A First Insight into North American Plant Pathogenic Fungi Armillaria sinapina Transcriptome

Armillaria sinapina, a fungal pathogen of primary timber species of North American forests, causes white root rot disease that ultimately kills the trees. A more detailed understanding of the molecular mechanisms underlying this illness will support future developments on disease resistance and management, as well as in the decomposition of cellulosic material for further use. In this study, RNA-Seq technology was used to compare the transcriptome profiles of A. sinapina fungal culture grown in yeast malt broth medium supplemented or not with betulin, a natural compound of the terpenoid group found in abundance in white birch bark. This was done to identify enzyme transcripts involved in the metabolism (redox reaction) of betulin into betulinic acid, a potent anticancer drug. De novo assembly and characterization of A. sinapina transcriptome was performed using Illumina technology. A total of 170,592,464 reads were generated, then 273,561 transcripts were characterized. Approximately, 53% of transcripts could be identified using public databases with several metabolic pathways represented. A total of 11 transcripts involved in terpenoid biosynthesis were identified. In addition, 25 gene transcripts that could play a significant role in lignin degradation were uncovered, as well as several redox enzymes of the cytochromes P450 family. To our knowledge, this research is the first transcriptomic study carried out on A. sinapina.


Introduction
In forest ecosystems, fungi represent a complex group of microorganisms that vary from uni-to multicellular entities. They are capable of developing in very different lifestyles, from saprophytes to plants, to animal and human pathogens, parasites and symbionts. Through their adaptative capacities, these organisms could play also roles as biocatalyst agents [1]. In addition, some fungi have the ability to interact with each other and with other ecosystem components in a multi-dimensional and complex way within the forest ecosystems [1][2][3][4].
Armillaria species are globally distributed in natural and planted forests of temperate, boreal and tropical ecosystems. Armillaria sp. are specific pathogens that colonize living trees and cause white rot. They are regarded among the most infamous and important phytopathogens of forest trees. More than enzymes involved in degradation of the cell walls of birch bark, and to potentially discover new genes involved in the biotransformation of betulin into betulinic acid, in an effort to valorize forest residues.

Fungal Culture and Growth Conditions
A. sinapina fungal culture was provided by the Biopterre laboratory (La Pocatière, QC, Canada). A. sinapina mycelium was maintained on Potato Dextrose Agar (PDA) medium, and stock cultures were stored at 4 • C. The identity of the A. sinapina isolate was confirmed by sequencing of the ribosomal internal transcribed spacer (ITS) DNA region, which displayed 100% sequence identity to A. sinapina isolate XY17_48 ITS NCBI sequence ID MH550358.1. To optimize the mycelial growth of A. sinapina, the culture medium of Yeast Malt Agar (YMA) was selected [20,[29][30][31][32]. Different growth condition parameters of liquid and solid culture of A. sinapina were tested in the presence or absence of betulin (15 µg/mL), as described previously [32].

Isolation of RNA
RNA was isolated from three biological replicates of the cultured mycelium of A. sinapina according to Fradj et al. [32]. Briefly, RNA was extracted from A. sinapina liquid culture of YMB medium supplemented with 0.2% DMSO (control) or 15 µg/mL of betulin (BET) and grown for 8 days on a rotary incubator at 150 rpm and 28 • C in the dark. The mycelium was filtered, frozen with liquid nitrogen, and then, placed in a 2 mL tube with 2 mL of TRIzol reagent (Fisher Scientific, Ottawa, ON Canada). The mycelia were crushed using 5 mm diameter stainless steel beads with the TissueLyser II (QIAGEN, Qiagen Retsch GmbH, Hannover, Germany) at a speed of 30 strokes per second for 3 min until a homogeneous sample was obtained. Tubes were incubated on ice for 5 min, then 400 µL of chloroform was added. Tubes were vigorously vortexed for 15 s and incubated in an ice bath for 3 min. After centrifugation at 13,000× g for 15 min at 4 • C, two phases were obtained; the clear upper phase of the supernatant, exempt of cellular debris, was recovered and transferred into a clean 1.5 mL microtube. RNA was concentrated using isopropanol [32]. RNA pellets were resuspended in 20 µL of nuclease-free water. RNA quality and quantity were determined based on absorbance ratios using the Nanodrop and by the Bioanalyzer (Agilent Technologies, Inc., Santa Clara, CA, USA). Samples with an RNA integrity number higher than 8 were selected for Illumina sequencing.

Transcriptome Sequencing, De Novo Assembly and Functional Annotation
Sequencing was performed using Illumina HiSeq 4000, PE 100 paired ends, at McGill University (Montreal, QC, Canada) and Genome Québec Innovation Centre (Montreal, QC, Canada) on cDNA libraries converted from isolated high quality mRNA. Low quality reads, reads with unassigned nucleotides and adapters were filtered out. Trimmomatic was used to remove the adapters, trim reads from the 3' end and filter all reads below 50 bp in order to obtain clean reads [33]. Then, normalization was performed to remove redundant reads in datasets without affecting its Kmer content [34]. De novo assembly of cleaned and standardized reads was performed using the Trinity (v2.6.5) assembler [35]. This method assembles the RNA-seq reads into full-length transcripts, called contigs or unigenes. The predicted unigenes were aligned and searched in the uniprot_sprot.trinotate_v2.0.pep public protein database using the BLASTX algorithm with an E-value cutoff threshold of 10-5 against the NCBI non-redundant (nr) protein database. Top BLAST hits were used for annotation of component/gene for each transcript. To quantify the gene transcript abundance, the raw RNA-Seq reads were mapped to assembled transcripts with Bowtie using the default parameters [36]. The gene transcripts' abundance was calculated as transcripts per kilobase million (TPM) [37,38].
Functional annotation of unigenes was performed using Trinotate (http://trinotate.github.io/), by aligning transcripts to the Swiss Institute of Bioinformatics databases (Swiss-Prot) with BLASTX and identifying protein domains (HMMER/PFAM), protein signal peptides, transmembrane domains prediction (signalP/tmHMM), and Clusters of Orthologous Groups (COG). Then, the functional annotation of unigenes was compared to curated annotation databases (EMBL UniProt eggNOG/GO Pathways databases). The Trinity assembly and the functional annotation of unigenes were integrated as an annotation report into an SQLite database.
Gene Ontology (GO) enrichment analysis for the unigenes was performed using Swiss-Prot. The WEGO 2.0 software was used to obtain GO functional classification and OmicsBox v.1.2.4 for the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses for all unigenes. The raw reads under the accession number PRJNA565538 were deposited to the Sequence Read Archive [39].

Accession Numbers
The sequences described in this paper have been deposited in the National Center for Biotechnology Information Sequence Read Archive (https://www.ncbi.nlm.nih.gov/sra/) under the accession number PRJNA565538.

Statistical Analysis
Modulations of transcripts levels with adjusted P-values ≤ 0.05 and a fold change (log2FC) ≥ 1 were designated as statistically significant differentially expressed genes (DEG). Differential analysis was done by Genome Québec with "DESeq2", using raw counts with "R" and its packages as for Inonotus obliquus and KEGG, which are identified with OmicsBox, which uses UniProt and KEGG databases [32].

Illumina Sequencing and De Novo Assembly
To better understand the molecular mechanisms underlying the differences in A. sinapina's transcriptome caused by the presence of betulin, RNA-Seq and de novo assembly were performed. RNA was extracted from biological replicates of cultured A. sinapina in the absence (control) or presence of the terpenoid substrate betulin (BET). A total of 170,592,464 raw paired reads were generated from all replicates (both conditions were analyzed together, Table 1). After filtering out low quality sequences, 166,006,772 clean reads were obtained, corresponding to approximately 97.3% of the total raw reads (Table 1). In view of the lack of availability of information (i.e., genome, transcriptome data) on A. sinapina, all clean reads obtained from the libraries were de novo assembled using Trinity. A total of 273,561 transcripts with an average length of 2051 bp, and a N50 length of 3523 bp were obtained (Table 1; Figure 1A). An evaluation of the size distribution showed that 96% of all annotated transcripts of A. sinapina measured more than 1 kb ( Figure 1B).

Functional Annotation of A. Sinapina Transcriptome
A total of 121,959 transcripts (44.6%) homologous to known proteins or conserved hypothetical proteins were obtained from the search using BLASTX. Next, 103,267 (37.75%), 61,272 (23.53%), 45,683 (16.70%), and 105,195 (40.39%) transcripts matched annotated sequences from the GO, COG, Pfam and KEGG databases, respectively. A total of 145,191 (53.07%) transcripts were annotated in this way (Table 2). Strikingly, 128,370 (46.92%) transcripts showed no similarity with known sequences from the databases, which suggests that unknown new genes have been uncovered from A. sinapina. Pathways databases). The Trinity assembly and the functional annotation of unigenes were integrated as an annotation report into an SQLite database. Gene Ontology (GO) enrichment analysis for the unigenes was performed using Swiss-Prot. The WEGO 2.0 software was used to obtain GO functional classification and OmicsBox v.1.2.4 for the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses for all unigenes. The raw reads under the accession number PRJNA565538 were deposited to the Sequence Read Archive [39].

Accession Numbers
The sequences described in this paper have been deposited in the National Center for Biotechnology Information Sequence Read Archive (https://www.ncbi.nlm.nih.gov/sra/) under the accession number PRJNA565538.

Statistical Analysis
Modulations of transcripts levels with adjusted P-values ≤ 0.05 and a fold change (log2FC) ≥ 1 were designated as statistically significant differentially expressed genes (DEG). Differential analysis was done by Genome Québec with "DESeq2", using raw counts with "R" and its packages as for Inonotus obliquus and KEGG, which are identified with OmicsBox, which uses UniProt and KEGG databases [32].

Illumina Sequencing and De Novo Assembly
To better understand the molecular mechanisms underlying the differences in A. sinapina's transcriptome caused by the presence of betulin, RNA-Seq and de novo assembly were performed. RNA was extracted from biological replicates of cultured A. sinapina in the absence (control) or presence of the terpenoid substrate betulin (BET). A total of 170,592,464 raw paired reads were generated from all replicates (both conditions were analyzed together, Table 1). After filtering out low quality sequences, 166,006,772 clean reads were obtained, corresponding to approximately 97.3% of the total raw reads ( Table 1). In view of the lack of availability of information (i.e., genome, transcriptome data) on A. sinapina, all clean reads obtained from the libraries were de novo assembled using Trinity. A total of 273,561 transcripts with an average length of 2051 bp, and a N50 length of 3523 bp were obtained (Table 1; Figure 1A). An evaluation of the size distribution showed that 96% of all annotated transcripts of A. sinapina measured more than 1 kb ( Figure 1B).   Of the 103,267 transcripts matched to known GO terms (Figure 2), the most abundant GO category was 'cellular component' with 48.0% (95,002 transcripts), followed by 'molecular function' with 39.1% (77,267 transcripts), and lastly, 'biological process' with 12.9% (25,492 transcripts). GO slim terms for molecular function included catalytic, binding, and transporter activities, reflecting the ability of A. sinapina to transport and metabolize diverse compounds. Cellular component terms consisted of 10 groups, in which the top three were cell, organelle and membranes ( Figure 2). Biology 2020, 9,153 6 of 20  The 61,272 (23.52%) unigenes collected from COG were classified into 25 categories ( Figure 3). A large number of transcripts (17,113) were classified in the functional category R (general function prediction only), followed by category E (amino acid transport and metabolism; 10,164 transcripts), and category G (carbohydrates transport and metabolism; 9009 transcripts) ( Figure 3). Interestingly, the functional category specialized metabolism, transport and catabolism (Q), represented 8% (6381 transcripts) of the COG annotated transcripts. The 61,272 (23.52%) unigenes collected from COG were classified into 25 categories ( Figure 3). A large number of transcripts (17,113) were classified in the functional category R (general function prediction only), followed by category E (amino acid transport and metabolism; 10,164 transcripts), and category G (carbohydrates transport and metabolism; 9009 transcripts) ( Figure 3). Interestingly, the functional category specialized metabolism, transport and catabolism (Q), represented 8% (6381 transcripts) of the COG annotated transcripts.
Furthermore, a total of 105,195 (40.39%) transcripts from A. sinapina transcriptome were assigned using the pathways reported in the KEGG ( Figure 4). Metabolism, genetic information processing, environmental information processing, cellular processes, and organismal systems were present. The most highly represented category of KEGG function was "the global and overview maps". Genes encoding specialized metabolite biosynthesis, carbohydrate metabolism, energy metabolism, lipid metabolism and nucleotide metabolism were also uncovered. These results indicate that active metabolic processes were ongoing in A. sinapina. Altogether, the results for GO, COG, Pfam and KEGG Furthermore, a total of 105,195 (40.39%) transcripts from A. sinapina transcriptome were assigned using the pathways reported in the KEGG ( Figure 4). Metabolism, genetic information processing, environmental information processing, cellular processes, and organismal systems were present. The most highly represented category of KEGG function was "the global and overview maps". Genes encoding specialized metabolite biosynthesis, carbohydrate metabolism, energy metabolism, lipid metabolism and nucleotide metabolism were also uncovered. These results indicate that active metabolic processes were ongoing in A. sinapina. Altogether, the results for GO, COG, Pfam and KEGG annotations suggest that most of the transcripts expressed in A. sinapina cultured cells took part in basic biological processes, such as metabolism and biological regulation.

Differential Expression Analysis
Differential expression analyses were performed between the generated databases of A. sinapina. The addition of betulin during A. sinapina cells culture led to 3943 differentially expressed genes (DEG). Forty-eight times more transcripts (3863) were up-regulated than down-regulated (80) in the presence of betulin (control vs. BET) ( Table 3). This suggests that A. sinapina undergoes a transcriptomic adjustment in the presence of botulin and activates the expression of a significant number of genes. In the top 50 up-regulated transcripts, several transcripts encoded proteins involved in lipid and carbohydrate metabolism such as ketoacyl-CoA thiolase, phosphoglycerate kinase, pyruvate decarboxylase, fumarate reductase, glycoside hydrolase, phosphopyruvate hydratase, and ceramide fatty acid hydroxylase (Tables 4 and 5). This indicates that A. sinapina responded to betulin by prompting the cell process towards basic metabolism to produce energy (catabolism) and growth. Interestingly, a methylsterol oxidase which catalyzes the Fe-dependent first step of the removal of the two C-4 methyl groups of 4,4, dimetylzymsterol in zymosterol biosynthesis from lanosterol, ranked at position 5 of the top up-regulated transcripts ( Table 4). This could indicate a feedback regulation step to process betulin. Moreover, a cytochrome P450 transcript whose activity has not been elucidated was up-regulated 9.91-fold (Table 4). Interestingly, P450 often catalyze reactions at bottleneck rate-limiting step, such as the lanosterol pathway. In addition, the conversion of betulin to betulinic aldehyde or acid, involves a redox reaction often performed by enzymes from the CYP450 family, appointing it as an interesting candidate, particularly for the biotransformation of betulin.     In the top 50 down-regulated genes from A. sinapina cultured cells, betulin treatment decreased the expression of transcripts involved in complex carbohydrates' transport and degradation such as pectate lyase, hexose transporter, glycosyltransferase, permease, and lactone oxidase (Table 5). This suggests that in the presence of betulin, A. sinapina cells slow down processes involved in the transport and in the degradation of carbohydrates, most likely to focus on lipids' and triterpenoids' degradation instead. This observation might indicate a shift in A. sinapina catabolism from sugars to betulin as carbon source.

Overview of Gene Expression with Biotechnological Relevance
In the presence of betulin, a significant number of gene transcripts involved in pathways leading to terpenoid biosynthesis, enzymes catalyzing various redox reactions and survival approaches acquired by white rot fungi were differentially regulated ( Table 6). As betulin is a triterpenoid retrieved in large amounts in white birch barks, special attention has been given to understand and identify enzymes involved in the triterpenoid metabolism.

Terpenoids
The analysis of enzymatic pathways of A. sinapina transcriptome with KEGG revealed a number of gene transcripts differently regulated in the presence of betulin including 23 transcripts encoding enzymes involved in terpenoid backbone biosynthesis, 3 transcripts encoding enzymes for sesquiterpenoid and triterpenoid biosynthesis, 12 transcripts for fatty acid biosynthesis and other terpenoid quinones. With regard to terpenoid backbone synthesis, genes encoding thiolase, hydroxymethylglutaryl-CoA lyase, phosphomevalonate kinase and phosphomevalonate synthase were identified. Transcripts for farnesyl diphosphate synthase, precursors of sesquiterpenoids (e.g., α-muurolene, trichothecene, protoilludene) and triterpenoids (e.g., lanosterol) have also been identified ( Figure 5). Muurolene synthase, trichodiene oxygenase and protoilludene synthase for sesquiterpenoid biosynthesis were also detected as well as squalene synthase and epoxidase (Table 6).

Discussion
Filamentous fungi are of considerable importance through their diversity, their ability to produce specialized metabolites, and their involvement in different bioprocesses. The focus of this research was on fungal phytopathogen A. sinapina and its possible application in various bioprocesses such as betulin biotransformation.
We obtained comprehensive transcriptome data from A. sinapina by analyzing transcripts from samples treated or not treated with betulin using Illumina sequencing and Trinity de novo

Cytochrome P450s
A total of 108 transcripts encoding cytochrome P450 were detected in the transcriptome of A. sinapina. Among them, 32 transcripts were up-regulated in the presence of betulin (Table 6). Interestingly, eight different transcript variants of the cytochrome P450 72A14 were found up-regulated following BET treatment (Table 6). In plants, the P450 CYP716A family, and specifically CYP716A15, is implicated in the synthesis of betulinic acid [40,41]. Potential functional domains of CYP72A14, similar to plant CYP716A15, may be involved in terpenoid metabolism. Using PROSITE and SMART in silico tools [42,43], we detected that, in addition to a common P450 domain, CYP716A15 has two similar potential transmembrane domains, whereas CYP72A14 has one. Both of these domains' sequences commonly display glycosylation, phosphorylation and N-myristoylation potential sites.

Enzymes Involved in Plant Cell Wall Degradation
In the pathogen-host relationship, it is important for the pathogen to use the host as a carbon source [7]. A. sinapina transcriptome revealed that the fungus contains key enzymes required for the degradation of plant host biomass. As such, decomposition of the plant cells require a group of enzymes that act in concert on cellulose, pectin and hemicelluloses (Table 6). Among them, gene transcripts involved in the degradation of cell walls that were highly up-regulated in A. sinapina transcriptome following betulin treatment included laccases, glucosidases, mannosidases, invertases, debranching enzymes, peroxidases, and glucanases (Table 6). Interestingly, two pectate lyases transcripts were down-regulated in the presence of betulin (Table 6). Moreover, as would be expected from white rot fungi, transcripts encoding cell wall degradation enzymes such as laccases, peroxidases, H 2 O 2 -generating oxidases and polysaccharide monooxygenase were present and up-regulated following the addition of betulin in A. sinapina transcriptome.

Discussion
Filamentous fungi are of considerable importance through their diversity, their ability to produce specialized metabolites, and their involvement in different bioprocesses. The focus of this research was on fungal phytopathogen A. sinapina and its possible application in various bioprocesses such as betulin biotransformation.
We obtained comprehensive transcriptome data from A. sinapina by analyzing transcripts from samples treated or not treated with betulin using Illumina sequencing and Trinity de novo transcriptome assembly. Using a similar method, A. L. Ross-Davis (2013) assembled and analyzed the transcriptome of A. solidipes [7]. Compared to their work, the average lengths and N50 of our contigs, scaffolding and unigenes were longer. In another research work on the saprophytic fungus Wolfiporia cocos, Illumina sequencing of the transcriptome yielded a total of 38,722,186 reads, which were assembled into 60,354 contigs with N50 of 765 bp. Here, we obtained 4-7 times more reads, ensuring a broad coverage and a more contiguous assembly, as confirmed by the average transcript length of 2051 bp. The average length of eukaryotic transcript genes is greater than 1 kb (1135-1695 bp in yeast, 1108-2667 bp in human) 103,267 (37.75%), 61,272 (23.53%), 105,195 (40.39%), 45,683 (16.70%) unigenes had similar sequences with the GO, COG, Pfam and KEGG databases, respectively. The most abundant GO categories and functional groups here ( Figure 2) were consistent with those of the transcriptomic analysis of A. solidipes [7]. In addition, 128,370 of our transcripts showed no similarity with known sequences based on available databases until now. Hence, we have discovered new genes from A. sinapina that had not been previously identified by other genomics researchers. The KEGG annotation indicated that the three most abundant categories were the metabolic pathways, biosynthesis of specialized metabolites, biosynthesis of antibiotics and carbon metabolism.
Treatment of A. sinapina cell culture with betulin revealed differential regulation of a large number of gene transcripts putatively involved in the development of white rot disease. Interestingly, 3863 transcripts were up-regulated, and 80 down-regulated in the presence of betulin. Transcripts of enzymes, such as cytochrome P450, lanosterol synthase, glycoside hydrolase, 3-ketoacyl-CoA thiolase 5 and C4-methyl sterol oxidase involved in the specialized metabolites' biosynthesis were retrieved in the top 50 up-regulated genes. P450s play essential roles in biological pathways, such as the biosynthesis of terpenoids and alkaloids. Because of their ability to introduce oxygen into non-activated C-H bonds, P450s are attractive tools for biotechnology applications such as biotransformation methods as an alternative to classical chemical synthesis, particularly in the field of triterpenoid oxidation [44]. Here, the CYP72A sub-family was found to be up-regulated in the presence of betulin. In plants, CYP72A is required for the metabolism of specialized triterpenoids that carry out species-specific functions, such as chemical defense response to specialized pathogens [45,46]. To date, no study has been published on the possible function of this sub-family and very little is known on P450 in general in fungi [44]. P450 enzymes often metabolize toxic substances into more polar and water-soluble compounds, and therefore, play an important role in detoxification, which may explain why they were up-regulated in the transcriptome of A. sinapina in the presence of antimicrobial betulin [47]. In addition, a number of gene transcripts of the P450 family identified in the transcriptome of A. sinapina and up-regulated by betulin treatment have not yet been functionally defined.
We also identified numerous transcripts encoding enzymes involved in terpenoids biosynthesis and proposed a reconstructed pathway. Interestingly, despite some differences (i.e., lanosterol synthase), this biosynthetic pathway of A. sinapina is very similar to the one present in plants. Thus, fungi, and probably plant pathogen ones, may have evolved to produce plant-like enzymes that enable them to better benefit from host's metabolites. In recent years, Basidiomycota white rot fungi have demonstrated their ability to degrade all components of lignocellulosic biomass, and have therefore, become promising candidates for biotechnological applications, such as the pretreatment of plant feedstocks for conversion to bioethanol, other biofuels and the production of value-added bioproducts. Genomic and transcriptomic studies of several fungi of the Basidiomycota Agaricomycetes family have revealed genes encoding enzymes that cause wood degradation [9]. Oxidation and depolymerization of lignin are considered a significant barrier to the industrial use of lignocelluloses from multiple sources such as wood residues. The lignin may be modified, and sometimes mineralized by extracellular oxidoreductases produced by white rot fungi, including lignin and manganese peroxidases, in conjunction with some enzymes generating hydrogen peroxide such as glyoxal oxidases or aryl alcohol oxidases [9]. The versatility of North species such as A. sinapina to biodegrade all structural components of plant cell walls is clearly shown in its transcriptome, which has a large number of genes associated to the degradation of plant cell walls, including enzymes that degrade lignin, cellulose, hemicellulose, and pectin [48]. Laccases, different peroxidases, several H 2 O 2 -generating oxidases and polysaccharide monooxygenase were present and up-regulated by the addition of betulin in A. sinapina transcriptome. Several of these enzyme types have been reported in other white rot fungi [49]. For example, a manganese-peroxidase was found in the fungal strain of C. subvermisposa with more than 11 isoforms [50].
Laccases are described in various organisms, from bacteria to insects, and have the ability to oxidize various phenolic substrates by reducing the oxygen in water [50]. Recently, a growing interest for polysaccharide mono-oxygenases uncovered in large numbers in gene analyses of white rot fungi has been shown [51]. They are a class of copper-dependent enzymes that play a role in boosting the degradation of polysaccharides in lignocellulosic biomasses. They are capable of weak hydrolysis of beta-glucan, carboxymethylcellulose, lichenin, wheat arabinoxylan and birch xylan. In addition, polysaccharide monooxygenases in conjunction with other cellulolytic enzymes stimulate the hydrolysis of lignocellulosic substrates, such as hydrothermally pretreated wheat straw [51,52]. Here, Cel16 was up-regulated in the presence of betulin.
The potential of A. sinapina in biomass deconstruction is also shown by the detection of various transcripts encoding proteins belonging to the glycosylhydrolase family, i.e., beta-glucosidase, alpha-xylosidase, probable exo-1.4 beta xylosidase, endo-1,4-beta-xylanase, bifunctional xylanase/deacetylase, xyloglucanase, polygalacturonase and exopolygalacturonase, endo-beta-1,4-glucanase D, alpha-xylosidase and alpha-glucosidase, arabinan endo-1,5-alpha-L-arabinosidase A and beta-mannosidase Regarding enzymes involved in cell wall modifications, it is noteworthy that there were more gene transcripts up-regulated than down-regulated. Only pectate lyases (PL) were down-regulated. The cell wall is an important barrier against pathogens that attack plants. Pathogens produce pectinases, polygalacturonases, pectins, PL and pectin esterases to digest homogalacturonans and rhamnogalacturonans [53]. In 2006 and 2013, studies indicated that secretion of PL and fungal pathogenicity could be influenced by nutritional factors, such as the presence of nitrogen in culturing medium. [54,55]. In addition, PL have an absolute requirement for Ca 2+ ions [56]. The negative regulation of PL could be explained by different factors, such as deficiencies in nitrogen or Ca 2+ in the culture medium, pH, temperature, and finally, by the presence of triterpenoids (i.e., betulin), which could be inhibitors to this enzyme [57].

Conclusions
In this study, we used A. sinapina, a root disease pathogen of hardwood hosts as a model to explore de novo assembly of a complete transcriptome and to assess changes in transcripts levels following betulin treatment using the Illumina paired-end sequencing method. The quality and the depth of our transcriptomic assembly is significantly improved, compared to other fungal transcriptome studies, and will serve as a basis for further transcriptomic research on North American Armillaria species. A total of 343,289 unigenes were obtained from A. sinapina mycelium following Trinity de novo assembly. Various gene transcripts orthologs, with important roles in the degradation of plant cell walls and involved in the production of metabolites, were identified and annotated. These results demonstrate that Illumina paired-end sequencing is a fast and cost-effective approach for the discovery of new genes in a non-model organism. This research may help to optimize applications exploiting A. sinapina and will facilitate the identification of candidate genes for uses in biotransformation, and also as an enzyme platform for the degradation and depolymerization of woody biomass to produce biofuels and a chemical product. Further work on the nature of the genes of the cytochrome P450 family and the non-annotated genes could reveal key elements in the relationship between A. sinapina and its plant host, and thus, its involvement in the life cycle of the forest ecosystem.