Conserved white-rot enzymatic mechanism for wood decay in the Basidiomycota genus Pycnoporus

Abstract White-rot (WR) fungi are pivotal decomposers of dead organic matter in forest ecosystems and typically use a large array of hydrolytic and oxidative enzymes to deconstruct lignocellulose. However, the extent of lignin and cellulose degradation may vary between species and wood type. Here, we combined comparative genomics, transcriptomics and secretome proteomics to identify conserved enzymatic signatures at the onset of wood-decaying activity within the Basidiomycota genus Pycnoporus. We observed a strong conservation in the genome structures and the repertoires of protein-coding genes across the four Pycnoporus species described to date, despite the species having distinct geographic distributions. We further analysed the early response of P. cinnabarinus, P. coccineus and P. sanguineus to diverse (ligno)-cellulosic substrates. We identified a conserved set of enzymes mobilized by the three species for breaking down cellulose, hemicellulose and pectin. The co-occurrence in the exo-proteomes of H2O2-producing enzymes with H2O2-consuming enzymes was a common feature of the three species, although each enzymatic partner displayed independent transcriptional regulation. Finally, cellobiose dehydrogenase-coding genes were systematically co-regulated with at least one AA9 lytic polysaccharide monooxygenase gene, indicative of enzymatic synergy in vivo. This study highlights a conserved core white-rot fungal enzymatic mechanism behind the wood-decaying process.


List of supplementary tables
. A and B mating type loci and related genes as manually annotated in the JGI databases. Table S3. Numbers of predicted B mating type genes and similar non-mating-type genes found in the genomes of the four Pycnoporus strains. Table S4. Identities and similarities between orthologous proteins linked to mating type or mating-type-like genes in Pycnoporus species. Table S5. JGI protIDs for the 60 protein models of class I peroxidase (CCP), class II peroxidase (short MnP, LiP, and typical and atypical VP) and HTP peroxidase genes in the genomes of three Pycnoporus species. Table S6. JGI protIDs for the copper radical oxidase gene models in the genomes of three Pycnoporus species.   Table S9. P. cinnabarinus BRFM 137 genome annotation related to protein secretion pathways. Table S10. P. cinnabarinus BRFM 137 genome annotation related to protein glycosylation pathways. Table S11. Propagation of functional annotations for predicted proteins involved in the secretory pathway, from P. cinnabarinus BRFM 137 to P. coccineus BRFM 310 and P. sanguineus BRFM 1264. Table S12. Propagation of functional annotations from P. cinnabarinus BRFM 137 to P. coccineus BRFM 310 and P. sanguineus BRFM 1264 predicted proteins involved in the glycosylation pathway.
Table S13. Number of predicted Cyt450 genes up-regulated after 3 day growth on Avicel, wheat straw, pine or aspen. Table S14. Gene count for predicted peptidases in the genomes of P. cinnabarinus BRFM 137, P. coccineus BRFM 310 and P. sanguineus BRFM 1264. Table S15. Numbers of peptidases identified in at least one secretome obtained from P. cinnabarinus BRFM 137, P. coccineus BRFM 310 and P. sanguineus BRFM 1264 grown on maltose, Avicel, wheat straw, aspen or pine. Table S16. List of proteins detected in the secretomes of P. cinnabarinus BRFM 137 grown on Aspen, Pine, wheat straw, Avicel or maltose. Table S17. List of proteins detected in the secretomes of P. coccineus BRFM 310 grown on Aspen, Pine, wheat straw, Avicel or maltose.
Table S18. List of proteins detected in the secretomes of P. sanguineus BRFM 1264 grown on Aspen , Pine, wheat straw, Avicel or maltose.
3 Table S19. List of the groups of genes (nodes) from P. cinnabarinus BRFM 137, P. coccineus BRFM 310 and P. sanguineus BRFM 1264 up-regulated in response to Avicel, and their transcription profile on wheat straw and aspen. Table S20. List of the groups of genes (nodes) from P. cinnabarinus BRFM 137, P. coccineus BRFM 310 and P. sanguineus BRFM 1264 up-regulated in response to wheat straw or aspen, not Avicel.                     Altogether, the B mating type and similar genes of P. cinnabarinus BRFM 137 were found on four distinct scaffolds (Table S2). Suspected B mating type genes reside all on scaffold 184836 at its upper end (196.2 kb total length). The two genes for the non-mating type Ste3-like proteins (Ste3. 3 and Ste3.4) are found at the upper end of scaffold 184788 (71.7 kb total length). Gene Phl5 is present on scaffold 184857 (35.7 kb) next to a gene Dam1 for an essential component of kinetochores (Buttrick & Millar, 2011). The orthologs of all these genes reside in one continuous DNA region of about 60 kb in length on scaffold 14 in P. coccineus BRFM 310 and on scaffold 718000065086 in P. sanguineus BRFM 1264, respectively (Table S2). This suggests that all these genes together represent the broader B mating type locus in Pycnoporus species with the actual B mating type genes (from 311212 to 325411 on scaffold 14 in P. coccineus BRFM 310; from 1774998 to 192157 on scaffold 718000065086 in P. sanguineus BRFM 1264) and any linked genes for non-mating-type pheromone-like peptides (upstream) and Ste3-like proteins (downstream). Because the respective groups of genes on the three scaffolds 184857, 184836 and 184788 in P. cinnabarinus BRFM 137 cannot be simply aligned up into a continuous larger DNA region, substantial rearrangements gene orders with translocations appear to have happened in the species. The situation in P. puniceus BRFM 1868 is similar where groups of genes are found on scaffold 31, 57 and 6, respectively. All potential genes with B mating type function in P. puniceus BRFM 1868 reside on scaffold 57 (Table S2).

List of supplementary figures
Regarding the continuous structure of their broader B mating type loci, P. coccineus BRFM 310 and P. sanguineus BRFM 1264 appear much more similar to each other than to the two other Pycnoporus species in support of the notion that these two species are closest related. Numbers of B mating type and similar non-mating-type genes (Table S3) and relative orders of Ste3, Ph and Phl genes differ between the four strains (Table S2), suggesting younger evolutionary activities in B mating type locus differentiation, possibly involved in B allele generation and species diversification.
In all four Pycnoporus species, there is an independent locus in the genome separated from the B locus for one pheromone-like peptide (the Phl2 gene in P. cinnabarinus BRFM 137 on scaffold 184858). In all species, this gene is neighbored by the mating-related gene Ste20 coding for a kinase acting in the mitogen-activated protein kinase (MAPK) pathway, the genes of which in some other basidiomycetes are found within mating type loci (Karos et al., 2000;Coelho et al., 2008;Kourist et al., 2015). Lack of close linkage of the Phl2-Ste20 DNA segment to the B mating type locus in the four Pycnoporus strains reinforces that the pheromone-like peptides of Phl genes have no mating type function. The primary and the predicted mature products of different Phl genes within a strain and between species are comparably well conserved (Table  S2). This contrasts the relatively little conserved Ph genes for typical B mating type pheromones. However, some of the processed pheromones appear to be shared between species (Table S2). Ph genes are intermingled in the B loci in all four species with genes for those Ste3 receptors that group in phylogenetic analysis with typical B mating type pheromone receptors and separately from non-mating-type Ste3-like proteins (Levasseur et al., 2014) while genes for non-matingtype Ste3-like proteins are not accompanied by any precursor genes for pheromones or pheromone-like peptides as is expected from observations in other species (Kües et al., 2011;Kües, 2015;Kues et al., 2015).
The same Phl-Ste20 constellation than in the Pycnoporus strains is found on scaffold 3 (models 184724 and 144172 encoded) in the also sequenced Trametes versicolor strain FP-101664 SS1 8 (Floudas et al., 2012) of the broader trametoid clade (Justo & Hibbett, 2011;Carlson et al., 2014). Unlinked on scaffold 8, T. versicolor has also copies of the two more strongly conserved non-mating type Ste3 genes (IDs 73179 and 170723) positioned downstream of its continuous B mating type locus (with 7 Phl genes and 5 pheromone precursor genes intermingled in between 3 B mating type pheromone receptor genes) while the gene for Dam1 in T. versicolor strain FP-101664 SS1 (ID 174288) was moved to another location (scaffold 12). Dam1, Ste3. 3 and Ste3.4 of T. versicolor were used for weighing in the trametoid clade in comparisons of encoded proteins between the different Pycnoporus species (Table S4). P. coccineus BRFM 310 and P. sanguineus BRFM 1264 showed for all proteins closer relationships as compared to the other species, similar as for the proteins from genes flanking the A mating type locus (Table S4).
Looking at sequence similarities of Mip1 and β-fg proteins, also these data suggests that P. coccineus and P. sanguineus are more closely related to each other than to the other two Pycnoporus species and to T. versicolor (IDs 138945 and 138946) and it appears that the data reflect the same phylogenetic relations between the four Pycnoporus species as formerly reported from ITS sequences, β-tubulin and a laccase (Lesage-Meessen et al., 2011). Importantly, the same tendencies of higher sequence conservation between P. coccineus and P. sanguineus as compared to the other species are again apparent in comparisons of proteins from HD1 and HD2 genes of the A mating type loci (T. versicolor IDs 187232 and 187235) and in comparisons of Ste3.1 and Ste3.2 proteins from the B mating type loci (T. versicolor IDs 73174 and 170716), (Table S4).

Peroxidase gene models in Pycnoporus genomes
A preliminary screening of the automatically-annotated genomes of Pycnoporus cinnabarinus BRFM 137, P. coccineus BRFM 310 and P. sanguineus BRFM 1264 was performed using the Search option ("peroxidase" as search term) at the JGI web-site. Then, a sequence-by-sequence exhaustive analysis was used to identify heme peroxidases (Table S15). The revision of the previously annotated genome sequence of P. cinnabarinus BRFM 137 evidenced the presence of thirteen sequences encoding heme peroxidases, which means two over those previously reported by Levasseur et al., 2014. Manual annotation of all the gene models identified was based on the highest sequence identities for each protein sequence derived from the predicted gene, multiple alignments with other 458 basidiomycete heme peroxidase protein sequences, and examination of theoretical molecular structures obtained by homology modeling using crystal structures of related peroxidases as templates and programs implemented by the automated protein homology modeling server "SWISS-MODEL" (Bordoli et al., 2008).
The heme peroxidases could be classified into three different groups as follows: i) Cytochrome c peroxidases (CCP) (1 model in each genome analyzed) belonging to Class I of prokaryoticorigin peroxidases; ii) Class II ligninolytic peroxidases, including members of the new subfamily of "short" manganese peroxidases (MnP) with both Mn-mediated and Mn-independent activity on low redox potential substrates (Fernández-Fueyo et al., 2014b), lignin peroxidases (LiP) able to oxidize high redox potential non-phenolic aromatic compounds, and typical and atypical versatile peroxidases (VP and VP-atypical) sharing catalytic properties of MnP and LiP, and iii) Hemethiolate peroxidases (HTP), some of them with peroxygenase and peroxidase activity (Hofrichter & Ullrich, 2014), differing from the above Class I and Class II peroxidases in the presence of a proximal cysteine (instead of a histidine) acting as the fifth heme iron ligand. No generic peroxidases (GP) able to oxidize low redox potential aromatic compounds in direct contact with heme were identified from these genomes.
Class II ligninolytic peroxidases could be annotated as LiP, MnP and VP on the basis of the presence or absence of only a few amino acid residues at the substrate oxidation sites (Ruiz-Dueñas et al., 2009) after homology modeling. In this respect: i) LiPs are characterized by harboring an exposed catalytic tryptophan homologous to Trp171 in Phanerochaete chrysosporium LiP-H8 and Trp164 of Pleurotus eryngii VPL; ii) MnPs are characterized by containing a Mn(II)-oxidation site near the internal propionate of heme formed by three acidic residues homologous to P. chrysosporium MnP1 Glu35, Glu39 and Asp179, and P. eryngii VPL Glu36, Glu40 and Asp175; and iii) VPs are characterized by presenting both the catalytic tryptophan and the Mn(II) oxidation site of LiPs and MnPs, respectively (atypical VPs contain an atypical Mn-oxidation site formed by one glutamate and two aspartate residues). Regarding MnPs, these peroxidases were annotated as members of the subfamily of short MnPs containing a short C-terminal tail like the exhaustively characterized short MnPs from Pleurotus ostreatus and Ceriporiopsis subvermispora (Fernández-Fueyo et al., 2014b,a).
A representation of the homology models obtained for the enzymes identified in P. coccineus BRFM 310, including key amino acid residues putatively involved in catalysis, is presented in Fig. S14 and Fig. S15 as an example of the homology models obtained for all the peroxidases identified. LiP, MnP and VP models were obtained using the Phanerochaete chrysosporium LiPH8 (PDB entries 1B80 and 1B82), Pleurotus ostreatus MnP4 (PDB entry 4BM1) and VP1 (PDB entries 4BLK and 4BLN), and Pleurotus eryngii VPL2 (PDB entries 2BOQ, 4FCS and 3FMU) crystal structures as templates. Regarding HTPs, the members of this superfamily identified in the Pycnopous genomes were modeled with the crystal structures of chloroperoxidase (CPO) (PDB entry 2CIW) from the ascomycete Leptoxyphium fumago and unspecific peroxygenase (UPO) (PDB entry 2YP1) from the basidiomycete Agrocybe aegerita. Unlike Class II peroxidases, differences in the amino acid residues of the heme environment were observed among the HTP models analyzed. One of the HTPs identified in each genome seems to be a CPO-type peroxidase containing Glu and His residues at the heme distal side involved in enzyme activation by H2O2 (Sundaramoorthy et al., 1995) (Fig. S16), whereas the other HTP models harbour Asn and His residues not present neither in CPO nor in UPO (in the latter Glu and Arg residues occupy homologous positions) (Piontek et al., 2013). This fact suggests putative differences in their catalytic properties which should be confirmed by heterologous expression and kinetic characterization of these enzymes. Finally, genes encoding dye-decolorizing peroxidases (DyP) were not identified in any of the Pycnoporus genomes analyzed as previously described in the P. cinnabarinus genome, a fact that is not frequent among white-rot basidiomycetes where this peroxidase family is widespread with only a few exceptions (Floudas et al., 2012;Ruiz-Dueñas et al., 2013).
A dendrogram showing sequence relationships between the peroxidases identified in the above Pycnoporus strains, and structural-functional classification of the ligninolytic peroxidases according to the presence of different catalytic sites in their theoretical molecular structure is shown in Fig. S16. Those peroxidases corresponding to the same isoenzyme in the three Pycnoporus strains have the same name and appear clustered together. The only exception is LiP4, which had been previously called LiP5 in P. cinnabarinus due to an error naming a VP as if it were a LiP (=LiP4) (Levasseur et al., 2014). As observed, in all cases the enzymes from P. coccineus and P. sanguineus are more related to each other than with those from P. cinnabarinus.
It is interesting to mention that among the P. cinnabarinus BRFM 137 peroxidase sequences deposited at JGI and included in this dendrogram, there are three (LiP3, ID# 7501; VP1, ID# 6829; and VP-atypical, ID# 6481) that needed manual curation to yield the sequences published by Levasseur et al.. On the other hand, model 1748658 from P. sanguineus BRFM 1264 was annotated as a putative non-functional LiP4. Its amino acid sequence converted into a structural homology model presents the general folding of a typical ligninolytic peroxidase. However it lacks key residues involved in enzyme activation by H2O2 (His and Arg residues at the distal heme side), in addition to amino acids in other key positions yielding an active enzyme. This model could be a natural pseudogene or the result of an error in the sequencing/assembling process.
The evolutionary relationships of Class I and Class II heme peroxidases from the Pycnoporus strains and other basidiomycetes are shown in Fig. S17. CCPs and hybrid ascorbate-cytocrome c peroxidases (APX-CCP) (the latter absent from the genomes of the Pycnoporus strains) form two clusters (A and B) clearly separated from class II peroxidases. Among these last, low redox potential GPs (formed by unclustered peroxidases and a few peroxidases grouped together into cluster C) have been described to be at the origin of the high redox potential ligninolytic peroxidases (Floudas et al., 2012), and at least one GP representative of this family is observed in most of the genome sequences of both white and brown-rot basidiomycetes. However, although this is the generality, GP-encoding genes have not been conserved in the Pycnoporus species' genomes, as also observed in other sequenced species (Dichomitus squalens, Trametes versicolor and Ganoderma sp) belonging, like those from the genus Pycnoporus, to the core Polyporoid clade (Ruiz-Dueñas et al., 2013). MnPs derived from ancestral GPs were the first ligninolytic peroxidases to appear by progressive incorporation of three acidic residues forming the Mnoxidation site (Floudas et al., 2012). In fact, there are atypical MnPs encoded in the Auricularia delicata, Fomitiporia mediterranea and Stereum hirsutum genomes that contain only two of the acidic residues representing intermediate evolutionary states between GPs and MnPs. Regarding the members of the MnP family, these appear grouped into three different clusters in the dendrogram. Long MnPs, including the classical enzymes from P. chrysosporium (Gold et al., 2000), and extralong MnPs grouped into cluster D were initially classified as two different MnP subfamilies (Floudas et al., 2012), although studies have demonstrated that both types of MnPs present similar catalytic properties and stability, and in consequence they are considered members of the same MnP family (Fernández-Fueyo et al., 2014a). Unlike long and extralong MnPs, short MnPs are grouped into two different clusters (E and F). On one hand, short MnPs comprised in cluster E are related to a group of atypical VPs (and a few VPs) from species of the core Polyporoid clade, including three atypical forms corresponding to the same isoenzyme in three of the four Pycnoporus strains analyzed. On the other hand, most of short MnPs, including those identified in the Pycnoporus strains analyzed, are intermixed with two VP/LiP intermediates from C. subvermispora (Fernández-Fueyo et al., 2012) and a few VPs from Spongipellis sp, Bjerkandera adusta, Bjerkandera sp. and Pleurotus species constituting cluster F. Finally, cluster G is characterized by containing LiPs (from B. adusta, Phlebia brevispora, Phlebia radiata, P. chrysosporium, Phlebiopsis gigantea and T. versicolor) and some VPs, including those from the three Pycnoporus strains.
The analysis of the position of the peroxidase genes in the different Pycnoporus genomes revealed that four of them (lip3, lip2, lip1 and mnp3) co-localize in the same syntenic block encompassing at least 11 genes. This peroxidase grouping was already described in P. cinnabarinus and in the closely-related T. versicolor (Levasseur et al., 2014). In a similar way, htp1, htp4, htp2 and vp2 cluster together on the same scaffold in the P. coccineus and P. sanguineus genome assemblies (htp1, htp2 and vp2 in P. cinnabarinus due to the absence of htp4 in its genome sequence).
Finally, the analysis of the evolutionary relationships of 117 basidiomycete HTPs (Fig. S18) confirmed that the enzymes identified in the Pycnoporus species belong to two different groups (clusters B and C in the dendrogram), probably exhibiting different catalytic properties, as previously predicted from the analysis of the amino acid residues located at the heme environment.

Copper Radical Oxidases
We identified in each genome seven genes coding for Copper Radical Oxidases (CROs) instead of 9 in the well-studied white-rot fungus Phanerochaete chrysosporium. Among them, three code for glyoxal oxidases (Table S21). GLX catalyze the two-electron oxidation of simple aldehydes and hydroxycarbonyls with the reduction of molecular oxygen to hydrogen peroxide. GLX thereby generate extracellular hydroperoxide that supports peroxidase activity and lignin degradation (Kersten & Kirk, 1987). For each predicted CRO in P. cinnabarinus we identified the orthologous gene in P. coccineus and P. sanguineus, including for the two characterized PciGLOX1 and PciGLOX2 enzymes (Daou et al., 2016).

Laccases
We identified five laccase, one multicopper oxidase and one ferroxidase coding genes in each genome (Table S22). The exon-intron structure of the lac genes was conserved with 11 introns in lac1, lac3, and lac4 and 13 exons in lac2 and lac5. The lac3 and lac1 genes were in tandem on scaffold 185007 from Pycci, scaffold 16 from Pycco and scaffold 7180000650860 from Pycsa, with intergenic regions ranging from 25,9 kb to 32,2 kb.

(GMC)-oxidoreductases
(GMC)-oxidoreductases were identified by using the CAZy annotation pipeline which is based on blastp search thresholds using characterized enzymes as queries and the conservation of HMM profiles (Lombard et al., 2014). To enhance protein function prediction, the proteins classified as AA3s in the CAZy database were further compared with 46 sequences of characterized fungal AA3s (Sützl et al., 2018). Phylogenetic trees were built using ClustalW for the alignment of the sequences and MEGA-CC (Kumar et al., 2012) for maximum likelihood analysis with 500 bootstrap values (Fig. S12,.

Small Secreted Proteins
Predicted secreted proteins were identified that met the five conditions : 1) a peptide signal predicted by SignalP, 2) predicted as secreted by TargetP and wolfpsort, 3) no Lys-Asp-Glu-Leu (KDEL) motif in C-terminal (prosite accession'PS00014'), and 4) 0 or 1 transmembrane helix found by TMHMM (Pellegrin et al., 2015). If there is one helix it should overlap the signal peptide. Small Secreted Proteins were predicted secreted proteins <300 aa long.

Secretory pathway
Genes coding for proteins involved in the secretory and glycosylation pathways were identified in Pycci using reciprocal best blast hits and e value<10 -10 with Aspergillus niger proteins as queries (Pel et al., 2007). Proteins involved in the secretion pathway included genes involved in protein entry into the endoplasmic reticulum (signal recognition, signal peptidase complex, translocation into the ER, protein folding into the ER), Protein misfolding (Unfolded Protein Response (UPR), ER associated Degradation (ERAD), proteasome), protein complex involved in protein transport (exocyst complex, SEC34/SEC35 complex, Trapp complex, COPI and COPII subunits), proteins involved in vesicle formation and docking (SNARE proteins, secretion related GTPases and interacting proteins, ER to Golgi and Intra-Golgi transport, Golgi to endosome transport, Vacuolar protein sorting, Cellular export and secretion; Table S23). Proteins involved in glycosylation pathways included proteins involved in the biosynthesis of nucleotide sugars for glycosylation events (UDP-Glucose, UDP-N-acetyl-glucosamine, GDP-mannose, UDPgalactose), transporters of sugar nucleotide donors (GDP-mannose, UDP-GlcNac, UDPgalactose, UDP-galactofuranose, Oligosaccharyltransferase subunits), synthesis of the dolicholphosphate linked ER-precursor Glc3Man9GlcNAc2, processing of the ER-precursor Glc3Man9GlcNAc2 after transfer to a polypeptide, Golgi mannosyltransferase, O-Glycosylation in ER, putative alpha-1,2-mannosidase with no homology to the MNS1/ER-alpha-1,2mannosidase family, GPI anchor biosynthesis, GPI-anchor transamidase complex (Table S24). Functional annotations were propagated from P. cinnabarinus predicted proteins to P. coccineus and P. sanguineus predicted proteins by homology search using orthoMCL with default parameters and the four genomes A. niger CBS 513.88, Pycci, Pycco and Pycsa (Tables S25-26).

Glutathione-S Transferases
The GST-coding genes were predicted with a combination of automated gene callers, and blastP tool using sequences from Phanerochaete chrysosporium that have been phylogenetically classified and functionally characterized for many of them (Meux et al., 2011(Meux et al., , 2013Mathieu et al., 2013;Thuillier et al., 2013Thuillier et al., , 2014Roret et al., 2015). The repartition of the sequences within the various classes was based both on phylogenetic relationship between P. chrysosporium and Pycnoporus sequences and active site comparison especially for GSTs from the Omega, GHR and Ure2p classes.
In fungi, seven GST classes have been defined (Morel et al., 2009). Among them, GSTs involved in translation (EFB class) and GSTs associated to membranes (MAPEG class) are generally not considered as true GST. Isoforms from the other classes (GSTO, GHR, Ure2p, GSTFuA and GTT2) are putatively involved in detoxification processes. The Pycci, Pycco and Pycsa genomes carry respectively 23, 34 and 35 gene copies. For comparison, Phanerochaete chrysosporium, Serpula lacrymans, Postia placenta and Trametes versicolor exhibit respectively 25, 30, 45 and 41 GST gene copies in their genomes (Morel et al., 2013). As other lignolytic fungi, GSTO, Ure2pA and GSTFuA are expanded in the three Pycnoporus species compared to GHR and Ure2pB classes. The expansion of GSTO class could be explained by a high number of serine containing GSTs. Contrary to the classical cysteine-containing GSTOs, which display a deglutathionylation activity (Meux et al., 2011), serine-containing GSTOs exhibit the classical glutathione transferase activity, known to be involved in detoxification pathways (Deroy et al., 2015). Since GSTOs are able to interact with wood extractives, which are wood-derived 13 molecules with potential antimicrobial activity, the serine containing GSTOs could protect Pycnoporus sp. from extractive toxicity.
However, these genes are not highly expressed in the culture conditions tested and are not induced by woody substrates. This is also the case for the GSTs from the other classes, except for one GSTFuA, that is induced in P. coccineus BRFM 310 during growth on pine. GSTFuA are fungal specific GSTs, that, additionally to their glutathionylation activity, have the properties to alternatively act as ligandins by binding wood extractive compounds at a L-site overlapping the glutathione binding pocket.
Noticeable was the difference between the global expression pattern of the GSTome for the three strains independently from the substrate. In P. cinnabarinus BRFM 137, the most strongly expressed GST gene codes for a predicted GTT2.1. GTT2.1s are specifically found in wood degraders. GTT2.2 expression is induced in P. chrysopsporium in presence of oak acetonic extracts and the recombinant protein is not active as a glutathione transferase but rather as a peroxidase (Thuillier et al., 2014). This GTT2.1 could thus be involved in oxidative stress rescue in P. cinnabarinus. Ure2pB genes are highly transcribed both in P. coccineus BRFM 310 and P. sanguineus BRFM 1264 but not in P. cinnabarinus BRFM 137. The physiological role of Ure2pB is still unknown but the P. chrysosporium ortholog is constitutively expressed and displays a deglutathionylation activity (Thuillier et al., 2013).
Altogether, these results show that the regulation of the GSTome differs in each strain despite very similar GST gene repertoires.

P450s
The RNASeq analysis showed several genes up-regulated or highly transcribed on (ligno)cellulose that had a cytochrome P450 predicted function according to KOGG classification. In this study, a gene was considered up-regulated if its log2 fold change was higher than 2 during growth on (ligno)cellulosic substrate as compared to maltose or if its log2 read count was higher than 12 on one of the tested substrates (cellulose, wheat straw, pine or aspen). The sequences of retrieved predicted cytochrome P450s were subjected to blastp similarity searches in the Fungal Cytochrome P450 Database (FCPD; last accessed 01 July 2018; (Park et al., 2008)).

Peptidases
Blastx searches of gene models from all three species of Pycnoporus against InterPro and MEROPS databases (Mitchell et al., 2015;Rawlings et al., 2016) followed by hand curation found 340 putative peptidases encoded by the P. coccineus BRFM 310, 354 encoded by P. sanguineus BRFM 1264, and 321 encoded by P. cinnabarinus BRFM 137. These numbers were in good agreement with the automated annotation results at the Joint Genome Institute where the respective peptidase counts are 360, 378, and 340. As in most basidiomycetes, the large majority of peptidase genes encode members of three MEROPS-recognized mechanistic classes: aspartic proteases, metalloproteases, and serine proteases (Table S27). Aspartic peptidase genes average just fewer than twenty percent of the total peptidases in basidiomycete genomes (Lilly and Gathman, unpublished analysis). Metallopeptidases and serine peptidases average 31 and 32 percent respectively. Pycnoporus species have a larger proportion of genes encoding aspartic peptidases, a situation found in Phanerochaete chrysosporium and Postia placenta among those species with published genomes.
Three families of peptidases, aspartic peptidase A1, serine peptidase family S10, and serine peptidase family S53 (including cross-listed S8/S53 members) are expanded in all three Pycnoporus species. Together these three families represent just under 30 % of all peptidase genes found in each genome. Members of these families (Type examples: A1 = pepsin, S10 = carboxypeptidase Y, S53=sedolysin) all are active only at acid pH. In addition, in most fungi the vast majority of these enzymes are secreted. A similar situation exists for the Pycnoporus species, where in P. sanguineus 89 % of the gene models for these families have signal peptides based on SignalP analysis; only slightly fewer (80%) have signal peptides in P. coccineus. In contrast, few of these proteins are predicted to have signal peptides in P. cinnabarinus, possibly due to lower quality of the gene models.
Proteomic data from the secretome of these species when grown on six different carbon sources found only members of the aforementioned expanded families. However, each species showed a different spectrum of secreted peptidases. The total number of peptidase gene products identified in the secretomes was highest for P. coccineus (31 peptidases identified), including 12 predicted S53 gene products. In P. cinnabarinus and P. sanguineus, across all media, one specific A1 peptidase contributed over half of the total A1 proteomic spectra, and the greatest number of spectra of any peptidase found. This peptidase also returned the highest number of A1 peptidase spectra on each different medium. In P. coccineus secretomes, one specific A1 and one peptidase assigned to the S8 and S53 family accounted for half of the spectra. A comparison of the sequences of the A1 peptidases genes from all three species shows that, in most cases, it is the products of the directly orthologous genes that are found in the medium. Similarly, the S10 peptidase genes found in the various media are, in most cases, products of orthologous genes (not shown). The paucity of secreted products of S53 genes in P. sanguineus and P. cinnabarinus makes a similar analysis difficult for them.
Overall expression data for peptidases in the three species in basal maltose medium revealed that members of all major mechanistic classes are expressed. However, expression was found for only a fraction of the total population of peptidase genes in each species (P. coccineus = 181 expressed/340 total; P. sanguineus = 97 expressed/354 total; P. cinnabarinus = 125 expressed/321 total). Serine peptidases represented the highest proportion of expressed peptidase genes in each species, followed by metalloprotease genes and aspartic protease genes. Transcription levels were studied on five different media, and expression levels compared to basal maltose medium. For those situations where the Log2FoldChange > |2|, most of the upregulated genes were members of the expanded A1, S10, and S53 peptidase gene families.     Phl non-matingtype genes 5 6 6 5      S8: AA3 gene transcription regulation and protein detection in the secretomes of Pycnoporus coccineus BRFM310. For some enzymes, activities could be predicted from a plylogenetic analysis with 46 characterized fungal AA3 GMC-oxidoreductases. Prediction for secretion was deduced from the presence of a predicted signal peptide, the absence of other trans-membrane domains and the absence of signal sequence for retention in the endoplasmic reticulum. Secretomes and transcriptomes were collected after three day growth on maltose (M); avicel (AVI), wheat straw (WS), Aspen (As) or Pine (Pi). Component of the TRAPP (transport protein particle) complex, which plays an essential role in the vesicular transport from endoplasmic reticulum to Golgi No hits found An04g08690 YDR108w TRS85 Subunit of TRAPPIII (transport protein particle), a multimeric guanine nucleotideexchange factor for Ypt1p, required for membrane expansion during autophagy and the CVT pathway; directs Ypt1p to the PAS; late post-replication meiotic role scf184863.g20 An15g00470 YDR407c TRS120 One of 10 subunits of the transport protein particle (TRAPP) complex of the cis-Golgi which mediates vesicle docking and fusion; involved in endoplasmic reticulum (ER) to Golgi membrane traffic scf185002.g59 An08g05190 YMR218c TRS130 One of 10 subunits of the transport protein particle (TRAPP) complex of the cis-Golgi which mediates vesicle docking and fusion; involved in ER to Golgi membrane traffic; mutation activates transcription of OCH1 No hits found An15g03010 YDR246w TRS23 One of 10 subunits of the transport protein particle (TRAPP) complex of the cis-Golgi which mediates vesicle docking and fusion; involved in endoplasmic reticulum (ER) to Golgi membrane traffic; human homolog is TRAPPC4 scf184829.g45 An14g06440 YDR472w TRS31 One of 10 subunits of the transport protein particle (TRAPP) complex of the cis-Golgi which mediates vesicle docking and fusion; involved in endoplasmic reticulum (ER) to Golgi membrane traffic scf184829.g75 An15g00060 YOR115c TRS33 One of 10 subunits of the transport protein particle (

Serine Peptidases
Peptidase S10 20 20 20 Peptidase S14 0 1 1   Table S16. List of proteins detected in the secretomes of P. cinnabarinus BRFM 137 grown on Aspen (As), Pine (P), wheat straw (W), Avicel (A) or maltose (M).    Table S19. List of the groups of genes (nodes) from P. cinnabarinus BRFM 137, P. coccineus BRFM 310 and P. sanguineus BRFM 1264 up-regulated in response to Avicel, and their transcription profile on wheat straw and aspen. For each node the mean of log2 read counts and log2 fold change are indicated. Nodes for which the mean log2 read count was ≥ 12 on cellulose and nodes for which the mean log2 fold change as compared to maltose was ≥ 2 were considered respectively as highly transcribed or up-regulated on the substrates. Non significant changes in transcript level as deduced from DESeq2 are indicated by log2 fold change = 0. All genes were identified after expert annotation. M: maltose, AVI: avicel, WS: wheat straw, Asp: aspen.    The dendrogram shows sequence relationships and structural-functional classification of the ligninolytic peroxidases (short MnPs have a Mn 2+ -oxidation site formed by two glutamate and one aspartate residues, shown on blue background; LiPs contain a catalytic tryptophan, shown on green background; VPs harbor the catalytic sites described for both MnPs and LiPs; and atypical VPs have an aspartate residue in the position occupied by one of the glutamic residues at the Mn 2+ -oxidation site, shown on pale blue background), and provides the residues homologous to L. fumago CPO/A. aegerita UPO heme distal residues (His105/Arg189 and Glu183/Glu196) (shown on orange background) and heme proximal residues (Pro28/Pro35, Cys29/Cys36 and Pro30/Pro37) (shown on pink background) identified in the HTPs. Sequence comparison as Poisson distances, and clustering using the UPGMA method and ''pairwise deletion'' option of MEGA5 . Numbers on branches represent bootstrap values (based on 1000 replications) supporting that branch. Only the values ≥50% are presented.