Comparative secretome analysis of Fusarium sp. Q7-31T during liquid fermentation using oat straw as a carbon source

Abstract

Q7-31T was screened from seven strains with high plant cell wall (PCW)-degrading enzyme activities and identified as Fusarium sp. based on the morphological characteristics and internal transcribed spacer (ITS) rDNA sequence analysis. The protein composition of the secretome produced by Q7-31T grown in a liquid medium with oat straw powders or glucoses as a carbon source was investigated using two-dimensional electrophoresis coupled with tandem mass spectrometry. A total of 28 protein species were identified among 115 spots that only existed or had a larger quantity (threefold or more) in the induction medium. As expected, most of them were potentially involved in PCW degradation. Six different superfamilies of glucoside hydrolases (GH5, GH7, GH10, GH13, GH18, and PL1) were identified, and proteases, oxidases, reductases, phospholipases, transferases, and proteins with unknown functions were also detected.

Introduction

Plant cell wall (PCW)-degrading enzymes are used in a wide array of industrial applications, in particular for the biomass conversion of fermentable monomeric sugars (Kuhad et al. 2011). The saccharification of lignocellulose by the bacterial or fungal enzymes has been a topic of extensive research since the middle of the last century (Pablo et al. 2013). The low specific performance of cellulases remains a major problem hampering the development of biomass resource conversion industries. A considerable number of attempts have been made to take advantage of enzyme and non-enzyme proteins isolated from biomass-degrading microorganisms in formulating improved enzyme mixtures (Himmel et al. 2010).

Several Trichoderma species can be cultivated on various agricultural waste products to produce lignocellulases (Sukumaran et al. 2009; Druzhinina et al. 2012). Due to its high secretion capacity and specific enzyme activity, Trichoderma reesei is currently the primary industrial source of cellulases. However, the lack of accessory enzymes such as hemicellulases and β-glucosidase in their secretome has prompted investigations into other fungal strains and enzymes that could potentially replace or supplement T. reesei cellulases (Martinez et al. 2008; Ravalason et al. 2012; Cázares-García et al. 2013).

Fusarium species as pathogenic fungi have a strong PCW-degrading ability. Many of them, such as Fusarium oxysporum, can produce multiple lignocellulases that have a synergistic effect on the hydrolysis of lignocelluloses (Gomez-Gomez et al. 2001). We have previously isolated seven Fusarium strains with a high level of PCW-degrading activity from the soil of the Qinghai-Tibet Plateau.

The proteomic investigation of filamentous fungi has been greatly limited by difficulties in protein extraction and lack of extensive fungal genome sequence databases (Carberry and Doyle 2007). The advent of robust protein extraction and separation technologies (e.g., 2-D gel electrophoresis [2-DE]; Miller et al. 2006), combined with advanced instruments (e.g., MALDI-TOF/MS [matrix-assisted laser desorption ionization time-of-flight mass spectrometry] and CapLC-ESI-MS/MS [capillary liquid chromatography/nanoelectrospray ionization-tandem mass spectrometry]) and genome sequence data, makes proteomic analysis possible in fungus-secreted enzymes (Abbas et al. 2005; Carberry and Doyle 2007). In this analysis, protein spots after 2-DE were identified, excised, and subjected to digestion with proteolytic enzymes such as trypsin. These peptide mixtures were then subjected to mass spectrometry (MS) separation, and the resultant peptide mass fingerprints were compared to the gene or protein sequence databases to facilitate protein identification (Resing and Ahn 2005; Oda et al. 2006). Annotated “reference maps” have been published for many species, with the idea of using them as standard comparisons for further 2-DE analysis (Herpoël-Gimbert et al. 2008; Adav et al. 2012).

In this study, we performed a comparative proteomic study of the enzymes secreted by Fusarium sp. Q7-31T grown in liquid medium with oat straw as a carbon source (induced culture), using 2-DE gels coupled with tandem mass spectrometry. The 2D map was used to compare the secretome composition of Q7-31T during liquid fermentation using glucoses as the sole carbon source (non-induced culture).

Materials and methods

Isolation and culture of Fusarium sp. Q7-31T

Fusarium sp. Q7-31T (JX102516) was isolated from the soil of the Qinghai-Tibet Plateau, and deposited in the China General Microbiological Culture Collection Center (CGMCC, collection accession number: CGMCC 3.17610). The pure isolate of Q7-31T was obtained through a series of subcultures on agar plates. The screening medium consisted of 0.6 % peptone, 0.1 % K₂HPO₄, 0.05 % MgSO₄·7H₂O, and 1 % glucose (pH 7.0). The plates were incubated at 28 °C for 3–5 days, and colonies that grew well were then inoculated on fresh agar plates. The fungal strains were inoculated on potato dextrose agar (PDA) plates at 4 °C. A loop-full of hyphae was removed from uniform colonies and inoculated into a basic liquid medium (BLM) with the same composition as the screening medium.

Preparation of oat straw

Dry oat straw was ground to a powder using a grinder. The powder was then boiled with water three times to remove glucose and chlorophyll extracts. Approximately 5 vol. of distilled deionized water was added to 1 kg of powder and then boiled for 30 min. The water was then decanted twice, and the powders were extracted with 75 % ethanol for 12 h and dried at 40 °C.

Enzyme assays and protein estimation

The activity of xylanase and CMCase was measured using the dinitrosalicylic acid (DNS) method as described previously (Miller 1959). PCW-degrading ability was measured using a DNS assay with 5 g/L oat straw power in 0.05 mol/L citrate buffer (pH 6.5) at 45 °C, with one unit defined as the amount of enzyme required to release 1 mol of glucose per minute from the appropriate substrate under assay conditions (Miller 1959). Protein concentration was determined according to the Bradford method using bovine serum albumin as the standard (Lowry et al. 1951).

Screening Fusarium sp. Q7-31T as a PCW-degrading fungus from seven strains

Q7-31T and the other six fungal strains were cut into plugs 4 mm in diameter, and inoculated centrally onto the plates containing PDA medium. After incubation at 25 °C for 7 days, plugs that were 8 mm in diameter were cut from the margin of actively growing colonies, and then inoculated into 60 mL inoculum medium [2 % glucose, 0.3 % peptone, in Mandels mineral salt solution: (NH₄)₂SO₄ 1.4 g/L, KH₂PO₄ 2.0 g/L, carbamide 0.3 g/L, CaCl₂ 0.3 g/L, MgSO₄ 0.3 g/L, FeSO₄ 5.0 mg/L, MnSO₄ 1.6 mg/L, ZnSO₄ 1.4 mg/L, and CoCl₂ 2.0 mg/L in distilled water)] (Eveleigh et al. 2009) in 150-mL Erlenmeyer flasks and incubated at 25 °C for 72 h. 0.5 % oat straw powder was used as the inducer in 100 mL of liquid medium (0.3 % peptone, in Mandels mineral salt solution) in 250-mL Erlenmeyer flasks inoculated with 10 mL hyphal suspensions from the inoculum medium and incubated on a shaker at 120 rpm and 25 °C. PCW-degrading activity of each culture medium was measured every 24 h using the DNS method and compared among the seven cell wall-degrading fungi stored in our lab [Fusarium sp. Q7-10 (Accession No. JX102517), Fusarium sp. Q7-21, Fusarium sp. Q7-31 (Accession No. FJ646593), Fusarium sp. Q7-42, Fusarium sp. QH24 (Accession No. JF803825) and Fusarium sp. QH101 (Accession No. JF911785) isolated from Qinghai–Tibet plateau, China].

Strain identification

Strain identification was based on standard morphological characterization and nucleotide sequence analysis of enzymatically amplified internal transcribed spacer (ITS) rDNA, including 5.8S rDNA. The cetyltrimethylammonium bromide (CTAB) method was used for genomic DNA extraction using mycelium grown on the PDA plates for 5 days (Xie et al. 2012). PCR was performed according to the procedure described by Van Burik et al. (1998), using 5′-GGAAGTAAAAGTCGTAACAAGG-3′ as the ITS5 forward primer and 5′-TCCTCCGCTTATTGATATGC-3′ as the ITS4 reverse primer. The sequences of the PCR products were compared with those available in the NCBI BLAST database (http://blast.ncbi.nlm.nih.gov).

Protein extraction and sample prior preparation

Proteins were extracted from fermented liquor using ammonium sulfate precipitation (20–90 %), followed by TCA-acetone precipitation. All extracts were then centrifuged at 13,000 g for 15 min at 4 °C. The proteins in the supernatant were precipitated and purified by using a 2D Clean-Up Kit (Amersham Biosciences), and then stored at −80 °C until analysis (Jiang et al. 2004).

2D gel electrophoresis

Precast immobilized pH gradient (IPG) strips 24 cm in length (pH 3 to 11, nonlinear) were used for the first-dimension isoelectric focusing (IEF). Proteins (500 μg) were loaded onto the IPG-strips by passive rehydration overnight at room temperature. All strips (three replicates per sample) were run together in an IPGphor equipped with an Ettan^TM IPGphor^TM 3 loading manifold. The strips were focused at 20 °C for a total of about 35 kVh. After IEF, the strips were sequentially incubated in a freshly prepared solution of 1 % dithiothreitol and 2 % iodoacetamide in 50 mM Tris–HCl (pH 8.8), 6 M urea, 20-% glycerol, and 2-% SDS (sodium dodecyl sulfonate) for 10 min. The second-dimensional sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was conducted on 8–16.5 % polyacrylamide gradient gels using an Ettan DALT II system for 30 min at 5 W/gel, and then for 5 h at 17 W/gel at 25 °C. After electrophoresis, the acrylamide gels were silver-stained for spot picking experiments and comparative analysis (De Moreno et al. 1985; Yan et al. 2000; Gitau et al. 2011).

Image digitalization and analysis

Stained gels were digitized using an image scanner (GS-800, Bio-Rad), and then images were processed with PDQuest (Bio-Rad). A differential analysis was first carried out among replicate gels in order to estimate the repeatability of the tissue map, and then the differential maps were subjected to PDquest analysis. All the software steps were manually verified in order to eliminate artifacts, splits and missed spots. For gel analysis, the volumes of spots were used, intended as arbitrary units assigned by the software. The isoelectric point and molecular weight were validated by calibration with internal standards as described previously (Gitau et al. 2011).

Protein identification

For protein identification, visible protein spots of interest in the induction medium were manually excised from the gels, destained with 15-mM K₃Fe(CN)₆ in 50-mM Na₂S₂O_3, washed with water and then stored in acetonitrile (Koc et al. 2001). The spots were subjected to overnight tryptic digestion (Wilm et al. 1996). Peptide mixtures were collected by extraction with acetonitrile followed by centrifugation. Peptides were subsequently acidified with 20 % trifluoroacetic acid (TFA), dried in a SpeedVac, resuspended in 0.2 % formic acid, and stored at −20 °C for tandem mass spectrometry.

Tandem mass spectrometry (MS/MS) analysis was performed on an XCT Ultra 6340 ion trap equipped with a 1200 high-performance liquid chromatography (HPLC) system. After loading, the samples were concentrated and desalted at 4 μL/min on a 40 nL enrichment column with 0.2 % formic acid. Then, the peptides were fractionated on a C₁₈ reverse-phase capillary column at a flow rate of 300 nL/min. Electrospray ionization (ESI) was performed under the following conditions: capillary voltage, 1730 V; dry gas, 5 L/min; dry temperature, 325 °C; trap drive, 100; skimmer 30 V; lens 1, −5 V; octopole RF amplitude, 200 Vpp; capillary exit, 90 V. The ion trap mass spectrometer was operated in positive ion mode. The trap ICC smart target was 300,000 units, and the maximal accumulation time was 100 ms. MS/MS was operated at a fragmentation amplitude of 1.3 V, and threshold ABS was 6,000 units and scan speed was 8,100 uma/s in MS and 26,000 uma/s in MS/MS scans. Peptide analysis was performed by scanning from m/z 250 to m/z 2,200 in AutoMS (n) precursor selection mode of the three most intense ions. We used dynamic exclusion to acquire a more complete survey of the peptides by automatic recognition and temporary exclusion of ions from which definitive mass spectral data had been previously acquired. Data analysis software was used to analyze MS/MS spectra and to generate a peak list that was introduced in the online version of Mascot MS/MS ion search software (http://www.matrixscience.com).

Results

Identification of Q7-31T

The mycelium of Q7-31T pure isolate was dry white villiform, and consisted of filamentous hyphae with septa (Fig. 1). A genomic DNA region corresponding to the partial sequence of the 18S ribosomal RNA gene, the internal transcribed spacer 1 (ITS1), the 5.8S ribosomal RNA gene, the internal transcribed spacer 2 (ITS2), and the partial sequence of the 28S ribosomal RNA gene was amplified by PCR using the genomic DNA isolated from Q7-31T as a template. Those nucleotide sequences were deposited in the GenBank database under accession number JX102516. The BLAST result showed that the analyzed DNA region of Q7-31T was 99 % identical to those of Fusarium and Gibberella (Fig. 2). Based on its morphological and molecular characterization, Q7-31T was identified as a strain of Fusarium species. During an extensive screening for lignocellulolytic soil fungi, Q7-31T was found growing on different kinds of natural lignocellulosic substrates, such as wheat bran, triticale bran, and oat straw, indicating that Q7-31T most likely produced well-balanced lignocellulases required to degrade these substrates. Therefore Q7-31T was chosen for further enzyme characterization.

Fig. 1

Morphological characteristics graph of the Fusarium sp. Q7-31T. a Reverse side of Q7-31T growing on PDA medium (digital image, DSLR Canon 450D, Japan); b Positive side; c Light microscope image shows filamentous mycelium with septa (phase-contrast microscope, Leica M80, Leica Microsystems, Wetzlar, Germany)

Fig. 2

Alignment of Internal Transcribed Spacer sequences. To show regions with variation among species, only 168 bases of Q7-31T were aligned with those of the closest species. Accession numbers for each reported species are in parentheses.

Enzyme assay during liquid fermentation

The enzyme assay indicated that the highest enzyme activities of xylanase, cellulase and PCW-degrading enzymes of Q7-31T were recorded after 3 days of incubation on the liquid fermentation media, as opposed to 7 days typically reported for many fungi (De Almeida et al. 2011; Marx et al. 2013 ). The myriad of different conditions reported for lignocellulolytic enzyme fermentation does not allow a proper comparison of the enzyme levels and activities for the different systems. Rezende et al. (2002) reported that Trichoderma harzianum produced 12.8 U/mL xylanase after 7 days of incubation. Isabelle Herpoël-Gimbert et al. (2008) showed that Trichoderma reesei produced 0.58 U/mg CMCase (carboxymethyl cellulose) and 0.43 U/mg FPase (filter paper enzyme). Particularly high levels of xylanase (25.3 U/mL), endocellulase (1.47 U/mL) and PCW-degrading enzymes were detected in the Q7-31T fermentation extract, compared with the other siz strains (Table 1). The oat straw powders used in liquid fermentation were degraded to a colloidal state, indicating that Q7-31T produced a well-balanced enzyme system required to degrade lignocellulosic feedstocks.

Table 1 Screening of fungal strains for PCW-degrading related activities

2-DE mapping of Q7-31T secretome

The 2-DE reference map of Q7-31T secretome was shown in Fig. 3. 2-DE enabled the separation of proteins, both induced and non-induced, over the entire pH 3–10 range and comprised proteins between 14 and 97 kDa. In total, 553 spots were identified in the liquid fermentation extracts, with 366 and 187 spots being specific to the induced and non-induced extracts, respectively. The most abundant single spots of the isoelectric series, which only existed or had a much larger quantity (3-fold or more) in the inducement medium (differential spots), were excised for identification as described previously, leading to 115 detected single spots with molecular masses (Mr) ranging from 22.2 to 80.4 kDa (Fig. 4a) and isoelectric points (pI) ranging from 4.2 to 9.0. Approximately 76 % of the total spots identified were acidic, with theoretical pI values ranging from 4 to 7, whereas the remaining 24 % were basic, with theoretical pI values ranging from 7 to 9 (Fig. 4b).

Fig. 3

Two-dimensional gel reference map of Q7-31T extracellular proteins. a Extracellular proteins in induction medium; b Extracellular proteins in non-induction medium; c Protein identities of differential proteins based on mass spectrometry (MS/MS). 2-DE was performed using a pH range of 3–10 in the first dimension. The protein loading was 500 μg and the gel was stained using the mass-compatible silver staining procedure

Fig. 4

Bar graph distribution of detected proteins based on two-dimensional gel. a. Molecular masses of the spots; b. Isoelectric points of the spots

Protein identification

Protein sequencing of the 115 differential spots MS/MS resulted in the identification of 28 protein species from the 40 spots, most of which (18 spots, 40 % relative to the total secreted) were enzymes involved in the glucan hydrolysis, including xylanases, cellulases, alpha-amylases, and chitinases (Fig. 3c, Table 2). However, no extracellular β-glucosidases were detected. Several metal-containing oxidases and other oxidoreductases potentially linked to lignin degradation were detected in the secretomes (6 spots). In addition to lignocellulolytic-related enzymes, proteases (10 spots), esterases (2 spots), transferases (2 spots), ATPase (1 spot) and hypothetical proteins with unknown function (1 spot) were also detected (Fig. 5).

Table 2 List of the protein identities of differential proteins based on tandem mass spectrometry (MS/MS)

Fig. 5

Pie chart distribution of protein identification based on protein quantity on the two-dimensional gel

According to the carbohydrate-active enzyme database (http://www.cazy.org), the detected glycosyl hydrolases (GHs) could be categorized into six different superfamilies, including GH5, GH7, GH10, GH13, GH18 and PL1 (Polysaccharide Lyases family 1) (Fig. 6). Both endoglucanases (EG) and cellobiohydrolases (CBH) from GH7 were found in the differential proteins. In addition, a total of six different proteases were detected, including M28, M14 and S53.

Fig. 6

Pie chart distribution of glycosyl-hydrolase identities of different superfamilies based on protein quantity on the two-dimensional gel

Discussion

In this study, we isolated and identified a PCW-degrading strain of Fusarium sp. Q7-31T from the soil of the Qinghai–Tibet plateau, China. The microbiology of Qinghai–Tibet plateau remains largely unexplored due to the high elevation and unique permafrost environment (Xie et al. 2012). The ITS rDNA sequence of Fusarium sp. Q7-31T showed a high identity with Fusarium oxysporum. The strains of Fusarium sp., which are commonly isolated from soil, deadwood, plant roots, and fungal biomass, have largely been studied as pathogenic fungi. In contrast to studies on Trichoderma sp., studies on the PCW-degrading enzymes of Fusarium sp. have been limited to the identification of proteins linked to its pathogenicity to fungi and plants. Therefore, further investigation was required to characterize the PCW-degrading enzymes expressed by the strain of Fusarium sp. Q7-31T.

MS has been established in recent years as a key tool for rapid and reliable protein identification. MS-based protein identification of organisms with fully sequenced genomes is very efficient, enabling the identification of hundreds to thousands of proteins in a short time. To our knowledge, however, the Fusarium sp. genome has not yet been sequenced. Thus, a major problem for the identification of proteins from Q7-31T is the limited information available in the protein databases. To enhance the identification rate, we conducted cross-species identification (Wilkins and Williams 1997) and searched the proteins against all species in the fungus kingdom.

In this study, the identification rate observed from protein spots cut to proteins being identified by tandem MS is approximately 35 %. It is expected that the identification rate without cross-species identification would be much lower. A total of 40 out of the 115 differential spots were identified from the 28 protein species (Fig. 3, Table 2). There are 22 spots of 16 protein species in the induction medium, indicating that a more complex extracellular enzyme system is needed for the utilization of oat straw. In addition, 75 spots could not find positive results in the identification, which might be associated with protein glycosylation (Abbas et al. 2005; Fryksdale et al. 2002).

Hydrolytic enzymes account for 78 % of the total proteins identified according to the spot quantity observed on the gel. Most of them (13 of 19 spots) were detected only in the induced medium. As expected, most of the proteins identified were glycoside hydrolases related to PCW degradation, and could be assigned to six superfamilies (Fig. 6). Xylanases of the GH10 family were the most abundantly secreted proteins (10 spots of 3 different proteins). Those proteins account for 51 % of the total glucoside hydrolases from Q7-31T, while CBH accounts for 60–70 % of the total glucoside hydrolases in the secretome of T. reesei (Sandgren et al. 2005). Xylanases act on lignin parts of a PCW, breaking it into little chains with a reducing or non-reducing end group. Xylan forms a sheath on each cellulose microfibril and is also zipped into the cellulose microfibrils during crystallization (Scheller and Ulvskov 2010). Thus, xylanases play an important role in producing reactive sites for cellulases and hemicellulases in PCW degradation (Himmel et al. 2010), and these functions were especially important during the degradation of the PCW. Two exoglucanases from GH7 and another two endoglucanases from GH5 and GH7 were also identified, but one of them, cellobiohydrolase I, was identified with a mascot score of 19, and thus should be considered provisional (Table 2). Alpha-amylase, pectatelyase, and chitinase were also detected in the secretome, but no β-glucosidase was identified in the differential spots. A previous study indicated that β-glucosidases were either intracellular, membrane-anchored, or played only a minor role in cellulose hydrolysis (Shallom and Shoham 2003). The production of a PCW degradation enzyme was shown to be dependent on fungus cultivation conditions and transcriptional regulation (Stricker et al. 2008). The content of the PCW enzyme system may be directly related to the utilization of oat straw as a carbon source.

We identified some major components of the protease system (Marx et al. 2013). Six proteases of three different families (M14, M28 and S53) were identified from ten spots. Four of them were present only in the inducted medium. This suggested that hydrolysis of protein played an important role during PCW enzymatic hydrolysis in Q7-31T. A large quantity and different kinds of proteases in the induced enzyme system were observed, and the proteases were the second highest in the system (Fig. 5). There are two possible reasons for this. Firstly, it may regulate the enzyme activity of hydrolysis, since there were different spots even in one kind of hydrolysis enzyme. For example, there were ten spots in xylanase of GH10 (Table 2). The same enzyme of different spots may act on different structures or reaction sites during the degradation of xylan. Secondly, it may decompose the hydrolysis enzymes in the induced medium due to the presence of a large amount of proteins (data not shown). There were more than 600 spots on the 2-DE gel (Fig. 3). Different hydrolysis enzymes worked at different times during the degradation of the PCW. Some enzymes may have supplied the reactive sites for the other enzymes, while some enzymes were decomposed after hydrolysis, as it may have interfered with the other enzymes given the limitation of space and time.

An ATPase from the AAA family was detected only in the induced medium, indicating that extra energy was needed in the degradation of the PCW. A 3-hydroxyisobutyrate dehydrogenase and a protein containing the 2OG-FeII domain were also found only in the induced medium, indicating that the fungal oxidation-reduction system was very active during the degradation of the PCW. The detection of lipase and acyltransferase showed that the saccharification of the PCW by fungus was an extremely complex process requiring the cooperation of different enzyme systems.

In summary, a PCW-degrading strain of Fusarium sp. Q7-31T was isolated and identified. A total of 28 extracellular protein species were identified form Fusarium sp. Q7-31T, most of which were involved in PCW degradation. Proteases, oxidoreductases and other enzymes made up a significant group. This suggested that fungal saccharification of a PCW required the cooperation of various enzymes systems with different biological functions. The results presented here indicated that Q7-31T, which grew particularly well with oat straw as a carbon source and produced a PCW enzymes cocktail with excellent xylanase and cellulase activities and glycosyl hydrolases from different families, had the potential to produce alternative enzymes for PCW degradation.