A time course analysis of the extracellular proteome of Aspergillus nidulans growing on sorghum stover

Hemicellulose degradation

Hemicellulose is the second most abundant fraction (25-35%) of lignocellulosic biomass [20]. Hydrolysis of hemicelluloses into simple sugars requires multiple enzymes including beta-1,4-endoxylanase, beta-xylosidase, alpha-glucuronidase, alpha-L-arabinofuranosidase and acetyl xylan esterase [26, 27]. Beta-endoxylanases cleaves the backbone of xylan and beta-xylosidase releases the xylose units from xylobiose and xylooligomers. Removal of xylan side chains is catalyzed by acetyl xylan esterases, ferulic acid esterases, alpha-L-arabinofuranosidases and alpha-D-glucuronidases [28, 29]. Xyloglucanactive beta-1,4-endoglucanase and beta-1,4-glucosidase cleaves xyloglucan backbone and beta-1,4-endomannanase and beta-1,4-mannosidase acts on (galacto-) mannan backbone [28, 30]. Twenty-eight different hemicelluloses degrading enzymes were identified in our study (Table 1). A total of twenty A. nidulans ORFs have been assigned to GH3 family with three of them being beta-xylosidases [31]. We found all three of these beta-xylosidases (AN8401 AN2359, AN2217). Beta-xylosidases (AN8401 and AN2217) were present in abundance on day 1 and showed a steady increase in spectrum count until day 14. Another beta-xylosidase (AN2359) was present on day 1, 3 and 7 but not on day 14. AN7275, a putative GH43 family xylosidase, only appeared on day 14. We found two of the potential three endoxylanses belonging to GH10 family (AN1818 and AN7401) and one out of the two belonging to GH11 family (AN3613). Beta-1,4-endoxylanase (AN1818) showed a dramatic increase in spectrum count from 101 on day 1 to 720 on day 14. AN3613 started with a high spectrum count on day 1 and then reached a plateau. AN7401 (which has a CBM) was present in lower abundance compared to the other xylanases. An alpha-1,4-galactosidase belonging to family GH27 and another belonging to GH36 (AN7152 and AN8138) were identified. AN7152 was found in high abundance on all 4 days, whereas AN8138 was only secreted on day 7 and 14. Eight proteins (AN2533, AN7781, AN8007, AN2534, AN7275, AN8477, AN10919 and AN7313) belonging to family GH43 were identified. Out of these eight proteins, AN7275 and AN8477 do not have a signal peptide. AN7781 was secreted on all days. Its spectrum count increased more than twice from day 1 to day 3 and then it was constant on day 7 and 14. We identified an arabinosidase from family GH53 (AN1571), and two other arabinosidases (AN7908 and AN2632) belonging to family GH62. We also found one enzyme belonging to family GH53 (endogalactanase), one from GH67 (alpha-glucuronidase), and one from GH93 (exo-arabinase). Most of the hemicellulases were secreted on all days except for nine hemicellulases, which were absent on day 1 but secreted on all other days. One xyloglucanase (AN5061) appeared on day 14 only whereas galactosidase (AN8138) was found on only day 7 and 14.

Cellulose degradation

The complete hydrolysis of cellulose requires random cleavage of internal bonds by endoglucanases, removal of cellobiose by cellobiohydrolases (exoglucanases), and release of glucose from cellobiose by beta-glucosidase [32–35]. Twenty-one different enzymes likely to be involved in cellulose degradation were identified in our cultures (Table 2). One beta-glucosidase (AN9183) belonging to family GH1 and seven beta-glucosidases belonging to family GH3 were detected. All of the family GH3 enzymes except AN2227 had recognizable signal peptide sequences. Beta-1,4-glucosidase (AN2828 and AN4102) showed high spectrum counts on all days, whereas AN7396 and AN5976 were absent on day 1 and day 14 respectively. Cellobiohydrolases are assigned to family GH6 and GH7. To date only family GH7 cellobiohydrolases have been characterized in A. nidulans[36] out of four predicted cellobiohydrolases belonging to family GH6 and GH7 from the A. nidulans genome sequence. We identified two cellobiohydrolases (AN1273 and AN5282) from family GH6 and two cellobiohydrolases (AN0494 and AN5176) belonging to family GH7. All showed a gradual increase in spectrum count from day 1 to day 14. Notably cellobiohydrolase (AN5176) showed high spectrum count of 234 on day 14 as compared to other cellobiohydrolases, whereas AN5282 was absent on day 1. We found one beta-1,4-endoglucanase (AN3418), from family GH7 and three beta-endoglucanases belonging to family GH5. Endoglucanases (AN1285 and AN8068) showed relatively low spectrum counts compared to the beta-endoglucanases from family GH7. A cellulase family protein (AN9166) was only present on day 3 and 14. We identified four beta-1,4-endoglucanases from family GH61 (AN10419, AN6428, AN3860 and AN3046) out of nine predicted ORFs for GH61 family proteins in the genome. Out of a total of twenty-one cellulases, seven enzymes were not secreted on day 1 whereas one putative endoglucanase AN2664 appeared only on day 14.

Pectin degradation

Due to the complex structure of pectin an array of enzymes are needed for its degradation [37, 38]. We identified twenty-two enzymes involved in pectin degradation (Table 3). The GH2 family contains beta-galactosidase, beta-mannosidase and beta-glucuronidases. We found two GH2 family members (AN2935 and AN2463). Out of these two GH2 family members, AN2395 showed increased spectrum count across all days. GH28 family consists of endo- and exo-polygalacturonases, rhamnogalacturonan hydrolases and xylogalacturonan hyrolases. Based on the A. nidulans genome nine ORFs have been assigned to this family. We identified three exopolygalacturonases (AN8761, AN8891, and AN10274) in the secretome. Out of three GH28 family exopolygalacturonases only AN8761 was secreted with high abundance on all days. So far GH78 family contains alpha-L-rhamnosidase exclusively. Seven A. nidulans ORFs have been assigned to this family. We found one alpha-rhamnoside from GH78 (AN7151) and it showed a gradual increased in spectrum count across all the days. We identified two pectate and two pectin lyases from PL1 family and two from PL3 family. Four A. nidulans ORFs have been assigned to rhamnogalacturonan lyases from family PL4. We identified two members of PL4. Two unsaturated rhamnogalacturonan hydrolases (AN7828 and AN9383), two pectin methylesterases (AN4860 and AN3390), and one rhamnogalacturonan acetylesterase (AN2528) were also identified. With the exception of members from GH2, GH78, GH105 and PL4, the pectinases were expressed at higher levels only in the early growth stages.

Starch degradation

Ten A. nidulans ORFs have been assigned to family GH31. We identified only three alpha-1,4-glucosidases from family GH31, two alpha-amylases from family GH13 and one amylase from family GH15 (Table 4).

Fungal cell wall remodeling enzymes

Sixteen enzymes, including chitinases and beta-1,3-endoglucanases involved in fungal cell wall modification/degradation were identified (Table 5). Though chitinases and beta-1,3-endoglucanases were present on day 1, the spectrum count of most of them increased notably on day 3.

Enzymes involved in various plant cell wall modifications

Table 6 reports four enzymes, two feruloyl esterases, a cellobiose dehydrogenase, and a tyrosinase potentially involved in modification of lignin and other phenolics. The cellobiose dehydrogenase may also be involved in modification of cellulose.

Mannan and bacterial cell wall degrading enzymes

Table 7 and Table 8 show seven mannan degrading enzymes and two bacterial cell wall degrading enzymes respectively.

Proteases

Table 9 lists eleven fungal extracellular proteases, the production of which increased over time. They included one metallopreotease (AN7962) one alkaline protease, one aspartyl protease and two hypothetical serine proteases, two serine carboxypeptidases (AN1226, AN2237, AN7121), two putative aminopeptidase, an alkaline protease (AN5558), a putative dipeptidyl aminopeptidase (AN6438), and a membrane peptidase (AN7159). Table 10 show thirteen esterases and lipases secreted by A. nidulans. Out of these 12 esterases and lipases, seven were secreted with high abundance throughout the time course, whereas the rest of them were only present on one or two time points.

Miscellaneous proteins

Total 164 proteins were identified in this category including several hypothetical proteins (Additional file 1: Table S1).

The presence of high enzyme activities in the extracellular fluid of A. nidulans led us to quantify the solubilized breakdown products from the sorghum in the extracellular filtrate (ECF). Soluble polysaccharides and oligosaccharides present in ECF were estimated by gas chromatography after depolymerization by methanolysis and conversion of trimethylsilyl methyl glycosides. Between 0.01% and 0.23% of the total dry mass of the initial sorghum was soluble as each sugar type (Figure 4). The galacturonic acid and glucose content fell dramatically on day 1 and day 2 compared to the control suggesting that these already soluble sugars were readily converted into a form that could be taken up by the fungus. The amounts of arabinose and xylose containing soluble material increased compared to the control, reflecting digestion of the xylan by the enzymes and perhaps a lower ability to take up the solubilized products. The amount of mannose increased fairly consistently from day 1 to day 14. Rhamnose and galactose levels remained relatively constant. The amount of sugars at the different time points was probably a result of the dynamic balance between enzymatic breakdown of sorghum polysaccharides and uptake of sugars by fungus.

To find out how effective the fungus was at digesting the sorghum polysaccharides, we quantified the amount of each sugar type remaining on the plate after various times of culture growth (Figure 5). Apart from a slight increase in glucose, which may have come from growth of the fungus or reflect experimental variability, the sorghum sugars decreased steadily. By day 7, 20% of the measurable glucose and xylose had been consumed and by day 14, 30% of these sugars had been utilized. The appearance of mannose and galactose, not present in the uninoculated sorghum, probably reflected the synthesis of fungal galactomannoproteins and mannans.

Growth conditions

Aspergillus nidulans Strain A78 was obtained from the fungal genetic stock center (FGSC) [56]. 3 g of ground sorghum (variety mix of AtX2752/RtX2783 and AtX2752/RtX430) was ground in a Thomas Wiley® Mini-Mill (Thomas Scientific, Swedesboro, NJ, USA) by passing through a 60-mesh screen. Fungal spores were counted using a hemacytometer (Spectrum Scientifics, Philadelphia, PA, USA) and adjusted to a million spores per milliliter in water. The ground sorghum was moistened with 6 ml water in a petri plate and autoclaved for an hour at 121⁰C. To these petri plates 15 ml of minimal media (pH 6.5; 95 ml of water, 5 ml 20X nitrate salts and 0.1 ml 1000X trace elements) and 1 ml of spore suspension containing 10⁶ spores was added. Then the petri plates were incubated at 37⁰C and 70% of relative humidity for 1, 2, 3, 5, 7 and 14 days. Uninoculated sorghum stover incubated in same conditions as inoculated stover was used as control. All the chemicals used in this study were purchased from Sigma-Aldrich (St. Louis, MO, USA), unless otherwise stated. Fungus grown on borosilicate 3 mm solid glass beads (Aldrich, Milwaukee, WI) in 1% glucose for 34 hours was used as a control for comparison of secretomes. A. nidulans cultured in 1% glucose for 34 hrs utilized only 1/4^th of the initial glucose and thus should not be undergoing starvation.

Sample preparations for microscopy

For scanning electron microscopy (SEM), fungal samples approximately, 5 mm squares, were cut from the plates and placed in the first fixative solution (2.5% glutaraldehyde in 0.2 M sodium phosphate, pH 7.2) and incubated for 2 h. The glutaraldehyde solution was removed and replaced with 0.2 M sodium phosphate buffer, pH 7.2 and samples were washed 3X for 20 min each. Post fixation was done in 1% osmium tetraoxide in dH₂O for an hour. Dehydration was done with a series of ethanol solutions (50, 70, 90, 95,100,100,100%), for 15 minutes each. Samples were critical point dried and mounted on aluminum stubs using silver paint adhesive followed by coating with gold/palladium. Samples were observed using an FEI quanta 600 electron microscopy (FEI, Hillsboro, OR, USA).

For transmission electron microscopy (TEM), the fixation was done as for the SEM samples. Post fixation was done in 1% osmium tetroxide /1.5% potassium ferricyanide in 0.1% sodium cacodylate buffer for 1 hour. Using a series of ethanol solutions (as described in SEM protocol), samples were dehydrated for 15 minutes each. The dehydrated samples were washed 2X with 100% propylene oxide and then incubation overnight in propylene oxide/polyresin/bed 812 overnight (Polysciences Inc., Warrington, PA, USA) followed by immersion in polyresin/ bed 812 and polymerized at 60⁰C. Samples were sectioned using a Leica EMCU 6 ultramicrotome (Midwest Lab Equipment, Iowa City, IA, USA). Sectioned samples were placed on nickel; carbon/formavar coated grids and stained with uranyl acetate and lead citrate and observed using a JEOL 2100 Transmission electron microscope (JEOL, Austin, TX, USA).

Estimation of fungal growth by chitin estimation

The entire plate contents of A. nidulans cultures after day 1, 2, 3, 5, and 14 was freeze-dried and then mixed thoroughly. One mg samples were transferred to vials with teflon lined caps and hydrolyzed in 3 ml of 6 N HCl, at 100⁰ C for 6 h. The HCl was evaporated overnight in a speed-vac [57]. Once the extracts were dried, 10 μl of 1 M ammonium hydroxide was added to neutralize any residual HCl. Then 10 μl of 100 nanomoles methyl glucamine was added as an internal standard, followed by 100 μl of 20 mg/ml potassium borohydride in DMSO and incubated at 40⁰C for 90 min to reduce hexosamines to hexosaminitols. After 90 min the reaction was stopped by adding 10 μl glacial acetic acid, followed by addition of 20 μl methyl imidazole and 200 μl of acetic anhydride and incubated for 10 min at room temperature to acetylate the hexosaminitols [58, 59]. The reaction was stopped by adding 500 μl water and the alditol acetates were purified by adsorbing them to a C18 sep-pak and desorbing them with methylene chloride to remove the polyphoenols. The dichloromethane was then evaporated to dryness. 25 μl of ethyl acetate was added to the dried sample, out of which, 1 μl was injected into the gas chromatograph (Agilent, Santa Clara, CA, USA). A standard curve was prepared using known amounts of glucosamine ranging from 10 to 250 nmol. The glucosamine content of 1 mg aliquots of dried A. nidulans grown in liquid media (49.50 μg) was used as a conversion factor to estimate milligrams of fungus per milligram of dry mass.

Estimation of enzyme activities by capillary zone electrophoresis (CZE) and dinitrosalicylic acid (DNS) assay

Extracellular extracts were collected by washing the whole petri plate contents with 5 ml of autoclaved water and filtering the extract using Whatman no 1 filter paper. Enzyme activities were detected semi-quantitatively by CZE using 8-aminopyrene-1,3,6-trisulfonate (APTS) labeled substrates (xylohexose, cellopentose, arabinoheptaose, mannohexose) [60]. 1 μl aliquots of extracellular extract and 2 μl of respective APTS labeled substrates were added to 22 μl of 50 mM ammonium acetate buffer of optimum pH for the respective enzymes to make a total volume of 25 μl. Samples were incubated for 1 h at 37⁰C and the progress of enzymes was followed by CZE (Bio-Rad, Hercules, CA, USA).

The DNS assay, developed by Sumner and Graham [61] for determination of reducing sugar, was used for quantitative determination of enzyme activity. We used a DNS reagent composed of 0.75% di-nitrosalicylic acid, 0.5% phenol, 0.5% sodium metabisulfite, 1.4% sodium hydroxide, and 21% sodium potassium tartrate [62, 63]. To 10 μl of aliquot of extracellular extract, 20 μl of 50 mM ammonium buffer of optimum pH and 20 μl of appropriate substrate (1%W/V) prepared in water, were added. The reaction was carried out in a 96 well PCR plate (Corning, NY, USA) and allowed to incubate for 30 minutes at 37⁰C. The reaction was terminated by addition of 40 μl of DNS mixture to each well and the plate was heated at 100⁰C for 5 min. We read the plate in a 96 well plate reader at 550 nanomoles (Tecan, San Jose, CA, USA). After subtracting substrate blanks from the reading enzyme activities were calculate from standard graphs.

Estimation of solublized sugars in extracellular filtrate

To identify and quantitate the breakdown products of the polysaccharides, oligosaccharides present in extracellular filtrate were estimated by gas chromatographic (GC) analysis of trimethylsilyl methyl glycosides. Methanolysis and derivatization were performed using the protocol of Chaplin [64] modified by Komalavilas and Mort [65]. A 25 μl aliquot of sample and 100 nanomoles of inositol as an internal standard were dried in 1 dram glass vials in speed vac. 200 μl of methanol 1.5 M of HCI was added to each vial followed by addition of 100 μl of methyl acetate and kept in a heating block at 80⁰C for minimum 3 hours. The vials were cooled, followed by addition of few drops of t-butanol and then evaporated under a stream of nitrogen at room temperatures. Butanol co-evaporates with the HCl, helping to remove the HCl without degrading sugars. 25 μl of 1:1:5 mixture of hexamethyldisilazane: trimethylchlrosilane: pyridine (TMS) was added to all samples. TMS in all the samples was evaporated using nitrogen gas. Dried samples were re-dissolved in 300 μl of isooctane. 1 μl of sample was injected in the gas chromatograph (Agilent, Santa Clara, CA, USA). Standards were made by taking aliquot of 100 nanomoles of each sugar standards and 100 nanomoles of inositol. From the integrator we calculated relative peak areas and used these to calculate how much of each sugar in sample. Area of sugar peak in sample/area of inositol peak in the sample/area of sugar peak in the standards/area of inositol in standards X 100 = number of nanomoles in the sample.

Liquid chromatography-tandem mass spectroscopy

The fungal cultures grown on sorghum for day 1, 3, 7 and 14 and fungal culture grown in 1% glucose were washed with 5 ml autoclaved water and the filtrates were collected. Concentration of total protein in these extracts was measured by Bradford assay. Small aliquots of the extracellular filtrates containing 50 ug of total protein were run on SDS-PAGE. Gel bands were excised, reduced with Tris (2-carboxyethyl) phosphine), alkylated with 2-Iodoacetamide, and digested for 16 hrs with 8 μg/ml trypsin using ammonium bicarbonate buffer. Peptides were extracted and analyzed by LC-MS/MS using an LTQ-Orbitrap XL hybrid mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA). For this analysis, an Eksigent spilt less LC pump (Eksigent, Dublin, CA, USA) was used to separate peptide populations on analytical C18 nano columns (Bruker-Michrom Inc., Auburn, CA, USA), with the column effluent being sprayed directly into a new objective pico view ion source. Using a “Big Three” MS/MS method, the Orbitrap analyzer collected an ultra-accurate (R = 60,000) scan of intact peptides, whilst the LTQ ion trap simultaneously performed MS/MS fragmentation analysis of the each of three most abundant peptides eluting in that chromatographic fractions. The LC-MS/MS raw files were used for database searching via the software application Mascot (v2.2.2 from Matrix Science, Boston, MA, USA) against sequences from A. nidulans as well as against a sorghum database [23, 66] . Searches were validated using Scaffold (Proteome Software Inc., Portland, OR, USA) and the Peptide Prophet algorithm [67]. Criteria for accepting protein identification, was only protein probability thresholds greater than 99% were accepted and at least two peptides needed to be identified, each with 95% certainty. Protein candidates containing similar peptides were grouped to fulfill the principles of parsimony. Search results were also checked for false discoveries using reversed decoy sequence databases, and no decoy sequences were detected, thus bringing our false discovery rate to zero. Changes in protein expression between samples from different time points were examined using the spectral count method [68]. The complete peptide reports from the Scaffold results for samples taken from days 1, 3, 7, and 14 are given in Additional file 3: Table S3, Additional file 4: Table S4, Additional file 5: Table S5, and Additional file 6: Table S6 respectively. SignalP was used to predict secretion signals in the identified proteins [69, 70]. Additional information for the presence of a signal peptide was obtained by accessing the following URL [25].

Estimation of residual sorghum hydrolysates by saeman hydrolysis

Saeman hydrolysis was done on the total biomass on the plates after 1, 7, and 14 days of fungal growth to estimate total sugars [30]. Three hundred mg samples of the thoroughly mixed freeze dried biomass from each petri plate were suspended in 3 ml of 72% sulfuric acid. The samples were incubated at 30⁰ C for 60 min with stirring every 5–10 min followed by dilution of sulfuric acid to 4% by adding 84 ml of water and autoclaved for an hour at 121⁰C. 2 ml aliquot of each sample was neutralized with calcium carbonate to pH 5–6. 25 μl of sample was processed for methanolysis, trimethylsilylation, and gas liquid chromatography (GLC) analysis as described above.