P. carnosa genome assembly and annotations are available via JGI Genome Portal MycoCosm  and DDBJ/EMBL/GenBank under the accession AEHB00000000.
Phanerochaete carnosa strain HHB-10118-sp and Phanerochaete chrysosporium strain RP-78 were obtained from the U.S. Department of Agriculture (USDA) Forest Products Laboratory (Madison, WI). These strains were maintained on YMPG agar plates under stationary conditions at 27°C. YMPG medium contained 2 g yeast extract, 10 g malt extract, 2 g peptone, 10 g glucose, 2 g KH2PO4, 1 g MgSO4·7H2O, and 15 g agar per 1 L in H2O.
Single-copy cdh and fet3/ftr1 genes of P. carnosa were amplified by PCR and sequenced to reveal an absence of allelic polymorphism . Specifically, PCR was performed using AccuPrime Pfx DNA polymerase and Reaction Mix (Invitrogen) with 160 ng of genomic DNA and 12.5 pmol of each primer in a 25 μL reaction volume. Primers for amplification of cdh (5'-TCKGARGCHGGVAAGAARGT-3' and 5'-GGVCCRATVCCGCTYTGGAA-3') were designed based on conserved sequences from six Basidiomycete fungi, and the PCR cycle was run as follows: 95°C for 9 min, 30 cycles of (95°C for 1 min, 50°C for 2 min, 72°C for 2 min), and 72°C for 15 min. Primers for amplification of fet3/ftr1 (5'-TGGACGATCTGGAACTTGTG-3' and 5'-TCTCACGGAAGACGATGAAG-3') were based on the corresponding P. chrysosporium sequence, and the PCR cycle was run as follows: 95°C for 9 min, 30 cycles of (95°C for 1 min, 65°C for 2 min, 72°C for 3 min), and 72°C for 15 min. Amplified sequences were cloned into the pCR2.1-TOPO plasmid (Invitrogen) and sequenced at the Analytical Genetics Technology Centre (Toronto, ON, Canada) or the Centre for Applied Genomics (Toronto, ON, Canada). Absence of clamp connections normally produced by heterokaryons during cell division  was also confirmed by microscopic visualization of P. carnosa mycelia.
Genome sequencing and assembly
The genome was sequenced using a combination of Sanger (4 kb, 8 kb, 40 kb paired end), 454 (Titanium unpaired; 3.7 kb, 5.2 kb, 6.1 kb, 14.5 kb Titanium paired end), and Illumina (3 lanes of 2x76 bp, 0.3 kb insert paired end) sequencing platforms ( Additional file 1: Table S1a). All general aspects of library construction and sequencing can be found at the JGI website . The Illumina data was assembled with the Velvet assembler (version 0.7.55; ) with a hash length of 61 and the following options; -ins_length 250 -scaffolding no -exp_cov 16 -cov_cutoff 5, to produce an assembly with a final graph with 51765 nodes, n50 of 2413, max 22432, total 37740533, using 73178660/94020728 reads. Contigs greater than or equal to 800 bp in length were shredded into 1000 bp fragments with 800 bp overlap to be used by Newbler assembler. After eliminating possible contaminant data, the combined set of velvet fragments, 454 and Sanger reads was assembled with the Newbler assembler, version 2.5-internal-10Apr08-1 with the following options; -fe reads2remove.FQC -consed -nrm -finish -info -rip -sio -g -ml 31 -mi 98 -e 48, to a final estimated assembled coverage of 58X, 1137 scaffolds with an N/L50 of 6/3.6 Mb, and 2687 contigs with an L/N50 of 248/45.2 Kb ( Additional file 1: Table S1b). One round of automated gap closure using our in house gapResolution tool  resulted in a final assembly with 2272 contigs with an N/L50 of 139/74.8 kb. Newbler assembled consensus EST sequence data was used to assess the completeness of the final assembly using alignment with 90% identity and 85% coverage thresholds. This resulted in 90.33% placement.
cDNA library construction, sequencing and assembly
Two clone cDNA libraries were constructed using RNA from P. carnosa mycelium grown in liquid YMPG medium and sequenced as described previously , with the minor difference that there were the two size ranges of cDNA (0.6 k-2 kb and >2 kb). The smaller cDNA insert library resulted in 9,971 ESTs for further processing and the larger insert library resulted in 9,530 ESTs. RNA purified from mycelia grown on mixed softwood was used to construct a 454 cDNA library according to the cDNA Rapid Library Preparation Method Manual (Rosch, Germany), yielding 1,139,862 ESTs for clustering and assembly. The entire set of 1,159,363 reads were assembled using Newbler (v2.3-PreRelease-6/30/2009) with default parameters resulting in 16,234 contigs greater than 50 bp. and 59,361 singlets.
The P. carnosa genome was annotated using the JGI annotation pipeline, which takes multiple inputs (scaffolds, ESTs, and known genes) and runs several analytical tools for gene prediction and annotation. Results were deposited in the JGI Genome Portal , a part of the integrated fungal resource MycoCosm  for further analysis and manual curation.
Genomic assembly scaffolds were masked using RepeatMasker  and the RepBase library of 234 fungal repeats . tRNAs were predicted using tRNAscan-SE . Using the repeat-masked assembly, several gene prediction programs falling into three general categories were used: 1) ab initio - FGENESH ; GeneMark , 2) homology-based - FGENESH+; Genewise  seeded by BLASTx alignments against GenBank’s database of non-redundant proteins (NR: ), and 3) EST-based - EST_map  seeded by EST contigs. Genewise models were extended where possible using scaffold data to find start and stop codons. EST BLAST alignments  were used to extend, verify, and complete the predicted gene models. The resulting set of models was then filtered for the best models, based on EST and homology support, to produce a non-redundant representative set of 13,937 gene models with characteristics described in Additional file 1: Table S2. This representative set was subject to further analysis and manual curation.
Predicted gene models were functionally annotated using SignalP , TMHMM , InterProScan , and BLASTp  against the National Center for Biotechnology Information nr database, and hardware-accelerated double-affine Smith-Waterman alignments  against  KEGG  and KOG . KEGG hits were used to assign EC numbers , and Interpro and SwissProt hits were used to map GO terms . Functional annotations of the representative set of genes are summarized in Additional file 1: Table S3 and Table S4. Multigene families were predicted with the Markov clustering algorithm (MCL ) using BLASTp alignment scores between proteins as a similarity metric.
Carbohydrate active enzymes
All P. carnosa protein models were subjected to a procedure combining BLAST and HMMer3 searches against sequence libraries and HMM profiles derived from the families of glycoside hydrolases, polysaccharide lyases, carbohydrate esterases, glycosyltransferases and carbohydrate-binding modules featured in the CAZy database [15, 69]. The models corresponding to glycoside hydrolase families GH1, GH2, GH3, GH5, GH6, GH7, GH9, GH10, GH11, GH12, GH16, GH28, GH31, GH43, GH45, GH51, GH53, GH55, GH61, GH74, GH79, GH115, carbohydrate esterase families CE1, CE15, CE16, carbohydrate-binding module family CBM1, as well as cellobiose dehydrogenase and cellulose-binding cytochrome b
562 were manually checked using TBLASTN against the P. carnosa assembly database . Gene models used for BLAST were obtained from the P. chrysosporium genome database [6, 71]. In the case of multi-domain proteins, sequences encoding CBM and its associated domains were separately used as separate queries. TBLASTN programs were performed with an expectation value of 1.0E-1, and all other settings at default values.
Protein models that were annotated as predicted sugar transporters and/or permeases in the P. carnosa genome portal v1.0 were used as queries for BLASTp against the NCBI protein database  and the P. chrysosporium genome portal version 2 to confirm these annotations. Annotations were considered accurate when either BLAST search gave an alignment to a predicted protein with an E-value of < 1e-10 and a score of ≥ 200. The P. carnosa genome was then re-searched using the P. chrysosporium genes selected above as queries to confirm that all of the gene models relevant to this analysis were selected.
For the phylogenetic analysis, multiple alignment was performed using MAFFT version 6 software  with the E-INS-i algorithm. The phylogenetic tree was then constructed from the multiple alignment using the bootstrapped neighbor-joining method (1000 bootstraps), and drawn using FigTree version 1.3.1 . This analysis included 28 gene models from P. carnosa and 25 gene models from P. chrysosporium, while partial gene fragments were removed.
The BLASTp algorithm available through the JGI Fungal Genomics Program website  was used with default settings to search Agaricomycotina gene catalogue proteins against reference proteins. Hits were then blasted against the NCBI database  with default settings, and aligned to the reference protein sequences using the tool at the Genestream Bioinformatics Resource server . Sequences were annotated to the reference protein when the bests hits to NCBI represented sequences of interest and the alignment showed at least 30% amino acid identity to the reference protein. Reference proteins were chosen based on biochemical evidence supporting their identity, and correspond to the following Genbank accession numbers: LO1 (laccase) LAC2_PLEOS, LO2 (peroxidases) LIG8_PHACH, LO3 (cellobiose dehydrogenase) CDH_PHACH, LDA1 (aryl alcohol oxidase) AAC72747, LDA2 (vanillyl-alcohol oxidase) VAOX_PENSI, LDA3 (glyoxal oxidase) AAA33747, LDA4 (pyranose oxidase) P2OX_PHLGI, LDA5 (galactose oxidase) XP_959153, LDA6 (glucose oxidase) XP_002910108, LDA7 (benzoquinone reductase) AAD21025, LDA8 (alcohol oxidase) AAB57849, methanol oxidase ALOX_PICAN, quinone reductase AF465406.
Initial determination of the putative cytochrome P450 gene models in P. carnosa was made by searching the JGI whole genome database for the term ‘P450’. The resulting putative sequences were subjected to BLAST analysis and searched for the presence of the conserved P450 signature domains namely, the oxygen-binding motif ‘EXXR’ and the heme-binding motif ‘CXG’. P450s that showed both the domains were considered authentic and were grouped into families and subfamilies based on the existing nomenclature criteria of > 40% amino acid homology for assigning a family and > 55% for a subfamily. The families were then grouped into P450 clans, a higher order level of nomenclature that represents a cluster of P450 families across species, grouped based on relationships that are beyond the family designations . P450 superfamily nomenclature rules were followed for assigning the clan, family and sub-family classification as earlier applied for P450ome classification in the model white-rot fungus, P. chrysosporium. P450s that did not have P. chrysosporium P450 homologues were annotated based on the phylogenetic alignment with other P450s on the phylogenetic tree; the tree was constructed using Mega 4 via the bootstrap UPGMA method . P450s showing both the conserved domains and a reasonable deduced protein length (≥ 330 aa) were used for the tree construction.
Metabolic network reconstruction
The P. carnosa metabolic network was reconstructed using version 15 of the Pathway Tools Software from SRI International Inc . This network was primarily constructed from the annotation using Enzyme Commission, Gene Ontology identifiers as well as name matching algorithms. The reconstructed metabolic network contained 1166 metabolites, 1630 enzymatic reactions and 11 transport reactions that were linked to 3292 enzymes. The completed network is available at .
Cultivation on model and wood substrates
P. carnosa and P. chrysosporium were grown in duplicate on modified Kremer and Wood medium containing 1.5% agar with 35 different model carbon sources  for 13 and 3 days, respectively, after which pictures of the plates were taken to compare colony diameter and thickness. Relative growth was determined by comparing the radius and density of the mycelia on a particular carbon source to that on D-glucose; relative growth profiles were then compared between the two fungal species. The extent of growth relative to plates containing glucose are categorized from high to non-detectable using the following designations: +++, ++, +, ±, -. No differences were observed between duplicates on any of the substrates, so only one of the pictures is presented. Monomeric and oligomeric carbon sources were used at a final concentration of 25 mM, while pure polymers were used at a final concentration of 1%. Crude plant biomass was used at a final concentration of 3%. This procedure allows consistent, qualitative assessment of colony development on many substrates; it has been performed routinely for more than 100 fungal species and showed good overall correlations with the genome content of these species [16, 18, 19].
The following hardwood and softwood species were used for wood cultivations: sugar maple (Acer saccharum), yellow birch (Betula alleghaniensis), trembling aspen (Populus tremuloides), red spruce (Picea rubens), white spruce (Picea glauca), balsam fir (Abies balsamea), and red pine (Pinus resinosa). All wood samples were obtained from New Brunswick, Canada. A 50 cm bolt at 80 cm and 130 cm trunk heights were cut from each wood species. Sapwood and heartwood sections were then separated, air-dried, and processed separately using a Wiley mill (Thomas scientific, NJ, USA). Resultant wood powder was sifted using 2 mm and 0.425 mm diameter mesh sieves and powder that passed through the 2 mm sieve but was retained by the 0.425 mm sieve was recovered. Four grams of wood powder were placed on top of 5 g vermiculite powder (< 1 mm dia.) in a glass petri dish measuring 9 cm in diameter; 20 mL of H2O was gently added to the dish, and the dish was then autoclaved for 20 min. A 0.5 cm dia. agar plug taken from the growing edge of P. carnosa or P. chrysosporium cultivated on YMPG agar plates was then transferred to the centre of each plate, and incubated at 27°C under stationary conditions. To maintain moisture content, 5 mL H2O was added to each plate every week during cultivation. Fungal growth was quantified by measuring the diameter of the fungal colony growing on each wood powder. Three replicate cultivations were prepared for each fungal and wood species.
An accelerated solvent extraction method (DIONEX, Application Note 335) was used to isolate wood extractives. To obtain standard samples for baseline analysis, 2 g of non-treated heartwood and sapwood samples from each wood species were mixed with approximately 0.45 g of diatomaceous earth, and then transferred to an 11 mL cell; the headspace in the cell was then filled with sand. Extraction was performed as follows; preheat 0 min, heat 5 min, static 5 min, flush 90%, purge 60 sec, cycles 5, pressure 1000 PSI, temperature 100°C, solvent 70% MeOH and 30% H2O (vol/vol). To obtain extractives from fungal-treated and control samples, each sample was mixed with solvent (70% MeOH, 30% H2O (vol/vol)) in a 10:1 ratio (solvent(mL):sample weight (mg)) and incubated for 24 h on an orbital shaker at room temperature, and then filtered. Supernatant was collected and a second extraction was performed on the filtered wood samples. Supernatants from the first and second extractions were combined, concentrated using a rotary evaporator, and then dried using a nitrogen evaporator.
Extractives were analyzed using Ultra performance liquid chromatography (UPLC). UPLC analysis was performed using a Waters Acquity Ultra Performance Liquid Chromatography equipped with a computer and Masslynx software, a binary solvent manager, a sample manager and an autoscan photodiode array spectrophotometer detector (PDA eλ). The UPLC was equipped with an Acquity UPLC BEH C18, 1.7 μm, (2.1 x 50 mm i.d.) reverse-phase analytical column from Waters housing a Van Guard BEH C18, 1.7 μm pre-column. All samples were dissolved in 70% acetonitrile: 30% H2O and diluted to a concentration of 10 mg/mL or to a minimum of 0.25 mL. Standards included gallic acid, methyl gallate, quercetin and rutin; 0.5 μL of each standard and 3 μL of samples were analyzed using gradient elution as shown in Additional file 1: Table S18. Column and auto-sampler temperature were maintained at 25°C. Two fixed detection wavelengths (280 nm and 350 nm) were used to monitor the eluting peaks. Resolved peaks were scanned by the photodiode array detector from 240 to 460 nm.
The F-C reagent method was used to calculate total phenolic concentration . A calibration curve was created using gallic acid at concentrations of 25 mg/L, 50 mg/L, 100 mg/L, 250 mg/L, and 500 mg/L in Milli-Q water. Briefly, 20 μL of sample, gallic acid standard or blank was transferred to a 2.0 mL cuvette; 1.58 mL of Milli-Q water and 100 μL of F-C Reagent were then added, mixed and incubated for 5 min. Subsequently, 300 μL of sodium carbonate was added, mixed and incubated for 2 h at room temperature. Absorbance was measured at 765 nm using a Beckman 800 series spectrophotometer.
Heartwood and sapwood of trembling aspen and red pine samples were collected before and after cultivation with P. carnosa and P. chrysosporium, and powdered using the mini-beadbeater-16 (Biospec products, USA); corresponding uninoculated controls were similarly processed. Two milligrams of wood powder were mixed with KBr (200 mg) and the mixture was pelletized using a die (1.3 cm diameter) and a hydraulic press. A Bruker Tensor 27 FT-IR was used to record the absorbance between 4000 and 400 cm-1 with a resolution of 4 cm-1. Spectra representing the average of 32 scans were corrected for atmospheric vapor compensation; baseline was corrected using the rubber band method (Opus software, v. 5.0). Spectra were normalized for unit-vector and mean-centred prior to the principal component analysis (PCA) using Unscrambler v. 9.7 software.