From simple and specific zymographic detections to the annotation of a fungus Daldinia caldariorum D263 that encodes a wide range of highly bioactive cellulolytic enzymes

Background Lignocellulolytic enzymes are essential for agricultural waste disposal and production of renewable bioenergy. Many commercialized cellulase mixtures have been developed, mostly from saprophytic or endophytic fungal species. The cost of complete cellulose digestion is considerable because a wide range of cellulolytic enzymes is needed. However, most fungi can only produce limited range of highly bioactive cellulolytic enzymes. We aimed to investigate a simple yet specific method for discovering unique enzymes so that fungal species producing a diverse group of cellulolytic enzymes can be identified. Results The culture medium of an endophytic fungus, Daldinia caldariorum D263, contained a complete set of cellulolytic enzymes capable of effectively digesting cellulose residues into glucose. By taking advantage of the unique product inhibition property of β-glucosidases, we have established an improved zymography method that can easily distinguish β-glucosidase and exoglucanase activity. Our zymography method revealed that D263 can secrete a wide range of highly bioactive cellulases. Analyzing the assembled genome of D263, we found over 100 potential genes for cellulolytic enzymes that are distinct from those of the commercially used fungal species Trichoderma reesei and Aspergillus niger. We further identified several of these cellulolytic enzymes by mass spectrometry. Conclusions The genome of Daldinia caldariorum D263 has been sequenced and annotated taking advantage of a simple yet specific zymography method followed by mass spectrometry analysis, and it appears to encode and secrete a wide range of bioactive cellulolytic enzymes. The genome and cellulolytic enzyme secretion of this unique endophytic fungus should be of value for identifying active cellulolytic enzymes that can facilitate conversion of agricultural wastes to fermentable sugars for the industrial production of biofuels. Supplementary Information The online version contains supplementary material available at 10.1186/s13068-021-01959-1.


Background
Because of the rapid development of our society in recent decades, our demands on liquid fuels have been increasing, and current reserves of fossil fuels will be exhausted in the near future. Thus, improvements in alternative energy have become imperative to sustain the further development of our society. First-generation starch-based

Open Access
Biotechnology for Biofuels *Correspondence: tho@gate.sinica.edu.tw 1 Institute of Plant and Microbial Biology, Academia Sinica, Taipei, Taiwan, ROC Full list of author information is available at the end of the article biofuels may supplement our energy requirement, but are not sustainable because of the use of food commodities as the starting raw material. Second-generation biofuels are produced from decomposition of lignocellulosic biomass without competition with our food resources and are thus more sustainable than first-generation biofuels. The major sources of second-generation biofuels are agricultural wastes such as rice straw and corn stover that contain high amounts of lignocellulose [1]. However, lignocellulose needs to be converted into sugars that can be further fermented into biofuels. Current lignocellulose can be decomposed by thermochemical and biochemical processes. The thermochemical routes, such as pyrolysis and gasification [2,3], decompose agricultural wastes into a mixture of gaseous and liquid forms of hydrocarbons. Although multiple compounds can be acquired by the thermochemical route, the processing requires high temperature and thus depends on high energy input. However, the biochemical route utilizes lignocellulolytic enzymes to degrade agricultural wastes into fermentable sugars, which requires lower temperature and produces relatively simple products [4].
Although the biochemical route requires lower energy input, the efficiency and stability of cellulases limits its use for long-term processing of agricultural waste. Furthermore, the cost of using cellulolytic enzymes to convert agricultural wastes into fermentable sugars is usually high [10]. To solve these problems, we need to discover an organism producing and secreting a high level of cellulases with good activity and stability. Most cellulases are produced by bacterial and fungal species, so researchers have been looking for highly efficient microbial cellulases to facilitate the cellulose decomposition process [11].
To achieve complete cellulose decomposition in industry, three types of cellulases have usually been used as a mixture from different fungal secretions because a single fungus is usually highly productive for only one or two types of cellulases [12,13]. Trichoderma reesei is one of the most extensively used cellulase producers, especially for EXOs; however, its BGLs account for only ∼ 0.5% of the total secreted cellulases [14,15]. Therefore, finding a novel organism that can produce and secrete all three types of cellulases would be beneficial. In addition to saprophytic fungi such as T. reesei, endophytic fungi have high potential for cellulase production and secretion. Recent studies of endophytes suggest that they produce volatile organic compounds applicable for biodiesel production, and the genomes of endophytic fungi encode a variety of carbohydrate-active enzymes (CAZymes) that may facilitate the degradation of lignocellulosic residues [16,17]. Although genomic analyses have provided high-throughput annotations of genes encoding potential cellulolytic enzymes, the expression of their protein products and their enzymatic activity of the secreted cellulases need to be investigated.
To efficiently verify cellulolytic enzyme activities, researchers have used various artificial substrates to identify specific cellulase activities. Filter paper (FP) assays have been extensively used for determining total cellulase activities [18]. Avicel and 4-methylumbelliferyl-βd-cellobioside (MUC) are commonly used as substrates for evaluating EXO activities [19]. 4-Methylumbelliferylβ-d-glucopyranoside (MUD) and p-nitrophenyl β-dglucopyranoside (pNPG) are used for assessing BGL activity [20,21]. However, some substrates may be used by more than one type of cellulase; for instance, both EXO and BGL can hydrolyze MUC, and their activities cannot be distinguished in test tube assays [19,[22][23][24]. Zymography may be a good alternative for screening potential microorganisms with high cellulase activities. Crude protein extracts from microorganisms can be directly loaded onto native polyacrylamide electrophoresis (PAGE) gels and cellulase activities can be easily visualized by adding colorogenic or fluorogenic substrates during staining [25]. Although both EXOs and BGLs can hydrolyze MUC, previous studies indicated that BGL activity is repressed by its product glucose via competitive inhibition [24], but EXOs do not respond to glucose inhibition [26].
Taking advantage of this glucose product inhibition property of BGLs, we have established a simple zymography assay that can clearly distinguish EXO activities without interference from BGLs. We discovered that the Daldinia caldariorum strain D263 fungus secretes a wide range of cellulolytic and hemicellulolytic enzymes, including EXOs, EGs, BGLs and xylanases. Our D263 genome assembly and annotation indicated that D263 may encode more than 100 potential cellulolytic enzymes, and the proportion among different types of CAZymes in D263 is distinct from that in T. reesei and A. niger, which have been used for commercial cellulase production. We also successfully identified bioactive cellulases in the D263 secretion by coupling zymography and proteomic analyses. This study demonstrated that combining specific zymography and mass spectrometry analysis provides a convenient method for identifying cellulolytic enzymes. With this approach, the D263 fungus was determined as a good source of these enzymes for facilitating second-generation biofuel production.

D263 produces and secretes a wide range of cellulolytic enzymes
To discover fungal species suitable for producing and secreting a wide range of highly bioactive cellulolytic enzymes, we screened more than 40 strains of ascomycetes by examining the cellulolytic enzyme activities in their crude secretion. Among the fungal strains, we found several fungal species/strains with a variety of cellulosic activity (Table 1 and Additional file 1: Table S1). The crude secretion of Chaetomella raphigera D2 (D2) contained high BGL activity, as shown in our recent studies [27]. However, we found no EG, EXO or xylanase activity in the culture medium of D2. Penicillium sp. YS-40 also exhibited considerable BGL activity but low xylanase activity and no EXO and EG activity. Fusarium proliferatum showed high EG, BGL and xylanase activity and moderate EXO activity. Aspergillus sp. RS-19, Neurospora sp. RS-6 and D. caldariorum D263 could hydrolyze filter paper (FP), so these three species could produce several types of cellulolytic enzymes that can work synergistically for cellulose decomposition. EXO, EG and BGL activity was high in Aspergillus sp. RS-19, but no xylanase activity was detected in its secretion. D. caldariorum D263 also showed all types of cellulolytic enzymes (i.e., EG, EXO and BGL), with BGL activity higher than that in Aspergillus sp. RS-19 (1.82 vs 0.42 U/mL). D263 secretion also contained high xylanase activity (2.69 U/ mL), which suggests that it has high potential for digesting hemicellulose. D263 showed relatively low activity on hydrolyzing the EXO substrate Avicel as compared with Aspergillus sp. (0.05 vs 0.18 U/mL), the commonly used fungal species for commercial enzymes. Hence, D. caldariorum D263 could be a good source for a complete set of cellulolytic enzymes or supplement of commercial cellulase because of its high BGL, EG and xylanase activity. We attempted to further mass-produce D263 cellulases in a 1-L fermenter. D263 crude secreted enzymes could hydrolyze carboxymethylcellulose (CMC) and pNPG, showing 0.29 and 0.11 U/mL activity, respectively, at 138 h. The D263 crude enzymes could also degrade FP, so they could completely digest crystalline cellulose (Fig. 1a).
To test whether D263 could hydrolyze agricultural waste into fermentable sugars, we incubated the crude enzymes of D263 with acid-pretreated sugarcane bagasse. Under the same FP unit, the secreted D263 cellulases could produce > 500 μg glucose from 10 mg sugarcane bagasse, although relatively lower than the commercial enzyme mixtures CTec3 and C1.5L (Fig. 1b). Interestingly, the mixture of D263 and C1.5L also showed a synergistic effect on decomposing sugarcane bagasse, (Fig. 1c). The commercial cellulase mixture C1.5L from T. reesei is high in EG and EXO activity, but is low in BGL activity [27], so the addition of D263 may act as a supplement of BGL activity that enhances the glucose production. In contrast, such synergism was not observed when mixing D263 secretion with CTec3 (Additional file 2: Fig. S1), consistent with Table 1 Cellulolytic enzyme activities of the examined ascomycete species One unit of enzyme activity (U) was defined as 1 μmol of product generated per minute, FP, filter paper; CMC, carboxymethylcellulose; pNPG, p-nitrophenyl β-dglucopyranoside  To further identify the bioactive cellulases in the D263, we used CMC zymography. D263 culture medium showed high EG activity (Fig. 2a, b). To distinguish EXO activities from BGLs, we developed an improved version of MUC zymography. In previous studies, it is known that the activity of BGLs is highly repressed by its product glucose [27][28][29]. Our previous research also showed that inhibition constants for glucose (i.e., K i glucose) of D2-BGL and N188 were 2.42 mM and 4.68 mM, respectively [27]. Therefore, we included 2% of glucose (111.1 mM) in the zymography assay in order to identify EXO activity without the interference of BGL activity. As a proof of our concept, we first used MUD zymography to confirm BGL activity in bona fide BGL (i.e., N188 and D2-BGL because of their sensitivity to high levels of glucose on zymography). On MUD zymography, D263, N188 and D2-BGL showed strong BGL activity but C1.5L and TRX only weak BGL activity ( Fig. 2c, d). A similar situation was observed with MUC as the substrate (Fig. 2e, f ). We also used a liquid assay (i.e., pNPG assay) to confirm that all five cellulase preparations had BGL activity, which was greatly inhibited with 2% glucose (Fig. 3a). As described above, activity observed from N188 and D2-BGL in MUD zymogram was sharply decreased in the presence of 2% glucose, so they were indeed BGLs and act as positive controls of BGL activity in our zymography. We have also showed that cellulase preparations C1,5L and TRX with expected high EXO activity but low BGL activity could serve as negative controls in our MUC zymography, since both of them showed similar signal intensities with and without glucose in MUC zymogram ( Fig. 2e, f ). Consistent with our zymography results, MUC liquid assay revealed that the MUCase activity was mostly derived from BGL activity in N188 and D2-BGL, which was severely reduced with the addition of glucose, while only a slight decrease was observed in C1.5L, D263 and TRX (Fig. 3b).
In addition to containing BGL activity, D263 cellulase preparation contained EXO activity, revealed on the MUC zymogram as bands that were insensitive to 2% glucose (Fig. 2e, f ). Therefore, BGL activity was highly repressed by the high concentration of glucose on both MUD and MUC zymograms. We applied this concept to reveal EXO activity with MUC zymography. For BGLs such as N188 and D2-BGL, signals in the MUD zymogram were also observed in the MUC zymogram because of the ability of BGL to also hydrolyze MUC (Fig. 2e). However, in the MUC zymogram, other positive signals were observed for cellulase preparations D263, C1.5L and TRX because of the presence of EXOs.
In the presence of 2% glucose, the MUC-positive signals in D2-BGL, N188 and the high-molecular-weight band of D263 were greatly diminished, whereas those of C1.5L, TRX and low-molecular-weight band of D263  Fig. 2 Addition of glucose helps specify the presence of exoglucanases on 4-methylumbelliferyl β-d-cellobioside (MUC) zymography. a Coomassie Blue staining of native PAGE for D263 crude enzyme. Novozyme 188 (N188) and Geobacillus sp. cellulase A (GsCelA) were negative and positive controls of endoglucanase activity, respectively. b D263 exhibits high endoglucanase activity. CMC zymography was used for detecting endoglucanase activity of the enzyme preparations. c, d D263 shows β-glucosidase activity. 4-Methylumbelliferyl-β-d-glucopyranoside (MUD) zymogram was used for detecting β-glucosidase activity. The addition of 2% glucose successfully represses β-glucosidase activity on the MUD zymography. Commercially available N188, Cellulast 1.5L (C1.5L), Chaetomella raphigera D2 β-glucosidase (D2-BGL) and laboratory-produced Trichoderma cellulase preparation (TRX) were used as positive and negative controls of glucose suppression of BGL activity. e, f D263 shows clear exoglucanase activity. MUC zymogram with 2% glucose addition was used for detecting exoglucanase activity of different cellulase preparations. The modified MUC zymography method can detect exoglucanase activity with little interference from β-glucosidase activity. The protein amount used in a-e was C1.5L, 2600 ng; N188, 150 ng; D263, 340 ng; TRX, 1100 ng; D2-BGL, 100 ng (for MUD assay) and 1000 ng (for MUC assay). All experiments were conducted on 8% native gel with at least three biological replicates ▸ remained largely unchanged (Fig. 2e, f ). The remaining MUC-positive signals in the presence of 2% glucose strongly suggest that EXO activity in the fungal cellulase extracts is not sensitive to glucose inhibition (Fig. 2f ). The results also demonstrate that most of the MUCase activity from D263, TRX and C1.5L resulted from EXOs. Recent reports suggested that some of the cellulase activity can be readily detected on denaturing gel zymograms [29]. Many hydrolytic enzymes can be denatured by the ionic detergent SDS, but renatured later if SDS is removed, which results in a zymographic technique allowing for the separation and detection of specific enzymes by SDS-PAGE with better resolution.
Therefore, we also examined the MUC zymogram by using SDS-PAGE to give a more precise estimation of protein molecular weight. The EXO activity in the MUC zymogram was only slightly affected by SDS (Fig. 4a-c). We still observed the glucose repression of BGL activity, and D2-BGL activity could not be detected on SDS-PAGE zymograms (Fig. 4a, b). This observation is consistent with our previous findings that Pichia pastoris-produced recombinant D2-BGL activity was sensitive to SDS and could not be renatured even after SDS was removed [30]. This finding is likely to due to unusual glycosylation of this enzyme in Pichia [30].
In this work, we have devised a simple, yet specific, zymographic detection of EXO secreted from the fungus D. caldariorum D263. With our improved MUC D263 genome encodes lignocellulolytic enzymes distinct from commercially used fungi From our zymography and liquid assays, D263 was able to degrade both cellulose and hemicellulose (Table 1, Figs. 1, 2). To identify highly bioactive cellulases that have high potential for industrial application, it is necessary to further purify the enzymes of interest from the D263-secreted cellulolytic enzymes. However, the purification and identification process can be laborious without knowledge of the D263 genome. To acquire sufficient genomic information, we have taken advantage of DNA sequencing technology to assemble the draft genome of D263. We obtained 195 contigs from our 454 sequencing, which could be further assembled into 110 scaffolds. The assembled genome size of D263 is 37.8 Mbp, with 10,477 genes annotated and a gene density of 277.4 genes/Mbp ( Table 2). The genome size and gene density resemble those of recently published endophytic fungal genomes. We have also calculated the GC content of D263 and previously reported endophyte genomes from Mycocosm database [31]. The GC content between D263 and Daldinia EC12 were quite similar and were slightly higher than the three strains of Hypoxylon spp. published in previous literature [17]. To further investigate the genes in D263 genome with potential for cellulose degradation, we have conducted Gene Ontology (GO) analysis to annotate the potential functions of D263 genes. The BLAST2GO tool enables us to decipher the functions of the potential genes based on their sequence similarities to the closest fungal species [32]. Our analysis revealed that around 10% of genes in the D263 genome (1046 genes) may encode hydrolases (Additional file 3: Table S2). Most of the lignocellulolytic enzymes are hydrolases, including glycoside hydrolases (GHs) that are essential for the degradation of various lignocellulosic residues. Because we observed that D263 produces a wide range of EG, EXO, BGL and xylanase activity, we compared the GO term distribution for D263 and the commercially used fungi T. reesei and A. niger (Fig. 5). All three fungi were enriched in glucosidases and mannosidases. The former is responsible for catabolizing complex carbohydrates into simple sugars, while the latter was essential for the formation of fungal cell wall. Interestingly, the D263 genome also showed specific enrichment of xylanases, which corresponds to the high xylanase activity observed in D263 secretion (Table 1). It is known that xylan is the most abundant non-cellulosic polysaccharides in plant cell wall, so the enrichment of xylanase genes and activity in D263 may provide efficient degradation of hemicellulose into five-carbon sugar xylose. Of note, both T. reesei and A. niger showed enrichment of sialidases, which were lacking in the D263 genome (Fig. 5). Sialidases are a group of GHs required for pathogen virulence. In microorganisms, sialidases function to hydrolyze sialic acid for nutritional purposes and are involved in adhesion and host cell infection [33]. Also, an Aspergillus fumigatus sialidase (Kdnase) may be responsible for its virulence and cell wall integrity [34]. However, sialic acid is barely present in plants [35,36], thus sialidases may not be needed for cellulose decomposition. To further identify the D263-specific cellulolytic enzymes, we have also compared the amino sequences of potential D263 O-glycosyl hydrolases in the Carbohydrate-Active enZYme (CAZy) database. GHs are widespread enzyme groups that hydrolyze glycosidic bonds rich in cellulose and hemicellulose [16,[37][38][39]. We found that the D263 genome encodes at least 103 GH proteins. To further validate the distinct features of D263 cellulolytic enzymes, we have compared the GH distribution of D263 to the commercial cellulase producing fungi T. reesei and A. niger (Fig. 6). All three fungi encode a large number of GH3, GH5 and GH16 enzymes. GH3 family proteins have a wide range of functions in cellulosic biomass degradation, including BGL and β-d-xylopyranosidase activity [40,41]. Although T. reesei contained more GH3 proteins compared to A. niger and D263, its BGL yield is the lowest among the three fungal species [27]. The results were supported by previous literature that T. reesei strain RUT-C30 cannot efficiently secrete BGLs to its environment, so low BGL activity was detected in its secreted enzymes [42]. GH5, on the other hand, is one of the important protein families with high EG and xylanase activity [43]. The GH16 family contains various β-galactanases and xyloglucan transglucosylases/hydrolases (XTHs) that may be required for xylan decomposition [16,44,45]. The enrichment of GH5 and GH16 family proteins is consistent with their high EG and xylanase activity in our observation (Fig. 1, Table 1). Interestingly, there are GH proteins that were specifically enriched in each fungal species.
The most presented members of GH family proteins are GH18 in T. reesei and GH28 in A. niger (Fig. 6). GH18 family proteins are enriched in chitinases, which were able to degrade the fungal cell wall and are essential for the mycoparasitic activity in Trichoderma spp. fungi [46]. A. niger produces GH28 polygalacturonases that belong to pectinases that were required for its virulence against plant tissues and were widely used in the food industry [47]. These two families were not enriched in D263, which is consistent with previous study that no endophytes tested showed pectinase activity and endophytes were usually considered as latent pathogens [48].
The GH61 family proteins, on the other hand, are most abundant in the D263 genome, but only a few existed in Aspergillus and Trichoderma genomes. GH61 enzymes are classified as auxiliary activity family 9 (AA9) lytic polysaccharide monooxygenases (LPMOs) and are involved in the cleavage of (1,4)-linked glycosidic bonds on the surface of crystalline plant polysaccharides, hence increasing substrate accessibility of other cellulases [49]. Also, co-expression of a LPMO in a commercial strain (i.e., T. reesei) enhanced the conversion of plant biomass [50]. A recent study showed that LPMOs may act synergistically with xylobiohydrolase to facilitate the degradation of xylans [51]. This specific enrichment of GH61 proteins in D263 may also facilitate the activity of its EG and xylanase activity, resulting in high activity observed in D263 secretion (Table 1). Prompted by this notion, we have assayed the activity of GH61 proteins in D263 cellulase preparation. In recent study, it was suggested that LPMO activity may be visualized on CMC zymography with the use of a reducing agent such as ascorbic acid [52]. However, our results indicated that no additional band was observed in the presence of 10 mM ascorbic acid on both native and SDS-PAGE gels (Additional file 2: Fig. S2A-D). Thus, the probable LPMO activity in the gel could be masked by the high EG activity of D263.
Our result also suggests that the activity of LPMOs may be significantly lower than that of EGs, and both types of enzymes showed similar molecular weight on zymography. Taken together, our genomic analyses suggest that D263 may encode a group of plant biomass-degrading enzymes that are distinct from A. niger and T. reesei, which makes this organism a novel source of cellulosic enzymes with potential of producing novel AA9 LPMOs useful for the biofuel industry.

Validation of cellulolytic enzymes secreted by D263
In addition to the discovery of new LPMO genes, our GO analyses revealed that the D263 genome may encode more than 50 cellulolytic enzymes, including at least 12 EGs, 4 EXOs, 9 BGLs, 19 LPMOs and 7 xylanases (Additional file 3: Table S2). The D263 EXOs belong to GH6 and GH7 families. Many members of these two families are known as secretory cellulose 1,4-β-cellobiosidases [53]. Although D263 has shown lower capability of Avicel degradation than Trichoderma and Aspergillus (Table 1), it indeed showed strong MUCase activity (Figs. 2f, 3b). Amino acid sequences of protein bands identified from zymograms of enzymatic activity can be determined by mass spectrometry, then used for annotation of genomic sequences encoding this enzyme. To identify the active cellulases secreted by D263, we partially purified the crude enzyme of D263 by using anion-exchange chromatography and detected MUC activities in each elution fraction. The fractions 15 to 22 contained proteins enriched with MUCase (Fig. 7a). We demonstrated high BGL and EXO activities on MUC zymograms corresponding to seven protein bands (Cel1-Cel7) on Coomassie Blue-stained gel (Fig. 7b-d). Our proteomic analyses indicated that the major proteins of Cel1 band were likely GH3 BGLs (genes 006152 and 001615), consistent with the observation that Cel1 showed obvious glucose inhibition during MUC zymography analysis (Fig. 7c, d, Table 3 and Additional file 4: Table S3). The most enriched protein in Cel2 was 1,4-β-d-xylosidase (gene 005328), consistent with our observation that D263 was enriched in xylanases (Table1, Fig. 5). However, although some of the peptides belong to Cel1 BGLs can also be identified in Cel2 (Table 2, Fig. 7 and Additional file 4: Table S3). The shift of molecular weight observed in the putative 1-4-β-d-xylosidase 005328 may be resulted from post-translational modification such as glycosylation. The enrichment of putative enzyme 005328 also supported the discovery that D263 showed enrichment of xylanases (Fig. 5). Cel3 to Cel5, on the other hand, correspond to EXOs of 42 to 49 kDa (genes 006394, 007734 and 010048), which clearly reflects to the MUCase positive signals in the MUC zymography with 2% glucose (Fig. 7c, d, Table 3 and Additional file 4: Table S3). Interestingly, the enriched fractions 19 and 22 also showed high EG activity in these molecular-weight regions, with one EG identified as the best matched protein of Cel6 (gene 002528) (Fig. 7e, f, Table 3, Additional file 4: Table S3 and Additional file 5: Table S4). Surprisingly, Cel7 was identified as only a partial fragment of endo-1,4-β-xylanase (gene 004812). It is known that the hydrolysis of xylan requires the orchestrated actions of endoxylanases and the β-xylosidases [54]. The identification of both 1,4-β-d-xylosidase (gene 005328) and endo-1,4-β-xylanase (gene 004812) suggests that D263 is capable of complete hydrolysis of xylan. Further investigation of D263 xylanases may be beneficial for discovering new and highly efficient xylan decomposition enzymes that can be used in industry. Interestingly, the identified cellulolytic enzyme may be observed from multiple major bands of both native and SDS-PAGE gels after our proteomic analysis, with their molecular weight slightly shifted from the molecular weight prediction based on amino acid sequences (Additional file 4: Table S3 and Additional file 5: Table S4). The differences may have resulted from post-translational modification of these cellulolytic enzymes, such as N-glycosylation and O-glycosylation. This phenomenon has been observed in many cases. Previous studies have shown that the C. raphigera expressed native D2-BGL has two glycosylation variants and the more extensively O-glycosylated form exhibited increased activity toward its natural substrate cellobiose [30]. The recombinant A. aculeatus AaBGL1, a 91-kDa β-glucosidase with 16 potential N-glycosidation sites, showed a smeared band between 135 and 140 kDa during SDS-PAGE analysis [55]. It was also reported that N-glycosylation is critical for the folding of Aspergillus terreus BGL when heterologously expressed in Pichia pastoris and Trichoderma reesei [56]. Future studies on the glycosylation variants of D263 cellulolytic enzymes will aid us in identification of critical glycosylation sites that are essential for cellulolytic enzymatic activity. From our identification of active EGs in the Endo1-Endo4 protein bands, we have found that both Endo1 and Endo3 were enriched in a potential EXO, gene 010048, which was also enriched in Cel4 and Cel5 in our SDS-PAGE zymography. Endo2 was enriched in gene 006394, which was identified as a potential EXO that is abundant in Cel3 (Fig. 7b, f, Table 3). Endo 4 contains a GH7 EG (gene 002528) that was also identified in Cel6 of our SDS-PAGE zymography. The results suggested that these EXOs and EGs were abundant in this molecular weight region. Since the mobility on a native gel is determined by both size and charge of a protein, it is plausible that a large protein with positive charge could migrate to a similar position as a smaller protein with less positive charge, which may also explain the slight size difference between proteomic analyses and native PAGE results. Interestingly, the proteins identified in Endo4 also include a GH30 endo-β-1-6 glucanase (gene 009220) and a potential GH61 LPMO (gene 001982). LPMOs exist in a relatively low amount as compared with most cellulolytic enzymes identified in the protein fractions and are not easily visible on zymography (Table 3, Additional file 2: Fig. S2, Additional file 4: Table S3 and Additional file 5: Table S4). Our proteomic analysis clearly indicates that the endophytic fungus D263 can produce and secrete a large variety of bioactive cellulases (Table 3). By conducting RNA sequencing analysis using our gene annotation for mapping, we will be able to identify the optimal conditions for induction of D263 cellulolytic enzymes, which will facilitate cellulose degradation and second-generation biofuel production.

Conclusions
The combination of zymography and mass spectrometry analyses provides a powerful tool for genome annotation and discovery of new cellulolytic enzymes usable in agricultural waste disposal. Here, we have discovered that the endophytic fungus D. caldariorum D263 is able to produce a wide range of cellulolytic enzymes that efficiently hydrolyzed acid-pretreated sugarcane bagasse. Our zymography revealed the presence of EXOs, EGs and BGLs in the secreted cellulolytic enzymes of D263. Addition of 2% glucose was essential for specific identification of EXOs without the interference of BGL activity in MUC zymogram. We have further identified the specific cellulolytic enzymes in D263 by integrative genomic and proteomic analyses. Our results also suggest that the AA9 LPMOs are more abundant in D. caldariorum D263 than that of commercially used fungi such as T. reesei and A. niger. Taken together, our research pipeline successfully proved that D263 has high potential for expressing and secreting a wide range of cellulolytic enzymes.

Purification of D263-secreted cellulolytic enzymes
The supernatant of D263 culture was filtered through a 0.45-μm filter, and the buffer exchange was performed with the binding buffer (50 mM Tris, pH 8) by using Vivaspin 20 (30 kDa MWCO, General Electric, USA). The anion-exchange chromatography involved using the ÄKTA-FPLC System (General Electric, USA) and a 5-mL HiTrap Q column (General Electric, USA). Elution was conducted with a linear gradient from 0 to 100% of elution buffer (20 mM sodium phosphate, 1 M NaCl, pH 8).
The eluted fractions with cellulase activity were separated on SDS-PAGE gels for evaluation of MUCase activity and identification of major cellulolytic enzymes by mass spectrometry.  [52]. In brief, 8% of native or SDS-PAGE gels containing 0.2% of CMC were soaked in 50 mM sodium acetate buffer, pH 5, with or without 10 mM ascorbic acid. The soaked gels were incubated at 50 °C for 16 h and the enzyme activity was revealed by staining with 0.2% Congo Red (Sigma-Aldrich, USA).

Cellulase activity assays in test tubes
For BGL activity assays, 100 μL of enzyme solution was mixed with 100 μL of 4 mM pNPG in 50 mM sodium acetate buffer, pH 5, with or without 2% (w/v) glucose. The reaction was performed at 55 °C for 5 min under agitation and was stopped by adding 600 μL of 1 M Na 2 CO 3 . The product concentration was determined by measuring the OD 405 of the final reaction solution with a BioTek GEN5 spectrophotometer. Serially diluted 4-nitrophenol solutions were used to establish the standard curve. For EXO activity assays, 100 μL of 1 mM MUC with or without 2% (w/v) glucose was mixed with 100 μL enzyme solution. The fluorescence was detected with a BioTek GEN5 spectrophotometer with excitation and emission wavelengths set at 360 and 450 nm, respectively. Serially diluted 4-methylumbelliferone was used for establishing the standard curve.

Genomic DNA isolation
Daldinia caldariorum D263 was cultured on PDA plate for 3 days, and the mycelium discs were transferred to MR medium with 0.5% α-cellulose and incubated for 5 days. The fungal mycelia were washed with distilled water and then ground into fine powder in liquid nitrogen. To extract genomic DNA, 12 mL urea extraction buffer (7 M urea, 0.3 M NaCl, 50 mM Tris-Cl, pH 8, 20 mM EDTA, 1% sarcosine, pH 8) was added to the ground mycelia and mixed by vortexing. A 12-mL amount of phenol:chloroform:isoamyl alcohol (25:24:1) was then added and mixed for 15 min at room temperature. The sample was then centrifuged for 10 min at 8000 rpm, 4 °C. The supernatant was transferred to a clean tube and mixed with 2 mL of 4.4 M ammonium acetate and 13 mL isopropanol. After centrifugation for 10 min at 8000 rpm, 4 °C, the supernatant was discarded and the pellet was washed twice with 1 mL of 100% EtOH. The pellet was air-dried and re-suspended in 200 μL ddH 2 O and heated for 1 h at 65 °C. After heating, 20 μL RNase A (10 mg/mL) was added and the sample was incubated at 37 °C for 30 min. After incubation, the sample volume was adjusted to 500 μL with ddH 2 O.

Genome sequencing and bioinformatic/proteomic analyses
About 400 μg isolated genomic DNA was used for library preparation, and 454 sequencing (Roche, Switzerland) was used to acquire sequence reads with average length of 373 bp. The draft genome of D263 was assembled by using Newbler V2.3, with mapping statistics shown in Additional file 1: Table S1. Blast2Go was used to identify enriched Gene Ontology (GO) terms in D263 genes [58]. GOBU was used to characterize the GO terms in the D263 genome [59]. KofamScan was used to predict the KEGG pathways of D263 genes [60]. The mapping statistics are shown in Additional file 6: Table S5. This Whole Genome Shotgun project has been deposited at DDBJ/ ENA/GenBank (Accession no. JAAOZU000000000). The GO terms and CAZy enzymes in T. reesei V2.0 and A. niger ATCC1015 were searched in the Joint Genome Institute (JGI) MycoCosm database [31,[61][62][63]. Hierarchical analysis was conducted based on Pearson correlation. For proteome analyses, protein bands showing cellulolytic enzyme activities were extracted from SDS-PAGE gels. The gel slices were trypsin-digested and underwent mass spectrometry. The amino acid sequences of annotated D263 genes underwent a Mascot search (Matrix Science, USA). Protein abundance was calculated by using ProteomeDiscoverer v2.4 (Thermo-Fisher, USA).