Quantitative Secretomic Analysis of Trichoderma reesei Strains Reveals Enzymatic Composition for Lignocellulosic Biomass Degradation*

Trichoderma reesei is a mesophilic, filamentous fungus, and it is a major industrial source of cellulases, but its lignocellulolytic protein expressions on lignocellulosic biomass are poorly explored at present. The extracellular proteins secreted by T. reesei QM6a wild-type and hypercellulolytic mutant Rut C30 grown on natural lignocellulosic biomasses were explored using a quantitative proteomic approach with 8-plex high throughput isobaric tags for relative and absolute quantification (iTRAQ) and analyzed by liquid chromatography tandem mass spectrometry. We quantified 230 extracellular proteins, including cellulases, hemicellulases, lignin-degrading enzymes, proteases, protein-translocating transporter, and hypothetical proteins. Quantitative iTRAQ results suggested that the expressions and regulations of these lignocellulolytic proteins in the secretome of T. reesei wild-type and mutant Rut C30 were dependent on both nature and complexity of different lignocellulosic carbon sources. Therefore, we discuss here the essential lignocellulolytic proteins for designing an enzyme mixture for optimal lignocellulosic biomass hydrolysis.

Lignocellulose, a major component of plant biomass produced by photosynthesis, is abundant, renewable, and sustainable; hence it is no surprise that lignocellulosic bioenergy has become an intense subject of investigation now. Compared with current fossil fuel, lignocellulosic biofuel offers better advantages, such as its renewable nature, the fact that it is environmentally friendly, and its potential to mitigate global warming and prevent fuel shortage; in addition it has no effect on the food chain. Lignocellulosic bioenergy is mainly obtained by hydrolysis of forest wastes, agricultural residues, hardwood, softwood, grasses, etc., and subsequent fermentation of resultant hydrolysates into either biofuel such as bioethanol or biohydrogen. Several researchers (1-3) have established pretreatment methods for lignocellulosic biomasses, and others have emphasized the advantages of enzymatic hydrolysis of biomasses into monomeric sugars for enhanced biofuel production (4). Despite establishment of ethanol production methods through fermentation of corn, starch, and sugarcane juice, production methods of lignocellulosic ethanol have yet to be optimized because of the lack of an optimal biomass hydrolyzing enzyme mixture. The bacterial strains belonging to Clostridium, Cellulomonas, Bacillus, Thermomonospora, and Thermobifida fusca and fungal species such as Aspergillus niger, Phanerochaete chrysosporium, and Trichoderma reesei have been known to produce lignocellulolytic enzymes (5)(6)(7). The filamentous fungus Trichoderma reesei is a well known efficient producer of cellulases and as such was developed as an industrial workhorse.
The cellulases and hemicellulases produced by T. reesei possess useful applications for various industries, such as the pulp and paper industries (8), textile industries (9), food and feed industries, and biofuel production. To date, optimized cultivation conditions, operational parameters, genetic mutation, and manipulation have been developed to enhance enzyme production potential (10). Morphological features, such as pellets, mycelial aggregates, and freely dispersed mycelia have been correlated to enzyme productions (11,12). Operational parameters, such as inoculum size (13,14) and agitation speed (15) were investigated as well. The pH-dependent secretory proteins of T. reesei, including QM6a, QM9414, Rut C30, and QM9414MG5 were profiled in cellulosic substrate (16). To enhance enzyme production potential, enzyme-inducing substrates such as cellobiose (17), lactose (18), sophorose (19), and cellulose (20,21) were tested. Despite these efforts, required breakthrough in an economical enzyme production to further its application in lignocellulosic biofuel have not yet been achieved.
Based on interpretations of experimental data from statistic models, several researchers have focused on the reconstitution of in vitro optimized enzyme cocktails for biomass hydrolysis (22,23). After achieving low cellulose hydrolysis yield from the reconstituted mixture consisting of endoglucanase, exoglucanase, and glucosidase, considerable attention was shifted to the understanding of basic biomass hydrolytic mechanisms and the discovery of essential proteins and roles of low abundant enzymes. To explore the mechanistic role of low abundant proteins, comprehensive accurate profiling of secreted proteins and their quantitative expressions with highly sensitive and advanced proteomic technology is emphasized.
The genome sequence of T. reesei has shed light on the diversity of hydrolytic enzymes secreted by this fungus (24); however, expression of these proteins and their expression level in response to various carbon sources remains to be elucidated. Although cellulase and xylanase production by T. reesei has been reported (25)(26)(27), literature on lignocellulose biomass hydrolysis by extracellular proteins (28) using proteomic technology have been rarely documented. The comprehensive secretome analysis by state of the art mass spectrometry proteomic techniques allows the identification and quantification of potentially essential lignocellulolytic enzymes secreted by T. reesei and/or its hypercellulolytic mutant (Rut C30) obtained from three mutagenesis steps: UV mutagenesis, N-nitrosoguanidine mutagenesis, and then repeated UV mutagenesis. In this study, different carbon sources, such as cellulose and lignocellulosic biomass derived from agricultural and forest wastes, were used, and quantitative expressions of lignocellulose-hydrolyzing proteins by T. reesei and its mutant were profiled using isobaric tags for relative and absolute quantitation (iTRAQ) 1 by nano-LC-MS/MS. The expression level of the secreted proteins by T. reesei QM6a and Rut C30 and their mechanistic role and the essential proteins required for in vitro reconstitution of enzyme mixture for lignocellulosic biomass were also discussed.

Microorganism Cultivation Conditions and Secretome Extraction-
This study used wild-type T. reesei QM6a (ATCC 13631) and mutant T. reesei Rut C30 (ATCC 56765), procured from the American Type Culture Collection and maintained following supplier's protocol. The strains were grown in potato dextrose broth, cell biomass was collected by centrifugation at 7000 ϫ g at 4°C (Beckman Coulter), then washed with sterilized MilliQ water, and later inoculated into a flask containing medium with composition 3.1 g/liter (NH 4 ) 2 SO 4 , 1.5 g/liter carboxymethyl cellulose, 1.5 g/liter NaCl, 1.2 g/liter KH 2 PO 4 , and micronutrients 0.300 g/liter MgSO 4 ⅐7H 2 O, 0.005 g/liter FeSO 4 ⅐7H 2 O, 0.075 g/liter MnSO 4 ⅐H 2 O, 0.03 g/liter CaCl 2 ⅐2H 2 O 0.002 g/liter COCl 2 , and 0.1 g/liter thiamin. Thereafter, the cell mass was collected by centrifugation at 7000 ϫ g and 4°C (Beckman Coulter) and used for inoculation of the test flask that contained the above mentioned medium with fibrous cellulose (Sigma; catalogue number C6663), corn stover, and saw dust (10 g/liter) as major carbon sources. The above basic medium with T. reesei QM6a served as control. The experimental design is outlined in Fig. 1. Each experiment contained two flasks for each substrate. To minimize biological variations, secretome from two flasks were pooled together and labeled as set A. Similarly, set B was prepared, and both sets (sets A and B) were processed separately. The test flasks were incubated at 30°C and 80 rpm for 120 h, and secreted proteins were collected by centrifugation at 7000 ϫ g, 4°C. The collected secretome was further clarified by filtration through a 0.2-m filter. Then it was concentrated by a freeze-drying technique using lyophilizer, and protein content was determined by the Bradford method.
Protein Digestion, Peptide Extraction, and Mass Spectrometric Analysis-The protein digestion, peptide extraction, and LC-MS/MS analysis were performed as described previously (5,29). In brief, protein from each test sample (150 g) was separated on SDS-PAGE, and the protein bands were visualized by staining with Coomassie Brilliant Blue. Each sample lane was sliced separately and washed with 75% acetonitrile containing TEAB (25 mM). The gel pieces were subjected to destaining with TEAB (25 mM) alone and TEAB with 50% acetonitrile. After complete destaining, the gel pieces were reduced with Tris 2-carboxyethyl phosphine hydrochloride (5 mM) and then alkylated with methyl methanethiosulfonate (10 mM). The protein digestion was performed by subjecting acetonitrile dehydrated gel pieces to sequencing grade modified trypsin (Promega, Madison, WI) digestion at 37°C for 20 h. The peptides were extracted using 50% ACN, 5% acetic acid and concentrated using concentrator (Eppendrof AG, Hamburg, Germany) for further iTRAQ reagent labeling.
iTRAQ Labeling and LC-MS/MS Analysis-The iTRAQ labeling of peptides obtained from the test samples with different carbon sources were performed using an 8-plex iTRAQ reagent Multiplex kit (Applied Biosystems, Foster City, CA) according to the manufacturer's protocol. The peptides from each substrate condition of set A were individually labeled with respective isobaric tags, incubated for 2 h, quenched with water, and vacuum-centrifuged to dryness. Similarly, samples of set B were labeled, and individual set A or B was analyzed. The labeling was as follow: 113, T. reesei Rut C30 control; 114, T. reesei Rut C30 cellulose; 115, T. reesei Rut C30 saw dust; 116, T. reesei Rut C30 corn stover; 117, T. reesei QM6a control; 119, T. reesei QM6a saw dust; and 121, T. reesei QM6a corn stover. The iTRAQ labeled peptides were reconstituted in buffer A (10 mM ammonium acetate, 85% acetonitrile, 0.1% formic acid) and fractionated by a ERLIC column (200 ϫ 4.6 mm, 5-m particle size, 200 Å pore size) with an HPLC system (Shimadzu) at a flow rate of 1.0 ml/min using an earlier optimized protocol (30). The HPLC solvents were buffer A (10 mM ammonium acetate, 85% acetonitrile, 0.1% acetic acid) and buffer B (30% acetonitrile, 0.1% formic acid). The 60-min HPLC gradient consisted of 0 -28% buffer B in 40 min, 28 -100% buffer B in 10 min, and 100% buffer B for 10 min. The HPLC chromatograms were recorded at 280 nm, and fractions were collected using automated fraction collector, concentrated using vacuum centrifuge, and reconstituted in 0.1% formic acid for LC-MS/MS analysis.
The LC-MS/MS analysis of fractionated labeled sample was performed with a QStar Elite mass spectrometer (Applied Biosystems/ MDS Sciex) coupled with online microflow HPLC system (Shimadzu). The resultant labeled peptides were separated on a home-packed nanobored C18 column with a picofrit nanospray tip (75-m inner diameter ϫ 15 cm, 5-m particles) (New Objectives, Wubrun, MA) coupled to the LC-MS/MS system at a constant flow rate of 0.3 l/min. Each sample fraction of one set was injected twice at two equal volumes and independently analyzed by the LC-MS/MS using 60-min gradient. The mass spectrometer was set in positive ion mode using Analystா QS 2.0 software (Applied Biosystems), data were acquired with a selected mass range of 300 -1600 m/z, and peptides with ϩ2 to ϩ4 charges were selected for MS/MS. The peptides above a five-count threshold were selected for MS/MS, and each selected target ion was dynamically excluded for 30 s with Ϯ30 mDa mass tolerance. Smart information-dependent acquisition was activated with automatic collision energy and automatic MS/MS accumulation. The fragment intensity multiplier was set to 20, and the maximum accumulation time was 2 s. The peak areas of the iTRAQ reporter ions reflect the relative abundance of the proteins in the samples.
Mass Spectrometric Data Search and Analysis-The data acquiring was carried out using Analystா QS 2.0 software (Applied Biosystems/ MDS Sciex). The peak list generation, protein identification, and peptide quantification were performed using ProteinPilot software 3.0 (revision number 114732; Applied Biosystems, Foster City, CA). The database search was performed against the genome project for the organism (http://genome.jgi-psf.org/Trire2/Trire2.home.html) database version 2.1 including 9129 predicted gene models. The database search for set A, set B, and combined search of set A and B were performed, and the mean values with standard deviations of the proteins that were common (set A and B) are reported in this study. The peptide identification was performed with Paragon algorithm (3.0.0.0, 113442) in ProteinPilot software, which was further processed with Pro Group algorithm where isoform-specific quantification was adopted to trace the differences between expressions of various isoforms. The search was done thoroughly where all cleavage variants were considered. Our defined parameters were: (i) sample type, iTRAQ 8-plex (peptide-labeled); (ii) cysteine alkylation, methyl methanethiosulfonate; (iii) digestion, trypsin; (iv) instrument, QSTAR Elite ESI; (v) special factors, none; (vi) species, none; (vii) specify processing, quantitate; (viii) identification focus, biological modifications and amino acid substitutions using DB, a concatenated target and decoy database of amino acid sequences of T. reesei proteins obtained from the genome project for the organism; and (x) search effort, thorough. The default precursors and fragment mass tolerances for QSTAR ESI MS instrument were adopted by the software. The peptide for iTRAQ quantification was automatically selected by Pro Group algorithm (at least one peptide with 99% confidence) to calculate the reporter peak area, p value, etc. To minimize false positive results, a strict cutoff for protein identification was applied with the unused ProteinScore Ն2, which corresponds to a confidence limit of 99%, and at least two peptides with 95% confidence were considered for protein identification. The resulting data set was auto bias-corrected. The false discovery rate was calculated using inhouse build program as 2 ϫ M d /(M d ϩ M t ), where M d represents the number of decoy matches, and M t is the number of target matches. The existence of signal peptide sequences was checked using the signal peptide prediction program SignalP version 3.0 (http://www. cbs.dtu.dk/services/SignalP).
Zymogram Analysis-The activity of the cellulases and glycoside hydrolases was determined by zymogram analysis on PAGE using 10% separating gel containing 1% Carboxymethyl cellulose (CMC). The gel was run in Mini Proteanா (Bio-Rad) gel running apparatus at 70 V for 120 -180 min. The gel was washed twice with Triton X-100. The renaturation/reactivation of the enzyme was carried out by overnight incubation of gel at 4°C in 2.5% Triton X-100. After overnight incubation, the gel was incubated at 42°C for 60 min in acetate buffer, pH 5.5. The gel was then stained with 0.1% Congo Red stain for 30 min and destained in 1 M NaCl for 15 min. The enzyme reacts with substrate incorporated in gel and produces distinguished color. The gel was photographed, and the bands were excised, destained, digested with trypsin, and identified by MALDI-TOF.

Functional Classification of Extracellular Proteins-Our
analysis focused on the secretome of T. reesei QM6a and mutant Rut C30 in cellulose and natural lignocellulosic biomass culture conditions using iTRAQ labeling. In total, 636 proteins (combined search of set A and B) were identified across both biological replicates using ProteinPilot (513 in replicate A and 530 in replicate B). We narrowed down the protein list by omitting proteins identified with single peptides because of their lower confidence level. The iTRAQ ratio Յ0.8 and Ն1.2 was assigned for down-and up-regulation, respectively. The peptide summary data are supplied as supplemental information. After applying the cutoff of unused protein score Ն2, which corresponds to a 99% confidence level and a false discovery rate of Յ1.0%, the proteins were sorted with N-terminal Sec-dependent secretion signal using SignalP 3.0 (31) and classified according to their biological functions into cellulases, hemicellulases, lignin-degrading proteins, peptidases, transport proteins, chitinase, phosphatases, transport, and hypothetical proteins (Table I and supplemental Tables  S1 and S2). The identification of intracellular proteins is suggestive of the presence of minor cell lysis, cell death, or secretion through unknown mechanisms. The data presented in Fig. 1C suggest significantly low (R 2 Ͼ 0.98) instrumental variation between duplicate analyses of one set of sample. Based on biological functions, 49.25% proteins were carbohydratases, which constitutes to 31.34% cellulose hydrolyzing and 17.91% hemicellulose degrading proteins (Fig. 1B). In addition to carbohydratases, peptidases and proteinases (22.01%) and lignin-degrading (13.43%) proteins were found to be expressed when T. reesei QM6a and Rut C30 were grown on cellulosic and natural lignocellulosic biomass as a major carbon sources (Fig. 1).
It is known that complete cellulose hydrolysis involves three major steps mediated by concerted activity by three different groups of enzymes, such as endoglucanases, cellobiohydrolases, and ␤-glucosidases. The first step in cellulose hydrolysis involves the splitting of cross-links between glucan chains by endoglucanases. In this study, we

Substrate-induced T. reesei Secretome
have quantified 18 different endoglucanases involved in the initial step of cellulose hydrolysis. Of the 18 endoglucanases, GH5 endoglucanase-2 (jgi͉Trire2͉120312͉), GH5 endoglucanase EG-1 (jgi͉Trire2͉122081͉), GH61 endoglucanase-4 (jgi͉Trire2͉73643͉), and GH30 endo-1,6-␤-D-glucanase (jgi͉Trire2͉3094͉) were significantly up-regulated in T. reesei QM6a and Rut C30 when corn stover and saw dust were used as major carbon sources (supplemental Fig. S1 and S2). The up-regulated GH74 endoglucanase (jgi͉Trire2͉49081͉) has iTRAQ ratios of 2.97, 1.87 (saw dust) and 4.32, 2.96 (corn stover) in T. reesei QM6a and Rut C30, respectively. T. reesei genome encodes eight endoglucanases with endo-1,4-␤-D-glucanase activities (one Cel7B, Cel12 and Cel45, two Cel5, and three Cel61) (24). Our data showed that all endoglucanases encoding genes were expressed except for Cel61 when lignocellulosic biomass was used as substrate for cultivation of T. reesei QM6a and Rut The degree of the gene or enzyme activation by metabolites or proteolytic proteins is poorly understood at present. Cellobiohydrolase catalyzes the second step in glucan chain hydrolysis through generation of cellobiose. The GH7 cellobiohydrolase I (jgi͉Trire2͉123989͉) with a high unused protein score (117.3) and iTRAQ ratios of 3.47 and 4.59 was highly expressed in T. reesei QM6a and Rut C30 (supplemental Fig. S3). The induced production of GH7 cellobiohydrolase I in Rut C30 could be due to genetic mutations that were described earlier (33,34). The strain T. reesei Rut C30 is known to have a deletion of ϳ2.5 kb in the cre1 gene that mediates glucose repression and frameshift mutation in the glucosidase II ␣ subunit gene (35)(36)(37). The T. reesei genome encodes only one GH7 cellobiohydrolase, whereas Aspergillus nidulans, Aspergillus fumigatus, A. oryzae, and Neurospora crassa encodes two GH7 cellobiohydrolase. However, T. reesei is a very efficient cellulose degrader as compared with these fungal strains, possibly because of their expression of unique GH proteins (see "Discussion"), includ- ing GH6 (jgi͉Trire2͉72567͉) and GH7 cellobiohydrolases (jgi͉Trire2͉123989͉), which have the ability to hydrolyze crystalline cellulose in the absence of endoglucanases. Furthermore, the comparative genome analysis has suggested that the number of genes encoding the proteins of the GH family, such as GH18, GH92, GH27, GH95, GH64, GH30, and GH89 are higher in T. reesei than other fungal strains such as A. nidulans, A. fumigatus, A. oryzae, Magnaporthe grisea, N. crassa, Fusarium graminearum, Candida albicans, Saccharomyces cerevisiae, Candida glabrata, Schizosaccharomyces pombe, Cryptococcus neoformans, and P. chrysosporium (24, 38 -44). Again, the necrophytic nature of T. reesei differentiates them from other fungal strains. The identified and quantified exo-glucanase GH6 exoglucanase 2 was also found to be up-regulated (supplemental Fig. S3).
The third step of cellulose hydrolysis, also an important regulatory step, involves the conversion of cellobiose into glucose by ␤-glucosidase. Thirteen glucosidases were quantified by iTRAQ and are presented in supplemental Fig. S4 and S5 along with their comparative expressions. Proteins, GH3 ␤-glucosidase Cel3b (jgi͉Trire2͉121735͉), GH17 glucan1,3-␤-glucosidase (jgi͉Trire2͉39942͉), and GH71 glucan endo-1,3-␣-glucosidase (jgi͉Trire2͉71532͉) were found to be up-regulated. The comparative iTRAQ ratios of ␤-glucosidase Cel3b suggest its significant induction (analysis of variance, p Ͻ 0.001) in T. reesei Rut C30, supporting its hypercellulolytic nature by eliminating the negative inhibition by cellobiose. In addition to exo-and endo-glucanases and -glucosidases, 12 cellulose hydrolyzing glycoside hydrolases, which belong to GH18 (jgi͉Trire2͉121355͉), GH24 (jgi͉Trire2͉109278͉), GH53 (jgi͉Trire2͉81541͉ and jgi͉Trire2͉ 59628͉), GH65 (jgi͉Trire2͉123456͉), GH76 (jgi͉Trire2͉74807͉ and jgi͉Trire2͉27395͉), and many others, were identified and quantified using iTRAQ. The majority of the cellulolytic proteins were highly expressed when T. reesei QM6a and Rut C30 were grown in saw dust and corn stover ( Table I). The hierarchical clustering of protein expression shown using gene patterning (45) had five different clusters of the protein that were highly expressed (Fig.  2). Our results demonstrated the different secretion patterns by wild-type QM6a and mutant Rut C30 (Figs. 2 and 3). The proteins clustered under cluster C1 were highly expressed in T. reesei QM6a in corn stover followed by QM6a saw dust, Rut C30 corn stover, and Rut C30 saw dust. The proteins of cluster C2 were highly expressed in QM6a saw dust followed by QM6a corn stover. The regulated proteins (red) suggest that T. reesei QM6a, being wild type, secreted a diverse range of hydrolytic enzymes, presumably to enable it to use on a wide range of carbon sources in the natural environment. The proteins of cluster C4 were highly expressed in Rut C30 saw dust. Thus, the iTRAQ-quantified and expressed protein clusters suggest protein expression as a function of strains, conditions, and complexity of substrates (Fig. 4).
Adhesion of cells to substrates has been considered as an important factor that enhances substrate hydrolysis caused by an increased enzyme concentration near the substrate, which eliminates competitors from liberated sugars already present in the culture medium. This study identified and quantified several substrate-binding proteins such as CBM1 cellulose-binding domain Cip2 (jgi͉Trire2͉123940͉), CBD3 carbohydrate-binding module family 13 (jgi͉Trire2͉ 105313͉), hydrophobin-2 (jgi͉Trire2͉119989͉), and hydrophobin-1 (jgi͉Trire2͉123967͉ and jgi͉Trire2͉73173͉) ( Table I).
iTRAQ Quantification of Lignin Depolymerizing Oxidoreductases Proteins-It is well known that the major components of P. chrysosporium lignin depolymerization system include multiple isozymes of lignin peroxidase that catalyze cleavage of C ␣ -C ␤ , whereas laccases are involved in aryl-C ␣ cleavage. T. reesei itself is not a major lignin-depolymerizing fungus. Surprisingly, laccase, a copper oxidase that catalyzes the oxidation of phenolics and aromatic amines, was expressed and quantified when cellulose, saw dust, and corn stover were used as major carbon sources (supplemental Table S1). Glyoxal oxidase (jgi͉Trire2͉124282͉), an important component of the lignolytic system, was detected with an unused protein score of 55.39 and 74 peptides. In addition to laccase and glyoxal oxidase, peroxidase/catalase (jgi͉Trire2͉70803͉), L-ascorbate peroxidase (jgi͉Trire2͉ 73523͉), copper/zinc superoxide dismutase (jgi͉Trire2͉ 65483͉), and several oxidoreductases were regulated when T. reesei and Rut C30 were grown on cellulose and natural lignocellulosic biomasses (supplemental Table S1). The temporal correlation of glyoxal oxidase, peroxidase, and oxidase substrate in cultures has been shown to exhibit close physiological connection (47,48). Thus, the identification, iTRAQ quantification, and regulation of laccase, proxidase, L-ascorbate peroxidase, oxidoreductase, glutathione reductase, and glyoxal oxidase highly support the lignin degradation potential of T. reesei QM6a and Rut C30. Supplemental Fig. S11 presents the hierarchical clustering of these proteins. Although laccase, lignin peroxidase, and manganese peroxidase were considered as lignin-degrading enzymes, according to Blanchette et al. (49), they are too big to penetrate the wood cell wall. Hence, microbes initially activate, easily diffusing oxidases and reactive radical generating enzymes that undergo series of complex cleavage reactions to participate in initial lignin depolymerization process (50). This agrees with our observation on the diverse range of secreted oxidoreductase proteins.
iTRAQ Quantification of Peptidases and Other Proteins-In addition to lignocellulolytic proteins, we have identified and quantified 54 peptidases, 6 chitinases, 8 phosphatase, transport proteins, and hypothetical proteins (supplemental Table S2), and their clustering pattern are shown in supplemental Fig. S12. Interestingly, the proteins that are grouped under cluster P1 and P2 were highly expressed in T. reesei QM6a when corn stover and saw dust were used as substrates, whereas proteins clustered under P3 and P5 were significantly up-regulated in T. reesei Rut C30. Although the exact role of peptidases in lignocellulose degradation has not yet been explored, its high expression and regulation suggest their possible involvement, and this will be further explained under "Discussion," as well as in the supplemental materials. Among the iTRAQ-quantified proteins, dipeptidyl peptidase (jgi͉Trire2͉21668͉) and amidase family proteins (jgi͉Trire2͉ 69863͉) were most abundant, with unused protein scores of 81.43 and 79.45, whereas aspartic proteinase (jgi͉Trire2͉ 81004͉), proteinase inhibitor I1 (jgi͉Trire2͉111915͉), ␣-N-acetylglucosaminidase (jgi͉Trire2͉58117͉), cyanophycinase (jgi͉ Trire2͉103039͉), aspartic endopeptidase (jgi͉Trire2͉119876͉), etc., were also significantly up-regulated. Furthermore, proteins belonging to chitinases, lipases, and phosphatases were expressed when lignocellulose was used as the major carbon source (supplemental materials).
Comparative Protein Expressions in T. reesei QM6a and Rut C30 -The mutagenized T. reesei Rut C30 was obtained by three mutagenesis steps, including UV mutagenesis (named the M7 strain); chemical mutagenesis (N-nitrosoguanidine) of M7, which generates the NG14 strain; and finally UV mutagenesis of NG14, which generates Rut C30 that possesses higher potential for secretion of extracellular proteins. When compared with the iTRAQ quantification data of these two strains, GH61 endoglucanase-cel61(jgi͉Trire2͉ 73643͉), GH7 cellobiohydrolase I (jgi͉Trire2͉123989͉), GH6 cellobiohydrolase II (jgi͉Trire2͉72567͉), GH3 ␤-glucosidase cel3b (jgi͉Trire2͉121735͉), glycosyltransferase family 20 (jgi͉Trire2͉77602͉), and other several GH proteins were significantly higher in T. reesei Rut C30 than in QM6a. This could be due to mutagenesis, because all mutagenesis treatments trigger DNA repair processes and enhance chromosomal recombination, resulting in chromosomal rearrangements. The higher production of unique enzymes (i.e., GH7 cellobiohydrolase I and GH6 cellobiohydrolase II) by Rut C30 that hydrolyzes cellulose in the absence of endoglucanases might be main reason for its hypercellulolytic nature. The iTRAQ protein quantitation data revealed that Rut C30 is more efficient in the production of extracellular protein, and this finding is also consistent with earlier reports (51). The genome sequence comparison revealed 223 single nucleotide variants, 15 small deletions or insertions, and 18 large deletions, leading to the loss of more than 100 kb of genomic DNA in T. reesei Rut C30 (34). The single nucleotide mutation and its impact on genes encoding secretory system remain unknown. However, the impact of some single mutations was mentioned earlier (33,34). The cellulose binding proteins, such as hydrophobin 1(jgi͉Trire2͉123967͉), CBD3 carbohydrate binding protein (jgi͉Trire2͉105313͉), and CBM1 cellulose-binding domain proteins (jgi͉Trire2͉123940͉) were significantly higher in Rut C30 than in QM6a. The high concentration of these proteins potentially enhances substrate binding, in turn increasing enzyme concentrations at the site, and thus increasing the rate of hydrolysis. The proteins involved in hemicellulose degradation such as endo-1,4-␤-xylanase 2 (jgi͉ Trire2͉549461), acetyl esterase (jgi͉Trire2͉103825͉), and pectin-degrading endopolygalacturonase (jgi͉Trire2͉103049͉) were also highly expressed in Rut C30.

DISCUSSION
As of now, little work has been done on the proteomic analysis of secretome by T. reesei using lignocellulosic biomasses as major carbon sources. However, most studies were focused on the production of enzyme, purification, and quantification using colorimetric methods (18,28). However, low levels of production, expression, and overlapping and/or novel enzymes and their activities cannot be detected by colorimetric enzyme activity assay techniques. Hence, we profiled a global view of the secretome of T. reesei in lignocellulosic medium using iTRAQ-based quantitative proteomics and compared substrate-induced protein expressions to understand the mechanism behind lignocellulosic biomass hydrolysis. We have quantified 65, 37, 34, and 54 cellulolytic, hemicellulolytic, and lignin-degrading proteins and peptidases, respectively, where all were significantly higher than the corresponding 52, 16, 14, and 16 proteins that were expressed in P. chrysosporium when cultivated in cellulosic or lignin culture medium (7). The significant expression of lignocellulolytic and proteolytic enzymes were noted in T. reesei secretome when compared with those reported by fungus P. sapidus (52) and P. chrysosporium (41). Using two-dimensional gel and MALDI-TOF-MS analysis of A. oryzae (wheat bran as a major carbon source) secretome, Oda et al. (53) identified 43 proteins from solid states and 37 from submerged cultures that had either lignocellulolytic or proteolytic activity. When the ascomycete Aspergillus flavus was grown in mineral salt medium containing quercetine glycoside rutin as the sole carbon source, Medina et al. (54) identified 12 enzymes that were involved in carbohydrate metabolism, 9 peptidolytic enzymes, and 10 different oxidoreductases. However, in our earlier study using fungus A. niger, we have demonstrated the expression and regulation of 30 carbohydrate active enzymes, 15 peptidases, and 12 oxidoreductases (5). Ten different carbon sources including xylan, birchwood, pectin, corn stover, and dried distillers' grains were used to map the secretome of F. graminearum (55). Although the genome of F. graminearum has been sequenced, there is still a significant number of extracellular proteins that have yet to be assigned to a specific function. The comparison of the secretion potential of cellulolytic and proteolytic enzyme from fungal strains of ascomycetes and basidiomycetes on cellulosic or lignocellulosic biomasses showed that T. reesei produced a significantly higher amount of GH proteins, which support its hypercellulolytic nature.
The importance of studying fungi in lignocellulosic biomass degradation and its possible applications in lignocellulosic bioenergy are also highlighted with an increasing number of genomes that have been sequenced thus far. About 18 different species including Aspergillus clavatus (44), A. flavus, A. fumigatus (42), A. nidulans (39), A. niger (43), A. oryzae (40), Aspergillus terreus, Botrytis cinerea (56), Chaetomium globosum (56), Coprinus cinereus (56), F. graminearum (56), Fusarium verticillioides (57), M. grisea (38), N. crassa (58), P. chrysosporium (41), Rhizopus oryzae (56), Sclerotinia sclerotiorum (56), and Stagonospora nodorum (56) have been sequenced and annotated, whereas several projects are still in the pipeline. Despite the importance of lignocellulose degradation by fungus and the availability of their sequenced genomes, relatively few proteomic studies have been reported, thus limiting the comparison of secretome among strains and species. Although the genome of these fungi encodes several lignocellulolytic genes, their expressions and protein production are dependent on several factors, including the type and nature of carbon sources. T. reesei is an efficient polysaccharide degrader, and one might expect it to contain a vast number of genes that encode for GH proteins. However, when compared with other fungal strains of Ascomycetes, T. reesei have lower GH encoding genes (200 GH) than in M. grisea (231 GH) and F. graminearum (243 GH). Other fungal strains belonging to Eurotiomycetes, including A. nidulans (247 GH), A. fumigatus (263 GH), and A. oryzae (285 GH) have higher GH encoding genes than in T. reesei (24). T. ressei genome harbors more GH genes than in white rot fungi P. chrysosporium, which is known for the complete degradation of all major components of the plant cell wall including lignin, cellulose, and hemicellulose (41).
The expression of the genes depends on the carbon sources. The six endoglucanases belonging to GH5 were expressed in P. chrysosporium (7), whereas only four corresponding GH5 proteins were identified and iTRAQ-quantified in T. reesei. However, in cellulosic culture conditions, secretory proteins belonging to family GH30, GH45, GH55, GH61, and GH81 were only produced by T. reesei, when compared with P. chrysosporium. The variable iTRAQ ratios of these expressed proteins in cellulosic and saw dust-and corn stover-containing medium showed dependence on carbon sources. Vanden Wymelenberg et al. (59) also found differential regulation of carbohydrate active enzymes in response to different carbon sources by P. chrysosporium. Three different GH30 endo-1,6-␤-D-glucanases (jgi͉Trire2͉ 3094͉, jgi͉Trire2͉64906͉, and jgi͉Trire2͉69736͉) that randomly hydrolyze ␤-1,6-linkages in glucans were expressed in T. reesei secretome and are differentially regulated in different carbon sources in both wild-type and mutant strains. Among the many families of glycoside hydrolases, the least characterized GH61 endoglucanase (jgi͉Trire͉64906͉) showed significantly high iTRAQ ratios when saw dust and corn stover were used as major carbon sources. Their high expressions potentially suggest their involvement in lignocellulose degradation, which open up doors for further studies. Again, CHBI (jgi͉Trire2͉123989͉) and CBHII (jgi͉Trire2͉72567͉) of T. reesei hydrolyze crystalline cellulose in the absence of endoglucanases and are thus different from CBHI and CBHII of P. chrysosporium (41) and other fungal strains. According to Medve et al. (60), substrate competition between two enzymes (CBHI and CBHII) of T. reesei at the active binding site is dominant (61). The contents of ␤(1,3)and (1,4)-glucans in cell wall varies with plant type, grasses, seaweeds, and grain bran (62); however, their presence in the cell wall validates the expression and up-regulation of GH55 exo-1,3␤-D-glucanase (jgi͉Trire2͉73248͉), which catalyzes the successive hydrolysis of ␤-D-glucose units from the nonreducing ends of 1,3-␤-Dglucans. Thus, this study highlights elevated secretion of cross-linked glucan chain-splitting endoglucanases, unique exoglucanases, cellobiose hydrolyzing glucosidases, and other GH family proteins that have been known to hydrolyze ␤-1,3or ␤-1,4-linkages. The iTRAQ data demonstrated the novelty of T. reesei through their expression of GH30, GH45, GH55, GH61, and GH81 proteins and potential of CBHI and CBHII. This may be due to chromosomal recombination or rearrangement during DNA repair or point mutations.
In T. reesei Rut C30, the iTRAQ ratio of GH7 cellobiohydrolase was 4.59 when corn stover was used as carbon source, whereas its corresponding ratios were 0.72 and 0.27 when cellulose and saw dust were used as carbon sources, indicating that expression was significantly induced by corn stover. Similarly, in T. reesei QM6a, the iTRAQ ratio 3.47 Ϯ 0.26 with corn stover as a substrate suggested its higher production compared with 2.94 Ϯ 0.269 in saw dust. The proteins such as GH5 endoglucanase EG-1 (jgi͉Trire2͉122081͉), GH61 endoglucanase-4 (jgi͉Trire2͉73643͉), GH30 endo-1,6-␤-D-glucanase (jgi͉Trire2͉3094͉ and jgi͉Trire2͉69736͉), GH45 endoglucanase V (jgi͉Trire2͉49976͉), and GH74 endoglucanase Cel74a (jgi͉Trire2͉49081͉) also showed variable expressions in corn stover and saw dust culture conditions. Thus, cellulolytic enzyme expressions were dependent on the complexity and the type of carbon sources.
The proteins belonging to oxidoreductase family, laccase, peroxidase/catalase, glyoxal oxidase, glutathione reductase, and glutathione S-transferase glyoxalase that plays an important role in lignin degradation were also detected and quantified. The protein contents in the plant cell wall and their composition, as well as their mechanochemical and regulatory roles, have been reviewed by Cassab (66), whose research emphasizes the protein as the predominant constituent of plant cell wall. According to Cassab (66), plant proteins play a major role in morphogenesis, in the ability to cross-link with the cell wall, and also in the formation of ␤-pleated sheets performing structural functional role, although another group of cell wall proteins are heavily glycosylated with arabinose and galactose. Thus, taken together with the plant cell wall composition, proportional protein content, and their possible cross-linkage with carbohydrates, iTRAQ-quantified peptidases and proteases validate their expressions during T. reesei growth on lignocellulosic biomasses. Again, the inclusion of peptidases and proteases are essential for the constitution of lignocellulosic biomass hydrolyzing enzyme mixture in enhancing the hydrolytic process by debranching polysaccharide linkage with plant cell wall protein.
The production of an economical and optimal lignocellulose-hydrolyzing enzyme mixture determines the success of the lignocellulosic biofuel industry. Although numerous research reports have concluded that the enzyme mixture with at least three types of enzymes including endoglucanase, cellobiohydrolase, and ␤-glucosidase were needed for complete hydrolysis of crystalline cellulose, this mixture is inefficient for complete digestion of cellulose. Researchers (22,23,67) have also tested ternary mixtures of purified cellulases, as well as multicomponent mixtures of pure endo-and exo-cellulases, endoxylanases, and debranching enzymes; however, the ultimate breakthrough has yet to be achieved. This could be due to a lack of knowledge on quantitative expression of proteins on natural lignocellulosic biomasses. Considering the T. reesei genome that encodes 200 GH proteins, 103 glycosyltransferases, 16 carbohydrate esterase, and several hemicellulolytic and lignolytic proteins, it will take several thousand experiments (different combinations of enzymes) to find the most efficient lignocellulose-hydrolyzing enzyme mixture. Hence, high throughput methods like quantitative proteomics are needed for systematic quantification of enzymes produced by T. reesei during lignocellulosic biomass degradation. This study established the global view of secretome during lignocellulosic biomass degradation, showing the expression and regulation of key enzymes. By knowing the expressed and most up-regulated enzymes, one can prioritize their inclusion while designing effective enzyme mixture for more efficient bioconversion of different types of biomasses. Considering the linkage pattern of cellulose with other noncellulosic polysaccharides, xyloglucans, pectins, and other biopolymers in lignocellulosic biomass and the predominant existence of ␤-(1,3)and ␤-(1,4)-D-glucans in plant cell wall (68,69), we emphasize that in vitro reconstituted enzyme mixture should contain ␤-(1,4)-, ␤-(1,3)-, and ␤-(1,6)-glucanases.

CONCLUSION
T. reesei, a filamentous fungus, not only possesses the capacity to secrete a large quantity of cellulases and hemicellulases but also acts as a host for the production of low cost enzymes useful for lignocellulosic bioenergy. This study utilizes natural lignocellulosic biomasses such as saw dust and corn stover as the major carbon sources to profile the expressions of lignocellulolytic proteins with high throughput isobaric tag for relative and absolute quantification (iTRAQ)based quantitative proteomic approach. We quantified lignocellulolytic proteins with their variable substrates-induced expressions. The biological functional classification of these quantified proteins revealed 31.3, 17.9, 13.4, 22.0, 6.3, 3.3, and 5.6% cellulases, hemicellulases, lignin-degrading proteins, peptidases, chitinases and phosphatases, and transport and hypothetical proteins, respectively. The expressions of these proteins revealed the need for low abundant auxiliary proteins for the efficient lignocellulosic biomass degradation. In addition to cellulases and hemicellulases, this study identified and quantified variable expressions of pectin-degrading proteins and peptidase and proteases in different substrates.