP algorithms, respectively (Emanuelsson et al., 2007).
Microarray experiments. From a combined assembly of the two cDNA libraries the 12,873 isotigs were identified and were used for the construction of 12-plex 135K oligonucleotide microarray (Nimblegen/Roche). Of those 12,214 isotigs could be used for probe/reporter construction and allowed up to 10 probes per isotig in tiled design. The microarray analyses were performed as single-label hybridizations. For each hybridization and each sample 10 g of total RNA was used for cDNA synthesis using the SuperScript Double-Stranded cDNA Synthesis Kit (Invitrogen) according to the manufacturer. For quality assessments the produced cDNA was analyzed on a 2100 Bioanalyzer and DNA 7500 kit (Agilent). For sample labeling the One-Color DNA Labeling Kit (NimbleGen/Roche) was used according to the manufacturer. After labeling,

Supporting Information
Organic matter degradation by ectomycorrhizal fungi 22 each sample received a Sample Tracking Control (NimbleGen/Roche) and hybridizations were immediately performed in a Hybridization System 4 (NimbleGen/Roche) for at least 16 h and according to the manufacturer. The following washing procedure was accordingly (NimbleGen/Roche) and the slides were finally scanned in an Agilent High-Resolution Microarray Scanner set at 2 m. The raw images were bursted and processed using the NimbleScan software v. 2.5 according to the manufacturer (NimbleGen/Roche). Normalized (log(2)-transformed) values were brought into the Omics Explorer ver. 2.2 (Qlucore) for PCA and statistical analyses. Hybridization signal were obtained for 12,214 isotigs.